Advances in Insect Physiology, Volume 32 (Advances in Insect Physiology)

Advances in Insect Physiology Volume 32 i This page intentionally left blank ii Advances in Insect Physiology ed...

Author: Stephen Simpson

219 downloads 1504 Views 5MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Advances in Insect Physiology

Volume 32

i

This page intentionally left blank

ii

Advances in Insect Physiology edited by S. J. Simpson School of Biological Sciences, The University of Sydney, Sydney, Australia Volume 32

ELSEVIER ACADEMIC PRESS

Amsterdam Boston Heidelberg London New York Oxford Paris San Diego San Francisco Singapore Sydney Tokyo iii

ELSEVIER Ltd The Boulevard, Langford Lane Kidlington, Oxford OX5 1GB, UK r 2005 Elsevier Ltd. All rights reserved. This work is protected under copyright by Elsevier Ltd, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier’s Rights Department in Oxford, UK: phone (+44) 1865 843830, fax (+44) 1865 853333, e-mail: [email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.elsevier.com/locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier’s Rights Department, at the fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. First edition 2005 British Library Cataloguing in Publication Data A catalogue record is available from the British Library. Academic Press An imprint of Elsevier 525 B Street, Suite 1900 San Diego, CA 92101-4495, USA http://www.academicpress.com

Academic Press An imprint of Elsevier 84 Theobald’s Road London WC1X 8RR, UK http://www.academicpress.com

ISBN-13: 978-0-12-024232-0 ISBN-10: 0-12-024232-X ∞ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in Great Britain.

iv

Contents Contributors

vii

Insect Immunity: An Evolutionary Ecology Perspective M. T. SIVA-JOTHY, Y. MORET, J. ROLFF

1

Antennal Movements and Mechanoreception: Neurobiology of Active Tactile Sensors ¨ RR E. M. STAUDACHER, M. GEBHARDT, V. DU

49

Eupyrene and Apyrene Sperm: Dichotomous Spermatogenesis in Lepidoptera M. FRIEDLA¨NDER, R. K. SETH, S. E. REYNOLDS

206

Index

309

v


vi

Contributors V. Du¨rr Abteilung Biologische Kybernetik und Theoretische Biologie, Fakulta¨t fu¨r Biologie, Universita¨t Bielefeld, Postfach 10 01 31, D-33501 Bielefeld, Germany

M. Friedla¨nder Department of Life Sciences, Ben Gurion University of the Negev, Beer Sheva, Israel

M. Gebhardt Lehrstuhl fu¨r Zoologie, TU Mu¨nchen, Lichtenbergstr. 4, D-85748 Garching, Germany Present address: Institut fu¨r Zoologie, Abteilung fu¨r vergleichende Neurobiologie, Universita¨t Bonn, Poppelsdorfer SchloX , D-53115 Bonn, Germany

Y. Moret Charge´ de recherche CNRS, Universite´ de Bourgogne, UMR 5561 Biogeósciences, Equipe Ecologie Evolutive, 6 boulevard Gabriel, 21000 Dijon, France

S.E. Reynolds Department of Biology & Biochemistry, University of Bath, Bath, UK

J. Rolff Department of Animal and Plant Sciences, University of Sheffield, Sheffield, Sheffield S10 2TN, UK

R.K. Seth Department of Zoology, University of Delhi, Delhi, India

M.T. Siva-Jothy Department of Animal and Plant Sciences, University of Sheffield, Sheffield S10 2TN, UK

E.M. Staudacher Center for Computational Biology, Montana State University, 1 Lewis Hall, Bozeman, MT 59717, USA Present address: Lehrstuhl fu¨r Zoologie, TU Mu¨nchen, Lichtenbergstr. 4, D-85748 Garching, Germany vii


viii

10

Insect Immunity: An Evolutionary Ecology Perspective Michael T. Siva-Jothya, Yannick Moretb and Jens Rolffa a

Department of Animal and Plant Sciences, University of Sheffield, Sheffield S10 2TN, UK Charge´ de recherche CNRS, Universite´ de Bourgogne, UMR 5561, Biogeósciences, Equipe Ecologie Evolutive, 6 boulevard Gabriel, 21000 Dijon, France

b

1 2 3

4

5 6

7

Introduction 2 Defence via behaviour 4 The insect immune system – boundary defence 5 3.1 The cuticle 6 3.2 The integumental epithelium 8 3.3 Sensilla 8 3.4 The digestive system 8 3.5 The spiracles and respiratory system 9 3.6 Reproductive tract 9 The insect immune system – haemocoelic defence 10 4.1 Clotting and wound closure 11 4.2 Self/Non-self recognition 12 4.3 Signal transduction 14 4.4 Effector systems – enzyme cascades and cytotoxins 16 4.5 Effector systems – antimicrobial peptides 17 4.6 Effector systems – haemocytes 17 Ecological immunology and variation in immune defence 21 5.1 Life history theory and the costs of immune defence 22 5.2 Specific relationships between hosts and parasites 30 Outlook 30 6.1 Memory in insect immunity? 31 6.2 High specificity, few receptors 31 6.3 Multiple infections 32 6.4 Plasticity of immune function 33 Conclusions 33 Acknowledgements 36 References 36

Abstract We review recent advances in our understanding of the mechanisms of insect immune defence, but do so in a framework defined by the ecological and ADVANCES IN INSECT PHYSIOLOGY VOL. 32 ISBN 0-12-024232-X DOI: 10.1016/S0065-2806(05)32001-7

Copyright r 2005 by Elsevier Ltd All rights of reproduction in any form reserved

2

MICHAEL T. SIVA-JOTHY ET AL

evolutionary forces that shape insect immune defence. Recent advances in genetics and molecular biology have greatly expanded our understanding of the details of the immune mechanisms that enable insects to defend themselves against parasites and pathogens. However, these studies are primarily concerned with discovering and describing how resistance mechanisms work. They rarely address the question of why they are shaped the way they are. Partly because we know so much about the mechanisms that it is now becoming possible to ask such ultimate questions about insect immunity, and they are currently emerging from the developing field of ‘ecological immunology’. In this review we first present an overview of insect immune mechanisms and their coordination before examining the key ecological/evolutionary issues associated with ecological immunity. Finally, we identify important areas for future study in insect immunity that we feel can now be approached because of the insight provided by combining mechanistic and ecological approaches.

1

Introduction

Insects are coelomate metazoans that have a dominant, open haemocoelic cavity in which the organs and tissue systems are suspended. This single pervasive and continuous body space is home to the last line of defence against pathogens and parasites: the insect’s immune system. A single, relatively large, fluid-filled body cavity has several advantages (see Willmer, 1990 for a comprehensive analysis). It is relatively efficient at distributing nutrients from the gut and collecting waste products; it provides a discrete environment for the evolution of large, complex tissues and organs and consequently allows independent growth of the gonads and provides a hydrostatic skeleton. These, and other, features have profound influences on an organism’s size, locomotion, life history and consequently ecology and evolution. Clearly, there are many aspects of insect biology that will influence and constrain the evolutionary paths that are available to this immensely successful taxon (see McGavin, 2001), but the consequences of, and constraints imposed by, an open haemocoel are core to understanding many aspects of the organisation and control of the insect immune system. Convention dissects the insect immune system into cellular and humoral components, a division which probably reflects the historical unfolding of our understanding of the vertebrate system as well as the practically constrained approach to studying insect immunity. Most studies of insect immunity using this conceptual dichotomy acknowledge (usually implicitly) that the approach is convenient, rather than biologically meaningful. With the advent of, and advances in, genomics our understanding of the mechanistic basis of insect immunity has changed dramatically in the last few years. Coupled with these insights of the immune machinery of Drosophila and Aedes is the development

INSECT IMMUNITY: AN EVOLUTIONARY ECOLOGY PERSPECTIVE

3

of an area of evolutionary biology that seeks to understand the basis for additive genetic variation in immune function. This microevolutionary perspective was initially focussed on sexual selection in vertebrates (Hamilton and Zuk, 1982; Folstad and Karter, 1992; Sheldon and Verhulst, 1996), but has more recently shifted its view to an ecological one that uses invertebrates as models (e.g. Schmid-Hempel, 2003; Schmid-Hempel and Ebert, 2003; Rolff and Siva-Jothy, 2003). The ‘mechanistic’ and ‘evolutionary’ approaches differ in several respects, but most importantly in how they deal with individual variation in immune traits and the kinds of pathogens they expose their model hosts to (see Hultmark, 2003). The mechanists necessarily remove individual variation from their systems because it is hard enough to isolate and identify mechanisms when the individuals are genetically similar. Consequently, what we know of immune mechanisms tends to come from genetically constrained models reared under ideal nutritional conditions in a relatively aseptic laboratory environment. This approach was, and is, a design necessity in addressing mechanistic questions. In contrast, evolutionary and ecological studies tend to use generalised assays of immune-function (for critiques see Siva-Jothy, 1995; Owens and Wilson, 1999; Ryder, 2003; Adamo, 2004) to address questions about the evolutionary maintenance of variation in immune systems, an approach that at best oversimplifies, and at worst ignores, the constraints imposed by and the meaning of quantitative measures of the underlying mechanisms. Clearly, the mechanistic and evolutionary/ecological approaches would, and have (see Kurtz et al., 2002a), benefited from a formal synergism. One core aim of this review is to examine insect immunity from a perspective that is integrated with ecology and evolution in the belief that synergism will offer insight. In reviewing the mechanistic components of the insect immune system, we have moved away from the humoral/cellular dichotomy and instead organised defence mechanisms from the viewpoint of how individuals are organised (by selection) to defend themselves: We structure this review by examining how individuals avoid the negative effects of pathogens and parasites. The first line of defence is behavioural avoidance, the second is boundary defence. Immunity is the last line of defence, and represents a collective ‘emergency service’ that the organism calls on when the standing precautions and defences fail. This scheme is developed from Schmid-Hempel and Ebert’s (2003) ‘defence components model’ in which they seek to (and succeed in) reconciling two disparate evolutionary approaches to understanding how hosts and parasites coevolve (Fig. 1). Such a structure is biologically relevant because clearly individual organisms are the units of selection on which pathogens and parasites act: change in immune genetics is only one response to that complex selection pressure. It is important to bear in mind that insect immune systems are also under selection from sources other than pathogens and parasites since certain components and systems have additional functions (e.g. phenoloxidase (PO)). We do not

4


Probability

Overcoming avoidance

Penetration

Escape recognition

Successful infection

1

1

1

1

0.75

0.75

0.75

0.75

0.5

0.5

0.5

0.5

0.25

0.25

0.25

0.25

0

0 1

2

0 1

2

0 1

2

1

2

Parasite types

FIG. 1 The defence component model as proposed by Schmid-Hempel and Ebert (2003). It shows a hypothetical host–parasite/pathogen interaction with three consecutive steps from encounter to successful infection. The probabilities on the y-axis are the probabilities of the parasite overcoming host defence at each step. For the sake of simplicity, a multiplicative model is assumed, whereby the probabilities of all three steps are multiplied to estimate the probability of successful infection. The most specific components of the system determine the overall outcome of the infection, and hence the specificity to the parasite (geno-) types.

deal with the biochemical, genetic or structural details of immune system components in this review since these topics are covered by a host of recent reviews, which are cited in the relevant sections.

2

Defence via behaviour

Behavioural mechanisms directed against the parasites in an organism’s environment is necessarily the first line of defence (e.g. Hart, 1997). Behavioural mechanisms for removing parasites are well known and range from simple, individual-based behaviours like dust-bathing in birds, through more complex interactions that often serve important social functions, such as grooming in primate groups. Such behaviours have even evolved into the complex interspecific interactions that exist on tropical reef cleaning stations where one species makes a living by removing ectoparasites from other species (e.g. Cote and Molloy, 2003). In insects defence behaviours directed at pathogens and parasites tend to be less well studied, but there are several good examples of how ecology and behaviour are used to reduce the risk, and effects, of parasitism. Several insects use acoustic signals to attract mates and certain sarcophagid and tachinid fly parasitoids use these cues to locate their singing cricket hosts (see Cade and Wyatt, 1984; Cade, 1991). A male cricket has a range of singing options with two extremes: sing loudly and attract females and parasitoids rapidly, or do not sing and do not reproduce but avoid parasitoids. In nature, male crickets utilise one of the two tactics. The high-risk, high-gain tactic where they sing and mate with as many attracted females as possible before the parasitoid strikes, or the alternative, low-risk, low-gain tactic where they remain silent and loiter near a singing male. This tactic has a much reduced risk of parasitism, but provides an


5

opportunity to sequester females that are attracted to the singing male. A simple, and probably common, behavioural mechanism for avoiding parasites is to disperse away from aggregations or populations that are likely to have a prevalence of parasites (e.g. Bischoff, 2003; and see Kurtz et al., 2002b for the way immune systems respond to such behaviour). Another simple, but very effective, behavioural weapon in a host’s battle with parasites is thermoregulatory behaviour. The thermal optimum of an insect host will often not be the same as the thermal optimum of a parasite and selection has favoured behavioural thermoregulation that elevates the host’s core temperature to disadvantage the parasite (see Thomas and Blandford, 2003). This ‘behavioural fever’ has been shown to be quite subtle, often being directed at specific pathogens (e.g. Adamo, 1998). Behavioural avoidance of parasites and pathogens is an understudied area but there is good reason to suppose it is an effective response against selection from predictable, relatively high-cost parasites or pathogen insults. Since many insect parasites enter with the host’s meal, it is likely that foraging behaviour will be under selection to reduce parasite exposure (while selection on parasites probably favours entry with food since resource acquisition by hosts cannot be compromised). Theory suggests that selection for parasite avoidance may even promote the evolution of eusocial behaviour (O’Donnell, 1997). It is beyond the scope of this review to address behavioural avoidance in detail, but we will reconsider this aspect in Section 7 when we reexamine the defence components model of Schmid-Hempel and Ebert (2003) (see Fig. 1) in light of this review.

3

The insect immune system – boundary defence

The second line in an insect host’s defence against pathogens and parasites is the outer body covering. This consists predominantly of a toughened cuticle forming a protective integument over the insect’s external surface. Even in the midgut, the one place where the insect’s external surface is formed by a relatively delicate epithelium, there is still a protective cuticular membrane (the peritrophic membrane) forming a static defence against the outside world (Peters, 1992). Although the integument forms a formidable barrier to the outside world, there are potential weak points in the intact external surface that parasites and pathogens might be expected to target (see Fig. 2). Moreover, once the largely physical barrier of the cuticle is breached, the epidermis is the next line of defence. The epithelium is likely to be a rather ineffective physical barrier, and appears to be the site of expression of a number of key immune effector systems. It is an oversimplification to regard the interface between an individual insect and the outside world as an inert barrier; in reality, ‘boundary defence’ is a combination of inert physical barriers that have a limited immune capacity.

6


FIG. 2 A schematic figure of the insect’s barrier defences, showing the regions where invasion is likely. The dotted boxes reveal detail within the integument.

3.1

THE CUTICLE

It is an axiom in entomology that insect cuticle is the key to understanding the phenomenal success of this taxon. The cuticle is a tough, flexible and waterproof barrier against the outside world and is formed by the basal epidermal cells (Neville, 1975). The outer surface of insect cuticle probably harbours a diverse and abundant microbial community (see Brey et al., 1986) even in the most aseptic habitats. However, given the propensity of many insects to live in microbial-rich environments, it seems likely that opportunistic infections will accompany each integumental breach. Very little data currently exist to indicate the frequency of such wounding events in natural populations: an irony given that the launch pad for host–pathogen studies, and insect immunity in particular, was Pasteur’s seminal study (see Brey and Hultmark, 1997) showing that frequent cuticular wounding in Bombyx mori was responsible for the transmission of the silkworm plague. The main features of cuticle that make it a good barrier are its thickness and its strength. The latter is achieved by cross-linking proteins in the exocuticle via melanisation and sclerotisation (Neville, 1975), processes that share their core enzymes with the immune system (see Section 4.4.1). The first part of the cuticle that a potential pathogen/parasite encounters is the outer, complex, but usually very thin, epicuticle. This layer is unlikely to


7

provide physical protection (with the possible exception of scale insects (Homoptera: Coccoidea) where it often consists of thick, complex stacks of wax) but may harbour some, as yet unquantified, antimicrobial activity. The procuticle (combined exocuticle and endocuticle) forms the next barrier, mainly because of its thickness and architecture. The only entomopathogens, which invade directly through the exoskeleton are fungi (Charnley and St. Leger, 1991) and studies of their invasion mechanics suggest several aspects of cuticular architecture are important in resistance. Perhaps the best studied fungus in this respect is Metarhizium anisopliae, which invades by combining physical and enzymatic processes. Hajek and St. Leger (1994) suggest, and David (1967) supports, the notion that resistance to fungal entomopathogens resides mainly in cuticular thickness, the degree of cuticular cross-linking within the cuticular laminae (i.e. cuticular strength) and the degree of sclerotisation in the cuticle. Moreover, pore-canals (narrow trans-cuticular ducts that transport material to the epicuticular surface (Neville, 1975)) may also afford a path-of-low-resistance for the diffusion of enzymes released by the invading fungi (Zacharuk, 1970). As well as its obvious physical characteristics cuticle also provides an active biochemical barrier. Brey et al. (1993) showed that bacterial challenge to abraded cuticle resulted in antimicrobial activity in the vicinity of the abrasion. PO activity (an immune effector system responsible for producing melanin, see Section 4.4.1) has also been detected in insect cuticle (Ashida and Brey, 1995), although, whether it is there in a structural context, or to afford defence is unclear. Regardless of why it is in the cuticle, the activity of this enzyme is directed towards pathogens in the cuticle: fungal germ tubes are melanised as they pass through procuticle before they entered the haemocoel (Golkar et al., 1993). Schal et al. (1998) have shown that there is an active association between the haemolymph and the cuticle: compounds in the haemolymph, probably residing in oenocytes, are readily transported to the outer epicuticular surface. Although these compounds were not immunologically active, it seems likely that the existence of such transport mechanisms means immunologically active compounds could also be easily transported to the surface of the cuticle (although no evidence exists for such a phenomenon at the time of writing). The remarkable physical properties of cuticle combined with its ability to respond biochemically to pathogens means that it is an extremely effective barrier to infection, preventing or slowing down pathogen invasion (St. Leger, 1991). The one external surface of the insect which has no immediate cuticular covering is the midgut. Here, the insect must balance the need to absorb nutrients across an extended surface in an efficient manner with the need to defend itself from non-self (with which the midgut is replete). Here, again insects rely on cuticle. As food passes into the midgut it is sleeved with a chitinous, porous ‘peritrophic’ membrane, which is permeable to nutrients and

8


enzymes, but affords the delicate midgut epithelium some physical protection from damage as well as protection from invaders (Richards and Richards, 1977; Peters, 1992). 3.2

THE INTEGUMENTAL EPITHELIUM

Once the cuticular barrier is breached, the pathogen either encounters the underlying cuticular epithelium or the haemolymph. The cuticular epithelium associated with the external integument appears to become immunologically active mainly upon wounding (Brey et al., 1993; Meister et al., 1994). Recent studies of gene expression in epithelia associated with the respiratory, digestive and reproductive systems reveal much of it to be immunologically active (see Tzou et al., 2000). Whether this indicates constitutive immune function in these tissues, or is indicative of persistent challenge by opportunistic pathogens via these physically relatively weak lines of defence is unclear. What is clear is that the epithelium underlying the insect’s external surface is capable of immunological activity. Because of its spatial situation it seems likely that this tissue boundary between the physical defences and the haemolymph not only plays a part in barrier defence, but is also likely to release ‘early-warning’ signals that recruit and activate haemolymph-based effector systems. 3.3

SENSILLA

Insect contact chemosensilla have an opening in the cuticle at their tip that provides access to the sensillum lumen, wherein lie the sensory dendrites, while olfactory sensilla are covered with abundant small pores. These openings are typically 0.2 mm in largest diameter (Chapman, 1998) and so represent potential entry points for small microorganisms. However, little evidence exists to indicate that active chemical-mediated immunity occurs at these sites. Semiporous barriers are, however, placed in these openings. Viscous fluids and fibrillar cuticular plugs (Shields, 1994) erected across these access points presumably exclude microorganisms, while allowing the passage of the important biochemicals. The internal lumen of chemosensilla is in turn isolated from the haemocoel by a barrier of specialised epidermal cells (Chapman, 1998), providing a further impediment to invasion. 3.4

THE DIGESTIVE SYSTEM

The digestive system probably offers most invasion opportunities for pathogens since it is the least well-defended region physically, and is constantly exposed to non-self (i.e. food). The need to digest food (often with the aid of symbiotic gut microbes that need a favourable environment) and efficiently absorb nutrients (achieved across the cuticle-free midgut) results in the least


9

well-defended region of the insect’s body in terms of barriers. The foregut and hindgut are physically protected to some extent, being lined with a relatively thin layer of cuticle (Chapman, 1998). The midgut, and its associated structures, present an unprotected epithelial surface to the outside world, often providing parasites with specific points of entry into the host (e.g. Han et al., 2000). Physical protection in this vulnerable region is afforded by the delicate cuticular peritrophic membrane. The cuticle lining in the foregut provides more than just a passive inert barrier; however, it is sloughed off when pathogenic bacteria attached to it (Binnington, 1993), a reaction that presumably denies the pathogens a foothold and results in passage of the sloughed off material into the chemically hostile midgut. The midgut epithelium is an immunologically active tissue that produces a host of defence peptides, including defensins, Gram-negative binding protein, chitinase-like protein, serine proteases and lectin-like protein (Lehane et al., 1997; Barillas-Mury et al., 2000; Tzou et al., 2000) as well as NO (e.g. Hao et al., 2003) and PO (Wilson et al., 2001). There is also strong evidence that lysozyme-like activity occurs through the midgut and the caecae (Daffre et al., 1994) but it is unclear whether this activity functions in digestion, or provides protection against bacteria. We know that the caecae are immunologically active in Drosophila, where they produce diptericin (Tzou et al., 2000). Another tissue that is unprotected by cuticle and is specifically targeted by parasites (e.g. Fries et al., 2001; Weiser and Zizka, 2004) as well as opportunistic infections (e.g. Franco et al., 1997) are the malpighian tubules. These structures are immunologically active (Sagisaka et al., 2001; Bao et al., 2003) and produce a range of antimicrobial peptides in Drosophila (Tzou et al., 2000). 3.5

THE SPIRACLES AND RESPIRATORY SYSTEM

An important potential site of entry into an insect host is the spiracles and tracheal system. As well as entomopathic nematodes that invade through the spiracles (e.g. de Doucet et al., 1998) and/or live in the tracheal system (e.g. Aikawa and Togashi, 2000), this route is likely to be used by opportunistic bacteria and fungal spores. Despite the potential ease of entry via this route there is very little information about how insects avoid infection through it apart from studies showing that the epithelium associated with the trachea is immunologically active (Tzou et al., 2000), and that there are intimate spatial relationships between haemocytes and trachea (Wigglesworth, 1965) suggestive of a defensive role. 3.6

REPRODUCTIVE TRACT

Much of the reproductive tract that comes into frequent contact with the outside world has a cuticular lining: the female’s genital tract and sperm storage organs are lined with cuticle, which stops at the junction with the oviducts

10


(Chapman, 1998). Because copulation and insemination provide pathogens with an opportunity for horizontal transmission one would expect the reproductive tract to be immunologically active. Studies of Drosophila have shown gender-specific expression of antimicrobial peptides in the reproductive tract epithelium (Tzou et al., 2000) and it appears that males also incorporate antibacterial peptides in their seminal fluid (Lung et al., 2001), presumably to afford their genetic investment some protection while it is in storage in the female’s spermatheca. Recent observations of the interactions between males and females during mating suggest that males may deliberately wound female genitalia in order to delay female remating, and thereby enhance the wounding male’s fertilisation success (see Crudgington and Siva-Jothy, 2000). The damage caused to the cuticular lining of the female’s genital tract is relatively extensive and stimulates a wound-healing response culminating in the production of melanic plugs (see Crudgington and Siva-Jothy, 2000). Of interest in this respect is the hemipteran family, the Cimicidae, or bed bugs. Males of this taxon utilise traumatic insemination and introduce their intromittent organ through the female’s abdomen wall and inseminate into her haemolymph (Carayon, 1966). These insects live in unsanitary conditions and recent studies have shown that females pay a large fitness cost associated with the introduction of bacteria during traumatic insemination (Morrow and Arnqvist, 2003; Reinhardt et al., 2003). This selection pressure appears to be so strong that female cimicids have evolved a unique immune-organ (the mesospermalege) in the region where males pierce and inseminate them (see Reinhardt et al., 2003).

4

The insect immune system – haemocoelic defence

Once an invader has breached the barrier defences, the insect has to produce a rapid and effective response that localises and neutralises the reproductive capacity of the pathogen or the growth potential of the parasite. This is best achieved by killing the invader (but see Sasaki and Godfray (1999) and Boots and Bowers (2004) for models that predict that hosts should produce no immune response to pathogens under certain, ecologically realistic, conditions). Insects rely solely on innate immune effector systems. Boman (1998) described these processes as ‘insect germ-line encoded anti-infection responses.’ This distinguishes these innate responses from the sophisticated immunity of vertebrates afforded them by the immunoglobulin gene superfamily. Insect immune systems have no specific immunoglobulin-based memory and are traditionally viewed as being relatively indiscriminatory when confronted with subtly different types of non-self. However, recent work on Crustacea suggests that invertebrate innate systems are capable of some remarkably specific immunological phenomena (Kurtz and Franz, 2003; Little et al., 2003) (see Sections 6.1 and 6.2). ‘Simple’ is often misinterpreted as evolutionarily


11

inferior to ‘complex,’ a viewpoint that falsely equates phylogenetic basality with functional inferiority. Insects are the most successful class of organism on the planet (see, for example, Gullan and Cranston, 2000) and their ‘simple’ immune system plays an important role in that success. The open haemocoel (cf. the closed circulatory networks of higher vertebrates) provides some advantages in terms of the function of the immune system. For example, mediators, effector systems and haemocytes can be more rapidly disseminated and organised. However, the open architecture also presents a problem when the insect is faced with systemic immune insult. An open body cavity facilitates rapid movement of infective agents through the host. Consequently, selection should favour the evolution of effector systems that rapidly and efficiently localise and neutralise invaders. One could argue that these needs render any acquired, or acquired-like, immune response pointless, since such responses are also characterised by their relatively slow response time. The following section is organised, where possible, according to the chain of temporal and organisational events that follow a systemic immune insult (i.e. a breach in the barrier defences). 4.1

CLOTTING AND WOUND CLOSURE

A septic wound presents a series of major physiological problems that must be dealt with rapidly. The major priority, particularly in holometabolous larvae where the haemolymph is under pressure, will be to plug the wound. This will prevent excessive blood loss and close the invasion route behind the outflowing haemolymph (which will, to some extent, flush out invaders in a hostile medium). The main reaction to blood loss is clotting, which also functions to immobilise, localise and begin neutralising pathogens that have entered via the wound. Clotting has been extensively studied in Crustacea (see Theopold et al., 2004 for review) where the reaction is triggered by pathogen-associated motifs (the so-called PAMPs – pathogen-associated molecular patterns) like lipopolysaccharide, peptidoglycans and b-1,3 glucan. Non-biotic stimuli for clotting probably also exist since clotting has non-defence roles as well (i.e. wound closure). The first physical changes that occur during clotting are an increased viscosity of the haemolymph and the inclusion of insoluble, glycosylated ‘sticky’ fibres which, in Drosophila, contain several clotting proteins including hemolectin and tiggrin (Scherfer et al., 2004). The production of these fibres, which adhere to each other and form a sticky net, begins to seal the wound, trap microbes and trap haemocytes (Gregoire, 1974), some of which are responsible for secreting the material that forms the fibres (e.g. Goto et al., 2003). Haemocytes are also attracted to/remain in the vicinity of the wound because the damaged epithelial cells near the wound release hemokinin (Cherbas, 1973), a compound that aids cell aggregation. PO usually becomes activated during

12


wound closure, particularly in the later stages once the ‘soft clot’ is established. Although this enzyme cascade is probably not involved directly in coagulation (Scherfer et al., 2004 but see also Li et al., 2002 and Cerenius and So¨derha¨ll, 2004 for an alternative perspective), it will kill invaders as well as melanise the material that constitutes the clot (Ra¨met et al., 2002a) thereby reestablishing an impermeable physical barrier. 4.2

SELF/NON-SELF RECOGNITION

The insect immune system recognises a range of non-self motifs, from wellcharacterised pathogen cell surface molecules including peptidoglycans, b-1,3 glucans, lipopolysaccharides and other sugar moieties (Theopold et al., 1999), collectively referred to as PAMPs. Insect hosts need to avoid reacting to self in the absence of immune challenge but, upon septic insult, must target non-self, and sometimes specific components of self, in order to neutralise the insult. For example, the haemocytes that encapsulate and isolate larger immune insults die (apoptose) and are melanised by PO, probably in the same way haemocytes in the vicinity of cuticular wounds die and are melanised during wound repair. Such reactions are probably mediated by signals of ‘altered self’ (e.g. Franc et al., 1999). The insect system’s ability to separate different types of non-self from each other will be relatively restricted (cf. vertebrates) because they lack immune-functional immunoglobulin superfamily proteins. Despite this, however, recent studies of invertebrate immunity in an ecological context show that invertebrate innate systems are capable of some remarkable feats of recognition (Kurtz and Franz, 2003; Little et al., 2003) (see Section 6.1). Exactly how this discriminatory capacity arises is currently far from clear. The organs and tissues in the haemocoel (with the exception of haemocytes) are covered by the basal lamina, or basement membrane. It is believed that the basal lamina is produced by haemocytes (Wigglesworth, 1956, Ball et al., 1987) because (a) there is an intimate association between haemocytes and the basal lamina during morphogenesis and rebuilding (Wigglesworth, 1973; Nardi and Miklasz, 1989; Nardi et al., 2001, 2003), (b) haemocytes are recruited to areas of basal lamina disruption during wounding (e.g. Lackie, 1988) and (c) haemocytes and basal lamina share immunogenic epitopes (Chain et al., 1992). One important function of the basal lamina in the context of self/non-self recognition may be to provide a uniform background signal of ‘self’ within the haemocoel, against which any non-self signal becomes more conspicuous. This notion is supported by the observation that termination of the encapsulation response (i.e. a coordinated haemocytic response to large non-self (see Section 4.6)) occurs when a basement membrane-like layer appears on the outside of the encapsulating, dead and melanised haemocytes (e.g. Liu et al., 1998; Pech and Strand, 2000).


13

Against this background of self, insects appear to distinguish non-self by relying largely on a host of pattern recognition peptides (see Janeway, 1989) that usually identify PAMPs. We concentrate our overview on two relatively well-studied groups of these pattern recognition peptides: the peptidoglycanrecognition proteins and the Gram-negative binding peptides. Peptidoglycan-recognition proteins are relatively conserved molecules that bind to peptidoglycans (a conserved, essential and unique component of the microbial surface) and thereby sense an infection (Dziarski, 2004). Drosophila has 12 peptidoglycan-recognition protein genes (Werner et al., 2003), not all the products of which function in alerting the immune system to the presence of invaders (see Mellroth et al., 2003). Insect peptidoglycan-recognition proteins have four main identified functions in terms of the immune effector systems they activate when bound to bacteria. (1) They activate the prophenoloxidase cascade (Yoshida et al., 1996; Kang et al., 1998; Takehana et al., 2002) by activating serine proteases. (2) They stimulate antimicrobial peptide production via the Toll and Imd pathways (see Gottar et al., 2002; Royet, 2004). (3) They appear to activate phagocytosis in some haemocytes (Ra¨met et al., 2002b), and (4) some peptidoglycan-recognition proteins seem to function to remove, or ‘clean up’, excess peptidoglycans in the haemocoel (Mellroth et al., 2003). Some peptidoglycan-recognition proteins are transmembrane proteins, the best studied of which is PGRP–LC. Mutants of PGRP–LC fail to respond to G, but not G+ bacteria (Choe et al., 2002; Ra¨met et al., 2002b). These phenomena are curious for two reasons. First, the peptidoglycans in G bacteria are concealed beneath the outer cell wall (e.g. Doyle and Dziarski, 2001). Second, G bacteria have a lipopolysaccharide-rich outer coating (G+ bacteria have no lipopolysaccharide). Since lipopolysaccharide is highly immunogenic, it is counter-intuitive that part of the recognition system that distinguishes G from G+ bacteria operates by detecting the concealed PAMP (see Leulier et al., 2003). The response, in Drosophila, of detecting G bacteria with PGRP–LC is the production and secretion of the potent antibacterial peptide diptericin, a member of the gloverin family of antimicrobial peptides (Bulet et al., 1999). Another well-characterised peptidoglycan-recognition protein is PGRP–SA, a soluble protein that has a high affinity for G+ bacteria in Drosophila. PGRP–SA mutants are unable to secrete drosomycin, a potent antifungal peptide, and do not respond to G+ infections, although they can clear fungal and G infections easily (Michel et al., 2001). The second important class of molecules that detect non-self are the Gramnegative binding peptides. As their name suggests, Gram-negative binding peptides detect and bind to G bacteria, principally targeting the lipopolysaccharide-rich and b-1,3 glucans component of the cell wall, resulting in the production of the potent antimicrobial peptides drosomycin, attacin and cecropin in Drosophila (Kim et al., 2000).

14


Other potential pattern recognition peptides are some thioester-containing peptides (TEPs) (which bear similarities to the vertebrate complement component C3) and hemolin (an immunoglobulin superfamily protein, Lanz Mendoza and Faye, 1999). An insect thioester-containing peptide (probably secreted in the fat-body, Lagueux et al., 2000) has been shown to act as an opsonin, promoting phagocytosis of G bacteria (Levashina et al., 2001) and suggesting it may have pattern recognition abilities. Studies of Manduca sexta hemolin have shown it to be an immune surveillance protein (Kanost et al., 2004) expressed in the gut of diapausing moths. It presumably affords them immune protection during this vulnerable life history stage (Lee et al., 2002). Hemolin also plays vital roles in development (e.g. Bettencourt et al., 2000, 2002) where precise mechanisms coordinating cell–cell recognition and interaction are as important as they are in immunity. This observation emphasises the important point that immune effector systems can be influenced by selection on other functions because of the tendency of these effector systems to be multifunctional. 4.3

SIGNAL TRANSDUCTION

Once non-self has been identified and signalled by conformational change in the detection molecules, the signal needs to be translated into an appropriate biological action (transduction). Soluble, humoral-based, signal transducers are responsible, among other things, for triggering the fast-acting ‘constitutive’ immune responses, the most important of which is prophenoloxidase (see Gorman and Paskewitz, 2001). The best understood of the humoral transducers are the serine proteases, a group of enzymes that mediate a range of physiological functions (Rawlings and Barrett, 1994). Immunologically functional serine protease proenzymes are activated by conformational changes in pattern recognition molecules (see above): the active serine protease then cleaves proenzymes in other controlling cascades (by targeting peptide bonds with a catalytic serine-containing domain). However, serine proteases (and other transducers such as ‘Persephone’, Ligoxygakis et al., 2002a,b) also activate the cell-based signal transduction pathways (see below) in response to microbial infection (see Hultmark, 2003) and so act as intermediaries for the activation of the slower responding ‘inducible’ defences as well. At the core of the insect immune response (Hultmark, 2003) are two cellbased signal transduction pathways referred to by the name of the transmembrane proteins that mediate them: Toll and Imd (Fig. 3). The biochemical details of these pathways have recently been reviewed (see Hultmark, 2003; Hoffmann, 2003): we will summarise the generic aspects of these pathways. Toll’s only known ligand is the protein Spa¨tzle but, because null Spa¨tzle mutants are less impaired at responding to microbial insult than are Toll mutants (Lemaıˆ tre et al., 1996), there are likely to be other Toll ligands. Cleavage


15

of Spa¨tzle by serine proteases (which were activated by certain G+ bacteria and/or fungi) in the haemolymph activates the Toll pathway (see Gobert et al., 2003; Weber et al., 2003 for details), resulting in the synthesis and secretion of the potent antifungal peptide drosomycin (Lemaıˆ tre et al., 1997) and the activation of haemocytes (Qiu et al., 1998). The Imd pathway is activated by G bacteria and/or fungi and is probably the principle regulator of inducible antimicrobial peptides directed at G bacteria and fungi (see Hultmark, 2003). Stimulation of the Imd pathway in Drosophila results in the synthesis and secretion of drosomycin, cecropin and diptericin. Stimulation of Imd also switches on the downstream JNK pathway (Sluss et al., 1996), a mitogen-activated protein kinase that forms the ‘frontend’ of vertebrate immune signalling pathways. In Drosphila this JNK activation results in the expression of cytoskeletal genes (Boutros et al., 2002) that are probably involved in wound healing (Ra¨met et al., 2002a).

Diptericin

G+ Fungi

Serine protease

Cecropin

Persephone

Drosomycin

Drosomycin Spätzle

GToll PGRP-LC

Imd

Nucleus Cactus/ Dif Relish JNK Cytoskeleton change

FIG. 3 A simplified schematic of the activation of the Toll and Imd pathways in Drosophila immunity.

16

4.4 4.4.1

MICHAEL T. SIVA-JOTHY ET AL EFFECTOR SYSTEMS

–

ENZYME CASCADES AND CYTOTOXINS

Phenoloxidase

Perhaps the most important constitutive immune effector system in insects is the tyrosinase (Chase et al., 2000) PO. This enzyme catalyses the initial steps in the production of the biopolymers melanin and sclerotin (see Sugumaran, 2002). As well as its core role in immunity, it also plays an ontogenetic role by the iterative production of melanin and sclerotin during exocuticle manufacture after ecdysis (Neville, 1975) as well as reproductive roles during the production of species-specific visual signals (e.g. True et al., 1999; Siva-Jothy, 2000). The production of PO from its inactive precursor PO is triggered via a serine protease cascade initiated by the detection of PAMPs (see So¨derha¨ll and Cerenius, 1998; Cerenius and So¨derha¨ll, 2004 for reviews). Immunological PO activity produces melanin, which is used to form one of two types of capsule around a pathogen. Cell-free, inert, melanotic capsules are found in a range of insects, including bumblebees (e.g. Allander and Schmid-Hempel, 2000) and mosquitoes (e.g. Gorman and Paskewitz, 2001). The second type of melanic capsule occurs in immune responses where haemocytes smother the invader and phenoloxidase activity melanises the resultant cell mass, forming a melanised cell-mass (e.g. Lackie et al., 1985). In both cases, the insect ‘externalises’ the intruder behind an inert and impermeable barrier. Interestingly, despite (a) the observation that PO activity is correlated with pathogen death and isolation and (b) the incredibly detailed dissection of the molecular mechanisms releasing and regulating the prophenoloxidase cascade (So¨derha¨ll and Cerenius, 1998; Cerenius and So¨derha¨ll, 2004), it remains to be empirically demonstrated how PO activity deals the fatal blow to the pathogen. It seems most likely that certain products of PO activity (quinones, phenols and reactive oxygen species) are utilised for their cytotoxic effects and that the consequently moribund (or dead) pathogen is finally smothered and externalised in the melanised capsule: if it is still alive the barrier will deprive it of oxygen and nutrients. An exciting recent development in the study of PO is a growing body of evidence that the cascade is not just a fast-acting blunt instrument of defence. It appears to be subtly integrated into other mechanisms, with suggestions that the ‘cross talk’ aids clotting (Li et al., 2002) and microbial peptide synthesis (Braun et al., 1998). 4.4.2

Nitric oxide

Nitric oxide is a soluble, highly reactive gas synthesised within cells by the enzyme nitric oxide synthase (NOS). Nitric oxide’s cytotoxic activity arises from its ability to combine with superoxide radicals and produce highly reactive peroxynitrite groups that are particularly effective at oxidising lipids.


17

Insect haemocytes are capable of generating nitric oxide in response to immune insult (Weiske and Wiesner, 1999) and NOS activity has been identified in the midgut epithelium of mosquitoes (Luckhart et al., 1998) and the cardiac valve (the junction between the foregut and midgut) of Tsetse flies (Hao et al., 2003). In both cases, NOS plays a defensive role, reducing the ability of the parasite to move though the NOS-active tissue and so gain access to the host. 4.4.3

Reactive oxygen species

The ‘respiratory burst’ is an NADPH oxidase-driven conversion of oxygen into the so-called ‘reactive oxygen species’, a group of highly reactive oxygen radicals. Reactive oxygen species are highly cytotoxic and have recently been identified in the haemolymph of immune-challenged insects (e.g. Whitten and Ratcliffe, 1999; Dettloff et al., 2001; Glupov et al., 2001). Because these radicals are unstable, and therefore very transitory, mechanistic detail about their production and control is lacking. It seems likely, however, that they are produced by haemocytes in the vicinity of microbial insult, and their synthesis and release are very tightly controlled in order to limit auto-reactive damage. 4.5

EFFECTOR SYSTEMS

–

ANTIMICROBIAL PEPTIDES

In comparison with other humoral immune effector systems, antimicrobial peptides are highly specific in their effects. However, that specificity comes with the cost of slow responsiveness (see Section 5.3). Drosophila shows three main structural groups of these peptides in seven distinct families (reviewed in Bulet et al., 1999). They are mostly relatively small, often membrane-bound, and are extremely effective at neutralising G+ bacteria, G bacteria and fungi in the haemolymph. They are synthesised mainly via signals transmitted through the Imd pathway (but the Toll pathway is also important) and are manufactured in the fat body (Hoffman and Reichhart, 2002), in haemocytes (Lowenberger, 2001) and in the epithelium (e.g. Tzou et al., 2000). Although these peptides are produced in quantity after microbial insult, and details of the pathways leading to the synthesis of these peptides are being rapidly uncovered (see Hoffmann, 2003 for review), we have little understanding of their coordinated role and mode of action. In a similar vein, almost nothing is known about insect immunological defences against viruses. There is some evidence that hosts produce proteins that interfere with viral replication (Wyers et al., 1995) , and one would additionally assume that boundary defences (Section 3) are critical. 4.6

EFFECTOR SYSTEMS

–

HAEMOCYTES

Our understanding of the insect haemopoietic system (see Lavine and Strand, 2002 for review) is derived largely from classifications based on haemocyte

18


morphology and/or behaviour. Almost nothing is known about the cell surface or genetic markers that control (and so define) haemocyte function (but see Chain et al., 1992; Mullet et al., 1993; Willot et al., 1994; Strand and Johnson, 1996; Lebestky et al., 2000). Recent studies by Hou et al. (2002) and Sorrentino et al. (2002, 2004) are, however, providing real insight into functional polymorphisms in haemocytes in relation to coordinated cellular responses to insult. Despite this, haemocyte lineage relationships and haemopoiesis remain, for the moment at least, poorly understood phenomena. One major drawback with the historical use of haemocyte morphology to define function types has been the abundance of morphology based names derived from studies of different species describing what are probably only a few distinguishable morphotypes. This problem is exacerbated by the fact that these studies are often conducted on monolayer preparations and the distinction between morphotypes made on qualitative criteria. For example, one commonly used division is the separation of immunologically active haemocytes into ‘granulocytes’ and ‘plasmatocytes’. The former, as their name suggests, contain granules or vesicles, the latter, by definition, do not, but additionally show spreading behaviour when in contact with a foreign surface. Most insect immunologists would accept this distinction. However, studies using quantitative techniques such as flow cytometry, where the size and granularity of thousands of haemocytes are measured, reveal a single population of haemocytes showing degrees of granularity (e.g. Chain et al., 1992). This is not to say that haemocytes do not have discrete functions in a coordinated immune response, rather that the morphological approach has some severe limitations in its ability to resolve those functions. This caveat being acknowledged, however, we can make certain generalisations about haemocytes on the basis of their morphology and behaviour. Small, cytoplasm-deficient cells are usually termed ‘prohaemocytes’ and are widely believed to differentiate into other haemocyte types. The best evidence for this function comes from studies showing that they (and other morphotypes) undergo mitosis (e.g. Gardiner and Strand, 2000) and from in vitro studies, which suggest that they differentiate into other haemocyte types (e.g. Yamashita and Iwabuchi, 2001). Haemocytes that contain granules and vesicles are usually referred to as ‘granulocytes’ but their behaviour varies across insect taxa (reinforcing the notion that morphology is not a good basis for inferring function). They are capable of phagocytosis in some insect species (Lavine and Strand, 2002) and are the first haemocytes to form an encapsulation response to a large invader (Pech and Strand, 1996) – quickly followed by other cell types. Pech and Strand (1996, 2000) show that the granulocytes attach and then apoptose (undergo programmed cell death). In this context, it is interesting that Chain et al. (1992) identified an epitope stuck to the contact surface that appeared to be released from blebbed granulocytes (Fig. 2d in Chain et al., 1992). The fact that they contain granules suggests granulocytes are involved in the synthesis and


19

storage of bioactive compounds involved in immunity, but little information exists about the details of this role and their coordination in an immune response. The other easily identifiable haemocyte morphotype is the plasmatocyte, a cell characterised by its propensity to spread on contact with a foreign surface. Plasmatocytes are capable of phagocytosis (e.g. Elrod-Erickson et al., 2000) and are the mainstay of the coordinated cellular attacks directed at haemocoelic intruders discussed below (see also Lackie et al., 1985). These cell types, and others (see Lavine and Strand, 2002), are not only intimately involved in the manufacture and secretion of many of the compounds already discussed, but also coordinate a number of distinct responses to septic insult. As with other components of the immune system, it is important to bear in mind that haemocytes have vital tasks other than defence: they are intimately involved with rebuilding during metamorphosis (e.g. Wigglesworth, 1965), cuticle manufacture (Sass et al., 1994) and basement membrane formation (Ball et al., 1987), among other things. Phagocytosis occurs when a haemocyte encounters and recognises a small (i.e. smaller than itself) pathogen. The pathogen is engulfed by the cell and is killed. Mammalian phagocytes kill the pathogens they engulf with nitric oxide (e.g. Nathan and Hibbs, 1991) and reactive oxygen species (e.g. Robinson and Badwey, 1994). Both of these effector systems have been associated with immune-challenged insect haemocytes (see Section 4.4) and probably perform a similar function. Phagocytosis can be promoted by certain cytokines (a thioester-containing peptide identified by Levashina et al., 2001 promotes phagocytosis of G bacteria), suggesting that pathogen-naı¨ ve haemocytes can be ‘switched on’ during an insult, presumably making haemocoel clearance more rapid. Targets that are too large for a single haemocyte to phagocytose (e.g. clusters of localised microbes) are smothered by layers of haemocytes which become melanised, a process known as nodule formation (because of the small, dark ‘nodules’ that appear within the haemocoel). Nodule formation requires that haemocytes must not only ‘recognise’ that phagocytosis is not an option, but must instead become adhesive and spread over the target as well as one another. The processes that control and regulate nodule formation are probably mediated by cytokines and cell adhesion molecules, and although nodule formation has not been mechanistically dissected, a number of candidate cytokines are known. The most potent insect cytokine known is plasmatocyte spreading peptide (Clark et al., 1997), a peptide which causes plasmatocytes to spread and externalise adhesion molecules (Strand and Clark, 1999). Plasmatocyte spreading peptide homologues have been identified from a number of Lepidoptera and some have similar activity (Wang et al., 1999; Strand et al., 2000). A Drosophila protein, peroxidasin, stimulates haemocyte adhesion and spreading (Nelson et al., 1994) and is similar to the well-characterised crustacean cell adhesion molecule peroxinectin (Johansson, 1999). Another

20


important aspect of nodule formation is the aggregation of haemocytes, a phenomenon stimulated by the soluble lectin-like protein hemocytin (Kotani et al., 1995). Once the haemocytes have been stimulated to stick and spread, and have formed a covering over the agglutinated microbes, the nodular cell mass is usually melanised by the activity of PO. The control of this process in the context of melanising self is a mystery, but the consequences are that the pathogens are neutralised and externalised. Once formed the nodules remain in the insect until it dies. Insects that are subjected to parasitoid attack face relatively large intruders in their haemolymph which need to be neutralised. In the case of parasitoid attack the consequences of failure to neutralise the parasite is host death, so the selection pressure for an effective response is strong. The response to large intruders is termed ‘encapsulation’ since the phenotypic consequence is usually spectacular (Fig. 4) and often visible without dissecting the insect. However, there is probably very little, if any, difference between cellular encapsulation and nodule formation apart from the scale of the process. Not surprisingly, most studies of insect cellular immunity focus on encapsulation because of its amenability to study. Haemocytes still need to identify the target, stick and spread and recruit other haemocytes to the task (see Lackie et al., 1985). Integrins, a class of vertebrate cell adhesion molecules, are probably involved in encapsulation (Pech et al., 1995; Lavine and Strand, 2001) and are expressed on the surface of haemocytes attached to a foreign surface (Nardi et al., 2003). Lavine and Strand’s (2003) recent work shows that at least one integrin plays an important role in regulating haemocyte adhesion during encapsulation. Encapsulation is a mainstay of ecological immunity assays because it is easy to measure the single phenotypic outcome of the coordination of several

FIG. 4 The cellular encapsulation response directed at a 1 mm length of nylon monofilament implanted in the haemocoel of an adult Tenebrio molitor beetle for 20 h. Slight melanisation is apparent over the surface of the encapsulating haemocyte mass covering the nylon. There are easily distinguishable areas where the melanisation is more intense.


21

branches of an individual’s immune system. Studies have revealed that the magnitude/speed of encapsulation is correlated with an individual’s haemocyte load (the total number of haemocytes in the haemocoel) and that haemocyte load is a variable that responds to selection from parasitoids (e.g. Kraaijeveld and Godfray, 1997; Kraaijeveld et al., 2001). As mentioned above (see Section 4.4.1), encapsulation does not always involve a cellular reaction to non-self: some insects produce a cell-free, homogeneous melanic capsule around the intruder. However, as with nodule formation, when there is a cellular response it is finally melanised, forming a dark, impermeable barrier around the insult, which remains in the host until death. Wound repair bears a lot of physical similarities to encapsulation: haemocytes are recruited to the critical site, become adherent and form a mass, which is eventually melanised. The process probably shares pathways and processes with encapsulation and nodule formation. However, wound healing is usually much more rapid than encapsulation (pers. obs.) and will have a more intimate association with clotting. Moreover, there appear to be distinct peptides associated with wound healing. Paralytic peptide 1, isolated from M. sexta, has been shown to speed up the cellular component of wound healing (Wang et al., 1999) while hemokinin is released by damaged cuticular epithelial cells and induces haemocytes to aggregate at the wound site (Cherbas, 1973).

5

Ecological immunology and variation in immune defence

Insect immunity was the exclusive domain of immunologists seeking to understand the mechanistic basis of immune effector systems. However, the last decade has seen the concepts of population biology, ecology and evolutionary biology combine with immunity to produce ‘ecological immunology’, one of the fastest growing fields of evolutionary ecology (Sheldon and Verhulst, 1996; Rolff and Siva-Jothy, 2003). This field of research examines how and why micro-evolutionary processes generate, and maintain, variation in immune effector systems and the coordinated host response to pathogens (SchmidHempel, 2003). Evolutionary ecologists based their reasoning on two main theoretical approaches. The first approach relies on the theory of the evolution of life history traits (Roff, 1992; Stearns, 1992) and assumes that the evolution and the use of immune defences are costly (Sheldon and Verhulst, 1996) (see Section 5.1). The second approach is based on arms-race models of coevolution (Van Valen, 1973), which propose that coevolution between hosts and parasites can lead to sustained oscillations in host genotype frequencies through negative frequency-dependent selection, favouring rare host genotypes (Haldane, 1949; Hamilton, 1980; Frank, 1991, 1993; Thompson and Lymbery, 1996: Peters and Lively, 1999, see Section 5.2). Since one of the biggest problems in combining two, or more, established areas of research is the loss of information through

22


language differences, we include a table of definitions of frequently used evolutionary terms (see Table 1). 5.1

LIFE HISTORY THEORY AND THE COSTS OF IMMUNE DEFENCE

How does variation in life history traits (the major features of an organism’s life cycle that determine fitness, e.g. size at birth, age at maturity, age-specific fecundity, survival rate) translate into variation in fitness among individuals? To examine this question, life history theory assumes the existence of trade-offs between traits that constrain the simultaneous evolution of two or more traits (Roff, 1992; Stearns, 1992). Immune defence should be viewed in this context since if immune defence only provided resistance to pathogens and parasites with no cost, then natural selection would have favoured universally perfect immunity. Since this is not the case (i.e. susceptibility persists), immune defence is probably costly and so is traded off against the need for investment in other important fitness traits: selection will favour individuals with an optimal balance between immune defence and other fitness traits. Several kinds of immune defence costs can be distinguished (Schmid-Hempel, 2003).

TABLE 1 Definitions of evolutionary terminology Terms used in the text Antagonistic pleiotropy Coevolution Frequencydependent selection Evolutionary trade-off Fitness Life history trait Physiological trade-off Adaptation (adaptive)

Definition The genetic correlation between traits is such that selection on one trait is opposed by the consequent selection on the second trait The joint evolution of two or more interacting species, each of which evolves in response to selection imposed by the other The fitness of a phenotype or a genotype varies with the phenotypic or genotypic constitution of the population Occurs when selection on one trait decreases the value of a second trait, i.e. a negative genetic correlation The average contribution of an allele or genotype to the next generation. Usually, only correlates can be measured A trait that is directly connected with fitness, such as development time, fecundity and viability Two or more traits compete for resources within a single organism A process of genetic change, owing to selection, whereby the average state of a character becomes improved with reference to a specific function

Reference Roff (1997) Futuyma (1998) Roff (1997) Stearns (1992) Futuyma (1998) Roff (1997) Stearns (1992) Futuyma (1998)


5.1.1

23

Evolutionary cost of immune defence

The evolutionary cost of immune defence relies on negative genetic covariance between a component of the immune system and another fitness-relevant trait of the organism or even another component of the immune system (Stearns, 1992). This phenomenon is assumed to result from antagonistic pleiotropy, where a gene that has a positive effect on one component of fitness (i.e. immune defence) has a negative effect on another. These genetic relationships between traits cannot be changed during the lifetime of the organism. Therefore, high expression of immune defence may negatively affect fitness by constraining other correlated fitness traits, especially in the absence of parasites or pathogens. These genetic trade-offs between immune defence and other fitness parameters are usually investigated through quantitative genetic estimation of trait covariance and selection experiments (see Table 2). Studies manipulate variation in host immune defence and then observe the correlated response in other important traits. For instance, Kraaijeveld and Godfray (1997) selected replicate lines of Drosophila melanogaster for increased resistance to the parasitoid wasp Asobara tabida and measured the correlated response on other important fitness parameters ranging from egg viability to female fecundity. Encapsulation ability was increased by 55% in five generations in their selection experiment. Compared to control lines, the resistant-selected lines were characterised by a twofold increase in circulating haemocytes and a reduced competitive feeding ability of the larvae under crowding. In contrast to this direct approach, other studies have selected for change in host traits and measured the corresponding change in immune defence. This ‘indirect’ approach was used by Koella and Boe¨te (2002) who selected lines of the mosquito Aedes aegypti for earlier or later age at pupation. They measured the extent to which selection changed the mosquito’s ability to encapsulate and melanise Sephadex beads. The authors obtained mosquito lines with early and late age at pupation and found that encapsulation ability, as well as adult body size, were positively correlated with age at pupation. The evolutionary cost of immune defence is assumed to affect the dynamic of resistant and susceptible genotypes in a host population according to parasite prevalence. Resistant host genotypes should only be maintained when parasites are abundant. Yan and Severson (2003) tested this assumption using the mosquito A. aegypti and the malaria parasite Plasmodium gallinaceum. The authors created experimental mosquito populations by mixing susceptible and resistant strains in equal proportions and then determined the dynamics of markers linked to loci for Plasmodium resistance and other unlinked neutral markers over 12 generations. They found that when a mixed population was maintained under parasite-free conditions, the frequencies of alleles specific to the susceptible strain at markers closely linked to the loci for resistance (QTL markers), as well as other unlinked markers, increased in the first generation and then fluctuated around equilibrium frequencies for all of those markers.

24

TABLE 2 Examples of studies of the cost associated with the evolution of immune defence Insect species

Selective regimes

Effects

Rothenbuhler and Thompson (1956) Sutter et al. (1968) Boots and Begon (1993) Yan et al. (1997) Ferdig et al. (1993) Kraajeveld and Godfray (1997) Fellowes et al. (1998)


(a) Select for increased resistance and measure corresponding changes in other traits Honeybee (A. mellifera) Increased resistance to bacterial Higher larval mortality disease Honeybee (A. mellifera) Increased resistance to bacterial Slower larval growth disease Indian meal moth (Plodia Selection for increased resistance Slower development, lower egg interpunctella) to granulosis virus viability, but increased pupal mass Mosquito (A. aegypti) Selection for increased resistance Decreased adult body size, to malaria parasite fecundity and longevity Mosquito (A. aegypti) Increased resistance to nematode Reduced reproductive success infections Fruitfly (D. melanogaster) Increased encapsulation to larval Reduced competitive ability parasitoids (Asobara tabida) Lower survival rate of larvae Fruit fly (D. melanogaster) Increased encapsulation to virulent larval parasitoids (Leptopilina boulardi)

References

in immune defence Correlated reduction of PO activity Earlier reproduction correlates with lower encapsulation response, the opposite for later reproduction reduced reproductive success

(c) Experimental competition between resistant and susceptible genotypes In parasite-free conditions Mosquito (A. aegypti) Mixing plasmodium-susceptible frequencies of susceptible alleles and resistant mosquito increased and under parasite populations in equal proportion exposure allele frequencies did and comparing frequencies of not change resistance and susceptible alleles after 12 generations under parasite-free or parasiteexposure conditions

Hosken (2001) Koella and Boe¨te (2002)

Yan and Severson (2003)


(b) Select for change in host trait(s) and measure corresponding change Dung fly (Scatophaga Selection for polyandry leading stercoraria) to larger reproductive organs Mosquitoes (A. aegypti) Selection for earlier or later age at pupation (i.e. age at reproduction)

25

26


Conversely, when the mixed population of mosquitoes was exposed to an infected blood meal every generation, allele frequencies at the QTL markers for resistance were not significantly changed. In other words, resistant genotypes are competitive only under parasite pressure. When parasite pressure was removed, resistant genotypes suffered from a lower competitive ability. 5.1.2

Physiological cost of immune defence

The physiological costs of immune defence results from resource-based tradeoffs between the immune system and other important functions. Assuming that the different functions of an organism compete for the same pool of resource, the allocation of resource to the immune system is expected to constrain other functions that are sustained simultaneously and vice versa. These resource costs have two components (Schmid-Hempel, 2003). First, the cost of maintenance of the immune system corresponds to the cost of keeping the machinery at a given level of readiness; second, the cost of using the immune system when responding to a challenge. The magnitude of resource trade-offs in the maintenance of the immune system is determined by constraints that result from the evolved physiology. However, maintaining immune defence is still a plastic trait (see Section 6.4) that shows variation influenced by individual decision. For instance, in the armyworm Spodoptera exempta, the basic level of PO activity in the cuticle, haemolymph and midgut is upregulated at high population density (Wilson et al., 2001). Similarly, mating activity is known to lead to non-resource-dependent downregulation of the immune function (Siva-Jothy et al., 1998; Rolff and Siva-Jothy, 2002). Measuring the resource cost of maintenance of the immune function is difficult, as many regulatory processes may interfere with it. For example, immune-depression under food stress (or an increase in other demanding activities) may reflect the occurrence of physiological regulation avoiding self-damage rather than a resource-based trade-off. As long as the regulatory mechanisms between functions are unknown, measures of the cost associated with the maintenance of the immune system will remain difficult to quantify. Unlike the cost of maintenance, the cost associated with producing an immune response is relatively easy to measure and has been the target of several studies (Table 3). An immune response is assumed to use up part of an organism’s energy budget. Demonstration of this cost consists of challenging a host immunologically and measuring the corresponding changes in other traits (including immune defence) compared to controls. For example, mosquitoes (Armigeres subalbatus), which have encapsulated micro-filarial parasites show reduced and delayed egg-laying (Ferdig et al., 1993). Similarly, fruit flies (D. melanogaster), which succeeded in encapsulating the eggs of the parasitoid wasp A. tabida during the larval stage, show reduced adult survival (Hoang, 2001).

Insect species

Protocol

Effects of treatment

(a) Nutrition and general stress (cost of the maintenance of the immune system) Bumblebee (B. terrestris) Restricted access to food in Reduction of the reproductive captivity success but no effect on encapsulation response Mealworm beetle Short-term nutritional Downregulation of the PO (T. molitor) deprivation activity, but rapid upregulation when beetles reaccess to food (b) Manipulation of the workload (cost of the maintenance of the immune system) Bumblebee (B. terrestris) Clipping wings to prevent Foraging bees show reduced foraging and flying encapsulation response Damselfly (Matrona basilaris)

Observation of activity in the wild

Fruit fly (D. melanogaster)

Increased reproductive activity

Mealworm beetle (T. molitor)

Comparing experimentally mated and unmated beetles

(c) Activation of the immune system (cost of the immune response) Mosquito (A. suballatus) Experimental infection with micro-filariae taken from mammalian host Bumblebee (B. terrestris) Antigenic challenge by injection of (LPS the surface molecules of Gram-negative bacteria) and latex beads

Reduction of the encapsulation response after copulation or oviposition Reduction of resistance against bacteria Mating reduces PO activity through juvenile hormone

References Schmid-Hempel and Schmid-Hempel (1998) Siva-Jothy and Thompson (2002)

Ko¨nig and Schmid-Hempel (1995); Doums and Schmid-Hempel (2000) Siva-Jothy et al. (1998) McKean and Nunney (2001) Rolff and Siva-Jothy (2002)

Reduced egg development owing to common biochemical pathway

Ferdig et al. (1993)

Reduced survival to starvation

Moret and Schmid-Hempel (2000)

27

(continued)


TABLE 3 Examples of studies of the physiological cost of immune defence

28

TABLE 3 Examples of studies of the physiological cost of immune defence (continued ) Insect species

Protocol

Effects of treatment

Damselfly (Mnais costalis)

Activation of the immune system by insertion of small nylon monofilaments

Siva-Jothy et al. (2001)

Fruit fly (D. melanogaster)

Infection by the parasitoı¨ d A. tabida

Negative correlation between PO activity and chronic burden of gut parasites (eugregarine trophozooites) Survivors of the parasitism had a reduced survivorship under both unstressed and stressed conditions

Bumblebee (B. terrestris)

Antigenic challenge by injection of LPS (the surface molecules of Gram-negative bacteria) Antigenic challenge by injection of LPS (the surface molecules of Gram-negative bacteria) Secretion of antibiotic compounds by the exocrine metapleural glands was prevented using nail polish to close them Activation of the immune system by insertion of small nylon monofilaments Antigenic challenge by injection of LPS (the surface molecules of Gram-negative bacteria)

Reduction of the reproductive success

Moret and Schmid-Hempel (2001, 2004)

Females show reduced number of eggs produced and ovarian total protein content Reduction of the respiration rate

Ahmed et al. (2002)

Reduced longevity under ad libitum feeding conditions

Armitage et al. (2003)

Reduced capacity of associative learning

Mallon et al. (2003a)

Mosquito (A. gambiae) Leaf-cutting ant (Acromyrmex octospinosus)

Honey bee (A. mellifera)

Hoang (2001)

Poulsen et al. (2002) MICHAEL T. SIVA-JOTHY ET AL

Mealworm beetle (T. molitor)

References


29

However, it is difficult to distinguish the cost of the immune response from the negative effect of the parasite in these experiments. The use of non-living and non-pathogenic immunogens (like nylon filaments, latex micro-beads or lipopolysaccharides (bacterial cell-surface molecules)) helps to avoid the potential confounding effect of parasitism. For instance, the immune response to an implanted nylon monofilament was shown to reduce longevity in the mealworm beetle T. molitor (Armitage et al., 2003). Bumblebee (Bombus terrestris) workers challenged with either lipopolysaccharides extracted from Escherichia coli, or bacteria-sized latex micro-beads show a reduced survival under starvation (Moret and Schmid-Hempel, 2000). The use of lipopolysaccharides as an immunogen recently helped to demonstrate a broad range of costs associated with the immune response (Table 3). In insects, the immune response to lipopolysaccharides is relatively specific and involves both the PO cascade and antimicrobial immune pathways. Bumblebee workers, which have been challenged with lipopolysaccharides, show an increased antibacterial activity but a reduction in PO activity (Moret and Schmid-Hempel, 2001), suggesting a trade-off between the two immune pathways (however, a better understanding of the physiological links between these immune pathways is required for a more robust conclusion). Lipopolysaccharide-challenged female mosquitoes (Anopheles gambiae) had a lower ovarian total protein concentration and produced fewer eggs (Ahmed et al., 2002). In the honeybee (Apis mellifera) producing an immune response to lipopolysaccharides has been claimed to negatively affect associative learning (Mallon et al., 2003a). In addition to resource-based trade-offs, physiological costs of immune defence also involve the self-damage caused to host tissues by the activated immune system. For instance, upon challenge, the activation of the PO cascade generates a variety of cytotoxic substances (Nappi and Ottaviani, 2000; Carton and Nappi, 2001) inside the open haemocoel of the insect. These molecules are toxic to pathogens, but may also cause cell damage and cell death in the host (Sugumaran et al., 2000). Fortunately for the insect host, mechanisms exist to limit or prevent self-reactivity in the open haemocoel. Some of these mechanisms are passive (e.g. melanin deposited during the encapsulation response serves as a trap for reactive oxygen species and helps to localise the immune response to the pathogen surface in Drosophila, Nappi et al., 1995). Other mechanisms are active such as the production of the serine protease inhibitor proteins that restricts PO activity to the site of infection in Drosophila (De Gregorio et al., 2002) and M. sexta (Zhu et al., 2003). These active mechanisms are also likely to be costly and therefore individual insects will have to balance the benefit of successful defence with the cost of self-reactivity. The life history consequences of self-reactivity are not yet known. However, assuming a cost to self-reactivity and/or its prevention for a particular component of the insect immune system, one would predict a switch to other, less costly, immune components when the prevalence of challenges is increased (Moret, 2003). This

30


maybe why locusts (Schistocerca gregaria) exposed to a high risk of infection exhibited greater antibacterial activity, while PO activity remains constant (Wilson et al., 2002). 5.2

SPECIFIC RELATIONSHIPS BETWEEN HOSTS AND PARASITES

Another approach to explain variable levels of immune defence in populations is that specific interactions between hosts and parasites themselves generate variable immune responses. Independent of any cost of immune defence, armsrace models of host–parasite coevolution (often referred to as the ‘Red Queen Hypothesis’ (Van Valen, 1973; Peters and Lively, 1999), suggest that parasites and pathogens become rapidly adapted to those host genotypes that are the most frequent in the population. This would favour rare host genotypes through negative frequency-dependent selection and would consequently maintain genetic variation among a host population. Such a coevolutionary dynamic (over the timescale of a few generations), where parasites and pathogens continuously track host defences in order to bypass them, should result in variable degrees, and success, of host defence. This hypothesis predicts parasites should become adapted to their local hosts (Hamilton et al., 1990; Ebert, 1994; Ebert and Hamilton, 1996; Imhoof and Schmid-Hempel, 1998; Lively and Dybdahl, 2000) and that parasites cannot infect different host types with the same efficiency (Jaenike, 1993; Ebert, 1998). However, the physiological mechanisms by which adapted parasites managed to overcome local host resistance remain unknown. Its existence suggests specificity in the innate system (see Sections 6.1 and 6.2). Studies from the host’s perspective have demonstrated that hosts also vary in the response of their specific immune responses when differentially susceptible to different parasite species, or different strains of the same parasite (SchmidHempel et al., 1999; Brown et al., 2001; Carius et al., 2001). Hosts can show both specific and non-specific responses to parasite infections. Investigations about the relationship between these two components of the immune system are rare since addressing this question requires the combination of the defence cost approach (see Section 5.1) with an understanding of the nature and degree of specificity in insect immunity (Fellowes et al., 1998; Webster and Woolhouse, 1998; Frank, 2000; Jokela et al., 2000).

6

Outlook

In the last section of this review, we examine topics that emerge from the synthesis between the mechanistic approach and the evolutionary ecological approach. These issues are mainly derived from research and theory in evolutionary ecology, but require an understanding of the underlying physiology.


6.1

31

MEMORY IN INSECT IMMUNITY?

Adaptive (acquired) immunity is restricted to vertebrates and comprises ‘antigenic specificity, diversity, immunologic memory, and self/non-self recognition’ (Goldsby et al., 2000). This is ‘unlike innate immune responses’ (Goldsby et al., 2000, p. 10). Goldsby et al. correctly assume innate responses are less mechanistically sophisticated than acquired responses, but equate this with a lack of functional sophistication. A recent study on copepods (Kurtz and Franz, 2003) demonstrated a remarkable degree of memory in invertebrate immunity. Copepods were infected with tapeworms and subsequently reinfected either with tapeworms that were genetically similar to the first infection, or genetically dissimilar to the first infection. Copepods reinfected with a genetically similar parasite were much more successful in clearing the infection. The immune response of the copepod was specific and was based on the primary infection. The mechanism remains unclear (it is unlikely that parasiteborne substances caused the differential infection success, because of the design of the study), but there are candidate compounds on which a mechanism for this ability might rest. For example, lectins occur in almost all animals; they are proteins that lack catalytic activity but bind to specific carbohydrates on cell surfaces (Marques and Barraco, 2000). Quantitative variation in different sugar motives (PAMPs – see Section 4.2) on the surface of the parasite might stimulate a specific quantitative response to a particular combination of sugars. This would constitute a type of dose-dependent recognition whereby bacterial strains which differed in their cell wall composition elicited a different, specific, combination of responses. Identifying the causal basis of this specificity is an important goal for ecological immunologists. 6.2

HIGH SPECIFICITY, FEW RECEPTORS

Two main pathogen receptor pathways are known from insects: Toll and Imd (Hoffmann, 2003) (see Section 4.3 and Fig. 3). They are assumed to be specific to either G+ or G bacteria (see however Gobert et al., 2003) and produce a rather coarse level of discrimination (in sharp contrast to the sophisticated specificity of vertebrate immunity). However, a study on bumblebees and their trypanosome parasites (Crithidia bombi) casts a different light on specificity in insect immunity. Mallon et al. (2003b) infected nine different colonies of the bumblebee B. terrestris with four different strains of C. bombi. All combinations were examined and the results show all host colonies differed in their susceptibility to the parasite. However, the response depended strongly on the pathogen isolate. There were no resistant or susceptible colonies, and the response depended on individual combinations. A similar study has been conducted on the water flea, Daphnia, and the picture that emerged from that study was the same. The combination of host clone and parasite strain (in this case a bacterium) was of central importance for the infection success of the

32


parasite (Carius et al., 2001). There is clearly a huge gap between our knowledge of the molecular mechanisms that enable differentiation between G+ and G bacteria and the results of these infection studies, which suggest the existence of a much higher degree of specificity (see Watson et al., 2005). Taken together with the findings on specific memory by Kurtz and Franz (2003) and the phenomenon of trans-generational transfer of immunity (Moret and Schmid-Hempel, 2001; Little et al., 2003), it seems likely that there are undiscovered mechanism(s) that allow insects to resolve different pathogens with relatively high resolution. Classical immunology has built up a large body of evidence that such specificity is unlikely to exist, but Hultmark (2003) recently highlighted the fact that most studies of immune function in Drosophila use non-pathogenic bacteria. Consequently, the immune phenomena identified might represent host responses to saprophytic microorganisms rather than responses to virulent infections. Furthermore, Oliver et al. (2003) recently demonstrated that facultative bacterial symbionts may additionally confer resistance in their hosts, making this issue even more complicated to resolve. 6.3

MULTIPLE INFECTIONS

Another problem with the way mechanistic studies are conducted is that the consequence of single infections is usually examined (but see Hurst et al., 2003; Hughes and Boomsma, 2004). Given the omnipresence of pathogens and parasites in the natural world, the most likely scenario is that concomitant infections are prevalent. A recent study by Hughes and Boomsma (2004) shows that avirulent microorganisms out-compete virulent parasites in simultaneous infections once the virulent parasite breaks down the host’s immune defence. Therefore, the ‘mix’ of the pathogen cocktail will be crucial to the infection (and host response) outcome. This observation poses considerable challenges for studies of insect immunity. How are concomitant infections dealt with by the host? Can resources for defence (e.g. essential amino acids), be depleted during these complex insults? How is the immune system upregulated after the first infection? An intriguing finding in the context of this last question is the enhanced resistance of mosquitoes against Plasmodium after prior systemic infection with bacteria (Lowenberger et al., 1999). If A. gambiae or A. aegypti were immune activated with bacteria before they obtained an infectious blood meal (either P. berghei or P. gallinaceum), they showed a significant reduction in parasite oocysts on the midgut. This finding is supported by the observation that insect immune responses can outlast the insult that stimulated them (Moret and Siva-Jothy, 2003). Signalling pathways and antimicrobial peptides are usually regarded as being highly conserved (Zasloff, 2002; Hoffmann, 2003). The fact that these pathways are conserved is surprising given the strong selection exerted by pathogens and the subsequent fast evolution of resistance genes (Hurst and Smith, 1999). However, as pointed out by Zasloff (2002) the use of


33

antibacterial peptides by hosts probably exploits a constraint in the design of bacterial cell walls. In contrast to multicellular organisms, bacterial cells are usually positively charged. Antimicrobial peptides bind to the charged component and destroy the cell wall. It is likely that evolutionary constraints are in place that prevent most bacteria from evolving resistance to this host response, but some bacteria, such as resistant forms of Serratia, have managed to reduce the concentration of negatively charged binding sites. 6.4

PLASTICITY OF IMMUNE FUNCTION

Insect immunology is traditionally a laboratory-based biological discipline. This constraint was probably imposed by the sophisticated and sensitive methodologies used to study it. Immunological studies also try to control conditions in order to reduce the variation in the studied immune trait. One effect of this tight control (and one reason for doing it) is that laboratory practitioners rarely observe phenotypic plasticity. Phenotypic plasticity is defined as ‘the property of a genotype to produce different phenotypes in response to different environmental conditions’ (Pigliucci, 2001). Among the best-known examples is Woltereck’s Daphnia, a species that produces a defensive ‘helmet’, in response to the odour of predatory fish. Phenotypic plasticity is likely to be a very important feature of immune defence. Table 4 lists studies that have measured components, or correlates, of immune defences in different environments and which suggest a role for phenotypic plasticity in immunity (although most of the cited studies were not examining phenotypic plasticity directly). The examples in Table 4 mainly look at haemocyte densities and PO: hardly any information is available on the plasticity of other components of the immune system. Overall, the picture that emerges suggests that immune defence in insects is highly plastic, although the adaptive value of this plasticity still needs to be demonstrated. Key questions are ‘Do the measured differences in defence traits translate into higher or lower survival and reproduction in the presence of parasites?’ and ‘Are the costs associated with maintaining and employing immune defence different in different environments?’ 7

Conclusions

We started off by describing the insect host’s defence (see Fig. 1) by behavioural means, via body surfaces, to the interior. Although the physiological, molecular and genetic understanding of the mechanisms of insect immunity has vastly increased, it has come at the price of stripping study organisms of their ‘natural’ environments. One aim of this review has been to integrate immunity with environment, and to achieve this end we conclude by extending the defence component model of Schmid-Hempel and Ebert (2003), to include and integrate these two different approaches (Fig. 5).

34


TABLE 4 Plasticity of immune function and resistance Species

Environment

Immune trait

Immune function Rhodnius prolixus

Diet

T. molitor

Chorthippus biguttulus S. gregaria Spodoptera

Population density

T. molitor

Population density

Lestes viridis

Time stress

Coenagrion puella Termites

Risk of predation and parasitism Social environment

D. melanogaster

Temperature

Encapsulation lower

B. terrestris

Temperature

Encapsulation

Reference Feder et al. (1997)

Diet

Haemocyte density, lysozyme, antimicrobial activity PO

Habitat

Phagocytosis

Population density

Antimicrobial activity Cuticular colour, PO, Encapsulation Cuticular colour

Kurtz et al. (2002b) Wilson et al. (2002) Wilson et al. (2001) Barnes and SivaJothy (2000) Rolff et al. (2004)

PO, Haemocyte density PO, Haemocyte density

Siva-Jothy and Thompson (2002)

Joop and Rolff (2004) Traniello et al. (2002) Fellowes et al. (1999) Benelli (1998)

As reviewed here and elsewhere (e.g. Hoffmann, 2003; Hultmark, 2003; see Nicolas Vodovar et al., 2005), we now have considerable knowledge of the mechanisms of insect immune defence but we still do not know what causes variation in immune defence (Schmid-Hempel, 2003). One relatively intensively studied source of variation is the examination of evolutionary and/or physiological costs of immune defence (see Section 5), but most of the other areas we highlight are relatively poorly studied. To illuminate the importance of combining immunological and ecological/ evolutionary perspectives, we will consider some scenarios from the extended defence component model (Fig. 5). The three major sources of variation considered here are host type, parasite type, and environment. We refer to host- and parasite type, respectively (rather than purely genotypes) as this also applies to species with plastic polyphenisms such as darker cuticles under higher densities (see Reeson et al., 1998; Barnes and Siva-Jothy, 2000). Our scenario is the most parsimonious as it only requires two host types, two parasite types and two environments, respectively. Despite this simplicity, the model produces eight different combinations at the three distinguished levels of host defence: behavioural avoidance, avoiding penetration by the parasite/pathogen


35

FIG. 5 The extended defence component model with the host (geno-) type and two environments. Shown are three steps in a hypothetical host–parasite/pathogen interaction. First, the parasite has to overcome avoidance behaviour by the host, then it has to enter the host by overcoming the external body walls and finally it has to overcome the immune defence. Shown are two hypothetical host (geno-) types (a and b) and two hypothetical parasite (geno-) types (A and B). The shading shows the probability of the parasite overcoming the different levels of host defence, for example, for the aA combination in environment 1 the probabilities are 40.8, 40.8, 40.2 and we calculated the probability of successful infections using intermediate levels, so here 0:9 0:9 0:3 ¼ 0:24. For simplicity, we assume a multiplicative model to calculate the probabilities of successful infections (see Schmid-Hempel and Ebert, 2003). More explanation may be found in the text.

and using the immune system. There are four important conclusions to be immediately drawn from this. First, the probability of infection for the same genotype depends on the environment, even within the same host-parasite-type combination (bB) (see Stacey et al., 2003 for a real example). Second, it is possible to invest differently in different levels of the defence system yet yield the same outcome (see, for example, aA and bA in Environment 2). Third, host-type b is more resistant in environment 1, but host-type a is on average more resistant in environment 2. Fourth, knowing the mechanisms is very important (level 3 ‘immune defence’ and to a lesser extent level 2 ‘penetration’) but variation also needs to be understood (see Schmid-Hempel, 2003). This latter view has recently been supported by a genetic study on variation in antibacterial immunity in D. melanogaster (Lazarro et al., 2004). They reported naturally occurring polymorphisms of genes involved in antibacterial immunity, primarily those genes that are related to recognition of pathogens and intracellular signalling.

36


In general, the combination of environment, host- and parasite-type determines the outcome of the interactions. From an immunological point of view, entering the host and establishing the infection are the important components. In conclusion, insect immune defence is an exciting field which provides applied benefits and gives valuable insights into developmental, genetic and evolutionary processes. Combining mechanistic understanding with an evolutionary and ecological overview will, we predict, be a fruitful union. Rephrasing Stephen Stearns (1998), we hope that ecological and evolutionary thinking will be regarded, and incorporated, as a useful tool in study of the physiology of insect immune defence and parasite resistance.

Acknowledgements We thank Dan Hultmark and Ulrich Theopold for making manuscripts available at very short notice, and all the attendees at the Volkswagen Stiftungfunded ‘innate immunity conference’ in Ploen for valuable discussions. Joachim Kurtz and Klaus Reinhardt gave valuable comments and shared their ideas with us. MS-J was funded by NERC and Leverhulme, YM by a Marie Curie fellowship and JR by an NERC fellowship.

References Adamo, S. A. (2004). Estimating disease resistance in insects: phenoloxidase and lysozyme-like activity and disease resistance in the cricket Gryllus texensis. J. Insect Physiol. 50, 209–216. Adamo, S. A. (1998). The specificity of behavioural fever in the cricket Acheta domesticus. J. Parasitol. 84, 529–533. Ahmed, A. M., Baggott, S. L. and Hurd, H. (2002). The costs of mounting an immune response are reflected in the reproductive fitness of the mosquito Anopheles gambiae. Oikos 97, 371–377. Aikawa, T. and Togashi, K. (2000). Movement of Bursaphelenchus xylophilus (Nematoda: Aphelenchoididae) in tracheal system of adult Monochamus alternatus (Coleoptera: Cerambycidae). Nematology 2, 495–500. Allander, K. and Schmid-Hempel, P. (2000). Immune defence reaction in bumble-bee workers after a previous challenge and parasitic coinfection. Funct. Ecol. 14, 711–717. Armitage, S. A. O., Thompson, J. J. W., Rolff, J. and Siva-Jothy, M. T. (2003). Examining costs of induced and constitutive immune investment in Tenebrio molitor. J. Evol. Biol. 16, 1038–1044. Ashida, M. and Brey, P. T. (1995). Role of the integument in insect defense: pro-phenol oxidase cascade in the cuticular matrix. Proc. Natl. Acad. Sci. USA 92, 10698–10702. Ball, E. E., Gert de Couet, H., Horn, P. L. and Quinn, J. M. A. (1987). Haemocytes secrete basement membrane components in embryonic locusts. Development 99, 255–259.


37

Bao, Y. Y., Mega, K., Yamano, Y. and Morishima, I. (2003). cDNA cloning and expression of bacteria-induced Hdd11 gene from eri-silkworm, Samia cynthia ricini. Comp. Biochem. Physiol. C. 136, 337–342. Barillas-Mury, C., Wizel, B. and Han, Y. S. (2000). Mosquito immune responses and malaria transmission: lessons from insect model systems and implications for vertebrate innate immunity and vaccine development. Insect Biochem. Mol. Biol. 30, 429–442. Barnes, A. and Siva-Jothy, M. T. (2000). Density-dependent prophylaxis in the mealworm beetle Tenebrio molitor L. (Coleoptera: Tenebrionidae): cuticular melanization is an indicator of investment in immunity. Proc. R. Soc. Lond. B 267, 177–182. Benelli, E. F. (1998). Ecological and adaptive aspects of immunocompetence in a social insect. Dissertation, ETH Zurich. Bettencourt, R., Assefaw-Redda, Y. and Faye, I. (2000). The insect immune protein Hemolin is expressed during oogenesis and embryogenesis. Mech. Dev. 95, 301–304. Bettencourt, R., Terenius, O. and Faye, I. (2002). Hemolin gene silencing by ds-RNA injected into Cecropia pupae is lethal to next generation embryos. Insect Mol. Biol. 11, 267–271. Bischoff, I. (2003). Population dynamics of the solitary digger bee Andrena vaga Panzer (Hymenoptera, Andrenidae) studied using mark-recapture and nest counts. Popul. Ecol. 45, 197–204. Boots, M. and Begon, M. (1993). Trade-offs with resistance to a granulosis virus in the Indian meal moth, examined by a laboratory evolution experiment. Funct. Ecol. 7, 528–534. Boots, M. and Bowers, R. G. (2004). The evolution of resistance through costly acquired immunity. Proc. R. Soc. Lond. B 271, 715–723. Boman, H. G. (1998). Gene-encoded peptide antibiotics and the concept of innate immunity: an update review. Scand. J. Immunol. 48, 15–25. Boutros, M., Agaisse, H. and Perrimon, N. (2002). Sequential activation of signaling pathways during innate immune responses in Drosophila. Dev. Cell 3, 711–722. Braun, A., Hoffmann, J. A. and Meister, M. (1998). Analysis of the Drosophila host defense in domino mutant larvae, which are devoid of hemocytes. Proc. Natl. Acad. Sci. USA 95, 14337–14342. Brey, P. T., Latge, J. P. and Prevost, M. C. (1986). Integumental penetration of the pea aphid, Acyrthosiphon pisum, by Conidiobolus obscurus (Entomophthoracea). J. Invert. Pathol. 48, 34–41. Brey, P., Lee, W., Yamakawa, M., Koizumi, Y., Perrot, S., Francois, M. and Ashida, M. (1993). Role of the integument in insect immunity: epicuticular abrasion and induction of cecropin synthesis in cuticular epithelial cells. Proc. Natl. Acad. Sci. USA 90, 6275–6279. Brey, P. T. and Hultmark, D. E. (eds) (1997). Molecular Mechanisms of Immune Responses in Insects. London: Chapman & Hall. Brown, M. J. F., Loosli, R. and Schmid-Hempel, P. (2001). Condition-dependent expression of virulence in a trypanosome infecting bumblebees. Oikos 91, 421–427. Bulet, P., Hetru, C., Dimarcq, J. L. and Hoffmann, D. (1999). Antimicrobial peptides in insects; structure and function. Dev. Comp. Immunol. 23, 329–344. Cade, W. H. (1991). Interspecific and intraspecific variation in nightly calling duration in filed crickets, Gryllus integer and G. rubens (Orthoptera, Gryllidae). J. Insect Behav. 4, 185–194. Cade, W. H. and Wyatt, D. R. (1984). Factors affecting calling behaviour in field crickets, Teleogryllus and Gryllus (age, weight, density and parasites). Behaviour 88, 61–75.

38


Carayon, J. (1966). Traumatic insemination and the paragenital system. In: Monograph of the Cimicidae (ed. Usinger, R.), pp. 81–166. Philadelphia: Entomological Society of America. Carius, H. J., Little, T. J. and Ebert, D. (2001). Genetic variation in a host-parasite association: potential for coevolution and frequency-dependent selection. Evolution 157, 408–420. Carton, Y. and Nappi, A. J. (2001). Immunogenic aspects of the cellular immune response of Drosophila against parasitoids. Immunogenetics 52, 157–164. Cerenius, L. and So¨derha¨ll, K. (2004). The prophenoloxidase-activating system in invertebrates. Immunol. Rev. 198, 116–126. Chain, B., Leyshon-Sorland, K. and Siva-Jothy, M. T. (1992). Haemocyte heterogeneity in the cockroach Periplaneta americana as revealed by monoclonal antibodies. J. Cell. Sci. 103, 1261–1267. Chapman, R. F. (1998). The Insects. Cambridge: Cambridge University Press. Charnley, A. K. and St. Leger, R. J. (1991). The role of cuticle-degrading enzymes in fungal pathogenesis in insects. In: The Fungal Spore and Disease Initiation in Plants and Animals (eds Cole, G. and Hoch, H.). NY: Plenum Press. Chase, M., Raina, K., Bruno, J. and Sugumaran, M. (2000). Purification, characterization and molecular cloning of prophenoloxidase from Sarcophaga bullata. Insect Biochem. Mol. Biol. 30, 953–967. Cherbas, L. (1973). The induction of an injury reaction in cultured hemocytes from saturniid pupae. J. Insect Physiol. 19, 2011–2023. Choe, K. M., Werner, T., Sto¨ven, S., Hultmark, D. and Anderson, K. (2002). Requirement for a peptidoglycan recognition protein (PGRP) in relish activation antibacterial immune responses in Drosophila. Science 296, 359–362. Clark, K. D., Pech, L. L. and Strand, M. R. (1997). Isolation and identification of a plasmatocyte-spreading peptide from the hemolymph of the lepidopteran insect Pseudoplusia includens. J. Biol. Chem. 272, 23440–23447. Cote, I. M. and Molloy, P. P. (2003). Temporal variation in cleanerfish and client behaviour: does it reflect ectoparasite availability?. Ethology 109, 487–499. Crudgington, H. and Siva-Jothy, M. T. (2000). Genital damage, kicking and early death. Nature 407, 855–856. Daffre, S., Kylsten, P., Samakovlis, C. and Hultmark, D. (1994). The lysozyme locus in Drosophila melanogaster – an expanded gene family adapted for expression in the digestive tract. Mol. Gen. Genet. 242, 152–162. David, W. A. L. (1967). The physiology of insect integument in relation to the invasion of pathogens. In: Insects and Physiology (eds Beament, J. W. L. and Treherne, J. E.). London: Oliver-Boyd. de Doucet, M. M. A., Miranda, M. B. and Bertolotti, M. A. (1998). Infectivity of entomogenous nematodes (Steinernematidae and Heterorhabditidae) to Pediculus humanus capitis De Geer (Anoplura: Pediculidae). Fund. Appl. Nematol. 21, 13–16. De Gregorio, E., Hun, S. J., Lee, W. J., Buek, M. J., Osaki, T., Kawabata, S. I., Lee, B. L., Iwanaga, S., Lemaitre, B. and Brey, P. T. (2002). An immune responsive serpin regulates the melanisation cascade in Drosophila. Develop. Cell 3, 581–592. Dettloff, M., Wittwer, D., Weise, C. and Wiesner, A. (2001). Lipophorin of lower density is formed during immune responses in the lepidopteran insect Galleria mellonella. Cell Tissue Res. 306, 449–458. Doums, C. and Schmid-Hempel, P. (2000). Immunocompetence in workers of a social insect, Bombus terrestris L., in relation to foraging activity and parasitic infection. Can. J. Zool. 78, 1060–1066. Doyle, R. J. and Dziarski, R. (2001). The bacterial cell: peptidoglycan. In: Molecular Medical Microbiology (ed. Sussman, M.), pp. 137–153. London: Academic Press.


39

Dziarski, R. (2004). Peptidoglycan recognition proteins (PGRPs). Mol. Immunol. 40, 877–886. Ebert, D. and Hamilton, W. D. (1996). Sex against virulensb: the evolution of parasitic diseases. Trends Ecol. Evol. 11, 79–82. Ebert, D. (1994). Virulence and local adaptation of horizontally transmitted parasite. Science 265, 1084–1086. Ebert, D. (1998). Experimental evolution of parasites. Science 282, 1432–1435. Elrod-Erickson, M., Mishra, S. and Schneider, D. (2000). Interactions between the cellular and humoral immune responses in Drosophila. Curr. Biol. 10, 781–784. Feder, D., Mello, C., Garcia, E. and Azambuja, P. (1997). Immune responses in Rhodnius prolixus: influence of nutrition and ecdysone. J. Insect Physiol. 43, 513–519. Fellowes, M. D. E., Kraaijeveld, A. R. and Godfray, H. C. J. (1998). Trade-off associated with selection for increased ability to resist parasitoid attack in Drosophila melanogaster. Proc. R. Soc. Lond. B 265, 1–6. Fellowes, M. D. E., Kraaijeveld, A. R. and Godfray, H. C. J. (1999). Cross-resistance following artificial selection for increased defense against parasitoids in Drosophila melanogaster. Evolution 53, 966–972. Ferdig, M. T. B., Spray, F. J., Li, J. and Christensen, B. M. (1993). Reproductive costs associated with resistance in a mosquito-filarial worm system. Am. J. Trop. Med. Hyg. 49, 756–762. Folstad, I. and Karter, A. J. (1992). Parasites, bright males, and the immunocompetence handicap. Am. Nat. 139, 603–622. Franc, N. C., Heitzler, P., Ezekowitz, R. A. B. and White, K. (1999). Requirement for croquemort in phagocytosis of apoptotic cells in Drosophila. Science 284, 1991–1994. Franco, A. M. R., Tesh, R. B., Guzman, H., Deane, M. P. and Grimaldi, G. (1997). Development of Endotrypanum (Kinetoplastida: Trypanosomatidae) in experimentally infected phlebotomine sand flies (Diptera: Psychodidae). J. Med. Entomol. 34, 189–192. Frank, S. A. (1991). Ecological and genetic models of host–pathogen coevolution. Heredity 67, 73–84. Frank, S. A. (1993). A model of inducible defence. Evolution 47, 325–327. Frank, S. A. (2000). Specific and non-specific defense against parasitic attack. J. Theor. Biol. 202, 283–304. Futuyma, D. (1998). Evolutionary Biology. Sunderland: Sinauer. Gardiner, E. M. M. and Strand, M. R. (2000). Haematopoesis in larval Pseudoplusia includens and Spodoptera frugiperda. Arch. Insect Bioch. Physiol. 43, 147–164. Glupov, V. V., Khvoshevskaya, M. F., Lozinskaya, Y. L., Dubovski, I. M., Martemyanov, V. V. and Sokolova, J. Y. (2001). Application of the nitroblue tetrazolium-reduction method for studies on the production of reactive oxygen species in insect haemocytes. Cytobios 106 (Suppl 2), 165–178. Gobert, V., Gottar, M., Matskevich, A. A., Rutschmann, S., Royet, J., Belvin, M., Hoffmann, J. A. and Ferrandon, D. (2003). Dual activation of the Drosophila Toll pathway by two pattern recognition receptors. Science 302, 2126–2130. Goldsby, R. A., Kindt, T. J. and Osborne, B. (2000). Kuby Immunology. New York: Freeman. Golkar, L., LeBrun, R. A., Ohayon, H., Gounon, P., Papierok, B. and Brey, P. T. (1993). Variation of larval susceptibility to Lagenidium giganteum in three mosquito species. J. Invert. Pathol. 62, 1–8. Gorman, M. J. and Paskewitz, S. M. (2001). Serine proteases as mediators of mosquito immunity. Insect Biochem. Mol. Biol. 31, 257–262.

40


Goto, A., Kadowaki, T. and Kitagawa, Y. (2003). Drosophila hemolectin gene is expressed in embryonic and larval hemocytes and its knock-down causes bleeding defects. Dev. Biol. 264, 582–591. Gottar, M., Gobert, V., Michel, T., Belvin, M., Duyk, G., Hoffmann, J. A., Ferrandon, D. and Royet, J. (2002). The Drosophila immune response against Gram-negative bacteria is mediated by a peptidoglycan recognition protein. Nature 416, 640–643. Gregoire, C. (1974). Hemolymph coagulation. In: The Physiology of Insects, Vol. 5 (ed. Rockstein, M.), pp. 309–360, NY: Academic Press. Gullan, P. J. and Cranston, P. S. (2000). The Insects. Oxford: Blackwell. Hajek, A. E. and St. Leger, T. J. (1994). Interactions between fungal pathogens and insect hosts. Ann. Rev. Entomol. 39, 293–322. Haldane, J. B. S. (1949). Disease and evolution. Ricer. Sci. 19 (Suppl), 68–76. Hamilton, W. D. (1980). Sex versus non-sex versus parasite. Oikos 35, 282–290. Hamilton, W. D. and Zuk, M. (1982). Heritable true fitness and bright birds – a role for parasites. Science 218, 384–387. Hamilton, W. D., Axelrod, R. and Tanese, R. (1990). Sexual reproduction as an adaptation to resist parasites (a review). Proc. Natl. Acad. Sci. USA 87, 3566–3573. Han, Y. S., Thompson, J., Kafatos, F. C. and Barillas-Mury, C. (2000). Molecular interactions between Anopheles stephensi midgut cells and Plasmodium berghei: the time bomb theory of ookinete invasion of mosquitoes. EMBO J. 19, 6030–6040. Hao, Z. G., Kasumba, I. and Aksoy, S. (2003). Proventriculus (cardia) plays a crucial role in immunity in tsetse fly (Diptera: Glossinidiae). Insect Biochem. Mol. Biol. 33, 1155–1164. Hart, B. L. (1997). Behavioural defence. In: Host– Parasite Evolution (eds Clayton, D. H., and Moore, J.), pp. 59–77. Oxford: Oxford University Press. Hoang, A. (2001). Immune response to parasitism reduces resistance of Drosophila melanogaster to desiccation and starvation. Evolution 55, 2353–2358. Hoffmann, J. A. (2003). The immune response of Drosophila. Nature 426, 33–38. Hoffmann, J. A. and Reichhart, J. M. (2002). Drosophila innate immunity: an evolutionary perspective. Nat. Immunol. 3, 121–126. Hosken, D. J. (2001). Sex and death: microevolutionary trade-offs between reproductive and immune investment in dung flies. Curr. Biol. 11, R308–R379. Hou, S. X., Zheng, Z., Chen, X. and Perrimon, N. (2002). The JAK/ STAT92E pathway in model organisms: emerging roles in cell movement. Dev. Cell 3, 765–778. Hughes, W. and Boomsma, J. (2004). Let your enemy do the work: within-host interactions between two fungal parasites of leaf-cutting ants. Proc. R. Soc. Lond. B 3 (Suppl), S104–S106. Hultmark, D. (2003). Drosophila immunity: paths and patterns. Curr. Opin. Immunol. 15, 12–19. Hurst, G., Anbutsu, H., Kutsukake, M. and Fukatsu, T. (2003). Hidden from the host: Spiroplasma bacteria infecting Drosophila do not cause an immune response, but are suppressed by ectopic immune activation. Insect Mol. Biol. 12, 93–97. Hurst, L. and Smith, N. (1999). Do essential genes evolve slowly?. Curr. Biol. 9, 747–750. Imhoof, B. and Schmid-Hempel, P. (1998). Single-clone and mixed-clone infections versus host environment in Crithidia bombi infecting bumblebees. Parasitology 117, 331–336. Jaenike, J. (1993). Rapid evolution of host specificity in a parasitic nematode. Evol. Ecol. 7, 103–108. Janeway, C. A. (1989). Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harbour Symp. Quant. Biol. 54, 1–13.


41

Johansson, M. W. (1999). Cell adhesion molecules in invertebrate immunity. Dev. Comp. Immunol. 23, 303–315. Jokela, J., Schmid-Hempel, P. and Rigby, M. C. (2000). Dr. Pangloss restrained by the Red Queen – steps towards a unified defence theory. Oikos 89, 267–278. Joop, G., and Rolff, J. (2004). Plasticity of immune function and condition under the risk of predation and parasitism. Evol. Ecol. Res. 6, 1051–1062. Kang, D., Liu, G., Lundstro¨m, A., Gelius, E. and Steiner, H. (1998). A peptidoglycan recognition protein in innate immunity conserved from insects to humans. Proc. Natl. Acad. Sci. USA 95, 10078–10082. Kanost, M. R., Jiang, H. B. and Yu, X. Q. (2004). Innate immune responses of a lepidopteran insect Manduca sexta. Immunol. Rev. 198, 97–105. Kim, Y. S., Ryu, J. H., Han, S. J., Choi, K. H., Nam, K. B., Jang, I. W., Lemaitre, B., Brey, P. T. and Lee, W. J. (2000). Gram-negative bacteria binding protein, a pattern recognition receptor for lipopolysaccharide and b1,3-glucan which mediates the signaling for the induction of innate immune genes in Drosophila melanogaster cells. J. Biol. Chem. 275, 32721–32727. Koella, J. C. and Boe¨te, C. (2002). A genetic correlation between age at pupation and melanisation immune response of the yellow fever mosquito Aedes aegypti. Evolution 56, 1074–1079. Koenig, C. and Schmid-Hempel, P. (1995). Foraging activity and immunocompetence in workers of the bumble bee, Bombus terrestris L. Proc. R. Soc. Lond. B 260, 225–227. Kotani, E., Yamakawa, M., Iwamoto, S.-I., Tashiro, M., Mori, H., Sumida, M., Matsubaru, F., Taniai, K., Kadono-Okuda, K., Kato, Y. and Mori, H. (1995). Cloning and expression of the gene of hemocytin, an insect humoral lectin which is homologous with the mammalian von Willebrand factor. Bioch. Biophys. Acta 1260, 245–258. Kraaijeveld, A. R. and Godfray, H. C. J. (1997). Trade-off between parasitoid resistance and larval competitive ability in Drosophila melanogaster. Nature 389, 278–280. Kraaijeveld, A. R., Limentani, E. and Godfray, H. C. J. (2001). Basis of the trade-off between parasitoid resistance and larval competitive ability in Drosophila melanogaster. Proc. Roy. Soc. London B 268, 259–261. Kurtz, J. and Franz, K. (2003). Evidence for memory in invertebrate immunity. Nature 425, 37–38. Kurtz, J., van der Veen, I. T. and Ryder, J. J. (2002a). Ecological immunity of arthropods – a thread of Ariadne?. Trends Ecol. Evol. 17, 204–205. Kurtz, J., Klappert, K., Schneider, W. and Reinhold, K. (2002b). Immune defence, dispersal and local adaptation. Evol. Ecol. Res. 4, 431–439. Lackie, A. M. (1988). Haemocyte behaviour. Adv. Insect Physiol. 21, 86–178. Lackie, A. M., Takle, G. and Tetley, L. (1985). Haemocytic encapsulation in the locust Schistocerca gregaria (Orthoptera) and the cockroach Periplaneta americana. Cell Tiss. Res. 240, 343–351. Lagueux, M., Perrodou, E., Levashina, E. A., Capovilla, M. and Hoffmann, J. A. (2000). Constitutive expression of a complement-like protein in Toll and JAK gainof-function mutants of Drosophila. Proc. Natl. Acad. Sci. USA 97, 11427–11432. Lanz Mendoza, H. and Faye, I. (1999). Physiological aspects of the immunoglobulin superfamily in invertebrates. Dev. Comp. Immunol. 23, 359–374. Lavine, M. D. and Strand, M. R. (2001). a-integrin subunits expressed in haemocytes of Pseudoplusia includens. Keystone Symposium on the Genetic Manipulation of Insects. Keystone Symposia, Silverthorne 63. Lavine, M.D. and Strand, M.R., (2002). Insect hemocytes and their role in immunity. Insect Biochem. Mol. Biol. 32, 1295–1309.

42


Lavine, M. D. and Strand, M. R. (2003). Haemocytes from Pseudoplusia includens express multiple alpha and beta integrin subunits. Insect Mol. Biol. 12, 441–452. Lazarro, B., Sceurman, B. and Clark, A. (2004). Genetic basis of natural variation in D. melanogaster antibacterial immunity. Science 303, 1873–1876. Lebestky, T., Chang, T., Hartenstein, V. and Banerjee, U. (2000). Specification of Drosophila hematopoietic lineage by conserved transcription factors. Science 288, 146–149. Lee, K. Y., Horodyski, F. M., Valaitis, A. P. and Denlinger, D. L. (2002). Molecular characterization of the insect immune protein hemolin and its high induction during embryonic diapause in the gypsy moth, Lymantria dispar. Insect Biochem. Mol. Biol. 32, 1457–1467. Lehane, M. J., Wu, D. and Lehane, S. M. (1997). Midgut-specific immune molecules are produced by the blood-sucking insect Stomoxys calcitrans. Proc. Natl. Acad. Sci. USA 94, 11502–11507. Lemaıˆ tre, B., Nicolas, E., Michaut, L., Reichhart, J. M. and Hoffmann, J. A. (1996). The dorsoventral regulatory gene cassette spa¨tzle/Toll/cactus controls the potent antifungal response in Drosophila adults. Cell 86, 973–983. Lemaıˆ tre, B., Reichart, J. M. and Hoffmann, J. A. (1997). Drosophila host defensb: differential induction of antimicrobial peptide genes after infection by various classes of microorganisms.. Proc. Natl. Acad. Sci. USA 94, 14614–14619. Leulier, F., Parquet, C., Pili-Floury, S., Ryu, J. H., Caroff, M., Lee, W. J., MenginLecreulx, D. and Lemaıˆ tre, B. (2003). The Drosophila immune system detects bacteria through specific peptidoglycan recognition. Nat. Immunol. 4, 478–484. Levashina, E. A., Moita, L. F., Blandin, S., Vriend, G., Lagueux, M. and Kafatos, F. C. (2001). Conserved role of a complement-like proteinin phagocytosis revealed by dsRNA knockout in cultured cells of the mosquito, Anopheles gambiae. Cell 104, 709–718. Li, D., Scherfer, C., Korayem, A. M., Zhao, Z., Schmidt, O. and Theopold, U. (2002). Insect hemolymph clotting: evidence for interaction between the coagulation system and the prophenoloxidase activating cascade. Insect Biochem. Mol. Biol. 32, 919–928. Ligoxygakis, P., Pelte, N., Hoffmann, J. A. and Reichhart, J. M. (2002a). Activation of drosophila Toll during fungal infection by a blood serine protease. Science 297, 114–116. Ligoxygakis, P., Pelte, N., Ji, C., Leclerc, V., Duvic, B., Belvin, M., Jiang, H. B., Hoffmann, J. A. and Reichhart, J. M. (2002b). A serpin mutant links Toll activation to melanization in the host defence of Drosophila. EMBO J. 21, 6330–6337. Little, T. J., O’Connor, B., Colegrave, N., Watt, K. and Read, A. F. (2003). Maternal transfer of strain-specific immunity in an invertebrate. Curr. Biol. 13, 489–492. Liu, C. T., Hou, R. F. and Chen, C. C. (1998). Formation of basement membrane-like structure terminates the cellular encapsulation of microfilariae in the haemocel of Anopheles quadrimaculatus. Parasitology 116, 511–518. Lively, C. M. and Dybdahl, M. F. (2000). Parasite adaptation to locally common host genotypes. Nature 405, 679–681. Lowenberger, C. (2001). Innate immune response of Aedes aegypti. Insect Biochem. Mol. Biol. 31, 219–229. Lowenberger, C., Kamal, S., Chile, J., Paskewitz, S., Bulet, P., Hoffmann, J. A. and Christensen, B. (1999). Mosquito–Plasmodium interactions in response to immune activation of the vector. Exp. Parasitol. 91 59–59. Lung, O., Kuo, L. and Wolfner, M. F. (2001). Drosophila males transfer antibacterial proteins from their accessory gland and ejaculatory duct to their mates. J. Insect Physiol. 47, 617–622.


43

Mallon, E. O., Brockmann, A. and Schmid-Hempel, P. (2003a). Immune response inhibits associative learning in insects. Proc. R. Soc. Lond. B 270, 2471–2473. Mallon, E. O., Loosli, R. and Schmid-Hempel, P. (2003b). Specific versus nonspecific immune defense in the bumblebee Bombus terrestris L. Evolution 57, 1444–1447. Marques, M. and Barraco, M. (2000). Lectins, as non-self-recognition factors in crustaceans. Aquaculture 191, 23–44. McGavin, G. C. (2001). Essential Entomology. Oxford: Oxford University Press. McKean, K. A. and Nunney, L. (2001). Increased sexual activity reduces male immune function in Drosophila melanogaster. Proc. Nat. Acad. Sci. USA 98, 7904–7909. Meister, M., Braun, A., Kappler, C., Reichhart, J. M. and Hoffmann, J. A. (1994). Insect immunity. A transgenic analysis in Drosophila defines several functional domains in the diptericin promoter. EMBO J. 13, 5958–5966. Mellroth, P., Karlsson, J. and Steiner, H. (2003). A scavenger function for a Drosophila peptidoglycan recognition protein. J. Biol. Chem. 278, 7059–7064. Michel, T., Reichhart, J. M., Hoffmann, J. A. and Royet, J. (2001). Drosophila Toll is activated by Gram-positive bacteria through a circulating peptidoglycan recognition protein. Nature 414, 756–759. Moret, Y. (2003). Explaining variable costs of the immune response: selection for specific versus non-specific immunity and facultative life history change. Oikos 102, 213–216. Moret, Y. and Schmid-Hempel, P. (2000). Survival for immunity: the price of immune system activation for bumblebee workers. Science 290, 1166–1168. Moret, Y. and Schmid-Hempel, P. (2001). Immune defence in bumblebee offspring. Nature 414, 506. Moret, Y. and Schmid-Hempel, P. (2004). Social life-history response to individual immune challenge of workers of Bombus terrestris L.: a possible new cooperative phenomenon. Ecol. Lett. 7, 146–152. Moret, Y. and Siva-Jothy, M. (2003). Adaptive innate immunity? Responsive-mode prophylaxis in the mealworm beetle Tenebrio molitor. Proc. R. Soc. London B 270, 2475–2480. Morrow, E. H. and Arnqvist, G. (2003). Costly traumatic insemination and a female counter-adaptation in bedbugs. Proc. R. Soc. Lond. B 270, 2377–2381. Mullet, H., Rattcliffe, N. A. and Rowley, A. F. (1993). The generation and characterisation of anti-insect blood cell monoclonal antibodies. J. Cell Sci. 105, 93–100. Nappi, A. J. and Ottaviani, E. (2000). Cytotoxicity and cytotoxic molecules in invertebrates. Bioessays 22, 469–480. Nappi, A. J., Vass, E., Frey, F. and Carton, Y. (1995). Superoxide anion generation in Drosophila during melanotic encapsulation of parasites. Eur. J. Cell Biol. 68, 450–456. Nardi, J. B., Gao, C. and Kanost, M. R. (2001). The extracellular matrix protein lacunin is expressed by a subset of hemocytes involved in basal lamina morphogenesis. J. Insect Phys. 47, 997–1006. Nardi, J. B. and Miklasz, S. D. (1989). Hemocytes contribute to both the formation and breakdown of the basal lamina in developing wings of Manduca sexta. Tissue Cell 21, 559–567. Nardi, J. B., Pilas, B., Ujhelyi, E., Garsha, K. and Kanost, M. R. (2003). Hematopoietic organs of Manduca sexta and hemocyte lineages. Dev. Genes Evol. 213, 477–491. Nathan, C. F. and Hibbs, J. B., Jr. (1991). Role of nitric oxide synthesis in macrophage antimicrobiol activity. Curr. Opin. Immunol. 3, 65–70. Nelson, R. E., Fessler, L. I., Takagi, Y., Blumberg, B., Keene, D. R., Ohara, O., Tanaka, A., Kotani, H., Miyajima, N. and Nomura, N. (1994). Peroxidasin: a novel enzyme-matrix protein of Drosophila development. EMBO J. 13, 3438–3447.

44


Neville, A. C. (1975). Biology of the Arthropod Cuticle. Berlin: Springer. Nicolas Vodovar, N., Vinals, M., Liehl, P., Basset, A., Degrouard, J., Spellman, P., Boccard, F. and Lemaitre, B. (2005). Drosophila host defense after oral insection by an entomopathogenic Pseudomonas species. Proc. Nat. Acad. Sci. USA 102, 11414–11418. O’Donnell, S. (1997). How parasites can promote the expression of social behaviour in their hosts. Proc. Roy. Soc. Lond. Ser. B 264, 689–694. Oliver, K. M., Russell, J., Moran, N. A. and Hunter, M. (2003). Facultative bacterial symbionts in aphids confer resistance to parasitic wasps. Proc. Nat. Acad. Sci., USA 100, 1803–1807. Owens, I. P. F. and Wilson, K. (1999). Immunocompetensb: a neglected life history trait or conspicuous red herring?. Trends Ecol. Evol. 14, 170–172. Pech, L. L. and Strand, M. R. (1996). Granular cells are required for encapsulation of foreign targets by insect haemocytes. J. Cell Sci. 109, 2053–2060. Pech, L. L. and Strand, M. R. (2000). Plasmatocytes from the moth Pseudoplusia includens induce apoptosis of granular cells. J. Insect Phys. 46, 1565–1573. Pech, L. L., Trudeu, D. and Strand, M. R. (1995). Effects of basement membranes on the behaviour of hemocyts from pseudoplusia includens: development of an in vitro encapsulation assay. J. Insect Physiol. 41, 801–807. Peters, A. D. and Lively, C. M. (1999). The Red Queen and fluctuating epistasis: a population genetic analysis of antagonistic coevolution. Am. Nat. 154, 393–405. Peters, W. (1992). Peritrophic membranes. Berlin: Springer. Pigliucci, M. (2001). Phenotypic Plasticity. Baltimore: John Hopkins University Press. Poulsen, M., Bot, A. N. M., Nielsen, M. G. and Boomsma, J. J. (2002). Experimental evidence for the costs and hygienic significance of the antibiotic metapleural gland secretion in leaf-cutting ants. Behav. Ecol. Sociobiol. 52, 151–157. Qiu, P., Pan, P. C. and Govind, S. (1998). A role for the the Drosophila Toll/Cactus pathway in larval hematopoiesis. Development 125, 1909–1920. Ra¨met, M., Lanot, R., Zachary, D. and Manfruelli, P. (2002a). JNK signaling pathway is required for efficient wound healing in Drosophila. Dev. Biol 241, 145–156. Ra¨met, M., Manfruelli, P., Pearson, A., Mathey-Prevot, B. and Ezekowitz, R. A. B. (2002b). Functional genomic analysis of phagocytosis and identification of a Drosophila receptor for E. coli. Nature 416, 644–648. Rawlings, N. D. and Barrett, A. J. (1994). Families of serine peptidases. Meth. Enzymol. 244, 19–61. Reeson, A., Wilson, K., Gunn, A., Hails, R. and Goulson, D. (1998). Baculovirus resistance in the noctuid Spodoptera exempta is phenotypically plastic and responds to population density. Proc. R. Soc. Lond. B 265, 1787–1791. Reinhardt, K., Naylor, R. and Siva-Jothy, M. T. (2003). Reducing a cost of traumatic insemination: female bedbugs evolve a unique organ. Proc. R. Soc. Lond. B 270, 2371–2376. Richards, A. G. and Richards, P. A. (1977). The peritrophic membranes of insects. Ann. Rev. Entomol. 22, 219–240. Robinson, J. M. and Badwey, J. A. (1994). Production of active oxygen species by phagocytic leukocytes. Immunology 60, 159–178. Roff, D. A. (1992). The Evolution of Life Histories: Theory and Analysis. New York: Chapman & Hall. Roff, D. (1997). Evolutionary Quantitative Genetics. New York: Chapman & Hall. Rolff, J. and Siva-Jothy, M. T. (2002). Copulation corrupts immunity: a mechanism for a cost of mating in insects. Proc. Natl. Acad. Sci. USA 99, 9916–9918. Rolff, J. and Siva-Jothy, M. T. (2003). Invertebrate ecological immunity. Science 301, 473–475.


45

Rolff, J., van de Meutter, F. and Stoks, R. (2004). Time constraints decouple age and size at maturity and physiological traits. Am. Nat. 164, 559–565. Rothenbuhler, W. C. and Thompson, V. C. (1956). Resistance to American foulbrood in honeybees. I. Differential survival of larvae of different genetic lines. J. Econ. Entomol. 49, 470–475. Royet, J. (2004). Drosophila melanogaster innate immunity: an emerging role for peptidoglycan recognition proteins in bacteria detection. Cell. Mol. Life Sci. 61, 537–546. Ryder, J. (2003). Immunocompetensb: an overstretched concept?. Trends Ecol. Evol. 18, 319–320. Sagisaka, A., Miyanoshita, A., Ishibashi, J. and Yamakawa, M. (2001). Purification, characterization and gene expression of a glycine and proline-rich antibacterial protein family from larvae of a beetle Allomyrina dichotoma. Insect Mol. Biol. 10, 293–302. Sasaki, A. and Godfray, H. C. J. (1999). A model for the co-evolution of resistance and virulence in coupled host–parasitoid interactions. Proc. Roy. Soc. Lond. B 266, 455–463. Sass, M., Kiss, A. and Locke, M. (1994). Integument and hemocyte peptides. J. Insect Physiol. 40, 407–421. Schal, C., Sevala, V. and Card, R. T. (1998). Novel and highly specific transport of a volatile sex pheromone by hemolymph lipophorin in moths. Naturwissenschaften 85, 339. Scherfer, C., Karlsson, C., Loseva, O., Bidla, G., Goto, A., Havemann, J., Dushay, M. S. and Theopold, U. (2004). Isolation and characterization of hemolymph clotting factors in Drosophila melanogaster by a pullout method. Curr. Biol. 14, 625–662. Schmid-Hempel, P. (2003). Variation in immune defence as a question of evolutionary ecology. Proc. Roy. Soc. Lond. B 270, 357–366. Schmid-Hempel, P. and Ebert, D. (2003). On the evolutionary ecology of specific immune defence. Trends Ecol. Evol. 18, 27–32. Schmid-Hempel, P., Puhr, K., Kruger, K., Reber, C. and Schmid-Hempel, R. (1999). Dynamic and genetic consequences of variation in horizontal transmission for a microparasitic infection. Evolution 53, 426–434. Schmid-Hempel, R. and Schmid-Hempel, P. (1998). Colony performance and immunocompetence of a social insect, Bombus terrestris, in poor and variable environments. Funct. Ecol. 12, 22–30. Sheldon, B. C. and Verhulst, S. (1996). Ecological immunology: costly parasite defences and trade-offs in evolutionary ecology. Trends Ecol. Evol. 11, 317–321. Shields, V. D. C. (1994). Ultrastructure of the uniporous sensilla on the galea of larval Mamestra configurata. Can. J. Zool. 72, 2016–2031. Siva-Jothy, M. T. (1995). Immunocompetensb: conspicuous by its absence. Trends Ecol. Evol. 10, 205–206. Siva-Jothy, M. T. (2000). A mechanistic link between parasite resistance and expression of a sexually selected trait in a damselfly. Proc. R. Soc. Lond. B 267, 2523–2527. Siva-Jothy, M. T. and Thompson, J. J. W. (2002). Short-term nutrient deprivation affects immune function. Physiol. Entomol. 27, 206–212. Siva-Jothy, M. T., Tsubaki, Y. and Hooper, R. E. (1998). Decreased immune response as a proximate cost of copulation and oviposition in a damselfly. Physiol. Entomol. 23, 274–277. Siva-Jothy, M. T., Tsubaki, Y., Hooper, R. E. and Plaistow, S. J. (2001). Investment in immune function under chronic and acute immune challenge in an insect. Physiol. Entomol. 26, 1–5. Sluss, H. K., Han, Z., Barrett, T., Davis, R. J. and Ip, Y. T. (1996). A JNK signal transduction pathway that mediates morphogenesis and an immune response in Drosophila. Genes Dev. 10, 2745–2758.

46


So¨derha¨ll, K. and Cerenius, L. (1998). Role of the prophenoloxidase-activating system in invertebrate immunity. Curr. Opin. Immunol. 10, 23–28. Sorrentino, R. P., Carton, Y. and Govind, S. (2002). Cellular immune response to parasite infection in the Drosophila lymph gland is developmentally regulated. Dev. Biol. 243, 65–80. Sorrentino, R. P., Melk, J. P. and Govind, S. (2004). Genetic analysis of contributions of dorsal group and JAK-Stat92E pathway genes to larval hemocyte concentration and the egg encapsulation response in Drosophila. Genetics 166, 1343–1356. Stacey, D., Thomas, M., Blanford, S., Pell, J., Pugh, C. and Fellowes, M. D. E. (2003). Genotype and temperature influence pea aphid resistance to a fungal entomopathogen. Physiol. Entomol. 28, 75–81. Stearns, S. C. (1992). The Evolution of Life Histories. Oxford: Oxford University Press. Stearns, S. C. (ed.) (1998). Evolution in Health and Disease. Oxford: Oxford University Press. St. Leger, R. J. (1991). Integument as a barrier to microbial infections. In: Physiology of the Insect Epidermis (eds Binnington, K. and Retnakaran, A.), pp. 284–306. Canberra: CSIRO. Strand, M. R. and Clark, K. D. (1999). Plasmatocyte spreading peptide induces spreading of plasmatocytes but represses spreading of granulocytes. Arch. Insect Biochem. 42, 213–223. Strand, M. R., Hayakawa, Y. and Clark, K. D. (2000). Plasmatocyte spreading peptide (PSP1) and growth blocking peptide (GBP) are multifunctional homologs. J. Insect Phys. 46, 817–824. Strand, M. R. and Johnson, J. A. (1996). Characterization of monoclonal antibodies to hemocytes of Pseudoplusia includens. J. Insect Physiol. 42, 21–31. Sugumaran, M. (2002). Comparative biochemistry of eumelanogenesis and the protective roles of phenoloxidase and melanin in insects. Pigment Cell Res. 15, 2–9. Sugumaran, M., Nellaiappan, K. and Valivittan, K. (2000). A new mechanism for the control of phenoloxidase activity: inhibition and complex formation with quinone isomerase. Arch. Biochem. Biophys. 379, 252–260. Sutter, G. R., Rothenbuhler, W. C. and Raun, E. S. (1968). Resistance to American foulbrood in honeybees. VII. Growth of resistant and susceptible larvae. J. Invert. Pathol. 12, 25–28. Takehana, A., Katsuyama, T., Yano, T., Oshima, Y., Takada, H., Aigaki, T. and Kurata, S. (2002). Overexpression of a pattern-recognition receptor, peptidoglycanrecognition protein-LE, activates imd/relish-mediated antibacterial defense and the prophenoloxidase cascade in Drosophila larvae. Proc. Natl. Acad. Sci. USA 99, 13705–13710. Theopold, U., Rissler, M., Fabbri, M., Schmidt, O. and Natori, S. (1999). Insect glycobiology: a lectin multigene family in Drosophila melanogaster. Biochem. Biophys. Res. Com. 261, 923–927. Thomas, M. B. and Blanford, S. (2003). Thermal biology in insect–parasite interactions. Trends Ecol. Evol. 18, 344–350. Thompson, R. C. A. and Lymbery, A. J. (1996). Genetic variability in parasites and host parasite interactions. Parasitology 112, S7–S22. True, J. R., Edwards, K. A., Yamamoto, D. and Carroll, S. B. (1999). Drosophila wing melanin patterns form by vein-dependent elaboration of enzymatic prepatterns. Curr. Biol. 9, 1382–1391. Tzou, P., Ohresser, S., Ferrandon, D., Capovilla, M., Reichhart, J. M., Lemaitre, B., Hoffmann, J. A. and Imler, J. L. (2000). Tissue-specific inducible expression of antimicrobial peptide genes in Drosophila surface epithelia. Immunity 13, 737–748. Van Valen, L. (1973). A new evolutionary law. Evol. Theor. 1, 1–30.


47

Wang, Y., Jiang, H. B. and Kanost, M. R. (1999). Biological activity of Manduca sexta paralytic and plasmatocyte spreading peptide and primary structure of its hemolymph precursor. Insect Biochem. Mol. Biol. 29, 1075–1086. Watson, F., Pu¨ttmann-Holgado, R., Thomas, F., Lamar, D., Hughes, M., Kondo, M., Rebel, V. and Schmucker, D. (2005). Extensive diversity of Ig-superfamily proteins in the immune system of insects. Science 309, 1874–1878. Weber, A. N. R., Tauszig-Delamasure, S., Hoffmann, J. A., Lelievre, E., Gascan, H., Ray, K. P., Morse, M. A., Imler, J. L. and Gay, N. J. (2003). Binding of the Drosophila cytokine Spa¨tzle to Toll is direct and establishes signaling. Nat. Immunol. 4, 794–800. Webster, J. P. and Woolhouse, M. E. (1998). Cost of resistansb: relationship between reduced fertility and increased resistance in a snail–schistosome host–parasite system. Proc. R. Soc. Lond. B 266, 391–396. Weiser, J. and Zizka, Z. (2004). Brachiola gambiae sp n. the microsporidian parasite of Anopheles gambiae and A-melas in Liberia. Acta Protozool. 43, 73–80. Weiske, J. and Wiesner, A. (1999). Stimulation of NO synthase activity in the immunecompetent lepidopteran Estigmene acraea hemocyte line. Nitric Oxide Biol. Chem. 3, 123–131. Werner, T., Borge-Renberg, K., Mellroth, P., Steiner, H. and Hultmark, D. (2003). Functional diversity of the Drosophila PGRP-LC gene cluster in the response to LPS and peptidoglycan. J. Biol. Chem. 278, 26319–26322. Whitten, M. M. A. and Ratcliffe, N. A. (1999). In vitro superoxide activity in the haemolymph of the West Indian leaf cockroach, Blaberus discoidalis. J. Insect Phys. 45, 667–675. Wigglesworth, V. B. (1956). The haemocytes and connective tissue formation in an insect Rhodnius prolixus (Hemiptera). Q. J. Microsc. Sci. 97, 89–98. Wigglesworth, V. B. (1965). The Principles of Insect Physiology. London: Methuen. Wigglesworth, V. B. (1973). Haemocytes and basement membrane formation in Rhodnius prolixus. J. Insect Physiol. 19, 831–844. Willmer, P. (1990). Invertebrate Relationships. Cambridge: Cambridge University Press. Willott, E., Trenczeck, T., Thrower, L. W. and Kanost, M. R. (1994). Immunochemical identification of insect hemocyte populations: monoclonal antibodies distinguish four major hemocyte types in Manduca sexta. Eur. J. Cell Biol. 65, 417–423. Wilson, K., Cotter, S. C., Reeson, A. F. and Pell, J. K. (2001). Melanism and disease resistance in insects. Ecol. Lett. 4, 637–649. Wilson, K., Thomas, M. B., Blanford, S., Doggett, M., Simpson, S. J. and Moore, S. L. (2002). Coping with crowds: density-dependent disease resistance in desert locusts. Proc. Nat. Acad. Sci. USA 99, 5471–5475. Wyers, F., Petitjean, A. M., Dru, P., Gay, P. and Contamine, D. (1995). Localization of domains within the Drosophila Ref(2) P protein involved in the intracellular control of sigma rhabdovirus multiplication. J. Virol. 69, 4463–4470. Yamashita, M. and Iwabuchi, K. (2001). Bombyx mori prohemocyte division and differentiation in individual microcultures. J. Insect Physiol. 47, 325–331. Yan, G. and Severson, D. W. (2003). Dynamics of molecular markers linked to the resistance loci in a mosquito–plasmodium system. Genetics 164, 511–519. Yan, G., Severson, D. W. and Christensen, B. M. (1997). Costs and benefits of the mosquito refractoriness to malaria parasites: implications for genetic variability of mosquitoes and genetic control of malaria. Evolution 51, 441–450. Yoshida, H., Kinoshita, K. and Ashida, M. (1996). Purification of a peptidoglycan recognition protein from hemolymph of the silkworm, Bombyx mori. J. Biol. Chem. 271, 13854–13860.

48


Zacharuk, R. Y. (1970). Fine structure of the fungus Metarhizium anisopliae infecting three species of larval elaterid (Coleoptera). III. Penetration of the host integument. J. Invert. Path. 15, 372–396. Zasloff, M. (2002). Antimicrobial peptides of multicellular organisms. Nature 415, 389–395. Zhu, Y. F., Wang, Y., Gorman, M. J., Jiang, H. B. and Konost, M. R. (2003). Manduca sexta serpin-3 regulates prophenoloxidase activating proteinases. J. Biol. Chem. 278, 46556–46564.

Antennal Movements and Mechanoreception: Neurobiology of Active Tactile Sensors Erich M. Staudachera, Michael Gebhardtby and Volker Du¨rrc a Center for Computational Biology, Montana State University, 1 Lewis Hall, Bozeman, MT 59717, USA b Lehrstuhl fu¨r Zoologie, TU Mu¨nchen, Lichtenbergstr. 4, D-85748 Garching, Germany c Abteilung Biologische Kybernetik und Theoretische Biologie, Fakulta¨t fu¨r Biologie, Universita¨t Bielefeld, Postfach 10 01 31, D-33501 Bielefeld, Germany

1 2

3

4

5

6 7 8

Introduction 50 The antennal motor system 51 2.1 Morphological types and model systems 51 2.2 Biomechanical and kinematic considerations 59 2.3 Efferent innervation of the antenna 65 2.4 Physiology of antennal motoneurons and muscles in insects 70 Sensory physiology of antennal mechanoreceptors 74 3.1 Sensory structures and transduction 74 3.2 Distribution of mechanoreceptors 84 3.3 Physiology of antennal mechanosensory neurons 100 Neuroanatomy of antennal mechanosensory and motor pathways 103 4.1 Mechanosensory neuropils 103 4.2 Connections to other parts of the CNS 109 4.3 Immunocytochemistry 113 4.4 Antennal motoneurons in insects 118 Central neurophysiology of antennal mechanoreception 122 5.1 Processing of antennal mechanosensory information by local brain interneurons 122 5.2 Descending antennal mechanosensory interneurons 126 Behavioural physiology of the antennal tactile sense 129 6.1 Passive sensing 130 6.2 Active sensing 143 Biomimetics and ‘antennal engineering’ 168 7.1 Steering insects and robots 169 7.2 Engineering of active tactile sensors 170 Conclusions 172 8.1 The antennal tactile sense of insects and Crustacea 172 Present address: Lehrstuhl fur Zoologie, TU Munchen, Lichtenbergstr. 4, D-85748 Garching, ¨ ¨

Germany. y Present address: Institut fu¨r Zoologie, Abteilung fu¨r vergleichende Neurobiologie, Universita¨t Bonn, Poppelsdorfer SchloX, D-53115 Bonn, Germany. ADVANCES IN INSECT PHYSIOLOGY VOL. 32 ISBN 0-12-024232-X DOI: 10.1016/S0065-2806(05)32002-9


50

E.M. STAUDACHER ET AL

8.2 Active and passive mechanical sensing 174 8.3 Levels of behavioural complexity 176 Acronyms and abbreviations 178 Acknowledgements 179 Note added in proof 180 References 180

1

Introduction

The tactile sense provides animals with vital information about their immediate environment, comprising cues on object location and surface properties. In insects and other arthropods, the antennae are probably the most important part of this tactile sense. The paired insect antenna is regarded as a true, segmented limb arising from the cranium, although there is still some debate about the precise number of segments in the head of the Mandibulata (Crustacea, Myriapoda and Insecta; Schmidt-Ott and Technau, 1992; Schmidt-Ott et al., 1994; Damen et al., 1998; Abzhanov and Kaufman, 1999; Boyan and Williams, 2002). Hennig (1986) uses the name Antennata as a synonym for the Mandibulata, expressing the fact that this taxon comprises all living arthropod species with antennae. The most ancient arthropods, the extinct trilobites, also carried antennae, which is why the 1st pair of antennae is typically considered to be the first pair of limbs, i.e. articulated appendages, in the ground plan of all arthropods. Accordingly, the Chelicerata are thought to have secondarily lost their antennae. Within the Mandibulata, the Crustacea carry two pairs of antennae: the antennules (first pair) are appendages of the deutocerebral segment and are considered to be homologous to the insect antenna, whereas the second pair of antennae are the appendages of the so-called tritocerebral segment (Hennig, 1986; Brusca and Brusca, 1990; Telford and Thomas, 1998; Boyan et al., 2002). Insects have lost the second pair of antennae during evolution. Antennae are equipped with various types of mechanoreceptors, including large numbers of exteroreceptive sensory hairs, but also proprioreceptors such as internal chordotonal organs, external hair plates, and campaniform sensilla. Being moveable, the antennae can actively sample the space surrounding the animal, thus aiding the localisation of obstacles, recognition of conspecifics or predators, active tracking of objects, or probing of surface structures. Because both passive and active movements can occur during tactile sensing, the tactile sensory system requires integration of proprioreceptive and exteroreceptive information in order to locate the stimulus source. Moreover, self-induced sensory stimulation may have to be predicted from motor commands, and compared with the actual sensory input. Therefore, understanding the neurobiology of the tactile sense of insects and crustaceans requires an integrative view of the motor physiology and biomechanics on the one hand, and of the

ANTENNAL MOVEMENTS AND MECHANORECEPTION

51

neurophysiology of the participating sensory and interneuronal structures on the other. In the light of this knowledge, the functional significance of the various system components can be interpreted for a range of natural behaviours. The last integrative view on insect antennae that included all known sensory modalities was given by Schneider (1964). Since then, much progress has been made, not only in the field of insect olfaction (e.g. Hansson, 1999) and the sensory physiology of mechanoreceptors (e.g. McIver, 1975, 1985; Zacharuk, 1985), but also in elucidating key components in tactile behaviours such as the localisation of obstacles or physical interactions in intraspecific communication. To illustrate this body of knowledge, and to extend and contrast the mentioned reports on mechanoreception, it is the aim of the present review to combine aspects of antennal biomechanics, mechanoreception, processing of tactile information, and the behavioural repertoire involving the antennal tactile sense. Each of these aspects will be treated in separate sections, concentrating on prominent animal model systems, such as the cockroach (mainly Periplaneta americana L. and Blaberus sp.), stick insect (Carausius morosus Br.), cricket (Gryllus sp., Teleogryllus sp. and Acheta domesticus L.), locust (Locusta migratoria L. and Schistocerca gregaria Forska˚l) and the honeybee (Apis mellifera L.). For comparison with a related animal group that employs very similar, though not homologous, organs in tactile sensing, we will expand our view on decapod crustaceans, notably crayfish and lobsters. Finally, we discuss the potential of active tactile sensing from a technical perspective, to point at a helpful mutual exchange between engineering and animal physiology.

2

The antennal motor system

To set the framework for discussing the anatomical and physiological results, the following sections present an overview of antennal morphology (Section 2.1); the basic physical properties of antennae with respect to tactile sensing (Section 2.2); the components of the contractile machinery (Section 2.3) and their physiological properties (Section 2.4). 2.1

MORPHOLOGICAL TYPES AND MODEL SYSTEMS

By far the most physiological studies have dealt with a small number of animal model systems, whose anatomical and physiological properties are considered to be representative for many related species. The following sections introduce the representative groups of insect model systems (Section 2.1.1). For comparison with an analogous tactile organ, Section 2.1.2 will then introduce the second antenna of decapod Crustacea.

52

2.1.1


Insect model systems

In principle, there are two types of antennae in insects: the segmented and annulated type, that differ primarily in articulation and musculature of the distal segments. In both types, the first two segments are called scape and pedicel. These two are followed by a varying number of further, distal sections (Matsuda, 1965; Weber and Weidner, 1974; Zacharuk, 1985; Brusca and Brusca, 1990; Boyan et al., 2002). 2.1.1.1 The segmented antennae of the Entognatha. Segmented antennae are found in Entognatha. They consist of scape, pedicel and a number of more or less uniform articulated segments. Antennal muscles arise from all except the last segment, thus all articulations are true joints. For example, detailed descriptions of the antennal musculature of the dipluran genera Japyx and Campodea, and of the collembolan genus Orchesella are given by Imms (1939). The Collembola have a tendency to reduce the number of antennal segments. The pedicel of segmented antennae lacks Johnston’s organ (Matsuda, 1965; Zacharuk, 1985), a chordotonal organ that is characteristic of the antennae of higher insects (see Section 3.2.3.3). Owing to the serial arrangement of segmental muscles in the flagellum of the Entognatha, repetitive sets of motoneurons presumably exist that innervate these muscles. To our knowledge, there is no information available concerning antennal motoneurons of these insects. 2.1.1.2 The annulated antennae of higher insects. Annulated antennae are found in the Ectognatha (bristletails, silverfish and winged insects). They consist of three functional segments (Fig. 1), the scape, pedicel, and the flagellum or clavola. Paralleled by the functional importance of a strong head–scape (HS) musculature, the Ectognatha have evolved an increasingly complex internal head skeleton for its proximal insertion: the tentorium. It consists of the posterior tentorium that is common to all Ectognatha, the anterior tentorium of the Dicondylia and the corpo-tentorium of the Pterygota (Koch, 2000). Apart from antennal muscles, some muscles of the mouthparts also insert on the tentorium. Currently, it is still debated whether scape and pedicel are homologous to coxa and trochanter, respectively, of an ancestral appendage (Haas et al., 2001; Boyan et al., 2002). The flagellum has a varying number of annuli or meristal segments, the morphology of which is not necessarily uniform (Weber and Weidner, 1974; Brusca and Brusca, 1990). The annuli are not true segments because they are not articulated by true joints, i.e. they do not contain muscles (Imms, 1939). Since the flagellum in its entirety essentially functions as a multisensory probe, it can be considered as a third segment in functional terms (e.g. Lee and Strausfeld, 1990; Kloppenburg et al., 1997; Du¨rr et al., 2001), if not in morphological terms. As the pedicel does not contain muscles and, thus, the flagellum cannot be actively moved relative to the pedicel, the connection between pedicel and flagellum will be called pedicel–flagellum (PF) junction, to


53

FIG. 1 Antennal joints and segments: (A) Schematic top view of an insect head. The antennae consist of three functional segments: the scape (Sc), pedicel (Pd) and an annulated flagellum. In all model systems discussed in this review, the annuli of the flagellum are approximately cylindrical and fairly uniform. The articulations head–scape (HS) and scape–pedicel (SP) are true joints (dark bars), and can be moved by musculature. The junction between the pedicel and the flagellum (PF, light bar) allows only passive deflection. (B) Crayfish and lobsters (Astacidea) and spiny lobsters (Palinurida) carry two pairs of antennae: The 1st pair are called antennules, the 2nd pair is called antennae. The basal segments of the antenna, coxopodite (C) and basipodite (B), connect to the multi-segmented endopodite that includes the long flagellum (distal to CF). In the Astacidea, the basipodite connects to a second, smaller branch, the exopodite,or scaphopodite (dotted). The latter is reduced in Palinurida. There are five functional joints: The two basal articulations ‘head’-coxopodite (HC) and coxopodite–basipodite (CB), as well as the distal articulations meropodite–carpopodite (MC) and carpopodite–flagellum (CF) are simple joints with one joint axis (dark bars). The BIM-joint of the Astacidea is a complex articulation involving the basipodite–ischiopodite and ischiopodite–meropodite joints of the endopodite and the basal joint of the exopodite (dotted rectangle, containing three light bars). In the Palinurida, the first three segments are fused. As a consequence, the CB-joint is absent and the BIM-joint is only formed by the ischiopodite and meropodite.

emphasise the difference to a true joint. Variations of the morphology and number of the flagellar annuli account for the different types described for annulated antennae (Table 1; e.g. filiform, setaceous, lamellate; for illustrations cf. Weber and Weidner, 1974; Zacharuk, 1985; Brusca and Brusca, 1990). The first, i.e. the most proximal, joint of the annulated antenna is the articulation of the scape with the head cuticle (HS-joint). The second joint connects the scape to the pedicel (scape–pedicel SP-joint, Fig. 1A). The HS joint is constructed differently in different insect groups. A hinge-joint occurs in locusts, crickets and stick insects (Gewecke, 1972b; Honegger et al., 1990a; Du¨rr et al., 2001), whereas a ball-and-socket joint is found in other insects. The axis of the hinge-joint is determined by two points of articulation of the scape with the head, thus restricting the movements of the scape to one plane.

54


TABLE 1 The morphology of the antennae of the insect model organisms Species Periplaneta americana L., female P. americana L., male

Type of antenna

Number of annuli

Filiform

139726

Filiform

135733

Acheta domesticus L., 13th instar

Filiform

150 255

Locusta migratoria, L. Schistocerca gregaria Forska˚l Carausius morosus Br.

Filiform Filiform

24 24

Filiform

40–46 40–46

Apis mellifera L., female Geniculate A. mellifera L., male Geniculate

10 11

A. mellifera L., worker

10

Geniculate

Length mm

Reference(s)

42

Schafer and Sanchez (1973) 42 Schafer and Sanchez (1973) 50 Toh (1977) 30 FudalewiczNiemczyk and Rosciszewska (1973) 14 Gewecke (1972) 14 Ochieng et al. (1998) 39 Slifer (1966) 35 Monteforti et al. (2002) Not reported Snodgrass (1956) Snodgrass (1956) 3.9 Esslen and Kaissling (1976) Slifer and Sekhon (1961) 2.4 Esslen and Kaissling (1976)

In locusts, a protrusion of the head cuticle, the antennifer (Snodgrass, 1935), contributes to form the lateral point of articulation which is more rigid than the median articulation point (Gewecke, 1972b). This configuration requires some initial contraction of the extrinsic antennal muscles, i.e. the ones moving the HS-joint, to tense the membrane of the median articulation before it can function as the second pivot of the HS-joint. In Locusta, this joint allows for 601 levation and 501 depression. The scape of Locusta is rotated laterally by 401 with respect to the sagittal plane such that, at rest, the scape points anterolaterally (Gewecke, 1972b). In crickets too, this plane of movement allows levation and depression of the scape, but is arranged parallel to the sagittal plane (Honegger, 1981; Honegger et al., 1990a). In stick insects, the axis of the HSjoint is rotated dorsally by 1491 relative to the horizontal plane, and medially by 101 relative to the frontal plane. The angular range of the HS-joint in walking stick insects is about 1001, as calculated from behavioural measurements by Du¨rr et al. (2001). The HS-joints of the cockroach (Periplaneta americana), the honeybee, ants, and the moth Manduca are different from the HS-joints of locusts, crickets and stick insects in that they work like a ball-and-socket-joint, allowing movements in more than one plane (Snodgrass, 1956; Kloppenburg, 1995; Ehmer and Gronenberg, 1997a; Kloppenburg et al., 1997). In Periplaneta,


55

the HS-joint is described as a ‘‘ball-joint-like structure’’ (Okada and Toh, 2001) that allows movement of the antenna in any direction (Okada and Toh, 2000, 2001). According to Ellerkmann (1994), movements of this joint cover about one-fourth of a semi-sphere in bees, when moving dorsally, ventrally or laterally. Medial movements do not cross the midline. The scape of Manduca is free to pivot into all directions (Kloppenburg et al., 1997). In all insect species where the SP-joint has been investigated, it is a hinge-joint which moves the pedicel, together with the flagellum, within a single movement plane (locust: Snodgrass, 1956; Gewecke, 1972b; cricket: Honegger, 1981; Honegger et al., 1990a; stick insect: Du¨rr et al., 2001; honeybee: Kloppenburg, 1995; ant: Ehmer and Gronenberg, 1997a; Manduca: Kloppenburg et al., 1997; see also Fig. 2). In the locust, one dorsal and one ventral point of articulation form the pivots of the hinge, allowing for 451 abduction and 451 adduction (Gewecke, 1972b). In crickets, the SP-joint is also composed of a dorsal and a ventral pivot, and allows horizontal movement of the pedicel/flagellum. Here, the pedicel locks against the scape during abductions at 901 lateral relative to the body long axis. Movements further laterally require the scape to be lifted. Abductions beyond 801–851 occur only rarely during natural antennal movements (Honegger, 1981). Therefore, in locusts and crickets, the two antennal joint axes are approximately orthogonal to each other. This is different in stick insects, where the axis of the SP-joint axis is rotated dorsally by 281 with respect to the horizontal plane. Therefore, the joint axes of the stick insect antenna are not orthogonal to each other. With the HS-joint axis being rotated by 1491, the smaller angle between the HS-joint and the SP-joint axes is 591 (Du¨rr et al., 2001; see also Section 2.2.1). In the honeybee and in ants, the SP-joint resembles an ‘elbow-like’, dicondylic hinge joint that allows flexion and extension in a single plane (Snodgrass, 1956; Ehmer and Gronenberg, 1997b). Kloppenburg et al. (1997) describe the movement of the SP-joint of Manduca as ‘‘more limited [with respect to the HS-joint], hinge-like’’. 2.1.2

Crustacean model systems

It is the 2nd pair of antennae that is typically referred to when speaking of the ‘crustacean antenna’. This convention will be kept throughout this review. Being the larger and more robust pair of antennae, it is this pair of appendages that is particularly relevant to the tactile sense of decapod crustaceans. Accordingly, this review will nearly exclusively cover the second pair of antennae in Crustacea. Most of the available information on crustacean antennal movements concerns clawed species of crayfish and true lobsters on the one hand (Astacidea), and the clawless spiny lobsters on the other (Palinurida). Within the Astacidea, all of the presented information has been collected on the marine lobster genus Homarus and on species of three families of crayfish, belonging to the European Astacidae (Astacus), the American Cambaridae (Cambarus,

56


FIG. 2 Kinematics of the antenna: (A) In the simplest case, an insect antenna can be modelled as a kinematic chain consisting of the head (box), two segments (heavy black lines) and two hinge joints (open circles) with single joint axes (dashed arrows). With respect to the head coordinate system (xH, yH, zH), the orientation of the HS-joint axis can be described by two Euler angles (jS, cS). The angle of the HS-joint is a. Assuming that the scape S is perpendicular to the HS-joint axis, the scape coordinate system (xS yS, zS) allows description of the SP-joint axis orientation by two further Euler angles (jPF, cPF). They determine the resting posture of the pedicel and flagellum (PF) that can be modelled as a single segment. The SP-joint angle is b. In some insect orders, e.g. in Hymenoptera, the HS-joint is a ball-and-socket joint with three degrees of freedom of rotation. The inserted coordinate system indicates the actions of the joints (italic) with respect to the head. (B) Crickets and locusts (e.g. Gryllus, Locusta) levate/depress the scape using a horizontal HS-joint axis and adduct/abduct the pedicel using a vertical SP-joint axis. In the stick insect (Carausius), the joint axes are rotated compared to the locust. In the cockroach (e.g. Periplaneta), the HS-joint mainly allows for adduction/ abduction (bold), but also levation/depression. To what extent the scape can also be


57

Procambarus and Orconectes) and the Australian Parastacidae (Cherax and Euastacus). Within the Palinurida, all studies concern the genera Palinurus and Panulirus. The decapod antenna consists of two basal segments, the coxopodite and basipodite, followed by two branches of segments: the lateral exopodite and the medial endopodite (Fig. 1B). The endopodite consists of four segments: the ischiopodite, meropodite, carpopodite and flagellum. With one exception, each pair of adjacent antennal segments forms a single joint between them. The exception is a complex joint formed by articulations of the basipodite, ischiopodite and meropodite (see below). The latter can be viewed as one functional joint which will subsequently be called the BIM-joint. Thus, the proximal to distal sequence of joints connecting the ‘head’ (cephalothorax) to the basal segments and the segments of the endopodite will be called HC-, CB-, BIM-, MC- and CF-joints (Figs. 1B and 2B). The exopodite of the Eumalacostraca, a taxon that contains the Decapoda, is the scale-shaped scaphopodite. The secondary reduction of this scaphopodite is a synapomorphy of the Palinurida (Hennig, 1986). Associated with this reduction is a different arrangement of the BIM-joint (see below). In the Astacidea, the proximal segment of the scaphopodite acts as the antagonist of the BIM-joint and, thus, is instrumental to rotational movements of the endopodite. The most detailed description of the anatomy and kinematics of the decapod antenna was given by Habig and Taylor (1982a) for crayfish of the genera Procambarus and Cambarus (Figs. 1B and 2B). The most basal HC-joint rotates the coxopodite in a vertical plane relative to the head. It is moved by two muscles that act as a depressor and a levator. The CB-joint rotates the basipodite in a horizontal plane relative to the coxopodite. It is moved medially by four adductors and laterally by one abductor.1 Two adductors of the CB-joint are bi-articulate muscles.

continued

rotated around its long axis is uncertain. The honeybee (Apis) has a geniculate antenna with a three-degree-of-freedom HS-joint. The pedicel is flexed/extended using the SPjoint. Note that segment lengths are not drawn to scale: the relative length of the scape is much exaggerated, and species differ greatly in flagellum length. Compared to the insect model systems, crayfish and lobsters have much more complex antennal kinematics. The HC- and CB-joints have limited action range. They abduct/adduct and levate/depress the endopodite, respectively. The BIM-joint rotates the endopodite around its long axis (bold arrow), thus slanting the movement plane of the flagellum. The two distal MCand CF-joints have nearly parallel joint axes, and act synergistically to flex/extend the flagellum. 1

Note that Habig and Taylor (1982a) use a different nomenclature than is used here, to account for the serial homology with leg musculature: depressor and levator are synonymous to Habig and Taylor’s remotor and promotor, respectively. Adductor and abductor are synonymous to Habig and Taylor’s depressor and levator, respectively.

58


The basipodite connects to two distal segments: the first segments of the endopodite and exopodite, respectively. Both articulations control a roll movement of the endopodite in an antagonistic manner. The remotor muscle of the basipodite rotates the exopodite medially (BE-joint) and presses it against the lateral surface of the meropodite. This causes the meropodite to rotate outward, i.e. the lateral side moves to a ventral position. The two ischioceritis muscles and the two meroceritis muscles of the BIM-joint synergistically cause an inward rotation of the entire endopodite. One meroceritis muscle is a bi-articulate muscle, spanning the entire BIM-joint. In essence, the complicated arrangement of the BE- and the BIM-joints makes them antagonistic components of one functional rotator joint, allowing the endopodite roll around its long axis by some 901. Together with the two basal joints, this allows all three degrees of freedom of rotation of the antenna, although the working ranges of the HC- and CB-joints are relatively small. The distal MC- and CF-joints have approximately parallel joint axes, allowing them to act as synergists with an expanded common flexion/extension range of the flagellum of almost 1801. Both joints are moved by a pair of flexor/extensor muscles. Upon full extension, the flagellum points straightforward, parallel to the body long axis. Full flexion rotates the flagellum to point rearward, where the arc described by the tip is a semi-circle above or lateral to the animal, depending on the rotation angle of the BIM-joint. Thus, the function of the three basal joints of the crayfish antenna is to orient the extension/flexion plane of the flagellum between a horizontal plane and a vertical plane. The musculature of the distal antennal segments in the crayfish Euastacus armatus is similar to that of cambarid species, with the addition of a small accessory extensor muscle spanning the MC- and CF-joint (Sandeman and Wilkens, 1983). For the distal segments of the lobster antenna (Homarus americanus), Sigvardt (1977) gives the joint angular ranges as 1041 for the MCjoint and 951 for the CF-joint. Other than in crayfish, the MC- and CF-joint axes are not parallel, but form an angle of 171. In its resting posture, the meropodite of the Homarus antenna is rotated 901 medially by the BIM-joint, so that the flagellum normally moves in a nearly vertical plane. In the more distantly related palinurid species, the major difference to the crayfish antenna is probably related to the reduction of the scaphopodite. In spiny lobsters the first three segments are fused (Balss, 1940; cf. Sigvardt, 1977), simplifying the complex BIM-joint to a single joint between the ischiopodite and meropodite. In the rock lobster Palinurus vulgaris, angular ranges of the three distal antennal joints are 801 for the inward/outward rotation about the oblique IM-joint axis (Clarac et al., 1976), and 601 each for flexion/extension about the parallel MC- and CF-joint axes (Clarac and Vedel, 1975). The IM-joint is moved by three internal rotator muscles and two external rotator muscles (Clarac et al., 1976). Both distal joints are moved by a pair of flexor/extensor muscles, and an additional accessory extensor muscle spans both joints (similar to Euastacus).


2.2

59

BIOMECHANICAL AND KINEMATIC CONSIDERATIONS

To allow a combined functional interpretation of anatomical, physiological and behavioural findings on antennal movements and mechanoreception, it is helpful to discuss some theoretical aspects of the various antennal ‘designs’ that have evolved (Section 2.2.1), but also the physical framework that the nervous system has to deal with when employing a flagellum as a tactile sensor (Section 2.2.2).

2.2.1

Kinematics, workspace and tactile acuity

Physiological limitations for tactile localisation of a contact point are set by three aspects. First, the spacing of sensilla on the flagellum (see Section 3.2), and the angular resolution of the proprioreceptive coding set the upper bound for the sensory resolution of surface structures and external coordinates of a contact point. Second, there is a limit to positioning accuracy set by the antennal musculature and its motoneurons, affecting the reliability of trajectories or sampling positions in repetitive contacts. Third, the properties of the kinematic chain, i.e. the sequence of joint axes and segments, determine the workspace of the flagellum, but also affects positioning accuracy within the workspace. Concerning the accuracy of proprioreceptive coding on the one hand (Sections 3 and 5), and positioning accuracy on the other (see Section 2.4), there is virtually no information as to how precise the nervous system can control these aspects. Concerning the significance of kinematic parameters on the sampling strategy, Du¨rr and Krause (2002) have offered a preliminary framework to assess the physical limitations set by the ‘construction’ of a given antenna (Fig. 2A). Following a suggestion on the functional significance of slanted joint axes in the stick insect Carausius morosus (Du¨rr et al., 2001), they show how the angle between the HS- and SP-joint, the inter-joint angle, influences the shape of the antennal workspace. In case of orthogonal joint axes, the workspace has a torus-like shape, where the length of the scape, determines the deviation from a sphere (the deviation being larger with increasing length of the scape). As the inter-joint angle becomes larger, the holes of the torus-shaped workspace become larger. Thus, the volume out of reach of the flagellum increases (Du¨rr and Krause, 2002; Krause and Du¨rr, 2004). On the other hand, provided that there is a limit of angular accuracy in each antennal joint (both sensory and motor), and noise in both joints is independent of each other, the impact of this inaccuracy on positioning decreases with the inter-joint angle (Du¨rr et al., 2001). In conclusion, slanting the joint axes to narrow the inter-joint angle gives rise to an out-of-reach zone, but increases tactile accuracy in the withinreach zone.

60


Slanting the axis of the flagellum on the SP-joint axis introduces an asymmetry into the torus-shaped workspace. The result of this asymmetry is that the out-of-reach zone narrows on the side to which the flagellum points, and widens on the other. Thus, assuming that the flagella of an insect were slanted laterally, i.e. diverge with respect to the midline, the widened out-of-reach zone would be located at the posterio-median side of the workspace (facing towards the head). Assuming bilaterally symmetrical antennae, the medially widened out-of-reach zone could be compensated by the presence of the contralateral antenna. It follows that the kinematic chain that describes the mobility of an antenna, also describes some physical limits that are likely to be of behavioural significance (Krause and Du¨rr, 2004). The kinematic chains of the model systems discussed in this review are illustrated in Fig. 2B. Among the three model systems with two hinge joints (locust, cricket and stick insect) the major differences lie in the length of the flagellum (shortest in the locust) and in the inter-joint angle (non-orthogonal in the stick insect). It is reasonable to assume that the axes of the stick insect are slanted for an evolutionary reason. However, whether or not stick insects benefit from an improved tactile acuity in a dorso-ventral direction, as the joint axis orientations would suggest, remains to be shown. Also, environmental constraints may bias the benefits of one construction over another. The ball-and-socket construction of the HS-joint in cockroaches and bees does not allow the full three degrees of freedom of rotation that a genuine ball joint, like the human shoulder joint, would allow. Rather, the degree of rotation around the axis of the scape is limited, due to the mechanical constraints of the joint membrane and by the small set of four HS muscles. In effect, the action range of these HS-joints may be more similar to that of a Cardan joint rather than a ball-and-socket joint, mainly allowing for levation/depression and abduction/adduction. The introduced redundancy does not add much mobility compared to the hinge-joint construction. One implication of the ability to rotate the flagellum around the long axis of the scape would be that a given point in the workspace could be approached from any angle, and that the contact surface could be adjusted. In a hinge-joint or Cardan construction, the contact surface is determined by the direction of movement. Although the range of rotation around the scape is not addressed explicitly by any of the available studies, Heran (1959) describes postural adjustments in the flying honeybee which imply such rotations. Thus, it is likely, that the honeybee can actively rotate the scape to touch an object with a particular part of the flagellar surface (see Section 3.2 for distribution of sensilla and Section 6.2.3 for tactile sensing in honeybees). Crayfish and lobsters can actively rotate the flagellum around its long axis by means of the BIM-joint. It acts to orient a movement plane, within which the MC- and CF-joints adjust the orientation of the flagellum. Owing to the small action ranges of the joints proximal to the BIM-joint, crayfish and lobsters have little freedom to adjust the movement plane other than by rotating it.


61

Therefore, the preferred contact surfaces of the flagellum in decapod crustaceans are the median and lateral surface (with respect to a resting posture with a horizontal movement plane). Even if a lobster antenna hits an opponent from above, i.e. by rapid extension of the MC- and CF-joints, it will hit him with the medial surface that is rotated by the BIM-joint to face downward. 2.2.2

Biomechanics of the flagellum

Antennal mechanoreception occurs in two ‘mechanical contexts’. In the ‘noncontact context’, the antenna is deflected or vibrated by flow or pressure oscillations of the surrounding medium. In the ‘contact context’, the antenna touches or is touched by a solid object, and is either bent by the applied force or oscillates upon release of the imposed deflection. Although both contexts have been studied, there is more detailed information available on non-contact deflections. All of the available information concerns the deflections of the flagellum and the corresponding torque or vibration transmitted to the PFjunction. This is because the flagellum has both the largest surface and the largest lever arm of all antennal segments. Moreover, it is the PF-junction that has been shown repeatedly to be the crucial site for mechanoreception in this context (Heran, 1959; Gewecke, 1970; Saager and Gewecke, 1989; Dreller and Kirchner, 1993a; see Section 6.2.3). In the ‘contact context’, under natural conditions, the site of object contact typically lies on the flagellum. Even in the honeybee, where the length of the scape relative to the total length of the antenna is much longer compared to the other model systems, the geniculate posture of the antenna makes contacts by the scape unlikely. Thus, it is important to understand the mechanical properties of the flagellum, and how they modify the mechanical stimulus received by the sensory structures. 2.2.2.1 Static properties. To our knowledge, there is no information about specific properties of the antennal cuticle. For instance, variations of cuticle thickness along a flagellum have not been measured. All of the studies available concern deflection properties of the flagellum. Antennal deflection induced by air flow is of behavioural relevance to all flying insects. In response to steady frontal air flow, the flagellum of the locust Locusta migratoria behaves similar to a glass cylinder of equal proportions (i.e. 13.9 mm length, 0.25 mm diameter and 3.17 mm2 exposed surface area). The torque increases with antennal angle, a, by less than sin(a), but by more than sin2(a), irrespective of air speed (Gewecke and Heinzel, 1980). However, the dependency of torque, T, on air speed, V, does not follow Newton’s law of aerodynamic drag (the relation is TV1.44 rather than TV2), indicating that frictional forces acting on the flagellum are not negligible. Rather, friction gains importance with increasing air speed. Moreover, because the estimated drag coefficient of the flagellum is always larger than that of an ideal, infinitely long cylinder, Gewecke and Heinzel (1980) argue that the force acting on the

62


flagellar tip is significant. As the locust decreases the antennal angle with increasing air-speed (see Section 6.1.3.1), it actively decreases the torque acting on the PF-junction and, thus, maintains the angular deflection at the base of the flagellum, d, within an almost linear range (dV1.2). In other words, the locust actively adjusts the antennal posture so that the restoring torque of the PF-junction approximately maintains the properties of a linear spring. The stiffness of the locust flagellum is not uniform in all directions (Gewecke and Heinzel, 1980). Pulling the tip of the flagellum into different directions by the same force causes forward deflections that are three times larger than the corresponding rearward deflections. Thus, the locust flagellum is least compliant to forces applied in the direction of air flow during flight. This direction also matches the locust’s preferred contact direction during active obstacle contacts: locusts often touch an object repeatedly with the medial surface of the flagellum (Saager and Gewecke, 1989). The above observations also apply to the flagellum of the honeybee, where stiffness is smallest in the ventral direction and largest in the dorsal direction (Heran, 1959). During flight, the honeybee holds its flagellum slightly elevated above the horizontal plane, a posture that exposes the ventral surface to dorsal deflections by the air current. Much like the locust, the honeybee actively decreases the angle between flagellum and air current with increasing flight speed (Heran, 1959). In a contact context, non-uniform compliance is advantageous, because an external force opposing extension of the flagellum may cause the flagellum to bend, whereas the same force would be resisted during flexion. This is likely to be of functional significance in the honeybee’s active tactile scanning behaviour. When presented with an object within the range of the antennae, honeybees scan the object with a movement sequence that is characterised by fast repetitive flexion of the SP-joint (Erber and Pribbenow, 2000). Owing to its non-uniform compliance, the flagellum remains stiff during object contacts, which is a necessary prerequisite to allow a sensory estimate of the contact point without taking the curvature of the flagellum into account. On the other hand, the lower compliance in the opposite direction reduces the chance of the antenna getting caught by surfaces or edges during extension. The stiffness of the flagellum of crayfish also is non-uniform. The flagellum of Cherax destructor is tapered, consisting of 220–250 annuli that increase in length and decrease in diameter towards the tip (Sandeman, 1989). The flagellum is oval in cross-section and exerts the largest mechanical resistance against medial deflection and the lowest mechanical resistance against dorsal deflection. Medial mechanical resistance is 0.0014 N m rad1 at the base of the flagellum and decreases by almost 1.5 orders of magnitude towards the tip. When bent passively, the curvature is confined to a small portion of the flagellum. With increasing force applied, the apex of the bend moves towards the base (Sandeman, 1989). In Procambarus spiculifer, the mechanical impedance (stiffness x mass) of the tapering flagellum decreases logarithmically with length (Taylor, 1975). If the tip of the flagellum is pressed against a flat surface,


63

the lateral deflection of any point between the most proximal contact point and the base of the flagellum is well described by an exponential function. An important mechanical aspect, which has been largely neglected by biologists, concerns the impact of the antennal shape on the bending characteristics of the flagellum. The importance of this is shown by an engineering study that investigates the difference of contact-induced bending as opposed to drag-induced bending of a tapered probe (Barnes et al., 2001). Transferred to the biological situation of a crayfish antenna, sensing the curvature at several locations along the flagellum is likely to be sufficient for a crayfish to tell whether the bend is caused by flow of the surrounding medium, or rather by physical contact with an obstacle. In an environment with still water, sensing a drag-induced bend must be due to active motion of the antenna, thus allowing the animal to distinguish between active and passive bending. A biological example of the importance of drag forces caused by the antennae is the speed-dependent adduction of antennae in queues of migrating spiny lobsters (Bill and Herrnkind, 1976). During long migration in water, reduction of drag forces can save a lot of energy. Indeed, a queue of five spiny lobsters walking at 0.35 m s1, reduce the summed drag of individuals by some 1.5 N. This is equivalent to more than 60% of drag reduction. Spiny lobsters also adapt their antennal posture to the walking speed of the queue. Compared to a mean abduction angle of 461, the queue of five adducts their antennae to 201, resulting in further drag reduction by 0.3 N. The speed-dependent adduction is reminiscent of antennal adduction of flying locusts and bees (see Section 6.1.3). 2.2.2.2 Dynamic properties. Three experimental paradigms have been applied to study the dynamic properties of the flagellum. The ‘contact context’ has been studied by the application of oscillations by means of a mechanical probe, or by vibration of the head with the flagellum being free to oscillate in air. The ‘non-contact context’ has been studied by exposing the flagellum to an air current of oscillating speed. Generally, studies have measured the passive properties of the flagellum of dead animals. Thus, it is important to note that the properties measured may be subject to active modification by the living animal, as has been shown to be the case in mosquitoes (Go¨pfert and Robert, 2001a). Imposing oscillatory movement upon either the tip or the base of an excised flagellum of Procambarus spiculifer causes travelling waves that take a characteristic form, depending on the contact site (Taylor, 1975). Deceleration of the travelling speed from base to tip sometimes shows profound non-linearities, which Taylor (1975) attributes, at least in part, to the unequal decrease of mass and stiffness from base to tip. Imposing oscillations on a locust’s head causes the flagellum to resonate at frequencies near 74 Hz (female) or 104 Hz (male). Apart from the sex difference, the resonance frequency also depends on age (being lower in young animals) and the direction of the oscillation (being largest in the horizontal plane; Heinzel and Gewecke, 1987). When superimposing the passive

64


oscillations with an air-current directed towards the head of the animal, the peak resonance frequency changes only little, but the amplitude of flagellar oscillations are strongly dependent on the antennal angle with respect to the current (oscillations being larger at large angles). Release of a passive static deflection at the PF-junction by angles between 0.11 and 11 (the sensitive range of the campaniform sensilla on the distal pedicel) causes the flagellum to return to its resting position with a time constant of 10 ms (Heinzel and Gewecke, 1987). If the PF-junction is deflected by a contact point at the middle of the flagellum, the return is aperiodic. If, however, the PF-junction is deflected by flagellar contact at the tip, the return time course is superimposed by damped oscillations at frequencies slightly below the resonance frequency. Thus, free oscillations after obstacle contact contain information about the contact location, a fact that has also been suggested to be exploited in technical applications of tactile sensors (Ueno and Kaneko, 1995; Ueno et al., 1998; see Section 7.2). In the honeybee, passive vibration of the flagellum via oscillation of the head produces peak amplitudes at frequencies between 240 and 300 Hz, which is close to the wing beat frequency of the bee (275 Hz; Heran, 1959). When exposing the flagellum of the locust to a sinusoidally modulated air current (Heinzel and Gewecke, 1987), the shape of the frequency tuning curve of the flagellum depends on whether air-particle displacement or air-particle velocity is kept constant. In the first case, the curve is of similar shape as when the head is vibrated, but the resonance frequency is shifted to higher frequencies by some 50 Hz. If, on the other hand, air current velocity is kept constant, the lower frequency range is less attenuated in comparison with passive vibration of the head. A natural example, where a gregarious locust is likely to encounter an air current of oscillating speed is during flight in a swarm. Indeed, the wing beat of a locust ahead can deflect the flagellum at amplitudes that are sufficient to stimulate the campaniform sensilla on the pedicel (Heinzel and Gewecke, 1987). In the honeybee, mechanical properties of the flagellum are suitable for passive transmission of airborne vibrations with high air-particle velocities (Kirchner, 1994). When exposed to vibration stimuli similar to those that occur near the abdomen of a waggle-dancing honeybee (Section 6.2.4.2), the flagellum follows the air particle movements like a stiff rod held by the pedicel. The amplitude of flagellar vibration is the same in all radial directions and increases linearly with air-particle displacement. The 10 Hz frequency component caused by the wagging movement of the dancing bee, as well as the 265 Hz frequency component caused by vibratory wing movements of the dancer, are transmitted with similar gain. The sound of the waggle dance is transmitted to the pedicel, where it is transduced by Johnston’s organ. As in the honeybee, the antenna of the fruit fly Drosophila melanogaster also acts to transmit airborne sound to Johnston’s organ inside the pedicel. The specialised antenna of the brachyceran flies has a bulbous flagellum with a feathered lever arm, the arista. Behavioural experiments show that female fruit


65

flies perceive the air current and sound produced during the courtship display of a male (Petit, 1958). The arista is an obligatory component for sound perception during courtship, while an intact pedicel, including Johnston’s organ, is not sufficient (Manning, 1967). Fixation of the PF-junction, without affecting the insertion of the arista, diminishes sound perception, showing that the whole of the flagellum acts in concert to transmit airborne sound in Drosophila (Manning, 1967). Recently, Go¨pfert and Robert (2001b) showed that the arista turns the flagellum around its long axis, thus rotating a chitinous hook inside the pedicel, much like a key’s bit inside a lock. The exerted movement inside the pedicel can be sensed. Interestingly, loading of the arista does not affect courtship success (Manning, 1967), suggesting that the resonance frequency can be altered with no impact on female’s ability to perceive male courtship signals, despite the fact that hearing, in a strict sense, should be affected. 2.3 2.3.1

EFFERENT INNERVATION OF THE ANTENNA

Muscle innervation in insects

An overview of the number of antennal muscles and their innervating neurons in Gryllus, Locusta, Carausius, Manduca and Apis is given in Table 2. The innervation of antennal muscles of the insect model systems, introduced in Section 2.1, reflects the basic innervation scheme commonly found in insect musculature. One skeletal muscle is typically innervated by two, sometimes more, excitatory motoneurons and one common inhibitory motoneuron (Hoyle, 1978; for review see Burrows, 1996). The insect antennal motor system differs from that of Crustacea (Section 2.3.2) in that insects lack common excitatory motoneurons. Also, insects have a higher total number of motoneurons innervating the antennal musculature. In addition to motoneurons, efferent modulatory dorsal unpaired median (DUM) neurons ramify on most insect skeletal muscles (for review, see Braünig and Pflu¨ger, 2001). The extrinsic, or tentorio-scapal, muscles of the HS-joint are located in the head capsule and have their proximal insertion points on the tentorium in the head. Distally they are attached to the basal cuticle of the scape via apodemes. The intrinsic, or scapo-pedicellar, muscles of the SP-joint originate on the cuticle of the scape and they are attached to the pedicellar base via apodemes (e.g. Imms, 1939). According to Gewecke (1972b), locusts have two extrinsic and two intrinsic antennal muscles. Imms (1939), on the contrary, depicts bipartite scapal levator and depressor muscles and an additional intrinsic muscle. In the locust antenna, the extrinsic muscles are innervated by the nerve Nervus musculus tentorio scapalis which originates separately from the main antennal nerve, N. antennalis, from the deutocerebrum (Gewecke, 1972b). The N. musculus tentorio scapalis splits into three branches which supply the two scapal muscles and a hair plate between the compound eye and the antennal base. Two nerves,

TABLE 2 Antennal muscles and muscle innervation of the insect model organisms 66

Number of excitatory motoneurons Species Gryllusa,b

Antennal joint HS SP

Locustab,c,d

HS SP

Carausiuse

HS SP

Manducaf,g

Apish,i,j

HS SP

Honegger et al. (1990a). Braünig et al. (1990). c Bauer and Gewecke (1991). d Gewecke (1972). e Du¨rr et al. (2001). f Fleming (1968). g Kloppenburg et al. (1997). h Kloppenburg (1995). i Erber et al. (2000). j Scha¨fer and Bicker (1986). b

Muscles

Fast

Levation Depression Adduction Adduction

3 2 1 1

5 5 1 2/3


1 1 1 1

9


Number of inhibitory motoneurons

Slow

1 (except M5b)

Total number of motoneurons

Number of DUM neurons

18

2

19

2

2 2/1

10

1

2 1 1 1

4 4–5 3 2–4

1

Adduction Adduction

5 2 2

X7 X5

Flexion Extension

4 1–2 1

9 1 3

13–16

X12

Probably none

15

2 E.M. STAUDACHER ET AL

a

HS SP

Movement


67

branching off the N. antennalis at variable positions, supply the intrinsic antennal muscles with motor innervation: one branch, the N. scapalis medianus, innervates the median adductor muscle, whereas the N. scapalis lateralis innervates the lateral abductor muscle (Gewecke, 1972b). The innervation of the antennal musculature is very similar in Carausius (Du¨rr et al., 2001) and Gryllus (Honegger et al., 1990a). For a schematic overview, see Fig. 3. The depressor muscles are innervated by the pure motor nerve N4 (cricket), which leaves the deutocerebrum via a root separate from the root of the main antennal nerve (N1). One of the cricket depressor muscles (M5a) receives additional innervation from the nerve N3, which branches off N1. The levator muscles are supplied by N3, which is a mixed sensory-motor nerve in crickets (Honegger et al., 1990a). The intrinsic scapo-pedicellar muscles receive their motor innervation by branch N4B and by N2, which is also a side branch of N1. N2 and N4 form a conspicuous anastomosis in the scape with single motoneurons projecting across it to innervate their specific target muscle. The motor axons either leave the brain via N4 and project through the anastomosis to the adductor muscle M6 or they leave the brain via N1 and N2 to innervate the abductor muscle M7. The motor branch of N2 contains also sensory axons in crickets and stick insects (Honegger et al., 1990a; Du¨rr et al.,

FIG. 3 Diagrammatic innervation pattern of antennal muscles in stick insects and crickets. The anastomosis in the nerve running between the abductor muscle and the adductor muscle (dotted line) contains the axons of motoneurons leaving the brain via N2 (light grey) and N4 (dark grey). The extension (broken line) of N3 to the depressor muscle (medium grey) is found in crickets only. DC, deutocerebrum.

68


2001). The branching pattern of these antennal nerves varies between different specimens in both species. In the honeybee, the extrinsic muscles are innervated by a motor nerve which branches off the main antennal nerve N1 in the scape. The intrinsic muscles are supplied by nerves that branch off N1 in the scape (Snodgrass, 1956). Similarly, the intrinsic antennal muscles of Manduca are innervated by small side branches of the main antennal nerve within the scape. However, the extrinsic muscles of Manduca do not have a distinct motor nerve, but are innervated by side branches of N1 (Kloppenburg et al., 1997). In locusts, the four antennal muscles are innervated by 19 motoneurons (Bauer and Gewecke, 1991) plus two DUM neurones with their somata in the suboesophageal ganglion (SOG) (Braünig, 1990, Braünig et al., 1990). Bauer (1991) performed immunohistochemical staining of antennal muscles and locust brains. She found proctolin-immunoreactive profiles on all antennal muscles, and up to 12 somata in regions of the deutocerebrum where antennal motoneurons reside. This suggests that proctolin is a modulator in the locust antennal motor system. Extra- and intracellular recordings have revealed that the extrinsic muscles of the locust, which consist of approximately 65 fibres each, receive slow, intermediate, fast and inhibitory innervation. Recordings from the median depressor muscle of the scape predominantly reveal slow motoneurons, whereas fast and slow excitatory junction potentials (EJPs) can be recorded in fibres of the lateral levator muscle. Bauer (1991) and Bauer and Gewecke (1991) interpret these findings in the light of possible functional specialisations: the depressor muscle must overcome the headwind drag during flight by a sustained force production, whereas the lateral levator muscle is phasically active during most resistance reflexes and, therefore, requires fast fibres. The inhibitory innervation of both muscles appears to be also biased, because IJPs were never observed in levator fibres. In crickets, the seven antennal (five extrinsic, two intrinsic) muscles are innervated by a total of 18 motoneurons. Backfills of single antennal nerves, immunohistochemical stainings and electrophysiological experiments show that the adductor muscle (M6) receives innervation by one fast, two slow and one inhibitory motoneuron. The abductor muscle (M7) is innervated, alternatively, either by two fast, two slow and one inhibitory motoneurons, or by three fast, one slow and one inhibitory motoneurons (Honegger et al., 1990a,b). Again, this indicates an asymmetric innervation of two antagonistic muscles of one joint. The inhibitory motoneuron is GABAergic (GABA: g-aminobutyric acid; Honegger et al., 1990b), the excitatory motoneurons are glutamatergic (Honegger et al., 1995). Fourteen of the 17 excitatory motoneurons can be labelled with an antiserum against the neuropeptide proctolin (Bartos et al., 1994), which modulates the contraction properties of antennal muscles (Allgaüer and Honegger, 1993, see Section 2.4). The two intrinsic muscles have been stained for their myofibrillar ATPase under different pH


69

conditions. Both, the adductor and the abductor muscles are subdivided into two subunits each, containing fast fibres in the periphery of the muscle and intermediate as well as slow fibres in the centre (Honegger et al., 1995). In addition, two suboesophageal DUM neurons project into all major antennal nerves and ramify on antennal muscles (Honegger et al., 1990b). In bees, the four extrinsic muscles that move the HS-joint, are controlled by nine excitatory motoneurons, the two intrinsic muscles in the scape are controlled by six excitatory motoneurons (Kloppenburg, 1995). An inhibitory neuron seems to be missing in the bee antennal motor system, based on immunocytochemical evidence (Scha¨fer and Bicker, 1986a). Work on the antennal muscles in bees has focussed on the two intrinsic muscles, since they control fast extension/flexion movements of the flagellum (reviewed by Erber and Pribbenow, 2000; see also Section 6.2.3). The extensor muscle contains 23 individual muscle fibres, which are innervated by three motoneurons (Erber et al., 2000). The flexor muscle consists of two clearly separated fibre groups, a dorsal group of 23 fibres and a ventral group of 20 fibres. A combination of backfills of antennal nerves and extra- and intracellular recordings have identified the ventral fibre group as a fast fascicle with probably one fast excitatory motoneuron. The ventral fibres are also of larger diameter than the dorsal fibres. A single motoneuron controls these fast flexor fibres (Erber et al., 2000). The extrinsic muscles of the antenna are innervated by a minimum of five to six motoneurons in bees (Ehmer and Gronenberg, 1997a). Additionally, two ventral unpaired median (VUM) neurons can be labelled by retrograde muscle stainings, which is similar to the situation in locusts and crickets (Braünig et al., 1990). These motoneurons apparently project via a separate nerve to the tentorio-scapal muscles. 2.3.2

Muscle innervation in crustacea

The musculature of the crayfish antenna is innervated by the main antennal nerve and a number of smaller side nerves, all of which leave the tritocerebrum (Procambarus: Habig and Taylor, 1982a; Cherax: Sandeman and Wilkens, 1983). Extracellular recordings from antennal motor nerves in the rock lobster P. vulgaris reveal four distinct units innervating the CF-extensor and three distinct units innervating the CF-flexor (Clarac and Vedel, 1975). Intracellular muscle fibre recordings reveal both excitatory and IJPs. Each one of the two extensor and flexor muscles is innervated by a ‘dedicated’ tonic motoneuron that terminates at only one muscle, and by the common inhibitor neuron, that innervates all muscles (Vedel, 1980). The CF-extensor receives additional innervation from a phasic excitatory motoneuron. Furthermore, the two extensor muscles share a common extensor motoneuron and the two flexors share a common flexor motoneuron. These common motoneurons reflect the synergistic action of the parallel MC- and CF-joints. The accessory extensor muscle is

70


innervated by a single motoneuron. From nerve recordings, Neil et al. (1982) conclude that at least 11 motoneurons innervate the musculature of the IMjoint in Palinurus. In the lobster Homarus americanus, each muscle of the distal antennal segments (meropodite and carpopodite) is innervated by a dedicated motoneuron (Sigvardt, 1977). The extensors are innervated by two common extensor motoneurons, while the flexors share only one common motoneuron. In the crayfish E. armatus, muscle innervation is almost the same as in the lobster Homarus (Sandeman and Wilkens, 1983). Other than in Homarus, where Sigvardt (1977) found no evidence for a common inhibitor neuron, IJPs can be observed in flexor muscles of Euastacus (Sandeman and Wilkens, 1983). Also, attenuation of EJPs in extensor muscles indicates the presence of inhibitory innervation. In spite of this difference, resistance reflexes in Euastacus and Homarus are very similar. 2.4

PHYSIOLOGY OF ANTENNAL MOTONEURONS AND MUSCLES IN INSECTS

As will be discussed in Section 6, the antennae of insects perform highly variable and adaptable movements in a variety of behavioural contexts. Several studies demonstrate ways in which the performance of the antennal musculature can be modified to obtain more flexibility in the power output and time course of muscular contractions. The functional role of antennal motoneurons and their relationship to contraction properties of their target muscles will be described first for ‘regular’ excitatory motoneurons (Section 2.4.1), followed by inhibitory innervation and other sources of efferent modulation (Section 2.4.2). 2.4.1

Excitatory innervation

Kloppenburg et al. (1999) have for the first time investigated the properties of ion channels of antennal motoneurons. Using in vitro and in vivo patch–clamp methods in a honeybee preparation, they identified four different voltage-activated currents including two K+-currents and two inward currents (Na+, Ca2+). These currents resemble those found in other insect neurons. The ratio between the currents, however, differed between different motoneurons. These differences provide the basis for future work that relates the different physiological properties of different motoneurons to their biophysical membrane properties. Studies on the physiology of antennal muscles are available for ants (Ehmer and Gronenberg, 1997b), locusts (Saager and Gewecke, 1989; Bauer, 1991; Bauer and Gewecke, 1991) and crickets (Honegger et al., 1990a; Allgaüer and Honegger, 1993; Bartos et al., 1994; Bartos and Honegger, 1997; Gebhardt, 2004). The work of Ehmer and Gronenberg (1997b) focuses on functionally relevant muscle properties, i.e. muscle fibre diameter, sarcomere length and


71

myofibrillar ATPase activity, to help explain the production of fast antennal movements. Their major finding is that the lengths of the sarcomeres negatively correlate with maximum speeds of an antennal retraction reflex in eight ant species belonging to three subfamilies. Highest speeds are observed in trap-jaw ants of the genera Odontomachus and Anochetus, which produce angular velocities of up to 12 200 and 13 2001 s1, respectively. Ants of these species have sarcomeres as short as 1.1 and 1.3 mm, respectively. The shortest sarcomeres are localised in the thickest and longest fibres in the periphery of the muscles. This fibre design with many, serially arranged, short sarcomeres should facilitate the generation of high contraction velocities. The peripheral fibres stain strongly against a pH-labile isoform of mATPase, suggesting that a fast mATPase is located in the peripheral fibres. Moreover, the antennal retractor muscle, which withdraws the antenna during the retraction movement, occupies almost 70% of the total volume of the antennal musculature in Odontomachus. This muscle is, therefore, pre-disposed for a high-power output. All these muscle properties show considerable variation between different muscles within a species and between species. Taking their data together, Ehmer and Gronenberg (1997b) conclude that the musculature of ant antennae is tuned to produce a large range of contraction velocities including the very fast movements produced by the large retractor muscle in trap-jaw ants. Saager and Gewecke (1989) report on different antennal reflexes of locusts, including tactile avoidance reflexes and resistance reflexes of the antennae during air-current-induced flight (see Section 6.1.2.1). They find that extracellular myogram recordings of the lateral and median scapo-pedicellar muscles reveal the activity of two to three and, occasionally, three to four units. The muscle potentials are different in size and shape, and their occurrence correlates with antennal movements. For example, the largest amplitude units occurring in the lateral muscle cause large single twitch contractions without complete tetanisation. Bauer and Gewecke (1991) report fast and slow EJPs, together with IJPs in scape muscle fibres. The latter have resting potentials around 40 mV. Fast and inhibitory innervations are different between both intrinsic muscles. In the lateral scape muscle, fast EJPs occur frequently, but IJPs are lacking, whereas in the median intrinsic muscle, slow, intermediate and, rarely, fast EJPs occur together with IJPs. Incremental open-loop electrical stimulation of the antennal nerve, together with registrations of muscular tension, reveal up to five steps of motoneuron recruitment in the force output. This suggests the presence of at least five motoneurons innervating a single muscle. In the median intrinsic muscle (adductor), evoked muscle twitches are facilitated above 10 Hz stimulation frequency; summation occurs beyond 40 Hz; and tetanic contractions are found in response to stimulation frequencies above 100 Hz. Tetanic contractions are followed by post-contractions of up to 5 s duration. The resulting forces are 106 N in magnitude. In the lateral intrinsic muscle (abductor),

72


facilitation, summation and tetanus occur at 4, 20 and 50 Hz, respectively. Forces measured from the lateral muscle are 105–104 N in magnitude. The different properties of the lateral and medial intrinsic muscles may match behavioural requirements in activation speed, in that the median muscle, with its slow innervation, maintains the prolonged antennal flight posture, whereas the fast lateral muscle is involved in fast avoidance reflexes. In locusts, intracellular recordings from scapal muscle fibres, combined with bath applications of proctolin, show that proctolin alone does not alter the resting membrane potential of the muscle fibres, nor does it change the amplitudes of spontaneous EJPs (Bauer, 1990). In contrast, proctolin modulates contractions evoked by excitatory motoneuron activity: both the amplitude and the duration of single contractions are enhanced by bath-applied proctolin. Additionally, proctolin induces a slow tension component underlying single contractions elicited by trains of electrical stimuli. Bauer (1991) also finds that electrically quiescent extrinsic muscles exposed to proctolin for hours respond with an increase of their basal tension in a concentration-dependent manner. Bauer (1991) concludes that such an increase in basal tension is useful whenever muscular long-term contractions are required, particularly during the fixed antennal flight posture. In crickets, the presence of fast and slow motor units was demonstrated by Honegger et al. (1990a). Extracellular stimulation of an adductor motoneuron (Ad1/2) leads to fast, twitch-like contractions of the antennal adductor muscle, whereas stimulation of another Ad1/2 motoneuron at stimulus frequencies above 40 Hz leads to slow adduction. Similarly, intracellular recordings from an abductor motoneuron of the group Ab2–4 reveal close correlation of single motoneuron action potentials with short antennal twitches, yet bursts of action potentials in a different Ab2–4 motoneuron are followed by slow antennal movements. Stimulation of the slow adductor motoneuron Ad3 causes slow force production above a firing frequency threshold of 30 Hz. In comparison, maximum tetanic forces of 3 105 N occur above 120 Hz stimulation frequency (Allgaüer and Honegger, 1993). Stimulation of fast motoneurons (e.g. Ad1) yielded single twitches with forces of 2–5 106 N, with maximum tetanic force at 11.5 105 N. Summation occurs in both muscles at stimulus frequencies above 40 Hz. Although there is some comparative detail available on the contraction forces produced by antennal muscles, the lack of knowledge about lever arms makes it difficult to relate forces to the kinematics of antennal movements, e.g. to predict the resulting torques. Hence, more information about the biomechanics of antennal joints would be desirable (see Section 2.2). 2.4.2

Inhibitory and modulatory innervation

Various modulators influence the action of antennal muscles (Allgaüer and Honegger, 1993). In extracellular recordings from a cricket antennal


73

neuromuscular preparation, the common inhibitor neuron of the antennal muscles is spontaneously active at an average frequency of 13.2 Hz while the antenna is at rest. During active antennal movements, the CI increases its firing rate before the onset of the activity of excitatory motoneurons. The average firing rate during antennal movements is 182 Hz with a maximum of 353 Hz. Evoked common inhibitor activity modulates the properties of slow contractions only: the peak force, contraction rate and resting tension decrease, while relaxation rate increases. For instance, CI activity starts lowering Ad3-evoked peak force of adductor muscle M6 at firing rates above 20 Hz. The higher the Ad3 activity, the higher the CI activity has to be to achieve a maximal force reduction (Allgaüer and Honegger, 1993). The CI effects can be blocked reversibly by application of picrotoxin, an antagonist of GABA. This supports the view that the CI is GABAergic (Honegger et al., 1990b). In contrast, DUM neuron activity has an effect on both slow and fast contractions. DUM neurones innervating antennal muscles are either silent or have low rates of spontaneous resting activity. DUM neuron activity does not generally precede bouts of antennal movements (Allgaüer and Honegger, 1993). In general, the effect of DUM neuron activity on force generation is weak and can only be elicited in 10% of the crickets tested. During DUM neuron stimulation at 10 Hz, a slow tetanic contraction, evoked by Ad3, is attenuated by an average of 25% of the peak force, and contraction speed decreases. However, forces of fast contractions are increased by an average of 11.9%, accompanied by an increase in contraction rate and a decrease in relaxation rate. In conclusion, DUM neurons have contrary effects on slow and fast contractions, and could, therefore, act to support fast force production by damping the effects of slow motoneuron activity. The transmitter of antennal DUM neurons is likely to be octopamine (Braünig, 1991; Spo¨rhaseEichmann et al., 1992), and application of octopamine at concentrations of 107–106 M onto the antennal musculature parallels the effect of DUM neuron activity (Bartos et al., 1994). In crickets, treatment of antennal muscles with proctolin results in the increase of forces of slow and fast contractions, accompanied by a decrease of relaxation rate (Bartos et al., 1994). Application of 108 M proctolin results in a force gain of up to 330%. An increase of the basal muscle tonus after proctolin exposure, as observed in locusts (Bauer, 1991), is absent in crickets. The behavioural significance of proctolin is further substantiated by Bartos and Honegger (1997). Using densitometric immunocytochemistry against proctolin, they show that proctolin stores in two antennal motoneurons, the slow adductor motoneuron Ad3 and the depressor motoneuron D5, are reversibly depleted during prolonged episodes of flight. During flight, crickets adopt a posture that includes an active and sustained forward position of both antennae, similar to what is known in locusts (Gewecke, 1972a). This suggests that proctolin is released onto antennal muscles during prolonged muscle contractions. Since motoneuron firing rates decrease in the course of a flight episode

74


and this decrease is accompanied by only small changes in antennal position, proctolin might play a decisive role in the maintenance of the antennal flight posture. Gebhardt (2004) tests this hypothesis by injections of anti-proctolin antiserum into the scape of flying crickets. The injections are followed by a reversible loss of the antennal flight posture, suggesting that the antiserum inactivated proctolin released onto the muscles. In conclusion, proctolin is indeed a component of the antennal muscular system that is necessary for the adoption of a sustained antennal posture as it occurs during flight. Though more studies are required to understand the basis of the differential intrinsic properties of motoneurons, it is apparent that the antennal musculature and its modulatory innervation can account for substantial plasticity of antennal movements. A variation in muscular structure alone contributes to the generation of variable movements. Modulatory substances, the release of which is controlled by efferent neurons, increase the dynamic range of muscle contractions beyond the limits set by the fast, slow and inhibitory innervation. The mechanisms used are efficient since a single modulator (e.g. octopamine) has differential effects across a muscle fibre population.

3 3.1

Sensory physiology of antennal mechanoreceptors SENSORY STRUCTURES AND TRANSDUCTION

In the following section, the term sensillum is understood as the whole assembly of cuticular structures, which may or may not be external, sensory cell(s), and other structures required to obtain a functional unit, which responds to mechanical and/or other stimuli. The insect antenna is a compound sensory organ carrying a large variety of receptors, including olfactory (Keil, 1982, 1989; Lee and Strausfeld, 1990; Steinbrecht, 1997; Shields and Hildebrand, 1999a,b), gustatory or contactchemosensory (Ru¨th, 1976; Schaller, 1978; Tichy and Loftus, 1983; Chapman, 2003), hygro- (Loftus, 1976; Tichy, 1979; Tichy and Loftus, 1990, 1996), thermosensory (Loftus, 1966, 1968, 1969; Altner, 1977b; Altner et al., 1977, 1981, 1983; Altner and Prillinger, 1980; Tominaga and Yokohari, 1981; Yokohari, 1981; Tichy and Loftus, 1987) and mechanosensory (Markl, 1962; Gewecke, 1972b; Schmidt, 1973; Altner and Prillinger, 1980; McIver, 1985; Zacharuk, 1985). The axons from all these receptor cells project directly to the deutocerebrum of the brain (cf. Zacharuk, 1985; Rospars, 1988). There are four basic types of mechanosensory structures: (1) hairs, (2) campaniform sensilla, (3) chordotonal organs and (4) strand organs or stretch receptors. All of these are found on and in various segments of insect antennae, and, can, thus, in principle function as proprio- and exteroreceptors. They can be found as single receptors or arranged in groups, e.g. to form hair plates (Markl, 1962; Gewecke, 1972b; Ehmer and Gronenberg, 1997b), Johnston’s organ (vande Berg, 1971) or


75

rings around a segment (Knyazeva, 1974; Toh, 1981), and their distribution on the antennae varies with sensillum type and insect species.

3.1.1

Mechanosensory hairs

A wide variety of mechanosensory hairs can be found on the insect antenna (Schneider, 1964; Altner and Prillinger, 1980; Zacharuk, 1985) and their morphology allows a crude classification. But only detailed studies of their fine structures led to descriptions that are adequate to infer their function (Altner, 1977a). Names like sensilla chaetica or sensilla trichoidea are used in different ways by different authors. Therefore, they do not allow a proper classification of the various hair types (cf. Table 1 of Altner and Prillinger, 1980). The two hair types with mechanosensory function are hairs without pores, but with flexible sockets (Section 3.1.1.1), and hairs with a terminal pore and a flexible socket (Section 3.1.1.2; Altner and Prillinger, 1980). 3.1.1.1 Tactile hairs. Hairs without pores, but with flexible sockets (Fig. 4) are often called sensilla chaetica (Toh, 1977, 1981), or sometimes sensilla trichoidea (Esslen and Kaissling, 1976). Their shank is of variable form, length and wall thickness, and sometimes they have longitudinal grooves reaching up to the tip. Their sockets are often asymmetric, allowing movement of the hair in only one direction, which is part of the mechanism underlying the directional responses of the sensory neurons. Usually, these hairs are innervated by one bipolar sensory cell. The inner dendritic segment of this neuron contains many mitochondria, as is the case in other sensory cells, a ciliary body and a rather short ciliary root. The outer dendritic segment contains arrays of parallel microtubules, which are connected by an amorphous substance at the distal end. Such a distal end is called a ‘tubular body’ (Thurm, 1964). In some cases, the tubular body is flattened, which provides another basis for directional sensitivity (Thurm et al., 1975). The proximal part of the outer segment lies within an extracellular space filled with receptor lymph. The outer dendritic segment, including the tubular body, is enclosed by cuticle, which is called the dendritic sheath. The distal part of the outer segment is attached to the base of the hair and surrounded by a cap, which probably consists of resilin (Thurm, 1964). The attachment site is on the side where the movement of the hair is most restricted. When the hair is moved in the preferred direction, the inner structures of the joint distort the tubular body of the sensory cell. This mechanical change in the outer dendritic segment is the basis for the mechano-electric transduction in insect sensory hairs (Thurm, 1964, 1965, 1983; Gaffal et al., 1975; French and Sanders, 1979, 1981). It is still debated, however, how significant the microtubules and the tubular body are for the transduction (Moran and Varela, 1971; Moran et al., 1977; Schafer and Reagan, 1981; Kuster et al., 1983; Zacharuk, 1985). We will not deal with the molecular basis

76


FIG. 4 Schematic view of the different elements comprising a mechanosensory hair, i.e. a mobile hair without a pore. In this and the following schemes, accessory cells are omitted for clarity.

of the mechanosensory transduction, but refer the reader to a recent review (Gillespie and Walker, 2001). 3.1.1.2 Contact-chemosensory hairs. Hairs with flexible sockets and terminal pores vary with regard to their length, and external structure (Fig. 5). They have been described under various names, such as sensilla chaetica, trichoidea, basiconica, styloconica (cf. Altner and Prillinger, 1980). Functionally, this hair type is considered to be mechanosensory and gustatory, i.e. contact-chemosensory. A


77

variable number of sensory cells, usually three to five, which send their dendrites into the hair shank, are considered to be chemoreceptive. As in the other hair type (see Section 3.1.1.1), only a single cell attaches to the base of the hair, and this neuron is considered to be mechanosensory. This cell is attached in the same manner to the hair base as in hairs without pores. Therefore, the mechanism for mechanosensory transduction is considered to be very similar to tactile hairs, if not the same.

FIG. 5 Schematic view of a contact-chemosensory or gustatory hair, with an emphasis on the mechanosensory neuron. Note, that this hair type has a single pore at the tip.

78

3.1.2


Campaniform sensilla

Campaniform sensilla (Fig. 6; Hicks, 1857) are externally recognisable sensory structures, which, like mechano- and chemosensory hairs, chordotonal organs and strand receptors, have been found and described in many locations on the bodies of many insect species (e.g. fly haltere: Pringle, 1948; Smith, 1969; cicada: Klein et al., 1988; Coleoptera: Kim and Yamasaki, 1996; moth: Schneider and Kaissling, 1957; cockroach maxillary palp: Pringle, 1938a,b). Some authors refer to single campaniform sensilla as Hick’s organ (Snodgrass, 1935), while others call a prominent group of campaniform sensilla on the distal pedicel Hick’s organ (Gewecke, 1972b). To avoid confusion, we will not use this term but simply refer to these structures as campaniform sensilla. Typically, an elliptic or round, dome-like structure, which sometimes has a small central pit, lies in the middle of a slight cuticular indentation. The elliptic shape of the cuticular structures provides the basis for the directional sensitivity (Heinzel and Gewecke, 1979; see Section 3.3) with the axis of greatest sensitivity aligned along the short axis of its ellipsoid shape (Thurm et al., 1975). Round campaniform sensilla are considered omni-directional. This sensory structure is usually innervated by a single bipolar sensory cell with a tubular body, which attaches to the central part of the cuticular dome. The outer segment is located within a receptor lymph space, and the tip, which

FIG. 6 Schematised section through a campaniform sensillum.


79

contains the tubular body, is enclosed by a cap, which is reminiscent of the situation in mechanosensory hairs. The so-called marginal sensilla have been described for the antennae of Periplaneta americana L. by Schafer and Sanchez (1973) and Toh (1977) and of Blatella (Campbell, 1972). According to Toh (1977), it seems as if these sensory structures, which are innervated by a single bipolar neuron with a tubular body, are just a morphological variation of campaniform sensilla. Pringle (1938a) was the first to report the responses of campaniform sensilla to pressure on the cuticle and to bending of a joint. Consequently, he called them ‘‘stress receptors’’. As in sensory hairs, the shape of the outer dendritic element is considered to be important for the directional sensitivity of the sensillum (Thurm et al., 1975). Given the many structural similarities of campaniform and hair sensilla, it is not astonishing that their transduction mechanism was proposed to be the same (Thurm, 1964; Thurm et al., 1975, but see Chapman et al., 1973). 3.1.3

Chordotonal organs

Chordotonal organs consist of arrangements of a variable number of scolopidia, a functional unit comprised of sensory and accessory cells (Fig. 7; for review see Field and Matheson, 1998). Sometimes, they are subdivided into distinct groups of scolopidia, which then are referred to collectively as scoloparia. Chordotonal mechanosensory neurons are viewed as completely internalised and modified homologues of the receptors associated with cuticular structures, i.e. hairs and campaniform sensilla (Schmidt, 1973). The receptor cells show some similarity with the mechanosensory cells described above, in that their bipolar neurons also have an inner and an outer dendritic element. The inner segment, however, is usually much longer than in the other sensory cells, and sometimes has a bulbous widening. Furthermore, the ciliary root extends from near the tip throughout the entire inner segment into the soma. The outer segment contains microtubules, like in other mechanosensory neurons, and often ends with a small dilatation (McFarlane, 1953; Howse, 1965; Howse and Claridge, 1970; vande Berg, 1971; Schmidt, 1973). In holometabolous insects, the distal end of the microtubular structure is dense and, thus, may be considered a tubular body (Schmidt, 1973). One or more sensory cells are usually associated with two accessory cells, the scolopale and the attachment cell, and only in a few cases a separate cap cell is described (Schmidt, 1973). Unlike the other mechanosensory neurons, chordotonal sensory cells are not directly attached to the cuticle, but indirectly through the attachment cell. Another specialisation of chordotonal organs is the scolopale cell, which surrounds a good part of the inner dendritic segment and the outer dendritic segment, nearly up to the distal end of the neuron. The invaginations of this cell form an extracellular space around the outer dendritic segment. The scolopale cell also contains the scolopale rods, tubular structures filled with

80


FIG. 7 Scheme of a scolopidial sensillum, as found in ordinary chordotonal organs and Johnston’s organ.

electron-dense material, which surround the dendrite of the neuron from the distal end of the inner segment to shortly proximal of the distal tip. The distal end of the sensory cell reaches into an electron-dense cap, which is surrounded by the attachment cell. The latter also encloses the distal end of the scolopale cell. Together, the scolopale rods, the cap and the ciliary roots modify the transduction mechanism of chordotonal sensory cells. The details of the precise role of the different structural elements are still a matter of debate (cf. Field and Matheson, 1998). Chordotonal sensory neurons are stretch- or relaxation sensitive. It has been suggested that transduction occurs due to either a resulting pinching (Thurm, 1965), just by stretching (Young, 1970) or by bending (Moran et al., 1977) of the distal end of the sensory cell. It is assumed that mechanically activated channels play a major role in these and the other insect mechanosensory cells. Although they have been described for vertebrate sensory cells some time ago (Guharay and Sachs, 1984; Kros et al., 1992),


81

mechanically activated channels have only recently been reported in insects (Torkkeli and French, 1999a,b). Cultured antennal mechanosensory cells from Manduca sexta, which presumably stem from Johnston’s organ, respond to positive and negative pressure perpendicular to the long axis of the dendrite (Torkkeli and French, 1999a,b). This is in accordance with the hypothesis that compression or expansion of their tips leads to transduction. A structural comparison of the mechanosensory and gustatory hairs with campaniform sensilla and chordotonal organs indicates that they may be homologues. The major similarities are of structural and developmental, rather than functional nature. Each of these sensilla has one or more sensory cells, all of which have ciliary structures. Furthermore, besides the sensory neuron, there are usually three accessory cells per sensillum. The cuticular structures ensheathing the outer dendritic segment of the sensory cells are secreted from the same cell type. During moulting, the structures of these sensilla undergo very similar sequences of events. These arguments, made by Schmidt (1973), lead to the general acceptance of the homology of these sensory structures. However, recently reported results from developmental and molecular genetics indicate that homologising these sensory structures may not be as simple. Some genes, that are important for the development of the peripheral nervous system determine which type of mechanosensory sensillum or olfactory organ will develop (Ato olfactory: Gupta and Rodrigues, 1997; Jhaveri et al., 2000; Ato mechanosensory: Bodmer et al., 1987; Brewster et al., 2001; Amos olfactory: Goulding et al., 2000; Jhaveri et al. 2000; Amos mechanosensory: Brewster and Bodmer, 1995; Brewster et al., 2001). However, there are two major differences between mechanosensory and olfactory organ development. First, olfactory sensilla develop from one to three secondary precursor cells, which are only associated with the olfactory progenitor cell (Sen et al., 2003). Second, the genes, which lead to the development of the different cell types are different from the genes activated during mechanosensory organ development (olfactory: Pros, Elev, Svp, cf. Sen et al., 2003). Thus, the different mechanosensory sensilla, including gustatory hairs, are more closely related to one another than to olfactory sensilla. In contrast to olfactory sensilla, all the mechanosensory organs develop directly from the sensory organ precursor (e.g. Brewster et al., 2001). The three key genes are Cut, Ato and Amos, and their activation leads to different types of mechanosensory structures. If Cut is expressed in the sensory organ precursor, external sensory organs, i.e. campaniform sensilla, mechanosensory or chemosensory hairs, or stretch receptors develop (Brewster et al., 2001). The activation of BarH1/BarH2 leads to campaniform sensilla, while hairs develop in their absence (Higashijima et al., 1992). A chemosensory hair is induced if Poxn is active, while a mechanosensory hair develops if it is absent (Dambly-Chaudiere et al., 1992; Nottebohm et al., 1992; Ghysen and Dambley-Chaudiere, 1993). This is in contrast to chordotonal organs and some stretch receptors, which are Ato dependent, and only develop in the absence of Cut (Blochlinger et al., 1991;

82


Bodmer et al., 1997; Brewster et al., 2001). Another group of stretch receptors is Amos dependent (Brewster and Bodmer, 1995; Brewster et al., 2001). From all this follows, that (1) the different mechanoreceptors are closer related to each other than to the various olfactory sensors, and that (2) campaniform sensilla, mechanoreceptive and gustatory hairs are closer related with each other than with chordotonal organs. The only indication of homology remaining is that in external and chordotonal organs, the neuron and one supporting cell, thecogen or scolopale cell, arise from the same secondary precursor lineage (Merritt, 1997; Younossi-Hartenstein and Hartenstein, 1997). 3.1.4

Strand receptors

Strand or stretch receptors are typically found in the joints of insect appendages (e.g. locust wing hinge; leg). In contrast to the other sensors (cf. Sections 3.1.1–3.1.3), they consist of multipolar neurons, some of which have central somata (leg: Braünig and Hustert, 1980). Their dendrites are usually embedded in a sheath of glial cells and connective tissue, but they are not associated with such obvious structures as scolopale cells. Unlike other mechanoreceptors, their dendrites do not contain specialised structures, like tubulary bodies or long ciliary roots that are visible in the electron microscope. Stretching of the free dendritic endings is supposed to induce transduction (Kuster et al., 1983). So far, an antennal strand receptor has been described only in the locust (Braünig, 1985; see also Section 3.3). The locust strand organ spans from the tentorium to the base of the scape and runs parallel to the levator muscle of the scape (Braünig, 1985). The length of the strand decreases with antennal elevation. In contrast to any other known antennal mechanoreceptor, the somata of the four to five receptor neurons are not located in the periphery. Rather, they are located in the medial protocerebrum, anterior to the central complex. These neurons arborise in the anterior dorsal deutocerebrum, i.e. close to, if not in, the dorsal lobe (DL). The finding of four somata in a similar location in the protocerebrum of a cricket indicates that a strand receptor may be present in the head of this species, too (Staudacher and Schildberger, 1999). 3.1.5

Mechanoreceptors of the crustacean antenna

Lobster and crayfish antennae bear many sensory structures, some of which may seem to be very similar to sensilla on insect antennae. Among these receptors are short smooth hairs, long smooth upright hairs, and short round peg hairs (Cherax destructor; Sandeman, 1989). Spine-like structures, smooth hairs, campaniform sensilla, and hydrodynamic receptors occur on the antennae of the rock lobster, Palinurus vulgaris (Vedel, 1985). However, there also are specialised hydrodynamic receptors and feathered hairs in crayfish (Tautz et al., 1981; Vedel, 1985), which are not innervated, but


83

aid in support movement detection by transmitting forces to nearby smooth hairs (Bender et al., 1984). Internal mechanoreceptors, i.e. chordotonal organs are found in all appendages of crustacea (Mill, 1976), including the antennae (e.g. Rossi-Durand and Vedel, 1981). Even though many of these receptors seem superficially similar to insect structures, they may not be if details are compared. For example, the mechanoreceptive hairs of insects and crustacea differ in the ultrastructure of the outer dendritic segment (Crouau, 1997). For chordotonal organs, details of fine structural differences to insect chordotonal organs have been reported (Howse, 1968). Because of the reported differences in ultrastructure, it is conceivable that the transduction mechanisms may also be different. It is, however, striking, that these receptor types appear to have similar functions, despite structural differences. Another example of the difficulty of ascribing the same function to superficially similar structures, are the crustacean campaniform sensilla. Here, the ultrastructure of what had been proposed to be campaniform sensilla led to a reclassification as bimodal sensilla and dermal glands (Gnatzy, 1984; Gnatzy et al., 1984; Schmidt and Gnatzy, 1984). The structures most likely resembling insect campaniform sensilla contain two scolopidial sensory cells, instead of one receptor cell with a tubular body (Schmidt, 1990). In both, crustacea and insecta the number of annuli of their flagella varies greatly, and is species dependent. Crayfish antennae have about 100 annuli, while rock lobsters have up to 400 (Tautz et al., 1981; Vedel, 1985). The bee has only 10, while a cockroach may have about 140 annuli (cf. Table 1). The situation is similar, if the numbers of mechanosensory structures on the flagella are compared. Rocklobsters have about 6000 such structures, while crayfish may only have about 2000 (Tautz et al., 1981; Vedel, 1985). About 1200 external mechanosensory structures can be counted on the flagellum of the bee, while there may be more than 7000 in a cockroach (cf. Table 8). Thus, a comparison of the mere numbers of external structures does not indicate a clear-cut trend in the difference in the numbers of sensilla on crustacean and insect antennae. A comparison of the number of sensilla per annulus leads to a different picture. Crayfish have 17 sensilla/annulus, rocklobsters 15 sensilla/annulus, while a cockroach has 50 sensilla/annulus and a bee 120 sensilla/annulus. This indicates that at least some insects have a higher density of sensory structures than some crustacea. However, a comparison based on external structures should be interpreted with caution, because the number of sensory cells associated with each sensillum may be very different. This is, for example, the case in crustacean olfactory sensilla, which can have up to 200 sensory cells, while insect olfactory hairs usually are associated with only 3–7 sensory neurons (Hallberg and Hansson, 1999). The situation seems to be similar for spider mechanosensory sensilla. Most spider mechanosensory hairs are innervated by three sensory cells, instead of a single neuron as in insects (Foelix, 1996; Barth, 2002). The so-called slit sensilla are elongated structures on the body of spiders. In contrast to campaniform sensilla, there are two dendrites running towards the exocuticle, only one of which enters a

84


canal and attaches to the membrane within the slit (Foelix, 1996). This sensory cell, however, has a tubular body, which is typical for arthropod mechanosensory cells (Barth, 2002). Functionally, campaniform and slit sensilla are very similar, and it would be interesting to know if they were homologous or analogous structures. 3.2

DISTRIBUTION OF MECHANORECEPTORS

The mechanosensory neurons of the antenna need to represent two different aspects of the world, (1) antennal movement induced by the animal itself, and (2) external stimulation of the antenna. The various sensory structures responding to mechanical stimulation have different characteristics, based on their morphological and anatomical properties, their different ways of responding to mechanical stimulation and their different response dynamics. Furthermore, animal species differ in their morphology, life styles and behaviour. These factors, together with differing receptor characteristics are likely to be reflected in the species-specific distribution of the various sensor types. Two variations of a general theme are conceivable: first, certain sensory structures will be distributed in very similar ways in different species, because they are especially suited to detect certain stimulus features, which are of behavioural significance to these species. An example could be the campaniform sensilla on the distal edge of the pedicel that are suitable to sense torque exerted at the PFjunction (cockroach: Schafer and Sanchez, 1973; cricket: Knyazeva et al., 1975, 1978; locust: Gewecke, 1972b; Knyazeva, 1974; stick insect: Rosciszewska and Petryszak, 1986). Second, other receptors will show larger differences in their distribution, because of species-specific tasks, e.g. sensory hairs on the tip of the bee antenna (Martin and Lindauer, 1966). 3.2.1

Hair plates

Hair plates, fields or rows are distinct arrangements of similar hairs in close proximity. On the antenna, they are located near the basal joints, close to the proximal margin of the distal segment. Due to their specific features, hair plates are discussed separately from other sensory hairs (see also Section 3.3). In Lepidoptera, the hair plates on the basal antennal segments were described early on by Bo¨hm (1911). Therefore, the hairs in these plates are sometimes called ‘Bo¨hm bristles’ (cf. Kloppenburg et al., 1997). We will, however, not use this term, but use the more general term ‘hair plates’ instead. Each hair plate is distinct from the surrounding tactile hairs and consists of a variable number of hairs (Table 3), which in all cases reported so far, are mechanoreceptive hairs, innervated by a single sensory cell (Schafer and Sanchez, 1973; Okada and Toh, 2001). The precise location of the hair fields reflects the species-specific arrangement of the joint axes, and indicates how

Periplaneta

Cricket

Locust

Head-scape joint axes

Ball joint: Okada and Toh (2000)

Hinge joint: 601 up and 501 down Gewecke (1972a)

Hinge joint: vertical plane: up/down; Du¨rr et al. (2001)

Ball joint; Snodgrass (1956)

Scapal hair plates

Ventral, dorsal, lateral (24, 25, 10 hairs; Petryszak, 1975) Dorsal, lateral, median (120, 60, 30 hairs; Okada and Toh, 2000)

Hinge joint: vertical plane: up/down; Honegger et al. (1990) Dorsal, ventrolateral, ventromedian (60, 20, 20 hairs; Knyazeva et al. (1975))

Dorsal, ventral, median, ventromedian; Gewecke (1972b)

One plate extending laterally and medially (90–120 hairs; Markl, 1962)

Scape-pedicel joint axes

Hinge joint: vertical plane: up/down Okada and Toh (2000)

Hinge joint: 451 lateral and median; Gewecke (1972a)

Pedicellar hair plates

Dorsal, ventral, dorsolateral; Schafer and Sanchez (1973), Rosciszewska and Petryszak (1986), Okada and Toh (2000)

Hinge joint: horizontal plane: lateral/median; Honegger et al. (1990) Gryllus: two hair plates; Kammerer and Honegger (1988) Acheta: lateral, median (15, 15 hairs; Knyazeva et al., 1975; Rosciszewska and Petryszak, 1986)

Medio-dorsal, dorsolatero-ventral, medioventral (16–20, 10–16, 16–18 hairs; Urvoy et al., 1984) Medio-ventral, ventro-latero-dorsal, medio-dorsal (Du¨rr, unpublished observation) Hinge joint: horizontal plane: lateral/median; Du¨rr et al. (2001) Ventro-dorsoexternal, lateralinternal, centroventral (14–22, 10, 7 hairs; Urvoy et al., 1984; Rosciszewska and Petryszak, 1986)

Lateral, mediodorsal, medioventral (25–30, 15, 25 hairs; Markl, 1962)

Lateral, median; Gewecke (1972b)

Carausius

Bee

Hinge joint: pedicel movement: up/ down; Snodgrass (1956)


TABLE 3 Hair plates and joint axes in the insect model organisms

85

86


antennal movements may be modulated by the activation of one or more hair plates (Kammerer and Honegger, 1988; Okada and Toh, 2000, 2001). The scape of Periplaneta has at least three hair plates (Petryszak, 1975; Okada and Toh, 2000), yet the descriptions of their location and the number of hairs in each plate differs greatly. Schafer and Sanchez (1973) report four hair plates, but no hair count. Petryszak (1975) reports ventral, dorsal and lateral hair plates with 24, 25 and 10 hairs, respectively. According to Okada and Toh (2000), there are dorsal, lateral and medial hair plates with 120, 60 and 30 hairs, respectively. These hairs are reported to have no pores (Schafer and Sanchez, 1973; Okada and Toh, 2001). The cricket scape has three hair plates (Rosciszewska and Petryszak, 1986). Knyazeva et al. (1975) report a dorsal scapal hair plate with some 60 hairs, plus one ventro-lateral and one ventromedial scapal hair plate with about 20 hairs. Four hair plates are reported for the locust scape. They are located dorsally, ventrally, ventro-medially and laterally (Gewecke, 1972b). According to reports by Urvoy et al. (1984) and Rosciszewska and Petryszak (1986), the scape of the stick insect Carausius morosus has three hair plates, labelled as medio-dorsal, dorso-latero-ventral and ‘medio-ventral’ with 16–20, 10–16 and 16–18 hairs, respectively.2 The scape of the honeybee inserts on the head with a ball joint (see Fig. 2). There is only one hair plate reported, which contains about 90–120 hairs (Markl, 1962). This plate, however, seems to extend laterally and medially around the base of the scape. Functionally, it may not matter, if one or more hair plates can be discerned. It is most likely that the movement of this joint is represented by the relative number and positions of the bent hairs. The pedicel is usually smaller than the scape and the movement axis of the SP joint is often nearly perpendicular to the axis of the HS joint (see Sections 2.1.1.2 and 2.3.1). It also bears hair plates near its proximal margin, adjacent to the SP-joint membrane. It is very likely, that they are comprised of the same type of mechanosensory hairs as the hair plates on the scape, i.e. they have no pores. The pedicel of Periplaneta bears three hair plates: dorsal, ventral and dorsolateral (Schafer and Sanchez, 1973; Rosciszewska and Petryszak, 1986; Okada and Toh, 2000). In Gryllus, two pedicellar hair plates have been described (Kammerer and Honegger, 1988). Acheta also has two hair plates: lateral and medial, with about 15 hairs in each (Knyazeva et al., 1975; Rosciszewska and Petryszak, 1986). However, the type of hair is not clear (Knyazeva et al., 1975, 1978). A very similar arrangement of pedicellar hair plates is reported for a locust. Gewecke (1972b) described lateral and medial hair plates on the pedicel. This may be due to a similar joint arrangement in these species (Section 2.1.1.2). 2

Comparison with another laboratory colony of Carausius morosus reveals a remarkably similar description of antennal hair plates, yet with reversed orientation of dorsal and ventral sides. Probably, the labelling and corresponding text description of the dorsal and ventral sides in Fig. 1 of Urvoy et al. (1984) were accidently swapped (Du¨rr, unpublished observation).


87

The three hair plates of Carausius are ‘ventro-dorso-external’, ‘latero-internal’ and ‘centro-ventral’ and have 14–22, about 10 and seven sensilla, respectively (Urvoy et al., 1984; Rosciszewska and Petryszak, 1986). For Apis, Markl (1962) describes three hair plates on the base of the pedicel. Like on the scape, these hairs appear to be mechanosensory (Schneider and Kaissling, 1957). The lateral hair plate with 25–30 hairs probably responds to an extension of the antenna, while the medio-dorsal and medio-ventral hair plates, with 15 and 25 hairs, respectively, respond to bending of the antenna (Markl, 1962).

3.2.2


Campaniform sensilla can be found all over the insect antennae in variable numbers and arrangements (Table 4). Their total number on the antenna may well be species-specific (cockroach: 140–400; Campbell, 1972; Schafer and Sanchez, 1973; Petryszak, 1975; cricket: Acheta: 238–285, Knyazeva et al., 1975, 1978; Gryllus: 13–16; Yildiz and Gebhardt, 2003; stick insect: 0–50; Slifer, 1966; Weide, 1960; honeybee: 44–48: Dietz and Humphreys, 1971; Esslen and Kaissling, 1976). Campaniform sensilla are most obvious on the distal edge of the pedicel, where they are usually arranged evenly in one or two rings around the rim. The number of rings, and sensilla per ring varies between species, but nothing is

TABLE 4 Campaniform sensilla in the insect model organisms Species Periplaneta americana L. P. americana L. P. americana L.

Acheta domesticus L. (Gryllus domesticus L.) Locusta migratoria L.

L. migratoria L. Carausius morosus Br. Apis mellifica L., worker A. mellifica L., drone

Named

Number of mechanosensory cells

Sensilla campaniformia (sensory domes) Campaniform sensillum Campaniform sensillum

Not reported

Sensilla campaniformia

One bipolar neuron


Campaniform sensilla Campaniform sensilla Sensilla campaniforme Sensilla campaniforme

Location

Reference(s)

Scape, pedicel (distal margin), flagellum (single, dorsal) Not reported

Schafer and Sanchez (1973)

1 Row of about 28 around distal part of pedicel Distal part of pedicel

Toh (1981)

70 Sensilla

Around distal pedicel

Gewecke (1972b)

Not reported

Two rows at distal pedicellus Soft cuticle of scapepedicel joint None identified with certainty Flagellar segment 1 and 3–10 Flagellar segment 3–10

Knyazeva (1974)

One with a tubular body One with a tubular body

Not reported N/a One (?) One (?)

Toh (1977)

Knyazeva et al. (1978)

Gewecke (1972b) Slifer (1966) Esslen and Kaissling (1976) Esslen and Kaissling (1976)

88


known about the functional relevance of these differences, e.g. their impact on angular resolution of deflection sensitivity at the PF-junction. Periplaneta has one ring with about 27 sensilla (Petryszak, 1975). In crickets, the number and arrangement of the campaniform sensilla are different in the two species. While Acheta has two rings with 60 sensilla, Gryllus has only one ring with 13–16 sensilla at the distal rim of the pedicellus (Knyazeva et al., 1975, 1978; Rosciszewska and Petryszak, 1986; Gryllus: Yildiz and Gebhardt, 2003). Two parallel rings are found in locusts (Knyazeva, 1974), in which Gewecke (1972b) counted about 70 campaniform sensilla. For Carausius, details are not yet published. In Apis, McIndoo (1914) mapped campaniform sensilla on the whole body, but not on the antenna. Thus, together with other sensory structures of the pedicel, the campaniform sensilla of the distal edge of the pedicel can probably play an important role for flight steering and maintenance (cockroach: Yagodin, 1980; locust: Gewecke, 1972a; see Section 6.1.3.1). Except for the sensilla on the distal pedicel, only a few campaniform sensilla are located on the first two antennal segments, and they are usually scattered over their surfaces (cockroach: Petryszak, 1975; Schafer and Sanchez, 1973; cricket: Knyazeva et al., 1975, 1978; stick insect: Weide, 1960). The only exception is the basal scape of Periplaneta, which contains a single row of campaniform sensilla (Rosciszewska and Petryszak, 1986). It is common across the five model species that the annuli of the flagellum bear only a small number of campaniform sensilla. In the cockroach the campaniform sensilla on the flagellum have been called ‘marginal sensilla’ (Campbell, 1972). There are usually one or two marginal sensilla on the dorsal surface of each annulus (Schafer and Sanchez, 1973; Petryszak, 1975). In Periplaneta, Acheta and Carausius, the situation is very similar (Knyazeva et al., 1975, 1978; Rosciszewska and Petryszak, 1986). They are found close to the distal edges of the annuli. Therefore, marginal sensilla may be well suited to respond to antennal bending, i.e. to deflections of one or more annuli relative to their neighbours. So far, their descriptions do not allow inferences about their directional sensitivity, and one may wonder, how well the small number of sensilla can cover the stimulus range for flagellar bending. It is conceivable, however, that the whole population of marginal sensilla is quite efficient in representing the direction and strength of antennal bending. A more complicated arrangement is reported for Apis. Esslen and Kaissling (1976) report eight sensilla on the first, four on the sixth, and 18 on the 10th annulus, while Dietz and Humphreys (1971) count none on the first, eight on the seventh, and 10 on the last annulus. All other annuli bear between one and four campaniform sensilla. Despite the differences in the number of sensilla, it seems clear that, first, the honeybee flagellum bears more campaniform sensilla per annulus than reported in other species, and second, that there is an obvious concentration of these sensilla at the tip. These observations probably have some significance, given the many ways bees use their antennae, e.g. in comb building (Martin and Lindauer, 1966).


3.2.3

89

Chordotonal organs

Chordotonal organs are common in insect antennae (Table 5). The most prominent one, Johnston’s organ, is a rather large and specialised example. Chordotonal organs associated with antennal movements are found in the head and the two basal antennal segments, but they have not been reported for the flagellum. In the honeybee, a chordotonal organ inside the head (Janet’s organ) connects the tentorium with the membrane of the HS joint (Janet, 1911). It is attached to the dorsal side of the joint membrane, suggesting that it may respond to upward movements of the scape. Janet’s organ has also been described in ants (Ehmer and Gronenberg, 1997b), but, so far, not outside the hymenoptera.

TABLE 5 Chordotonal organs in the insect model organisms Species P. americana L.

P. americana L.

Named Johnston’s organ

Number of mechanosensory cells

Location

Not reported

Pedicel

About 150

Completely surrounding the distal pedicel Ventral hypodermis of distal pedicel Medial and lateral in scape

A. domesticus L. (G. domesticus L.)

Connective chordotonal organ Chordotonal organ


Chordotonal organ


Johnston’s organ

12 scoloparia with four to seven scolopidia ea.

Pedicel

A. domesticus L. (G. domesticus L.) L. migratoria L.

Johnston’s organ Chordotonal organ

Not reported

Antennae

Not reported

L. migratoria L.

Chordotonal organ

Not reported

L. migratoria L.

Johnston’s organ

Six–seven scoloparia

Proximal scape to proximal pedicel; lateral and medial Proximal to distal pedicel; medial Proximal to distal pedicel; medial

About 50 One scoloparium ea., medial with 10, lateral with five scolopidia Five scolopidia

Upper twothirds in pedicel

Reference(s) Schafer and Sanchez (1973) Toh (1981)

Toh (1981) FudalewiczNiemczyk and Rosciszewska (1973) FudalewiczNiemczyk and Rosciszewska (1973) FudalewiczNiemczyk and Rosciszewska (1973) Knyazeva et al. (1978) Gewecke (1972b) Gewecke (1972b) Gewecke (1972b)

90


3.2.3.1 Chordotonal organs of the scape. For some species, chordotonal organs are described in the scape. Distally, they are attached to structures of the SP joint. Depending on the location of the attachment site(s), they respond to antennal movement in this joint. In Periplaneta, one organ with four scoloparia with a total of 20 scolopidia reaches from the medio-dorsal scape into the proximal pedicel (Petryszak, 1975). Thus, it probably responds to downward flexion of the pedicel. The 13th instar of Acheta possesses a medial and a lateral scoloparium with 10 and five scolopidia, respectively (Fudalewicz-Niemczyk and Rosciszewska, 1973). A lateral and a median chordotonal organ reach from the proximal edge of the scape to the ventral edge of the pedicel. Thus, they are suitable to respond to medial and lateral rotation of the pedicel (Gewecke, 1972b). 3.2.3.2 Ordinary chordotonal organs of the pedicel. Movements in the PFjunction can be registered by small chordotonal organ(s), the number of which may be species-specific. In contrast to Johnston’s organ (Section 3.2.3.3), the smaller, ‘ordinary’ chordotonal organs can only respond to certain movements, because of their restricted attachment sites. The pedicel of Periplaneta houses two chordotonal organs (Petryszak, 1975; Toh, 1981). The medial organ consists of three scoloparia with 6–10 scolopidia in each, while the more distally situated organ has 5 scolopidia. Toh (1981) mentions a single connective chordotonal organ with 50 scolopidia and two neurons in each of these. However, he does not describe the location of this organ in the pedicel. One chordotonal organ with five scolopidia in one scoloparium is reported for 13th instar Acheta (Fudalewicz-Niemczyk and Rosciszewska, 1973), and possibly two chordotonal organs exist in the pedicel of adult Gryllus (Yildiz and Gebhardt, 2003). The locust pedicellar chordotonal organ reaches from the medial wall of the pedicel to the joint membrane between pedicel and flagellum (Gewecke, 1972b), and, thus, may be suited to respond to lateral movement. 3.2.3.3 Johnston’s organ. Johnston’s organ varies in size from species to species. Because it reaches around the whole PF-junction in a ring-like fashion, it can respond to flagellar movements in all directions. In Periplaneta, Johnston’s organ encircles the inner surface of the PFjunction (Petryszak, 1975; Toh, 1981). According to Petryszak (1975), this organ consists of 12–14 scoloparia with 4–5 scolopidia in each, while Toh (1981) counted 150 scolopidia with three receptor neurons in each. In the 13th instar of Acheta, Johnston’s organ consists of 12 scoloparia with 4–7 scolopidia each (Fudalewicz-Niemczyk and Rosciszewska, 1973; Knyazeva et al., 1975). Unfortunately, details are missing for adult crickets (Knyazeva et al., 1975), and it is not clear if sensory cells are being added during postembryonic development to antennal chordotonal organs. For the more complicated tibial chordotonal organs of the Teleogryllus front leg, Ball and Young (1974) report that sensory cells are added during postembryonic development,


91

while Klose (1996) reports for the same species that these chordotonal organs are complete at the end of the embryonic development. For a pedicellar chordotonal organ in a cockroach, the number of neurons increases by about 10 during postembryonic development (Bloechl and Selzer, 1988). The scolopidia of the locust Johnston’s organ can be subdivided into a lateral and a medial portion, each of which contains six to seven scoloparia, each one containing a few scolopidia. They attach at the soft membrane of the PF joint, just distal to the ring of campaniform sensilla (Gewecke, 1972b). In the honeybee, the scolopidia of Johnston’s organ run towards the joint membrane of the PF-junction (McIndoo, 1922). It is not clear, whether the differences in the number of chordotonal organs reported for the different species reflects the difficulties in obtaining reliable counts of these delicate structures or, whether they are real, and of functional significance. Unlike studies on sex-specific differences, comparative studies with regard to this functional question are missing. The sex-specific difference in the number of sensory cells of Johnston’s organ in mosquitoes is of functional significance, because males use this organ to hear the wing beat frequency of the females (e.g. Risler, 1977; Go¨pfert et al., 1999), but a comparative study on sex-differences in other species is lacking. 3.2.3.4 Chordotonal organs in crustacea. Chordotonal organs of the crustacean antennae have received much attention (e.g. Clarac and Vedel, 1971, 1975; Sigvardt, 1977; Rossi and Vedel, 1980; Vedel, 1980; Rossi-Durand and Vedel, 1981; Vedel and Monnier, 1983; Sandeman, 1985). In the crayfish C. destructor, two chordotonal organs sense movement of the distal three segments (Sandeman, 1985). The proximal chordotonal organ spans the IM-joint. The distal chordotonal organ spans the MC-joint with an additional side branch spanning the CF-joint. This is virtually identical in Homarus americanus (Sigvardt, 1977) and in P. vulgaris (Clarac and Vedel, 1975). The proximal chordotonal organ of Cherax (Sandeman, 1985) is also in the same location as the myochordotonal organ in Palinurus (Vedel and Monnier, 1983), but lacks the accessory muscle of the latter. Sensory units of the chordotonal organs are directionally selective and discharge in a tonic or phasic-tonic manner, coding a mix of position, velocity and acceleration, which is similar for insects (cf. Section 3.3). Yet, Sandeman (1985) also documents a unit the average spike rate of which codes the logarithm of the angular velocity of the IM-joint. The modulation of tonic discharge rate with joint angle may be as small as a 1 Hz increment per 101. However, given the range fractionation properties of chordotonal organs (Matheson, 1992), it is impossible to use the response properties of single afferents to assess the overall angular resolution of the proprioreceptive information that is available to the animal. In the rock lobster P. vulgaris, the sensory neurons of the MCF chordotonal organ (that spans the MC- and the CF-joint) can be categorised according to their static or dynamic response characteristic, according to their location in

92


the organ, and according to their sensitivity to extension and/or flexion of the MC- and CF-joint (Rossi and Vedel, 1980). The ratio of static and dynamic units is 7:3. Among the static units, some selectively code angular position of the MC-joint or the CF-joint, others respond most strongly if the two joints are moved into the same direction (homo-directional units) or in opposite direction (heterodirectional units). The joint-selective units can be assigned to a proximal and a distal cell group in the organ. The homo- and heterodirectional units are assigned to a central cell group. An analog case cannot be made for antennal chordotonal organs in insects, because (1) no known insect chordotonal organ spans two joints (cf. Field and Matheson, 1998), and (2) the physiology of the antennal chordotonal organs is not known well enough. However, it is known that a chordotonal organ in the locust hindleg contains five populations of receptor cells, each of which is sensitive to different parameters of the stimulus or combinations of stimulus parameters (Kondoh et al., 1995). 3.2.4

Sensory hairs

Purely mechanosensory (Table 6) and bimodal mechano-/chemosensory hairs (Table 7) account for a notable number of sensilla on insect antennae. As mentioned earlier, these hairs can have a variety of shapes and lengths. Their numbers and specific arrangement on the antennae, especially on the flagellum are quite variable in different species. It is also important to note obvious sexual dimorphisms in the number of sensilla (e.g. Periplaneta: Lambin, 1973; Schafer and Sanchez, 1973, 1976). In the case of Periplaneta, the numbers reported by different authors vary greatly, and, thus, should be used and interpreted with caution. Environmental factors and population density are reported to have an impact on the number of chemosensory sensilla of locust antennae and mouthparts (Greenwood and Chapman, 1984; Chapman and Lee, 1991; Rogers and Simpson, 1997). Recently, it has been shown that the numbers of purely mechanoreceptive hairs on the femora of locust hindlegs are also affected by such factors (Rogers et al., 2003). Some of these changes in locusts (Greenwood and Chapman, 1984; Chapman and Lee, 1991; Rogers and Simpson, 1997) are supposed to be due to changes in the chemical, rather than due to changes in mechanosensory input (Rogers and Simpson, 1997). In some cases, the number of contact-chemosensory hairs changes, and, as a consequence, the number of mechanoreceptive neurons does too (Rogers and Simpson, 1997). It is, however, not known if, and how, such changes affect the sensory neuropils in the deutocerebrum. Forming additional sensory structures on any part of the insect body requires, among other prerequisites, changes in the cuticle. Therefore, new sensilla can be acquired only when insects moult (e.g. Shepherd and Murphey, 1986; Ka¨mper, 1992). Thus, changes in external structures may only be expected in hemimetabolous insects. Consequently, the number of sensilla is most


93

TABLE 6 Mechanosensory hairs in the insect model organisms

Species Periplaneta americana L.

Named Sensilla chaetica A (Hair plate sensilla) Sensilla chaetica B Stout hairs

Number of mechanosensory cells per sensillum

Location

Reference(s)

One

Head-scape and scapepedicel joints


One with a tubular body Not reported

Scape, pedicel

Toh (1981)

Scape

Middle sense hairs

Not reported

All segments


Smallest hairs

One

All segments

Locusta migratoria L. Carausius morosus Br. C. morosus Br.

Borsten ( ¼ bristles) Tactile hairs

Not reported

Scape, pedicel

One with a tubular body One with a tubular body Not reported

Flagellum

FudalewiczNiemczyk and Rosciszewska (1973) FudalewiczNiemczyk and Rosciszewska (1973) FudalewiczNiemczyk and Rosciszewska (1973) Gewecke (1972b) Slifer (1966)

Flagellum

Not reported

Flagellum

P. americana L. Acheta domesticus L. (Gryllus domesticus L.) Acheta. domesticus L. (G. domesticus L.)

NP sensilla

Apis mellifica L.

Sensillum trichoideum B1

A. mellifica L.

Sensillum trichoideum B2

Last annulus

Monteforti et al. (2002) Esslen and Kaissling (1976) Esslen and Kaissling (1976)

likely fixed in adults of holometabolous species, such as bees. In holometabolous insects, environmental changes may have effects, but these will probably only affect structures within the central nervous system (CNS). 3.2.4.1 Distribution along the flagellum. Usually, two to five long and a variable number of intermediate hairs are found near the distal edges of the scape and the annuli of the antenna of Periplaneta (Table 8; Petryszak, 1975). Intermediate hairs are also evenly scattered on the pedicel and form rings around the annuli of the flagellum. The number of rings increases with distance from the base of the antenna (Petryszak, 1975). It remains to be shown if the hairs closest to the distal edge of an annulus respond to relative deflection of adjacent annuli. Furthermore, the number of contact-chemoreceptors increases from antennal base to the middle of the flagellum, i.e. the 50th–75th annulus (male: 130; female: 100 sensilla), and then drops towards the tip (male 70; female: 30 sensilla; Lambin, 1973). A similar distribution was reported for another cockroach

94


TABLE 7 Gustatory hairs in the insect model organisms

Species Periplaneta americana L. P. americana L. P. americana L. P. americana L. Schistocerca gregaria Forska˚l Carausius morosus Br. Apis mellifica L. A. mellifica L.

Named

Number of mechanosensory cells per sensillum

Sensilla chaetica B (Thick-walled chemoreceptors) Sensilla chaetica (B)

One

Sensillum chaeticum Tp (terminalpore)-sensilla Sensillum chaeticum TP sensilla

One with a tubular body One with a tubular body One with a tubular body One with a tubular body One with a tubular body

Hakenborste (¼ hooked bristle) Sensillum trichoideum D

One with a tubular body

Not reported

Location

Reference(s)

All segments, especially flagellum Flagellum, scarce on scape and pedicel Flagellum


Flagellum

Schaller (1978)

Flagellum

Ochieng et al. (1998) Monteforti et al. (2002) Martin and Lindauer (1966)

Last annulus Flagellum Flagellum

Toh (1977, 1981) Ru¨th (1976)

Esslen and Kaissling (1976)

species (Schaller, 1978). This distribution is in contrast with some, but not all of the other model organisms, where the number of sensilla often increases towards the tip of the flagellum (Weide, 1960; Apis: Martin and Lindauer, 1966; Esslen and Kaissling, 1976). For Acheta, two large hairs are reported to be located on the distal scape (Fudalewicz-Niemczyk and Rosciszewska, 1973), while medium and small hairs are scattered on the surfaces of scape and pedicel (Fudalewicz-Niemczyk and Rosciszewska, 1973). In addition, medium length hairs form circles around the first few annuli, but seem to be irregularly placed on more distal annuli (Fudalewicz–Niemczyk and Rosciszewska, 1973). Mechanosensory hairs are reported to form regular rings around the annuli near their distal edges (Rosciszewska and Petryszak, 1986). In the cricket Gryllus bimaculatus, the number of sensory hairs increases with distance from the antennal basis, and this seems to be similar in females and males (Fig. 8; Staudacher, unpublished observation). This seems to be similar in another cricket species, Teleogryllus commodus, where no structural differences between male and female flagella are reported (Rence and Loher, 1977). This arrangement is somewhat reminiscent of the situation in the cockroach, where each annulus carries a ring of contact-chemosensory hairs. The fact that different types of hairs (mechanosensory in a cricket and gustatory in a cockroach), are arranged in a similar way along the flagellum, could reflect a

Species



Periplaneta americana L., female 400

Chordotonal organs

Johnston’s organ

300

6500

49 000

400

10 000 6640

46 200

300

6500

87 000

400

15 000

4 210

Acheta domesticus L., 13th instar

185

Olfactory sensilla 29 300

1 with 50 150 Scoloparia scoloparia, each each with 3 with 2 scolopidia scolopidia

P. americana L., sex unspecified

Strand receptors

6410

P. americana L., male 400

Gustatory hairs

4 59

Reference(s) Schafer and Sanchez (1976) Schafer and Sanchez (1973) Lambin (1973) Schafer and Sanchez (1976) Schafer and Sanchez (1973) Lambin (1973) Toh (1977, 1981)


TABLE 8 Number and distribution of sensory structures in the insect model organisms

Okada and Toh (2000) Petryszak (1975)

3 Organs with 10, 5 and 5 scolopidia, respectively

FudalewiczNiemczyk and Rosciszewska (1973)

12 Scoloparia, each with 4–7 scolopidia

(continued )

95

3 with 8 12–14 scoloparia, each scoloparia, each with 5–20 with 4–5 scolopidia scolopidia

96

TABLE 8 Number and distribution of sensory structures in the insect model organisms (continued ) Species


A. domesticus L., adult


Gustatory hairs

Chordotonal organs

Johnston’s organ

May be four

4 70

40–50

4000–4500

600–700

Staudacher and Schildberger (1998)

4–5 sensory cells

Ochieng et al. (1998) Braünig (1985) Weide (1960) Slifer (1966) Monteforti et al. (2002)


295 gregarious 315 solitary

Carausius morosus Br.

Reference(s)

Gewecke (1972b)

6–7 Scoloparia, each with a few scolopidia

Schistocerca gregaria Forska˚l

Olfactory sensilla

Knyazeva et al. (1975) Rosciszewska and Petryszak (1986)

4 110

Gryllus bimaculatus DeGeer Locusta migratoria, L.

Strand receptors

Apis mellifera L., male

42

388

130

18971

A. mellifera L.,

Slifer and Sekhon (1961)

worker: bee 1

44

995

100

4781

bee 2

48

1271

132

4870

A. mellifera L., unspecified

Snodgrass (1956) Esslen and Kaissling (1976)

4 155–190

Esslen and Kaissling (1976) Dietz and Humphreys (1971) Markl (1962)


Rosciszewska and Petryszak (1986) Urvoy et al. (1984)

446–54

97

98


difference in the relative importance of tactile and contact-chemoreceptive senses for these animals. This possibility may not be too surprising, e.g. given the sexual dimorphism of the number of olfactory structures in cockroach and bee (Schafer and Sanchez, 1973; Esslen and Kaissling, 1976). In contrast to the mechanosensory hairs on the first two segments (Gewecke, 1972b), only little is known about mechanosensory hairs on the flagellum of the locust antennae. Slightly fewer contact-chemosensory hairs are found on the flagellum of gregarious than solitarious animals (Ochieng et al., 1998). The number of contact-chemosensory hairs is the highest on the first annulus, with a second maximum at the tenth annulus. Distal to this second peak, the number declines steadily towards the tip of the antenna (Ochieng et al., 1998). Weide (1960) counts 4000–4500 sensilla chaetica and trichoidea on the flagellum of Carausius, amounting to about 85% of the total number of sensory structures he reported. It is not clear, however, how many of these sensilla are mechanosensory or gustatory. Slifer (1966) describes two types of mechanosensory hairs with different distributions along the flagellum: long, heavy hairs

FIG. 8 SEM pictures of different sections along the antenna of a cricket (G. bimaculatus De Geer) exemplify the distribution of sensory structures along the antenna. With increasing distance from the base of the antenna, more sensors are located on each annulus. (a) Scape (Sc), pedicel (Pd) and first two annuli (A1, A2). (b) Six annuli from the first quarter of the antenna. (c) Four annuli from the second quarter of the antenna. (d) Tip of the flagellum. Figs. 5a–c: same scale. Fig. 5d has a different scale.


99

with blunt tips are found on the proximal third of the antenna, whereas short, delicate ones with sharp tips occur on the distal two-thirds of the antenna. In addition, there are thick-walled chemoreceptor pegs with a terminal pore and a flexible socket, which are inclined towards the tip of the antenna (Slifer, 1966). Their distribution along the antenna has not been reported, but Weide (1960) notes that the number of sensilla in general increases along the flagellum and reaches a maximum at the tip. It would be interesting to know if these two classes of mechanoreceptive hairs respond differently to mechanical stimulation and, if so, in which way. This could help shed more light on the way the antenna is used for obstacle avoidance by this animal (see Section 6.2.1). The different sensilla of bee antennae are, in general, most abundant on the tip (Esslen and Kaissling, 1976). Like in the stick insect, two types of mechanosensory hairs are reported for the honeybee flagellum. Furthermore, these two classes are also distributed in very distinct ways along the antenna. The first class, sensillum trichoideum B1, is absent on the first two annuli, increases in number until the fifth or sixth annulus, and reaches the highest number at the tip (Esslen and Kaissling, 1976). Sensors of the other class, sensillum trichoideum B2 (the ‘hooked hair’ or ‘‘Hakenborste’’ of Martin and Lindauer, 1966), only occur on the tip (Esslen and Kaissling, 1976). Contactchemoreceptive hairs, sensillum trichoideum D (or trichoideum D II; ‘‘spearlike hair’’ or ‘‘Spiessborste’’ of Martin and Lindauer, 1966), occur in high numbers on the first and around the sixth annulus, but are most abundant at the tip (Dostal, 1958; Esslen and Kaissling, 1976). Drones and workers are similar regarding the distribution of the different mechanoreceptive structures, but the numbers of sensilla may be slightly different. This difference, however, is not as large as for the olfactory sensilla, which are reported to be 19 000 in drones compared to only about 4800 in workers (Esslen and Kaissling, 1976). Besides the gradients of sensilla along the flagellum, the form of the tip of the honeybee antenna is also very peculiar. In both drone and worker, the last annulus has a flat, oval area that faces posterio-laterally (‘‘sensory plate’’ of Martin and Lindauer, 1966; Esslen and Kaissling, 1976). Furthermore, most mechanosensors are concentrated close to the tip on the anterio-lateral and medial side (Esslen and Kaissling, 1976). Here, a special arrangement of the second class of mechanosensory ‘hooked hairs’, and contact-chemosensory ‘spear-like’ hairs is reported. A single gustatory hair is surrounded by a number of mechanosensory hairs, which are arranged in a circle; and there are three such rings at the tip (Martin and Lindauer, 1966). This special arrangement of sensory structures probably is very significant for the precision of combcell building by workers (Martin and Lindauer, 1966), and may be also for pattern recognition on flower petals (see Section 6.2.3). It is not clear if another specific arrangement of sensory hairs and/or the concentration of sensory structures at the tip of the antenna play a role for discerning microstructures on the surfaces of flowers (Kevan and Lane, 1985). Noteworthy, though not yet convincingly explained, is the fact that the ventral side of the honeybee flagellum, i.e. the side

100


facing the scape when flexed, is nearly devoid of sensory organs (Dietz and Humphreys, 1971). 3.2.4.2 Considerations about functional significance. In insects, sensory hair plates are present on the scape and pedicel, and most of the model insects have a ring of campaniform sensilla around the distal edge of the pedicel. In some species, the distribution of mechanosensory and gustatory hairs shows a maximum near the middle of the flagellum, but, most often, a second maximum of sensillum density is located at the tip of the flagellum. Furthermore, the highest concentration of sensory structures occurs on the very tip of the flagellum. Such a concentration, and in some cases, special pattern of sensillum arrangement is found across species. The number of chordotonal organs and the number and distribution of gustatory hairs as well as other sensory structures varies between species. In some cases, there is also sexual dimorphism. Different receptors respond to different ways of mechanical stimulation, e.g. stress in the exoskeleton, bending of the flagellum in a certain joint, or bending of one or more hairs due to touch, and are distributed in different patterns over the antenna. Thus, the activity pattern of a specific group of antennal sensilla can represent only one aspect of the movement of the antenna or external stimulation of only one part of the antenna. Only the pattern of activity of all the different antennal sensors across the antenna can fully represent movement or stimulation of the antenna as a whole. However, in principle, all antennal mechanosensory structures can function as both, proprio- and exteroreceptors, depending on the source of the transduced force. Moreover, because the antennae are actively moved, the representation of the external space by active antennal sensing is ambiguous. A somatotopic arrangement of afferent fibres in the CNS alone will not be sufficient to represent the complex stimulus pattern arising from self-motion-induced and external stimulation of the antennal sensors. Possible mechanosensory self-stimulation due to active movement cannot easily be discerned from an external stimulation, thus, implying the necessity for central modification or gating of the sensory input, e.g. the postulated ‘‘re-afference principle’’ of Holst and Mittelstaedt (1950; see also Section 5.25.2). Moreover, biomechanical properties of the antenna may also help to discriminate between exteroreceptive and proprioreceptive cues (cf. Sections 2.2.2.1 and 7.2). 3.3

PHYSIOLOGY OF ANTENNAL MECHANOSENSORY NEURONS

The mechanical stimuli detected by the antennae include tactile stimuli, proprioreceptive signals and graviception (Horn, 1983; Horn and Bischof, 1983; Section 6.1.1), which involves a variety of internal and external mechanosensory neurons. The plethora of anatomical data on the structure of antennal sensilla (Section 3.1) or the knowledge about olfactory coding (see Shields and Hildebrand, 2001, or Stocker, 2001, for recent reviews) is contrasted by the


101

small number of electrophysiological studies of antennal mechanosensors. In this section, we will discuss studies that focus on aspects of proprioreceptive physiology relevant for behaviour. Bending of the locust flagellum is detected by a ring of some 70 campaniform sensilla around the distal end of the pedicel. The directional selectivity of each sensillum follows a cosine function of deflection angle, with excitation being strongest if the flagellum is bent away from it, i.e. by increasing the longitudinal strain across it (Heinzel and Gewecke, 1979). If the flagellum is deflected towards the sensillum, i.e. compressing the cuticle, the response ceases. The step response of each sensillum consists of a phasic–tonic receptor potential and spike train. Both the plateau of the receptor potential and the spike frequency increase proportionally with increasing deflection of the flagellum until a saturation level is reached at 1.31. An antennal strand receptor organ is characterised by Braünig (1985). This receptor is present in locusts and, probably, in crickets (cf. Staudacher and Schildberger, 1999). The strand spans the HS-joint and is innervated by four to five sensory neurons with central cell bodies (see Section 3.1.4). Depression of the HS-joint leads to an elongation of the strand which is linear between 01 (HS-joint completely elevated) and 901; further depressions yield little or no additional increase in length (Fig. 9A). A full depression over 1501 corresponds to a 1.5-fold increase of the resting (01) length of the strand. The sensory neurons of the strand receptor organ are tonically active at rest. Elongation of the strand leads to an increase in the firing frequency of at least three discernible units. Accordingly, relaxation of the strand leads to a decrease in firing frequency (Fig. 9B). Allgaüer (1989) describes the responses of sensory neurons of the scapal chordotonal organ (sCO) of crickets (and of unidentified pedicellar sensory neurons) to forced deflections of the antennal base, as revealed by intracellular recordings. As the y-shaped sCO spans the SP-joint, it is well suited to monitor position and horizontal movement of the pedicel. The sensory neurons encountered by Allgaüer (1989) are position-, velocity-, and acceleration-sensitive to various extent and, therefore, resemble the sensory neurons of other chordotonal organs (e.g. Hofmann et al., 1985; Matheson, 1990). Accordingly, ablation of the sCO leads to a decrease of the precision of horizontal antennal tracking movements (Honegger, 1981) and to a decrease of the overall angular range of the movements (Kammerer and Honegger, 1988). The scapal hair plate of cockroaches (Periplaneta americana) plays a major role for detecting objects in the horizontal plane (Okada and Toh, 2000; Okada et al., 2002; see Section 6.2.2). Each hair of the hair plate sits in a moveable socket with a V-shaped ridge restricting the proximal movements of the hairs and thus imposing directional selectivity. Tungsten electrode recording from the base of individually deflected hairs reveals no resting activity (Okada and Toh, 2001). Upon increasing deflection, hairs start to fire in a phasic–tonic manner beyond a deflection threshold of 61. The hairs can be kept activated for as long

102


FIG. 9 (A) The length of the connective tissue strand of the antennal strand receptor depends on the angle of the HS-joint. Inset explains the definition of the angle. The length of the strand increases almost linearly with increasing depression of the scape in the range between 01 and 901 in three locusts (different symbols). (B) Extracellular recording from strand receptor neurons (upper traces, recordings; lower traces, stimulus monitors in Bi, ii, iii). (Bi) The firing frequency increases upon elongation of the receptor strand (E) and decreases upon shortening (S in Bii). The firing rate of the neurons reflects the sinusoidal stimulation of the receptor strand (Biii). Adapted from Braünig (1985), with permission from Wiley and Sons, Inc.

as 10–20 min. Both the phasic and the tonic components of the response are dependent on the amplitude of deflection: the phasic, but not the tonic component increases with increasing deflection velocity. Phasic spike rates exceed 200 Hz upon large and fast deflections. The hair plate is subdivided into medial, dorsal and lateral portions, which consist of 30, 120 and 60 hairs, respectively


103

(Okada and Toh, 2001). Most importantly, the hairs are covered and thus deflected by the cutaneous membrane of the HS-joint in a way that depends on the angular position of the HS-joint. Depending on their relative location, different afferents will be bent by a different degree. The pattern of activated hairs and the strength of their activation is, thus, suitable to code for horizontal and vertical HS-joint angles for sustained periods of time. This is the case in the cockroach HS-joint investigated by Okada and Toh (2001), but also in other body parts, e.g. the cervical bristle fields of the bee (Thurm, 1963). The number of hairs, their respective angle of bending and the distribution of the bent hairs across the hair plates accounts for an ‘across-fibre-pattern’ (Pfaffmann, 1955), representing a population code of antennal posture. The studies discussed above cover campaniform sensilla, strand receptors, chordotonal organs and hair plates. The behavioural relevance of other antennal mechanoreceptive neurons, such as the mechanosensory or gustatory hairs, remains so far to be studied. The picture emerging from these studies implies that the physiology of antennal mechanoreceptive neurons does not fundamentally differ from that of mechanoreceptive neurons of other body parts, e.g. the legs. The ring-like arrangement of the campaniform sensilla of the distal pedicel can be considered to be an antennal specialisation reflecting the connection between the pedicel and the flagellum which can be deflected into all directions. Other specialisations, which can be expected, such as a tuning of chordotonal neurons to high movement velocities (see Section 6.2.1), have still to be investigated.

4

Neuroanatomy of antennal mechanosensory and motor pathways

4.1 4.1.1

MECHANOSENSORY NEUROPILS

Insects

In contrast to the antennal olfactory pathways (Boeckh and Tolbert, 1993; Hildebrand, 1996; Laurent, 1996; Menzel and Mu¨ller, 1996), antennal nonolfactory afferent projections have not been adequately elucidated, despite the behavioural importance of the arthropod tactile sense (cf. Gewecke, 1975; Pelletier and McLeod, 1994; Du¨rr and Krause, 2001; see Section 6). Thus, the part of the deutocerebrum outside the antennal lobe (AL) can only be described as a diffuse neuropil, anatomically and metaphorically. Some regions of the diffuse neuropil are often referred to as ‘antennal mechanosensory and motor center’ (AMMC, Rospars, 1988; Homberg et al., 1989), which does not take into account that there are a number of different mechanosensory structures at different locations along the antenna. So far, only a few studies provide data on antennal mechanosensory afferents and their arborisations in the DL of the deutocerebrum of a locust, and these form a basis for ascribing

104


functional significance to these areas (Gewecke, 1979; Braünig et al., 1983; Braünig, 1985). A single study in Gryllus includes afferent fibres from the flagellum and shows three distinct areas of termination of antennal receptors: the AL, the DL and the ventral area of flagellar afferents (VFA; Staudacher and Schildberger, 1999). Antennal sensory afferents, regardless of their modality, project towards the deutocerebrum via side branches of the two main branches of the antennal nerve N1 (see Fig. 10; Gewecke, 1972b; Braünig, 1985). Receptors from the flagellum enter the brain exclusively via the antennal nerve (Nervus antennalis; N1). In the pedicel, two additional side branches contain the axons of Johnston’s organ, the pedicellar hair plates, and other sensory structures. The axons of the sensory structures of the scape, scapal hair plates and chordotonal organs, enter N1 via separate side branches in the scape. The Nervus antennalis lateralis and Nervus antennalis medianus (N1l and N1m, respectively) branches fuse to a joint N1, either within the scape or shortly after entering the head capsule. The antennal stretch receptor enters N1 via a separate nerve within the head. Variations of these branching patterns may be expected in other species, because this picture is based on the situation in the locust. In all insects, the antennal nerve N1 enters the deutocerebrum from ventral and more or less lateral, with respect to the embryonic neuroaxis (cf. Boyan et al., 1993). At the point of entry into the deutocerebrum, the antennal nerve contains the axons of all the antennal sensory cells and some motoneurons. Immediately upon entering the brain, however, the axons divide into several bundles, or tracts, projecting to different areas in the deutocerebrum, where they terminate in distinct neuropils. To date, there is not a single systematic study about the branching patterns of the axons from different parts of the antenna and/or different modalities in any single insect. Thus, a general picture about the gross neuroanatomy of the sensory neuropils and the origin of the axon end terminals in the different neuropils can only be assembled from studies in different species (bee: Pareto, 1972; Suzuki, 1975; Arnold et al., 1985; locust: Gewecke, 1979; Braünig et al., 1983; Braünig, 1985; cricket: Staudacher and Schildberger, 1999; cockroach: Staudacher, unpublished). However, because it seems that reported differences in the numbers of tracts are likely to be due, at least in part, to the experimental approaches (for example, in the honeybee, four, six, and seven tracts have been discerned by Pareto, 1972; Suzuki, 1975 and Arnold et al., 1985, respectively), the emerging picture should be regarded with some caution. In all insects studied so far, a large portion of the fibres entering the deutocerebrum, namely those of tracts one to four (T1–4) terminate in the glomeruli of the (olfactory) AL (Fig. 11; e.g. moth: Rospars and Hildebrand, 1992; butterfly: Kim et al., 1990; honeybee: Pareto, 1972; Suzuki, 1975; Arnold et al., 1985; cricket: Staudacher and Schildberger, 1999; locust: Hildebrand, 1996; Laurent, 1996; cockroach: Distler and Boeckh, 1996; fly: Strausfeld, 1976). Most of the afferents branching in the AL are considered to be olfactory. Tracts one to four


105

FIG. 10 Schematic view of the antennal nerve and the side branches, which run towards the various sensory structures inside and on the surface of the locust antenna. The nerves and branches are named according to Gewecke (1972b) and Braünig (1985).

separate from the other afferents early on, i.e. ventrally. Fibres from tracts one and two run towards the core of the AL. From there, they enter the glomeruli of the AL, where they arborise (Apis mellifera L.: Pareto, 1972; Suzuki, 1975; Gryllus bimaculatus: Staudacher and Schildberger, 1999). Tract three fibres run along the posterior margin of the AL to enter some glomeruli from there (bee: Pareto, 1972; Suzuki, 1975). The fourth tract also runs along the posterior margin of the AL, and some of its fibres also enter the glomeruli, while others form a small, not very dense neuropil between antennal and dorsal lobes (bee: Suzuki, 1975). In the cricket Gryllus bimaculatus (Staudacher and Schildberger, 1999), and the cockroach Gromphadorhina portentosa (Staudacher, unpublished observation) the situation seems to be very similar, indicating that these tracts may be homologous. Upon entry of the antennal nerve into the deutocerebrum, tracts five and six separate from the others and bypass the AL. They are reported to run posterior to the AL in the honeybee (Suzuki, 1975), but dorsal to the AL in the cricket

106


FIG. 11 Comparison of the antennal neuropils in the deutocerebrum of (a) cricket and cockroach (Staudacher and Schildberger, 1999; Staudacher, unpublished) and (b) the honeybee (Pareto, 1972; Suzuki, 1975; Arnold et al., 1985; Maronde, 1991). The similarities in the layout of the deutocerebrum are obvious. However, there are also differences. Tracts T6III and T6IV have not been described for a cricket or cockroach, while the T7 (vfa area) has not been described in the bee. Tracts running from the antennal nerve to the different neuropils are named T1—T7. Broken lines indicate the borders of the deutocerebrum. Note, that the embryonic neuroaxis, which is tilted by about 901 relative to the head axis, is used in both schemes.

(Staudacher and Schildberger, 1999). Tract five contains sensory cells, which terminate and arborise in an area specifically called the DL. Tract six neurons have arborisations in the DL, but also send collaterals into the posterior protocerebrum (bee: T6I and T6III of Suzuki, 1975; Arnold et al., 1985) or into the anterior deutocerebrum (cricket, cockroach: T6I). Furthermore, the same cells have collaterals that project into the SOG (T6II). In the bee, the T6II fibres have some bilateral arborisations in the SOG (Pareto, 1972; Maronde, 1991), and some ipsilateral collaterals reach into the more posterior SOG (T6IV of Maronde, 1991). In the cricket and a cockroach, these collaterals stay ipsilaterally in the dorsal to medial SOG (Staudacher and Schildberger, 1999; Staudacher, unpublished). The branching patterns of the fibres in tract six are similar to what has been reported for afferents from the scapal and pedicellar hair plates, the campaniform sensilla of the distal pedicel and Johnston’s organ in locusts (Gewecke, 1972b; Braünig et al., 1983). It remains to be shown, however, that these receptor groups have the same arborisation patterns in


107

cricket, bee, and in other insects, too. Recent work in progress on Periplaneta attempts to do this (Baba and Comer, 2003). So far, there is only a single description of an additional tract seven, found in the cricket G. bimaculatus (Staudacher and Schildberger, 1999). Tract seven fibres separate from the main bundle ventrally to run in a posterior–medial direction. Numerous small, closely packed fibres terminate and form a neuropil of dense appearance with conspicuous parallel bands in the ventral posterior deutocerebrum. This neuropil is, thus, posterio-lateral to the AL, and posterioventral to the DL. In the cricket, and the cockroach Gromphadorhina portentosa (Staudacher, unpublished observation), this is the third antennal neuropil in the deutocerebrum. Because the afferents terminating in this area are presumed to arise from the flagellum, this neuropil is called VFA (Staudacher and Schildberger, 1999). The striped appearance of this neuropil may be an indication for a somatotopic arrangement of the receptor terminals. This would not be surprising, since many insect first-order neuropils have been shown to receive orderly afferent projections with regard to location or functional properties (e.g. auditory afferent fibres; Ro¨mer, 1983; Ro¨mer et al., 1988), leg hair afferents in Drosophila (Murphey et al., 1989), in the locust middle leg (Mu¨cke and Lakes–Harlan, 1995; Newland et al., 2000), locust hind leg (Newland, 1990; Burrows and Newland, 1993) and the cricket cercal system (Jacobs and Theunissen, 1996; Paydar et al., 1999). Under the assumption of a somatotopic arrangement, one might speculate on the behavioural role of the VFA neuropil (cf. Camhi and Johnson, 1999). However, only a detailed analysis of the VFA anatomy and physiology will elucidate, if this area has a functional significance for active tactile sensing. The emerging picture seems to be that the insects investigated so far are very similar with regard to tracts T1–4, which mainly contain axons terminating in the AL (Fig. 12). There seem to be some slight differences with regard to tracts T5 and T6, which could reflect species-specific differences. Another point is that tract T7 and the associated VFA area have not been described in the honeybee. Could this reflect the difference in length between honeybee and cricket or cockroach antennae, or be linked to the different ways in which these insects use their antennae? Some hints may also come from a comparison with crustacea. 4.1.2

Crustacea

The insect antenna is homologous to the crustacean antennule, which seems to be less important as an active tactile organ than the crustacean antenna (which lacks a homologue in insects). Therefore, it is interesting to see if the functions of the two crustacean head appendages and their different usage have an impact on brain anatomy. One question would be: ‘do structures related to the antenna show analogies to the insect deutocerebrum?’ Afferents from the first appendage, the antennule, project to the deutocerebrum. All olfactory afferents run towards the olfactory lobe, while the other

108


FIG. 12 Schematic representation of how different types of sensilla from various locations along the antenna project via various tracts to the different neuropils in the CNS, especially the deutocerebrum. Arrows indicate projections of the sensory cells, while the labels above the arrows indicate the receptor cell locations along the antenna. Hypothesized projections are marked with question marks (?).

afferents project into the lateral and median antenna I neuropil (Sandeman et al., 1992; Schmidt and Ache, 1992). Non-olfactory afferents from the antennule enter the deutocerebrum via the nerve of antenna I, and project into the bi-lobed lateral antennular neuropile (LAN; Schmidt et al., 1992). Thin and medium-sized axons are usually oriented along the long axis of this neuropil, and many send perpendicular branches into the neuropil, which stay within the same lobe, and are regularly spaced (50–70 mm; Schmidt et al., 1992). Motoneurons project only into the ventral areas of the LAN (Schmidt and Ache, 1993). In the median LAN, they branch in a conspicuous pattern with regularly spaced side branches (50–70 mm), while the arborisations are not as regular in the lateral LAN (Schmidt and Ache, 1993). Physiological results confirm what was suggested by the anatomical results, which is that the LAN is an area of integration of mechanosensory and olfactory information, and is a ‘lower’ motor centre for the antennule (Schmidt and Ache, 1996). The antenna II neuropil is located in the tritocerebrum, i.e. the primary projection area of the antennal afferents; has an elongated tubular appearance. It receives input from antennal sensory neurons and projections from motoneurons and various interneurons (Sandeman et al., 1992). In Procambarus clarkii, about 2000 sensory hairs from the flagellum project to the brain (Tautz et al., 1981). Afferents run in four tracts along the margin of the AL, but with regularly spaced (25–35 mm), perpendicular branches reaching into the lobe. Ascending and descending interneurons, and some local interneurons show similar arborisation patterns in the AL (Tautz and Mu¨ller-Tautz, 1983). This could be interpreted as a somatotopic arrangement in this neuropil. Motoneurons, which supply the muscles of the second antenna, enter the tritocerebrum through a variety of nerve branches. They send regularly spaced branches into the AL, and collaterals into the diffuse areas of the tritocerebrum


109

(Habig and Taylor, 1982a,b). Sandeman et al. (1993) report a striking observation. They compared the structures of the antennal neuropils across different Decapod species. In the crab, which has the shortest antenna, the neuropil is least well structured. The neuropil is highly structured in species with the longest antennae, both of which are shrimps. An intermediate situation is found in crayfish, the antennae of which are shorter than in shrimps and longer than in crabs (Schram, 1986; Sandeman et al., 1993; Ruppert and Barnes, 1994). It may prove useful to quantify such correlations with regard to body vs. antenna length, distribution of mechanosensory structures, etc. in both crustacea and insects. Knowledge about the different life styles, and use of the antennae may also prove to be useful in this regard. The compilation of such diverse information may lead to a better understanding of how brain structures are modified through evolution. The afferents from both crustacean antennae project to highly structured brain areas, which may be regarded as integration centres for antennal motor control. But their structure also indicates somatotopic arrangements of afferents and the branches of other neurons (Habig and Taylor, 1982a,b; Tautz and Mueller-Tautz, 1983; Schmidt and Ache, 1993), which may be well suited for roles in active tactile sensing. Two of the projection areas of antennule afferents in the deutocerebrum are laminated and receive both mechanosensory and chemosensory inputs (Schmidt et al., 1992; Schmidt and Ache, 1996; Strausfeld, 1998). Insects like firebrats (Thermobia) and silverfish (Lepsima) have uniform flagellar segments and laminated neuropils in the deutocerebrum, which seem to be structurally very similar to the LAN and medial antennular neuropile (MAN), but these insects do not have an olfactory lobe (Strausfeld, 1998; Strausfeld et al., 1998). Is there a link between these areas and the VFA, not only structurally, but also functionally? Another question arises from the correlation of antennal length and the degree of order in the tritocerebral neuropils of Crustacea (Sandeman et al., 1993). Could the difference with regard to the VFA between cricket and cockroach, on the one hand, and the honeybee on the other, indicate a similarity of these cases with respect to the length of the flagellum?

4.2

CONNECTIONS TO OTHER PARTS OF THE CNS

Mechanosensory stimulation of the antenna and the corresponding activity of mechanosensory afferents are important in a wide variety of behavioural contexts. Thus, it is not surprising that input of mechanosensory afferents is integrated with input from other sensory modalities through a variety of connections within the CNS. Anatomically, this is manifested by the overlap of afferents and many different types of neurons like motoneurons, local interneurons, some of which project to other brain neuropils, and ascending and descending interneurons.

110

4.2.1


Connections with motoneurons

The finding that mechanosensory afferents overlap with antennal motoneurons is not surprising. In the cricket, the locust and the stick insect, antennal motoneurons project to the dorsal deutocerebrum, where they may overlap with some, probably mechanosensory, antennal afferents (Honegger et al., 1990a; Bauer and Gewecke, 1991). In Manduca and Apis, afferents and motoneurons project to the same area in the deutocerebrum (Kloppenburg, 1995; Kloppenburg et al., 1997). Double labelling experiments have further substantiated this overlap in the bee (Kloppenburg, 1995). As yet, it is not clear if this overlap means direct connections between those classes of cells. 4.2.2

Connections with local brain neurons

A variety of local brain neurons have been described, which connect the antennal mechanosensory areas with the AL (Section 4.2.2.1) and various other brain regions (Section 4.2.2.2). Moreover, there also are connections to other parts of the CNS (Section 4.2.3). 4.2.2.1 Antennal lobe (AL). Some neurons recorded in the AL respond to mechanical stimulation. Because nothing is known about direct mechanosensory input to the AL, this is interpreted as evidence for indirect connections between olfactory and mechanosensory neuropils in the deutocerebrum (Waldow, 1975, 1977; Boeckh et al., 1983, 1984; Kanzaki et al., 1989; Zeiner and Tichy, 1998; Hill et al., 2002). For some of these neurons in cockroach, direct connections to other neuropils can be ruled out (Boeckh et al., 1984). A number of projection neurons of the male moth AL respond to clean air stimulation, which should only affect the mechanosensory afferents, but none of these neurons has deutocerebral branches outside the AL (Kanzaki et al., 1989). Similarly, bee AL interneurons respond also to stimulation with air alone, even though they may not have branches outside the olfactory neuropil (Flanagan and Mercer, 1989). However, only little is known about neurons which actually have projections in both neuropils. A distinct class of bee AL interneurons (‘‘D(M)Ho’’) with branches in the AL, but also in other areas of the deutocerebrum, responds to olfactory and mechanosensory stimulation (Flanagan and Mercer, 1989). Within this group, the morphology is known for only one neuron, which has arborisations in the glomeruli of the dorsal AL and part of the mechanosensory areas in the cockroach (Malun et al., 1993). This neuron sends a collateral into the outer antenno-cerebral tract, and arborises en route (Malun et al., 1993). It is obvious, that all these cells take part in the integration of olfactory and mechanosensory input, but their precise role and impact on the behaviour of the animal is, as yet, elusive. 4.2.2.2 Other brain regions. Maronde (1991) reports three areas of overlap between antennal afferents and visual interneurons from the lobula, namely (1)


111

T6I, (2) T6III and (3) the DL ( ¼ T5). These three areas get input from sensory structures of the scape and pedicel, and from visual neurons of the lobula. These interneurons have one arborisation area in the ipsilateral lobula and blebby branches in either the ipsilateral, contralateral or both DLs. They show a variety of responses to visual stimulation (Hertel and Maronde, 1987). Connections like these may have functional roles in behavioural responses like the direction-sensitive antennal response, which can be elicited by stimulation of the ipsilateral compound eye (Kloppenburg and Erber , 1995). 4.2.3

Connections to ascending and descending neurons

The antennal mechanosensory areas are connected to other parts of the CNS via cells ascending from, and neurons descending to the ventral nerve cord. Some of these interneurons have also been labelled in immunohistochemical studies, and will, therefore, be mentioned later (see Section 4.3). Some DUM neurons from the SOG project to the antennal mechanosensory and motor system, and may have significant arborisations in the deutocerebrum. Some of the cells, which were described in the locust (SA 2, SA 4, SA 5 and SAD 1), have rather obvious arborisations in the dorsal deutocerebrum (Braünig, 1991). One of these cells, SA 5, has branches in both the olfactory and mechanosensory areas of the deutocerebrum. They may be octopaminergic and have input regions in the SOG and output areas in the brain. This may indicate, that these neurons are involved in neuromodulation of activity in their target areas of the brain (Braünig, 1991), but direct physiological evidence is missing. Arborisations of two other, non-DUM, ascending neurons, one suboesophageal neuron and another cell with unknown soma location, have been shown in the cricket AMMC (Horseman et al., 1997). The ascending input of these, and other cells is reported to have a rather strong impact on the function of the antenna (Horseman et al., 1997). Even more striking is the finding of release sites of a suboesophageal DUM neuron onto antennal muscles that have been reported for locust and cricket (Braünig et al., 1990). These indicate a direct modulation of muscle activity by ascending input. A large number of ipsi- and contralateral descending brain neurons (DBNs) have been described (e.g. Strausfeld et al., 1984; Ho¨rner and Gras, 1985; Brodfuehrer and Hoy, 1990; Gronenberg and Strausfeld, 1990; Kanzaki et al., 1991; Bo¨hm and Schildberger, 1992; Hensler, 1992). Some of these cells have branches in the mechanosensory areas of the brain, and probably receive input from antennal mechanosensory neurons, but may also have output regions in this area (Okada et al., 2003). The antennal mechanosensory input is integrated with other modalities (Ho¨rner and Gras, 1985; Staudacher, 1998a,b, 2001; Staudacher and Schildberger, 1998, 1999) and relayed to the thoracic and abdominal ganglia of the ventral nerve cord, where descending neurons usually branch in the dorsal to medial neuropil areas (Ho¨rner, 1992; Staudacher and Schildberger, 1998; Staudacher, 2001).

112


A variety of descending neurons in the locust get ocellar input, and input in the dorsal protocerebrum and deutocerebrum. Three of these cells (DNI, DNM, DNC; Griss and Rowell, 1986) are of major importance for locust flight stabilisation (Reichert and Rowell, 1986; Reichert, 1993). Another large descending neuron with branches in the non-olfactory areas of the deutocerebrum is TCG (Bacon and Mo¨hl, 1983), which has a homologue in Periplaneta (Gundel and Penzlin, 1995). This neuron is wind sensitive, which may in part be due to input from hairs on the head of the locust, but it also responds to antennal movement (Bacon and Mo¨hl, 1983). It would be interesting to know, if these cells receive input from Johnston’s organ and the campaniform sensilla on the distal pedicel, and how important these inputs are for their functional role. Most descending neurons have at least some branches in the mechanosensory deutocerebrum, but most striking are those that have rather large arborisations in these areas. In Periplaneta, a large contralateral descending neuron (DMIa-1) has a number of branches in the medial protocerebrum, but major arborisations in an area that may resemble the VFA in crickets (Burdohan and Comer, 1990, 1996; Ye and Comer, 1996). This neuron is reported to play a major role in a cockroach escape reaction, which is triggered by tactile and wind stimulation of the flagellum (Comer et al., 1994; Stierle et al., 1994). A possible homologue of this cell (DBNc2-2) has been described anatomically, but never tested physiologically (Staudacher, 1998b). Six contralateral and two ipsilateral descending neurons with major branches in the mechanosensory areas of the deutocerebrum were described in the cricket (Staudacher, 1998a,b; Staudacher and Schildberger, 1999; Gebhardt and Honegger, 2001). All these cells have branches in other brain areas and, thus, can be considered multimodal, but their main arborisations are in the nonglomerular areas of the deutocerebrum. A detailed description of the branching areas of two ipsilateral (DBNi1-2 and DBNi2-1) and one contralateral (DBNc2-3) cell demonstrates a correlation with the arborisations of simultaneously labelled antennal afferents (Staudacher and Schildberger, 1999). The three neurons show differential overlap with the afferents. DBNi2-1 branches overlap exclusively with afferents in the DL (T5 and T6). The branches of the other two cells overlap only with afferents in the VFA. Here, the contralateral descending neuron arborises in the entire neuropil, while the ipsilateral descending cell overlaps with the afferents at median depth in the neuropil (Staudacher and Schildberger, 1999). These results illustrate the need for more detailed studies on afferent projection areas, because to date it is uncertain if and how the afferent inputs are spatially separated in the deutocerebrum. Thus, one can only speculate about the functional significance of the differences in the branching patterns of the interneurons. Like other inputs to the CNS, the mechanosensory afferents are connected to a variety of areas in the CNS. There are local deutocerebral connections to the AL and other parts of the deutocerebrum. Via brain interneurons, the


113

mechanosensory neuropils are also connected to visual neuropils, the lobula and the medulla. A large number of descending neurons has a few branches in this area of the deutocerebrum. Eight descending cells have major arborisations there, which seem to be restricted to very specific sub-regions of the mechanosensory neuropils. A number of ascending cells from the SOG and the ventral nerve cord project into this region. One has to be very cautious in assigning directions of information flow to these connections, because it has been shown in a couple of cases, that input and output sites may be close to each other, that information can travel in two directions (Hughes and Wiersma, 1960; Watson and Burrows, 1983; Watson and Pflu¨ger, 1989; Malun, 1991; Homberg, 1994). Nevertheless, the variety of connections show, that mechanosensory input from the antenna is integrated on a variety of levels with other input to, and activity in the CNS. It is an exciting task now, to study how these anatomically defined connections affect the control of the antennal movements, and how significant they are for the integration of antennal mechanosensory responses with ongoing behaviour or changes in behaviour. 4.3

IMMUNOCYTOCHEMISTRY

So far, only relatively basic studies of the neuroanatomy of antennal mechanosensory afferents are available, and the corresponding antennal mechanosensory circuits in the deutocerebrum (DC) are not known in detail. Immunohistochemistry, if reported in sufficient detail, and combined with knowledge of the underlying circuits may permit the formulation of hypotheses about the functional properties of these pathways. Some immunohistochemistry has been done on insect antennae (Section 4.3.1), on motoneurons (Section 4.3.2), and other neurons with arborisations in the brains of some of the model organisms (Sections 4.3.3). 4.3.1

Antennal mechanoreceptors

4.3.1.1 Acetylcholine. The most abundant transmitter in the insect nervous system (Breer, 1981; Breer and Sattelle, 1987), acetylcholine (ACH), is also present in antennal sensory cells. Indirect evidence for cholinergic cells has been provided by immunohistochemistry of the enzyme acetylcholine esterase (ACHE) in the bee brain. In this study, the glomeruli of the AL were labelled, but a much stronger stain was reported for the DL, including the posterior DC ( ¼ dorsal DC), fibres projecting into the tritocerebrum (TC), and the SOG (Kreissl and Bicker, 1989). The latter projections may be T6I and T6II, which were described by Suzuki (1975; Section 4.1.1). A similar result is reported for a locust, where the antennal nerve, AL and AMMC stain strongly with an ACHE antibody (Rind and Leitinger, 2000). A recent study using in situ hybridisation of an ACH receptor subunit in the bee brain provides more indirect evidence that ACH is the transmitter of mechanosensory afferents. Here, the

114


DL is weakly stained, a hint towards the presence of neurons, which are postsynaptic to cholinergic cells (Thany et al., 2003). Further evidence comes from an autoradiography study that shows that a-Bungarotoxin, which specifically blocks nicotinic acetylcholine receptors, strongly stains the antennal and the DL of the bee (Scheidler et al., 1990). Most important, however, these studies are in accord with results reported for moth antenna. Here, a number of mechanoreceptor neurons on the antennae of Manduca contain ACHE, which is widely accepted as an indicator for cholinergic neurons (cf. Stengl et al., 1990). These comprise the mechanosensory hairs on the scape and the pedicel, Johnston’s organ in the pedicel, mechanosensory hairs (sensilla chaetica), on the annuli of the flgellum, and possibly campaniform sensilla on the tip of the antenna of male moths (Stengl et al., 1990). Thus, even though not all of these results stem from the same species, and even though many other species have not been studied, it may be stated that acetylcholine probably is the predominant transmitter of the mechanosensory afferents. 4.3.1.2 Serotonin. In Periplaneta, a few serotonergic axons are described at the base of the antennal nerve, and these can be traced into the DL (Salecker and Distler, 1990). They might be the axons of neurons from internal mechanosensory organs, if the result of a study on locust leg sensilla can be transferred: both, external and internal mechanosensory neurons of the locust leg can be stained with an ACHE antibody, suggesting ACH as a transmitter (Lutz and Tyrer, 1988). Moreover, internal and external receptors can be discerned by their immunoreactivity with antibodies against serotonin, which only stain internal receptors (Lutz and Tyrer, 1988). Such a method could, in principle, be used to identify and discern the branching areas of internal and external antennal mechanoreceptors. This could be an important tool to gain more detailed insight into the functional organisation of the mechanosensory areas in the deutocerebrum. 4.3.1.3 Nitric oxide. Sensory afferents from hairs, chordotonal organs and campaniform sensilla in locust leg show immunoreactivity against soluble guanylyl cyclase-a, which indicates that these neurons might be targets of interneurons using nitric oxide (NO) as a modulator (Ott et al., 2000). Nitric oxide synthase (NOS), which is found in NO-releasing cells, generates NO and can indirectly be stained by NADPH diaphorase. In cockroach and locust there is weak staining in the non-glomerular DC, and there are no individually discernible fibres (Ott and Elphick, 2002). The latter result could mean that antennal mechanosensory fibres are not the target of NO releasing interneuron, provided it could be shown, that they are not stained with antibodies against soluble guanylyl cyclase-a. 4.3.1.4 Taurine. So far, no purely taurinergic neurons (Bicker, 1992) have been described. Taurine is colocalised with octopamin, GABA, and therefore considered a neuromodulator, rather than a transmitter (Stevenson, 1999).


115

During bee larval development, all the different types of receptor neurons on their antenna react with an antibody against the amino acid taurine (Eichmu¨ller and Scha¨fer, 1995). In the last larval stage, cells in the scape and pedicel, and their axons, which can be traced into the deutocerebrum, can be stained with taurine antibodies. Taurine-like immunoreactivity (taurine-like IR) then decreases during the pupal stage. At pupal ecdysis, taurine–IR is reported for cells in the flagellum: both types of mechanoreceptive hairs (trichoid B1 and B2) and gustatory hairs (trichoid D: 5–7 cells) are stained at this stage (Eichmu¨ller and Scha¨fer, 1995). The finding that all the receptor types show taurine-like IR during development contrasts with the results of a study on adult bees (Scha¨fer et al., 1988). In adult bees, the glomeruli of the AL and some unidentified neurons show taurine-like IR. However, only weak taurine-like IR and no discrimination of single branches or cells is reported for the DL, while no presumed mechanosensory axons are stained (Scha¨fer et al., 1988). 4.3.2

Antennal motoneurons and muscles

The antennal muscles are innervated by excitatory and inhibitory motoneurons, the numbers of which are species specific (cf. Table 2; Sections 2.3 and 2.4). Immunohistochemical studies reveal similarities regarding excitatory and inhibitory transmitters, and modulatory substances despite the differences in the innervation patterns of antennal musculature. 4.3.2.1 Glutamate. Glutamate, the putative excitatory transmitter of insect motoneurons, and a precursor of GABA, shows a staining pattern similar to GABA in the bee brain. But the intensity of the labelling is significantly weaker in GABAergic neurons, and, thus, Glutamatergic and GABAergic neurons can be discerned (Bicker et al., 1988). In the bee, some large axons from the DL, which show glutamate-like IR, project into the antennal nerve, and are presumed to be motoneurons. Ten of these axons are stained in antennal motor nerve cross-sections and three distinct profiles in the sensory nerve (Bicker et al., 1988). The number of profiles counted in the immunohistochemical study (13; Bicker et al., 1988) is not expected, because 15 motoneurons are described (Kloppenburg, 1995), and there is no indication for inhibitory, GABAergic motoneurons in the bee (cf. below; Scha¨fer and Bicker, 1986b). The transmitter of the two remaining motoneurons is unknown. Seventeen excitatory motoneurons are described for the cricket (Braünig et al., 1990; Honegger et al., 1990a). Two of these motoneurons are glutamatergic (Bartos et al., 1995), but according to a recent study, this number could be higher (Schu¨rmann et al., 2000). An ongoing study is reported to show, that locust antennal motoneurons are glutamatergic (cf. Homberg, 2002). All these publications strongly indicate that glutamate is the excitatory transmitter of many insect antennal motoneurons.

116


4.3.2.2 Gamma-aminobutyric acid. GABA is an important inhibitory transmitter in insect CNS (Bicker et al., 1985; Scha¨fer and Bicker, 1986b; Honegger et al., 1990b). In Manduca, some axons were stained in the antennal mechanosensory and motor tract, and indicate the presence of an unknown number of GABAergic, inhibitory motoneurons (Hoskins et al., 1986). Thus, it remains to be shown, how many of the at least 12 motoneurons (Kloppenburg et al., 1997) in Manduca are inhibitory. The bee DLs are homogeneously stained, and a number of perikarya in the rind nearby are marked with GABA antibodies (Scha¨fer and Bicker, 1986b). However, no GABAergic stain is reported in the antennal nerve of the bee (Scha¨fer and Bicker, 1986a). This could mean that, unlike cricket (Honegger et al., 1990a), locust (Bauer and Gewecke, 1991), and stick insect (Du¨rr et al., 2001), each of which has a single inhibitory motoneuron, the bee has no common inhibitor. In the cricket, the antennal common inhibitor neuron was filled retrogradely, and then double labelled with an antiGABA (Honegger et al., 1990a), and in a double labelling experiment shown to contain Proctolin, like 13 other cricket antennal motoneurons (Bartos et al., 1995; Schu¨rmann et al., 2000). 4.3.2.3 Proctolin. Proctolin seems to play an important role in the modulation of the insect antennal motor system by prolonging muscle contraction (Section 2.4.1; Bartos and Honegger, 1997). In crickets, proctolin has been shown to be a co-transmitter in 14 of the 18 antennal motoneurons (Bartos et al., 1994, 1995; Bartos and Honegger, 1997). In the locust, where it has similar effects as in the cricket, immunoreactive terminals are found on the antennal muscles, and some somata are stained in the rind of the deutocerebrum (Bauer, 1991). Thus, one may conclude that Glutamate is the main excitatory, and GABA the inhibitory transmitter of antennal motoneurons. So far, proctolin is the only modulator shown to be present in, and important for the antennal motor system. In some cases, precise numbers of motoneuron axons stained in immunohistochemical experiments are missing or do not seem to account for all of the neurons expected to be found. Therefore, it is possible that other transmitters are present. Further studies are needed to elucidate this issue. In an immunohistochemical study on ACH, large profile neurons, which project into, and through the antennal nerve, and, thus, may resemble motoneurons, were not stained (Kreissl and Bicker, 1989). Thus, at least in the bee, ACH can be excluded from being a transmitter of the antennal motoneurons. 4.3.3

Other immunohistochemical results

The other immunohistochemical studies focus on describing the distribution of a given antigen within the brain. In most cases, they do not contain sufficiently detailed information to make them valuable for a detailed understanding of active tactile sensing. However, they contain hints with regard to transmitters


117

and modulators, which could be valuable for future experiments. Thus, they will be mentioned without lengthy discussions. Local GABAergic (Manduca: Homberg et al., 1987) and serotonergic cells (locust: Tyrer et al., 1984), octopaminergic (for review: Stevenson and Spo¨rhaseEichmann, 1995; bee: Mercer et al., 1983; Kreissl et al., 1994; locust: Konings et al., 1988; Stern et al., 1995; Stern, 1999) are described. In the bee, serotonin decreases and octopamine increases visually induced movements of a directionspecific antennal reflex (Erber and Kloppenburg, 1995). Several dopaminergic neurons have been identified (locust: Wendt and Homberg, 1992; bee: Scha¨fer and Rehder, 1989; physiology, bee: Flanagan and Mercer, 1989; cockroach: Waldow, 1975). A number of DBNs have been stained with immunohistochemical methods. Some are reported to be glutamatergic (bee: Goodman, 1981; Bicker et al., 1988), others to be GABAergic (tritocerebral dwarf, locust; Tyrer et al., 1988), some serotonergic (bee: Rehder et al., 1987; PI(2), locust: Williams, 1975; Tyrer et al., 1984; Hensler, 1988; PM3, Manduca: Homberg and Hildebrand, 1989), and some octopaminergic (bee: Kreissl et al., 1994; locust: Williams, 1975; Bacon and Tyrer, 1978; Bacon and Mo¨hl, 1983; Konings et al., 1988). Other studies describe branching patterns of ascending, and yet unidentified neurons. GABAergic (cricket: Schoch et al., 2001, 2002), and histaminergic fibres (cricket, bee: Bornhauser and Meyer, 1997), ascending branches of vasopressin-like immunoreactive neurons (locust: Thompson and Bacon, 1991; Thompson et al., 1991), leucokinin IR cells (cockroach: Naessel et al., 1992) are described. Allatostatin immunoreactivity is widespread (Stay et al., 1994; cricket, cockroach: Schildberger and Agricola, 1992; cricket: Neuhaüser et al., 1994; locust: Vitzthum et al., 1995, 1996). FMRFamide (cockroach: Verhaert et al., 1985; locust: Myers and Evans, 1987; Remy et al., 1988; Manduca: Homberg et al., 1990) and pigment-dispersing hormone IR (stick insect: Homberg et al., 1991) are reported. Gastrin-cholecystokinin does not stain branches in the DC (cockroach, locust: Tamarelle et al., 1988, 1990). 4.3.4

Immunohistochemistry summary

The results of all the neuroanatomical studies show a growing body of knowledge about the antennal mechanosensory system. However, to understand this system on a functional level, and to gain insight about the impact of the system on the behaviour of the animal, more details are needed. It seems to be of major importance to systematically compare the projection patterns of single receptor types from different locations along the antenna, and to do this across species. On the basis of this knowledge, descriptions of connections to other areas of the brain and the ventral nerve cord can, if reported in sufficient detail, reveal the overlap between interneurons and very specific areas of the afferent layer. Only, then will it be possible to understand the functional significance of the connections between mechanosensory afferents and different areas of the

118


CNS. Information gained from immunohistochemical studies needs to be more detailed with regard to mapping the precise areas of the stain, and more attempts are needed to identify single neurons. Such detailed results could provide testable hypotheses for neuroethological experiments. 4.4

ANTENNAL MOTONEURONS IN INSECTS

The peripheral innervation pattern of insect antennal motoneurons is outlined in Section 2.3.1 with an emphasis on the comparative layout of antennal nerves. Here, we concentrate on the aspects of the anatomy of antennal motoneurons within the deutocerebrum (or AMMC). Data are available for crickets (Honegger et al., 1990a), locusts (Bauer and Gewecke, 1991), honeybees (Kloppenburg, 1995), ants (Ehmer and Gronenberg, 1997b), the tobacco hornworm (Kloppenburg et al., 1997) and stick insects (Du¨rr et al., 2001). Moreover, Baba (2000) shows the reconstruction of one antennal motoneuron of the cockroach Periplaneta americana. Several important characteristics of the central motoneuron anatomy, sometimes shared by two or more species, emerge from these studies, as discussed in the following section. To enable inter-species comparisons, the embryonal neuroaxes will be used as an anatomical frame of reference (Boyan et al., 1993). 4.4.1

General organisation

All antennal motoneurons investigated so far, are restricted to the deutocerebral hemisphere ipsilateral to the muscles innervated, though dendrites may be located in the tritocerebrum in the cricket (Honegger et al., 1990a) or may almost reach the SOG of the fused brain in the honeybee and in Manduca (Kloppenburg, 1995; Kloppenburg et al., 1997). This resembles the unilateral layout of thoracic leg motoneurons, e.g. in the locust (for review, see Burrows, 1996). As a consequence of this unilateral layout, bilaterally co-ordinated antennal movements must be controlled by bilateral interneurons (cf. Section 5.2). Since antennal movements are sometimes co-ordinated with posterior body appendages like the pro-thoracic legs (Horseman et al., 1997; Du¨rr et al., 2001), again, interneurons have to connect between antennal and leg motor centres (see Sections 4.2 and 5.2). 4.4.2

Are antennal motoneurons ‘‘identified neurons’’?

Most of the antennal motoneurons studied so far have been characterised on the population level. In the honeybee, for instance, antennal motoneurons are difficult to discriminate because of their similar central branching patterns, but groups of motoneurons are discernible on the basis of their soma clusters (Kloppenburg, 1995). The situation is similar in Manduca, where the central ramifications of the motoneuronal dendrites overlap to a great extent. Only the


119

intrinsic muscle motoneuron somata tend to cluster in one region of the deutocerebrum (Kloppenburg et al., 1997). Bauer and Gewecke (1991) found the variation in the soma locations to be too large to identify motoneurons by their soma position, though the HS-joint and the SP-joint motoneuron somata tend to be separate in the dorsal (‘posterior’ after Bauer and Gewecke, 1991) deutocerebrum. In crickets, the antenno-motor system contains five morphologically identifiable individual motoneurons and five groups of two to three morphologically very similar motoneurons. The two ‘structural sets’ (Honegger et al., 1990a) of the intrinsic muscle motoneurons can further be discriminated on the basis of their physiological properties, but two of the three Ab2–4 abductor motoneurons are indistinguishable (Honegger et al., 1990a). Du¨rr et al. (2001) also describe the antennal motoneurons as individuals, with two sets of morphological twin motoneurons. It is to be expected from other insect motor systems that all of the antennal motoneurons could probably be individually identified upon an increase of knowledge of their properties, e.g. physiology, co-transmitter content or target muscle-fibres.

4.4.3

Soma positions

The study of Kloppenburg (1995) clearly demonstrates that, in the honeybee, somata of antennal motoneurons cluster in six groups. These clusters are located anterior and posterior to the DL and contain 15 somata. The numbers of somata per cluster vary between individuals. The three clusters of the extrinsic muscle motoneurons for instance, contain a maximum of nine somata with two somata consistently located in the posterior dorsal cluster; the remaining two clusters may contain three and four or two and five somata, respectively. Ehmer and Gronenberg (1997b) depict two sets of extrinsic antennal motoneurons of two single scapal muscles. The somata of these motoneurons lie in close vicinity to each other, suggesting that ant antennal motoneurons also have clustered somata. In Manduca, clustering of antennal motoneuron somata depends on the joint innervated: the motoneuron somata of extrinsic muscles are scattered over a wide area in the dorso-anterior soma band between the deuto- and protocerebrum, whereas the motoneurons of intrinsic muscles are concentrated in the dorso-lateral region of the soma band (Kloppenburg et al., 1997). In stick insects, the motoneuron somata are grouped into three areas, located in the lateral, the ventro-anterior and the ventro-posterior deutocerebrum (lateral, posterio- and anterio-dorsal, respectively, relative to the body-long axis in Du¨rr et al., 2001). The somata of functionally different motoneurons controlling different antennal joints and/or movement directions are not located in separate soma clusters (Du¨rr et al., 2001). Most of the cricket antennal motoneuron somata are located in the lateral deutocerebrum, in the vicinity of the root of the antennal nerve N1. Some

120


motoneuron somata lie in the ventral deutocerebrum (Honegger et al., 1990a). As in locusts (Bauer and Gewecke, 1991), no information on the formation of soma clusters is available. Some attention has been paid to the question of how the arrangement of neuronal somata in clusters may reflect a relationship of the adjacent neurons, e.g. a grouping according to a functional affinity. The antennal motoneuron somata clusters are unlikely to mirror a functional relationship, as the composition of stick insect soma clusters demonstrates. On the other hand, extrinsic and intrinsic muscle motoneuron somata of the honeybee separate in different clusters, though the somata of both motoneuron types lie close to each other in one area (Kloppenburg, 1995). Although the presence of distinct clusters does not appear to be a necessary feature of insect antennal motoneurons, the organisation of some antenno-motor systems is reminiscent of the findings of Siegler and Pousman (1990b) on the arrangement of metathoracic excitatory motoneurons into stereotyped clusters delimited by a glial sheath, which reflect a developmental rather than a functional relationship (Siegler and Pousman, 1990a). Thus, developmental studies may clarify, for example, whether the lateral band of motoneuron somata in crickets represents a ‘cluster’ in the sense that these neurons are the progeny of one neuroblast. 4.4.4

Dendritic arborisation pattern

All antennal motoneurons investigated so far have their dendritic arborisations in a part of the deutocerebrum outside the AL, the ‘AMMC’ (antennal mechanosensory and motor centre: Rospars, 1988; Homberg et al., 1989; or ‘DL’: Suzuki, 1975). Honegger et al. (1990a) extend this neuropil by adding the dorsal tritocerebrum which contains motoneuronal projections, but no antennal mechanosensory projections (‘AMC’). In honeybees and in Manduca, the dendrites of the motoneurons form a welldefined lateral, semi-circular border of the DL, where highest fibre densities occur. Fibre densities are reduced medially towards the esophageal foramen. The motoneuronal branchings end dorsal to the AL, and some fibres extend posteriorly close to the SOG (Kloppenburg, 1995; Kloppenburg et al., 1997). Taken together, the antennal motoneurons appear to constitute a common projection field with a similar projection pattern of individual motoneurons. Similarly, in ants the median portion of the DL receives no motoneuronal projections, though antennal motoneurons occupy two-thirds of the DL (Ehmer and Gronenberg, 1997b). The antennal motoneurons of crickets and stick insects resemble each other in that some individuals can be identified on the basis of their morphology in both taxa. The motoneurons of both insects will, therefore, be discussed in a greater detail in the following paragraphs, although a direct comparison of the arborisation patterns of potentially homologous motoneurons is hindered by the different brain morphologies of crickets and stick insects (Du¨rr et al., 2001).


121

The levator motoneuron branchings of Carausius are restricted to the lateroposterior (‘antero-lateral’ in Du¨rr et al., 2001) and to the medio-anterior to latero-anterior (‘postero-medial’ to ‘postero-lateral’ in Du¨rr et al., 2001) parts of the AMMC. A circular patch in the latero-anterior AMMC is devoid of levator motoneuron arborisations. Depressor motoneurons ramify in lateral parts of the AMMC (Du¨rr et al., 2001). Branchings of adductor motoneurons extend from the latero-anterior AMMC to its centre, whereas abductor motoneurons arborise mainly laterally in an area similar to that occupied by the depressor motoneurons. The abductor motoneurons also show a circular patch devoid of stained profiles. Thus, despite sharing some common projection areas, functionally distinct motoneurons differ in arborisation patterns specific for a group, e.g. the levator motoneurons with unique arborisations in the central AMMC. In crickets, the anatomy of the antennal motoneurons is detailed by Honegger et al. (1990a). The two levator motoneuron groups L1/2 and L3-5 have overlapping dendrites (branches 1) that extend medially into the AMMC. In the depressor motoneurons, D1 and D4 overlap with many fine branches in the lateral AMMC, and the secondary neurites of D2/3 overlap with those of D5. Branch 1 of D4 overlaps with branches 1 and, partly, branches 2 of D2/3. Depressor and levator motoneurons differ in the extent of their branchings along the dorso-ventral axis of the brain in that depressor motoneurons extend more ventrally than do levator motoneurons. The intrinsic muscle motoneurons (see Fig. 13) ramify slightly more anteriorly at the level of N1 with respect to the extrinsic muscle motoneurons which have their main arborisation at the level of N4. Within the adductor motoneurons, Ad1/2 and Ad3 have one branching area in common (branch 1 of Ad3). Interestingly, the antagonistic Ad1/2 and Ab1 resemble each other in their general morphology. Stick insects lack the adductor twin pair that is present in crickets (Du¨rr et al., 2001). The resulting overall picture suggests that functionally related motoneurons often share projection areas, and that the projection areas of functionally different motoneurons can be displaced along one axis in crickets. Individual motoneurons within a functional group, however, are nevertheless characterised by unique dendritic branches. Thus, crickets and stick insects on the one hand, and bees, Manduca and possibly ants on the other, seem to differ in the ‘individuality’ of their antennal motoneurons. Whether this is due to specific differences in the projection patterns of antennal mechanosensory afferents and interneurons is unknown. Further possible explanations may be provided by differences in the general structure of the dorsal deutocerebrum. Bees, for instance, have a very well-marked DL with a compact structure (Suzuki, 1975), that may concentrate dendritic ramifications to a larger extent than in crickets and, possibly, stick insects, which have a less well defined and more diffuse dorsal deutocerebral neuropil (Honegger et al., 1990a).

122


FIG. 13 Motoneurons innervating the adductor muscle M6 of the cricket (Gryllus sp.). Left: schematic layout of antennal muscles and nerves. 4a, b, c, levators; 5a, b, depressors; 6, adductor; 7, abductor; N1, antennal nerve; stippled line: scape-pedicel border. Right: the morphological twins Ad1/2 project to M6 via N2. Ad3 innervates M6 via N4B and N2A through the anastomosis between N2 and N4. c, circumoesophageal connective. Asterisks mark dendrites ramifying in the same brain area. After Honegger et al. (1995) with permission of the authors.

5

Central neurophysiology of antennal mechanoreception

The aim of this section is to provide an overview of the central neurophysiology of the processing of antennal mechanosensory information in insects. Section 5.1 will cover the local processing of antennal mechanosensory information; Section 5.2 concentrates on the physiology of descending brain interneurons, which convey antennal mechanosensory information to the ventral nerve cord. 5.1

PROCESSING OF ANTENNAL MECHANOSENSORY INFORMATION BY LOCAL BRAIN INTERNEURONS

It has to be expected from the organisation of the circuitry for generating and controlling leg movements in insects that local interneurons, intrinsic to the deutocerebrum, play a major role in the primary integration of sensory information from the body appendages. The emerging picture is that presynaptic inhibition of sensory neurons, interactions of excitation and inhibition in local networks, the involvement of non-spiking interneurons and reflex reversal shape the final motor output (e.g. Burrows, 1992; Ba¨ssler, 1993; Wolf and


123

Laurent, 1994; Wolf and Burrows, 1995; Ba¨ssler and Bu¨schges, 1998; Bu¨schges and El Manira, 1998). The knowledge of the antennal sensory-motor circuitry has not yet grown to an extent that allows us to draw a similarly detailed picture for the antennal control circuits. Local brain interneurons have been characterised in crickets, which process antennal proprioreceptive information and visual information from the compound eyes (Gebhardt, 1996). These local interneurons belong to at least seven morphological types, some of which are intrinsic to one deutocerebrum (Fig. 14B) and others of which extend into other brain neuropils, like the contralateral deutocerebrum (Fig. 14C), the tritocerebrum (Fig. 14D), the protocerebrum (Fig. 14E–G) or into the optical neuropils (Fig. 14H). The vast majority of them are unimodally driven by passive deflection of the SP-joint in an angular range between 01 and 1001 (01 representing the body long axis). The responses of the interneurons depend on the angular velocities of deflections, their directions and on the angular position of the SP-joint. Some interneurons exhibit a high specificity of their responses to certain joint angles, e.g. the interneuron shown in Fig. 15, which responds best with phasic bursts of spikes to abductions between 01 and 131, i.e. slightly lateral to the body long axis (Gebhardt, 1996). Another interneuron that directly connects visual neuropils to the AMMC, as found in bees (Maronde, 1991), responds best to forced antennal adductions in the range between 01 and 601 and to wide-field ‘‘light on’’-stimuli (Gebhardt, 1996). Local spiking and non-spiking interneurons, which are able to drive movements of the SP-joint upon current injection, reside in the dorsal-most area of the AMMC (Gebhardt, 1996). Thus, there is evidence that antennal proprioreceptive information is processed diversely in the AMMC: on the one hand, some local interneurons simply convey information from antennal sensory neurons, as estimated from the coding properties of scapal chordotonal organ sensory neurons (Allgaüer, 1989). On the other hand, some local and intersegmental interneurons are precisely tuned to a very narrow range of movement parameters, thus indicating very selective proprioreceptive input and/or extensive computations of proprioreceptive information in the deutocerebrum. Boeckh et al. (1984) summarise data on the convergence of olfactory, mechanosensory and thermal information from the antenna to deutocerebral, local and projection interneurons in the cockroach Periplaneta (O+M neurons). As displacement of the antennal base does not drive O+M neurons, the sources of the mechanosensory response are probably flagellar campaniform sensilla which are activated by relative movements between flagellar segments (Ernst and Boeckh, 1983). Zeiner and Tichy (1998) demonstrate that 40% of local and projection interneurons sampled in the central AL, are responsive to deflection of the flagellum and to olfactory stimuli (lemon oil). The interneurons either respond to sinusoidal flagellar deflections in a cycle-to-cycle manner or display habituating phasic–tonic responses to a whole stimulus train. Interactions

124


FIG. 14 Sketch of morphological types of antennal mechanosensory interneurons in the cricket brain. (A) Schematic drawing of one brain hemisphere: P, protocerebrum; L, lobula; M, medulla; AL, antennal lobe; vfa, ventral area of flagellar afferents; c-vfa: contralateral vfa; coc, circumoesophageal connective. Diagrammatic representation of the morphologies of different types of mechanosensory brain interneurons: local interneurons of the deutocerebrum (B); interneurons, which connect the deutocerebra of both hemispheres (C); tritocerebral interneurons with arborisations in the ipsilateral deutocerebrum (D); interneurons with somata and dendrites in the deutocerebrum and an axon ascending into the ipsilateral lateral protocerebrum (E); protocerebral interneurons with arborisations in one deutocerebrum (F); local interneurons of the protocerebrum (G); interneurons, which connect the deutocerebrum with the ipsilateral optical neuropils (H).

between mechanical and olfactory stimuli in the neuron population tested were observed frequently (61%). In the majority of neurons, the interaction consists of a response enhancement, in the others of a response depression. Homberg (1984) and Schildberger (1984) report on extrinsic mushroom body interneurons in the honeybee and in the cricket Acheta, respectively. These local brain interneurons respond to combinations of olfactory, visual, acoustical, gustatory and antennal mechanosensory stimuli in a complex way. In the honeybee, one antenno-cerebral tract interneuron, which represents a commonly encountered response type, is inhibited by different scents, with a trailing pronounced rebound excitation, but is strongly excited by antennal mechanical stimulation (touching the antenna with a glass probe). Other antenno-cerebral interneurons are not responsive to mechanical stimulation of the antenna. At the level of extrinsic a-lobe interneurons, visual responses join in, and antennal mechanosensory responses (touch of the ipsilateral antenna) are preserved at least in one out of six interneurons tested (Homberg, 1984).


125

FIG. 15 (A) Intrinsic deutocerebral interneuron responds to forced deflections (upper trace, stimulus monitor) of the SP-joint with EPSPs and spikes (lower trace, five superimposed sweeps of the intracellular recording). (B) The maximum response (normalised spike count per ramp) occurs selectively for abductions slightly lateral to 01 (¼ body long axis). Note, there is a small offset of 41 at the turning points of the movement. Insets illustrate the direction of the forced antennal movements. (Data from Gebhardt, 1996.)

The response patterns encountered in crickets are also complex. Schildberger (1984) notes that interneurons may change their response pattern during an experiment. Nevertheless, antennal mechanosensory responses are dominant in local projection neurons and mushroom body lobe neurons, with approximately 86% and 68%, respectively, responding to odourless air puffs to the antenna. Two bilateral, serotonin-immunoreactive interneurons, which ramify in various neuropils including the mushroom bodies, are assumed to provide feedback from the protocerebrum to the AL (Periplaneta: Salecker and Distler, 1990; Manduca: Sun et al., 1993). These neurons were recorded by Hill et al. (2002) in the moth Bombyx mori. Air puffs to the antennae carrying odours and air puffs without odour elicit identical excitatory responses in these neurons, which led Hill et al. (2002) to the conclusion that the response is of mechanosensory nature. In an attempt to sample the variety of modalities coded for in extrinsic output neurons of the mushroom bodies of Periplaneta americana, Okada et al. (1999) grouped recordings from neurons in the ventral pedunculus and b-lobe into sensory, movement-related and sensory-motor units. There is no clear difference in the distribution among these groups. The majority of mechanosensory extrinsic neurons in the pedunculus are multisegmental in that they received inputs from various parts of the body. For example, Okada et al. (1999) present recordings from a mechanosensory unit that is activated by tactile stimulation of the ipsilateral antenna and legs, but not of contralateral appendages. Multimodal sensory units were common. In general, the activity

126


of most mechanosensory units was related to tactile or air-current stimuli to the antennae. Movement-related units appeared to be specialised to various aspects of movement, such as onset or direction of movement, or the movement of the front legs. Often, these interneurons are not activated by proprioreceptors of the respective limbs, but are thought to be related to the generation or control of a movement command. Okada et al. (1999) also describe a sensory-motor unit recorded at the pedunculus-lobes junction of the mushroom body that is activated by tactile stimuli to both antennae and by movement of the ipsilateral antenna, thus integrating exteroreceptive and proprioreceptive information. These findings suggest that antennal mechanosensory information received in the deutocerebrum from antennal sensory neurons is not only unimodally processed in the AMMC, but also combined with other modalities (olfaction) in first-order interneurons (projection interneurons). The multimodal responses of mushroom body output interneurons indicate that antennal mechanosensory information is also preserved in higher brain centres. Its pronounced representation in interneurons of Acheta (Schildberger, 1984) emphasises its significance. Since local interneurons of the cricket AMMC never responded to olfactory stimuli (Gebhardt, 1996), and since tactile stimulation of the flagellum or air puffs to one antenna were used in many studies on insect olfaction, one can hypothesise that proprioreceptive information from the antennal base is processed separately from tactile information from the flagellum. The latter appears to converge with olfactory information at an early stage of processing. Staudacher (1998a,b) and Staudacher and Schildberger (1999) suggest that flagellar sensory neurons project to the neuropil VFA separate from the dorsal deutocerebral neuropils in the cricket, which might support this hypothesis. 5.2

DESCENDING ANTENNAL MECHANOSENSORY INTERNEURONS

The mechanosensory information sampled by antennal mechanosensory neurons has to be distributed to motor centres in the ventral nerve cord to adaptively adjust the behaviour of the animal. Approximately 200–230 descending brain interneurons have been found in the cricket (Staudacher, 1998) and in the cockroach (Okada et al., 2003). Some of these have been characterised with respect to their sensory physiology and their influence on locomotion (e.g. cockroach: Burdohan and Comer, 1990, 1996; cricket: Bentley, 1977; Boyan and Williams, 1981; Ho¨rner and Gras, 1985; Richard et al., 1985; Bo¨hm and Schildberger, 1992; Gras and Ho¨rner, 1992; Hedwig, 1996; Staudacher and Schildberger, 1998; Gebhardt and Honegger, 2001; Staudacher, 2001; locust and grasshopper: Rowell and O’Shea, 1976a,b; Bacon and Tyrer, 1978; Mo¨hl and Bacon, 1983; Rowell and Reichert, 1986; Hensler and Rowell, 1990; Hensler, 1992; Hedwig, 1994; Manduca: Milde and Strausfeld, 1990; Kanzaki et al., 1991). A recent review discusses the behavioural significance of


127

descending brain interneurons with a focus on the control of stridulation, walking and flight (Heinrich, 2002). In the studies cited above and in related work, much attention has been paid to the analysis of the visual sense and its behavioural impact. However, antennal mechanosensory inputs to a descending brain interneuron have been described in the context of flight control. In locusts, the tritocerebral commissure giant (TCG) interneuron receives inputs from five bilateral wind-sensitive hair fields, from the antenna and the compound eye (Bacon and Tyrer, 1978). Bacon (1980) describes how the relative weight of these inputs is shifted in three different insects. The TCG of locusts receives strongest inputs from the numerous head hairs, but weak antennal inputs. In Gryllus campestris, which has many fewer head hairs, but long antennae, no head hair inputs to the TCG are present and strongest inputs arise from the antenna (bending the flagellum), similar to the mantid Sphodromantis lineola, the TCG of which responds weakly to flagellar stimulation. Interestingly, the TCG of all three species displays a conserved directionality with best responses for downward deflections of the antenna. The output connections of the TCG and, hence, its flight control function remains unaffected by the shifting sensory inputs. An antennal mechanosensory pathway mediates escape responses in cockroaches (Comer et al., 1988; Burdohan and Comer, 1990; Keegan and Comer, 1993; see Section 6.1.3.3). Extracellular recordings from one neck connective reveal a correlation between the increased spiking activity in the connective contralateral to the stimulated antenna and the direction of the escape turn. Using intracellular recordings and stainings, two descending interneurons have been identified, which contribute to this increased spiking activity (Burdohan and Comer, 1990, 1996; Ye and Comer, 1996). One of them, DMIa-1, is a contralaterally descending brain interneuron, which extends at least as far as the first abdominal ganglion. DMIa-1 responds to mechanical stimulation of the flagellum ipsilateral to the cell body. The response consists of a phasic burst of action potential with a latency of 6–8 ms. The action potential travel down the axon of DMIa-1 with a conduction velocity of 4.6 ms1. An analogous descending interneuron has recently been described in the cricket (Mathenia et al., 2003). A second large interneuron, DMIb-1, has its cell body in the SOG and also extends at least to the first abdominal ganglion. Mechanical stimuli to either antenna give rise to brief phasic responses with latencies of 8 ms (antenna ipsilateral to the soma) and 10 ms (ipsilateral antenna) in DMIb-1. Its conduction velocity is 4.7 ms1. The receptors feeding into these descending interneurons appear to be located in the antennal base, as immobilisation of the antennal base abolishes spiking, whereas ablation of the flagellum has no effects (Comer et al., 2003). The sensitivity of both neurons to flagellar displacement is high: deflection by 0.05 and 0.5 mm reliably triggers spikes in DMIa-1 and DMIb-1, respectively. Their sensitivity for wind stimuli, however, is rather weak compared to the sensitivity of the cercal system. DMIa-1 displays a relatively weak directionality which is slightly emphasised for lateral

128


stimuli, whereas DMIb-1 has a much more pronounced directionality for stimuli from the front for both antennae. Thus, the two DMI neurons constitute a rostral escape circuit which detects tactile stimuli and, therefore, complements the cercal system (see Section 6.1.3.3). To address the question, which of the parameters of the activity of the descending interneurons correlates with the turning direction and the turning angle, Ye and Comer (1996) performed extracellular neck connective recordings in tethered cockroaches walking in an open-loop paradigm. They found that the turning direction can be predicted from the latency difference for the first spike, and less reliably, from the spike count difference between both connectives. Cockroaches turn away from the touched antenna, towards the side of the first DMI spike, and towards the side of higher neuronal activity. The amplitude of the turn, however, co-varies with the spike count difference alone. Accordingly, when electrically stimulating one neck connective, turns are directed towards the side stimulated and the turning angle could be increased by either increasing the stimulus voltage or by increasing the pulse frequency in trains of stimulus pulses. Hence, a thoracic comparator seems to analyse the bilateral pattern of neuronal activity, evaluating latency and spike count differences differentially for the control of turning direction and turning amplitude. Several studies deal with descending brain interneurons in crickets, but their focus is on sensory modalities other than antennal mechanoreception: they concern auditory (Boyan and Williams, 1981), visual (Richard et al., 1985), wind-evoked escape (Ho¨rner, 1992) or the effect of various sensory stimuli on walking control (Ho¨rner and Gras, 1985). Ho¨rner and Gras (1985) have recorded from local and descending brain interneurons of Acheta, while presenting a variety of sensory stimuli to the crickets. The interneurons recorded belong to different morphological types with branchings in various brain neuropils except the mushroom bodies. Responses to mechanical stimulation of the antenna (wind streams with 0.8 ms1) are found in 25% of the interneurons sampled. Other effective modalities which elicit responses in single interneurons, are mechanical stimulation of the cerci (41%), visual stimuli (19%) and acoustical stimuli (15%). Antennal responses appear to be widely present in a descending multimodal interneuron set, indicating the behavioural relevance of antennal mechanosensory information. Bo¨hm and Schildberger (1992) report on different descending brain interneurons and their responses to visual, acoustical, antennal and cercal (air puffs) stimuli and their involvement in the control of walking. They depict two descending interneurons, which apparently have dendrites in or, at least close to, the deutocerebrum. One of them, an ipsilaterally descending interneuron has a dendrite reaching into the lateral protocerebrum and another projecting towards the deutocerebrum. Despite this pattern of branching, this interneuron responds, only to omit acoustic stimuli. Another interneuron with a dendrite close to the deutocerebrum and a contralaterally descending axon also


129

responds unimodally to directional visual stimuli. Whether the absence of antennal responses in both interneurons is due to their dendrites located outside the mechanosensory portion of the deutocerebrum, remains unknown. Two ipsilaterally and six contralaterally descending brain interneurons with a dendrite in the mechanosensory deutocerebrum have been characterised by Gebhardt (1996), Staudacher (1998a,b) and Staudacher and Schildberger (1999), five of which have been recorded intracellularly by Gebhardt and Honegger (2001). The ipsilaterally descending interneurons project at least to the first abdominal ganglion. All of these interneurons respond phasically to forced deflections of the SP-joint; one (DBNi1–2, nomenclature according to Staudacher, 1998b) is also driven by visual inputs from the ipsilateral compound eye. The antennal mechanosensory responses of all these interneurons are directionally selective, e.g. with DBNi1–2 responding best to deflections between 301 and 601 lateral (01 being the body long axis), whereas other interneurons respond best close to lateral or medial extreme positions of deflections. Thus, it appears that these interneurons represent the total angular range of SP-joint movements in a fractionated manner. Owing to only fast deflection, but never static SP-joint posture, elicit spikes, these interneurons might serve the detection and localisation of external objects ahead of the cricket. For such a task, a mechanism must exist to discriminate between excitation resulting from external stimuli or from self-induced stimulation, i.e. active antennal movements. Such mechanisms were proposed by Holst and Mittelstaedt (1950; the ‘re-afference principle’) and by Sperry (1950; the ‘corollary discharge’). Poulet and Hedwig (2002, 2003) have recently shown that a corollary discharge mechanism modulates the auditory O-interneuron in the prothoracic ganglion of crickets. Similarly, the synaptic inputs to DBNi1-2 are suppressed in a gradual manner depending on the strength of an antennal motor command. This suppression is of a central nervous origin, since it persists in a preparation in which one antenna was rendered motionless and, thus, peripheral feedback loops were not active. It can be hypothesised that the antennal mechanosensory information conveyed to the posterior ventral cord is filtered to remove the self-induced response component. The resulting, so far unknown, behaviour modulated by the descending interneuron would then be influenced by external stimuli only.

6

Behavioural physiology of the antennal tactile sense

Having reviewed the neurophysiology of antennal mechanoreceptors and corresponding neuropils in the insect brain, the following section gives an overview of the various behavioural functions of antennal mechanoreception. It is hard to draw a line between mechanical stimuli in general and tactile stimuli in particular. One reason for this is that, in spite of the neuroanatomical description of different neural tracts (see Section 4.1), it is still difficult to

130


determine the extent to which proprioreceptive and exteroreceptive information is being processed separately in the deutocerebrum. Another reason is that tactile contact information is typically accompanied by information about active or passive movement of the antenna. As a result, mechanoreceptor signals other than those associated with the actual object contact, will qualify the information in terms of the behavioural context in which antennal contact has been made. Thus, from a behavioural perspective, proprioreceptive and exteroreceptive cues should not be separated. The following sections will deal with a wide range of behavioural situations and consider many different submodalities of mechanoreception. Some of the examples, e.g. the role of the antennae in flight or hearing, will appear to be not of primary concern to tactile sensing, either because the behavioural situation excludes tactile contacts (as in flight), or because no contacts are necessary (as in hearing). However, because antennae are involved in several aspects of steering in locomotion, some of which involve tactile contacts and some of which do not, a comparison of these behavioural contexts can shed light on general control mechanisms. Similarly, hearing in honeybees involves mechanoreceptors that are most likely stimulated in behavioural contexts where tactile information is relevant. Moreover, honeybee hearing is known to be of particular importance in dance communication, a behavioural situation in which the animal is also likely to make tactile contact. Finally, tracking and orienting movements of antennae will be treated in cases that are likely to increase the likelihood of tactile contact, although they must not necessarily lead to tactile contact. In the following sections, behavioural functions of the antennae will be grouped according to whether the sensory act is typically characterised by passive displacement (Section 6.1), or rather by active antennal movements (Section 6.2). 6.1

PASSIVE SENSING

Behavioural contexts in which antennal mechanoreception is characterised by passive sensing concern the role of the antennae in the sense of gravity (Section 6.1.1), reflex actions that remain restricted to the antennal joints (Section 6.1.2), and primarily passive sensory functions in locomotion, as in steering and escape reactions (Section 6.1.3). 6.1.1

Graviception

Unlike decapod crustaceans, which possess a statocyst in the basal segment of the antennule (e.g. Hennig, 1986), insects have no dedicated gravity sensor. Rather they infer their body orientation relative to gravity from proprioreceptors of the limbs or of the body segments. The contribution of antennal mechanoreception in sensing the gravity vector has been demonstrated for


131

behavioural contexts of gaze stabilisation in crickets, and negative geotaxis in stick insects. The cricket Gryllus bimaculatus uses sensory cues from antennae and cerci to sense the roll angle around the body long axis relative to an upright posture. This can be investigated in stationarily walking crickets that are held by a support, and rotated together with the substrate on which they walk. If the substrate (e.g. a ball) is held by a bearing, only the appendages can signal a deviation from gravity. In this situation, passive roll by 7901 causes the walking cricket to counter-roll its head by some 7101 in males, and by 7151 in females (Horn and Bischof, 1983). Walking is a necessary condition for this response, indicating context-dependent gating of the neural pathway. Antennectomy at the level of the mid-scape reduces the compensatory head roll to 751, with no sex difference remaining. Additional removal of the cerci abolishes the response completely. Ablation of a single antenna or of a single cercus is sufficient to reduce compensatory head roll, and various combinations of cut antennae and cut cerci reveal that the impact of the cerci is larger than that of the antennae. Also, the impact of any combination of removed appendages is smaller than the linear sum of the individual effects (Horn and Fo¨ller, 1985). If crickets are rotated while the substrate is not held by a bearing, there is additional stimulation of leg proprioreceptors. In this situation, the head counter-roll angle increases 2.5- to 3-fold compared to the ball-on-bearing situation. As this response is stronger than that elicited by the leg proprioreceptors alone, antennal and cercal graviceptive information appear to facilitate the effect of leg proprioreceptors (Horn and Fo¨ller, 1985). The source of antennal proprioreceptive information that elicits the head counter-roll response is unknown. A first potential source could be that of hair plates at the SP-joint signalling passive horizontal deflection of the pedicel. In this case, antennectomy at the scape would preclude the ability to detect the deviation from the horizontal plane. Indeed, unilateral antennectomy has little effect on crickets that walk on a horizontal surface (Horn and Fo¨ller, 1985), a result that has been used to show that sensory cues from the antennae and cerci cause no tonic head counter-roll. However, if the SP-joint was involved in graviception, a putative tonic effect would be zero when walking on a horizontal plane. A second potential source of information could be proprioreceptors of the HS-joint that detect passive vertical deflection of the scape, i.e. a deviation from the vertical plane. In this case, antennectomy causes a reduction of antennal mass and of the corresponding static downward force acting on the HS-joint. The stick insect Carausius morosus shows negative geotaxis, an orientation behaviour that mainly depends on leg proprioreceptors (Ba¨ssler, 1971) but involves supplementary information of the antennae. When walking on a vertical plane in the dark, stick insects normally tend to walk upward. If the leg proprioreceptors are intact, antennectomy has a negligible effect on body

132


orientation. However, if the animal’s own weight is compensated for by an upward-pulling force, intact animals perform equally well as before, but animals without antennae no longer show geotactic behaviour (Wendler, 1965). Using a similar approach, but recording the preferred body orientation at rest, Ba¨ssler (1971) demonstrated that the main sensors of gravity in the stick insect are the hair plates of the thorax-coxa joints of the legs. Again, antennectomy of otherwise intact animals does not affect the magnitude of the orientation angle relative to gravity, but increases the number of angles with reversed sign, i.e. animals facing downwards. In animals with shaved coxal hair plates, the antennae are sufficient to signal the correct sign, so animals preferably face upward. Applying an upward-pulling force to reverse the gravitational force vector sensed by leg proprioreceptors without affecting the orientation sensed by the antennae, the number of positive geotactic responses increases in intact animals. It is only after antennectomy that negative geotaxis becomes apparent again, i.e. with animals now facing downward. Ba¨ssler (1971) concludes that the hair plates of the thorax-coxa joint determine the magnitude of the orientation angle relative to gravity, while the antennae only indicate the sign of gravity. Furthermore, the impact of the antennae on the sign must be larger than that of the leg proprioreceptors. As in the head counter-roll response of crickets, the source of antennal sensory information in stick insect geotaxis is still unknown. What is interesting about the mentioned studies is that both crickets and stick insects typically move their antennae during walking (see Section 6.2.1). To what extent active antennal movement interferes with or is important for the antennal role in graviception remains to be investigated. 6.1.2

Avoidance, assistance and resistance reflexes

Imposing passive movement on antennal joints, touching the flagellum at rest, or resisting antennal movements with a stop causes reflex activity in antennal muscles. Depending on the effect of the motor pattern in relation to the stimulus, these reflexes act to avoid contact, assist or resist imposed movement. Reflexes are integral parts of all behaviours that involve antennal movements, and it is clear that they are not due to unmodifiable, hardwired feedback circuits, but rather involve both central and sensory modulation. In general, antennal reflexes bear many similarities to leg reflexes, including state-dependent reflex reversal (e.g. see Ba¨ssler and Bu¨schges, 1998). 6.1.2.1 Antennal reflexes in insects. Locusts (Schistocerca gregaria) respond to passive or active contact between their flagellum and an object either by moving the antenna away from the object, or by a sequence of abduction and adduction movements that can last for several seconds (Saager and Gewecke, 1989). The latter sequence has been called a reflex chain, but because the trigger for the transition from the avoidance reflex to the subsequent resistance reflex


133

is unknown, it may well be an active exploratory movement sequence that requires more central control by the brain than would be expected in case of a reflex chain. In fact similar spontaneous oscillations, which the authors observed in ‘‘very active’’ animals, suggest that it may be a centrally patterned movement that does not rely on sensory information. Finally, because the movements reported by Saager and Gewecke (1989) were restrained to the SP-joint, possible interactions with the HS-joint that could be involved in exploratory movements were prevented. The avoidance reflex can be elicited by flagellar deflections as small as 0.051. In a standing animal, the movement of one antenna in Schistocerca is not neurally coupled to that of the other. This is not so during flight, when the antennae assume a symmetrical posture, pointing medio-laterally. During flight, the response to passive deflection of an antenna turns into a resistance reflex of that antenna. Moreover, ‘‘mediad deflection of one antenna with a needle causes lateral movement of the other’’ (Saager and Gewecke, 1989, p. 525). The reflex reversal can be elicited by a frontal air-current that triggers abduction of the flagellum, much like the tactile avoidance response, which turns into an active adduction as soon as the flight muscles are activated. During flight, co-activation of adductor and abductor muscles maintains a stable posture of the antenna (within 11 deviation). Sinusoidal deflection of the flagellum in this behavioural context always elicits a resistance reflex that acts to stabilise the posture by increasing the force as expected from a negative feedback loop. After switching off the air current, the animal stops flying, whereupon the antennal reflex reverses back to an avoidance reflex. Saager and Gewecke (1989) further demonstrate that fixing the SP-joint, thus eliminating the sensory input from the hair plates at this joint, abolishes the tonic component of the scape muscle electromyogram (EMG). In this case, only phasic avoidance reflex activity remains. On the other hand, by immobilisation of the PF-junction, i.e. reducing the stimulation of pedicellar campaniform sensilla, the tonic EMG component is increased. This can be attributed to the resistance reflex. Cockroaches of the genera Periplaneta and Blatella show a tactile antennal avoidance reflex if the antenna is touched with a dry glass needle at more than 1 cm away from the head. At more proximal contact points, the insects move the maxillary palps towards the source of stimulation (Frings and Frings, 1949; cf. Cornwell, 1968, p. 114). Since Comer et al. (1994) use very similar stimuli to elicit escape behaviour in cockroaches, it has to be noted that such manual stimulus presentations can vary considerably, even if the same tool is used for stimulation. On the basis of observations on the cockroach Periplaneta americana, Okada et al. (2002) suggest that the HS-joint hair plates subserve a direction-selective negative feedback loop that pulls the HS-joint towards a median position around 401. This interpretation is supported by two findings following ablation of the HS-joint hair plates: first, the effective workspace of the flagellum widens (-40 to 1301 vs. -50 to 1501) and abduction–adduction

134


cycles lengthen. Second, far lateral contacts during adduction of the antennae are more often followed by abduction than before. Object contacts during abduction are not affected, however. According to the hypothesis of Okada et al. (2002), the tendency to adduct increases with increasing abduction angle, and an avoidance reflex in response to object contacts with far lateral objects is overridden by a resistance reflex. 6.1.2.2 Reflex types in crustaceans. Studies on rock lobsters and crayfish have contributed much insight into the sensory-motor link between antennal proprioreception and reflex movements. In the rock lobster Palinurus vulgaris, mechanical stimulation of the MCF-chordotonal organ can elicit both a resistance- and an assistance reflex (Clarac and Vedel, 1975). Induced reflex activity is restricted to the four dedicated extensor/flexor motoneurons, to the accessory extensor motoneuron and to the common inhibitor. The common extensor and common flexor motoneuron as well as the phasic CF-extensor motoneuron do not contribute (Vedel, 1980). During a resistance reflex, the tonic motoneurons that would assist the imposed movement stop firing. Extensor reflex activity to imposed flexion is velocity-dependent, whereas flexor reflex activity to imposed extension is not. Passive movement of the CF-joint causes reflex activity also in the muscles of the neighbouring MC-joint, but this inter-joint effect is weaker than the intra-joint effect on the CF-musculature. Assistance reflex activity occurs in extensor motoneurons when the animal becomes spontaneously active. In this case, extension of the flagellum elicits extensor activity and a corresponding decrease in the antagonistic flexor activity. The result is a positive feedback action upon extension. The assistance reflex habituates and its gain is strongly velocity-dependent. Imposed extension also causes velocity-dependent reflex activity in the common inhibitor. Furthermore, assistance reflex activity causes history dependence in that a subsequent resistance reflex to imposed flexion is no longer velocity-dependent. Vedel (1980) suggested that the behavioural role of the extension assistance reflex is to support voluntary extension of the flagellum during the threat and defence behaviour of Palinurus. In the lobster Homarus americanus, passive flexion at the MC- or CF-joint causes a resistance reflex involving the two dedicated extensor motoneurons and one of the common extensor motoneurons (Sigvardt, 1977). Conversely, extension of these joints causes activation of the dedicated flexor motoneurons. Accordingly, passive flexion and extension of the CF-joint elicits a resistance reflex in the crayfish Cherax destructor (Sandeman, 1989). As in Palinurus, these reflexes are mediated by the bi-articular MCF chordotonal organ. Tactile hairs on the flagellum of the crayfish C. destructor subserve an avoidance reflex. Brushing the medial surface of the flagellum elicits inhibitory potentials in the tonic extensor motoneuron, whereas brushing the lateral surface causes an increase of the spike rate in the same neuron (Sandeman, 1989). As extension of the flagellum is equivalent to movement towards the medial


135

direction, the resulting movement is directed away from the stimulus. Bending of the flagellum of C. destructor has a pronounced effect on tonic motoneurons whose activity leads to the relief of the imposed tension. With regard to movement direction of the flagellum, this is an assistance reflex, but the strain sensors are part of a negative feedback loop. 6.1.2.3 Reflex modulation in crustaceans. All studies on crustacean antennal reflexes have emphasised the presence of modulatory effects. In P. vulgaris, the motoneuron activity associated with the resistance reflex is dependent on ambient temperature and on ascending influences from the uropods and pereiopods (Clarac and Vedel, 1975). In crayfish (Euastacus armatus), reflex-like extension of the antennae occurs in response to tactile stimulation of the leg bases; to passive rotation of the body; or as part of the escape behaviour (Sandeman and Wilkens, 1983). In Euastacus, flexion of the antenna occurs in response to tactile stimulation of the uropods or telson. Also, motoneuron activity associated with the resistance reflex can be modulated by electrical stimulation of almost all ipsilateral brain nerves and of the contralateral optic nerve and connective, indicating the complexity of the neural circuits involved. Tactile stimulation of the branchiostegite enhances the resistance reflex by facilitation of motoneuron activity. The reverse effect occurs during voluntary leg movements. These are accompanied by a reduction of tonic motoneuron activity that is typically associated with the resistance reflex, and by a reduction of co-activation of flexor and extensor motoneurons (Sandeman and Wilkens, 1983). 6.1.3

Steering of locomotion

Several insect orders employ their antennae in orientation and steering behaviours, ranging from wind-directed orientation (anemotaxis or anemomenotaxis) in airborne (Section 6.1.3.1) and terrestrial locomotion, to tactile course control during running (Section 6.1.3.2). Furthermore, tactile contacts and, possibly, air-current-induced deflections of the flagellum can trigger fast turning movements in escape or defence behaviour (Section 6.1.3.3). 6.1.3.1 Course control during flight. The use of antennae in control of flight speed is not restricted to insect orders with highly specialised antennae, such as flies, but is also well documented in locusts and bees. For example, tethered flying honeybees decrease their wing beat amplitude in response to a frontal air current. Antennectomy, or preventing the air current from reaching the antennae abolishes this response (Heran, 1957). At an air speed of 7 m s1, the antennae cause a decrease of wing beat amplitude by 351, compared to 101 in the antennectomised animal (Heran, 1959). In free flying bees, antennectomy prevents light-dependent reduction of flight speed that can be observed in bees flying out of a dark room towards a bright window.

136


When an air current blows towards the head of a tethered flying bee, both antennae move inward, requiring an adduction and an outward roll in the HSjoint, but also a flexion of the SP-joint (Heran, 1959). Both of these components increase with increasing speed of the air current. Various immobilisation and ablation experiments show that deflection of the flagellum is the adequate stimulus, and that the movements of each antenna are elicited by receptors located in the ipsilateral pedicel. Moreover, information of one antenna does not cross the midline to affect the posture of the contralateral antenna, in spite of the presence of a neural connection via the dorsal SOG (Maronde, 1991). Heran (1959) further suggested that it is the flagellar vibration at the PFjunction that is decisive, rather than its static deflection. Finally, because the air flow experienced during a yaw turn of the flying bee causes an outward deflection of the outer antenna, which in turn leads to a decrease in the wing beat amplitude, Heran (1959) proposed that antennal proprioreceptors close a negative feedback control loop that assists course stabilisation during flight. In flying locusts (Locusta migratoria), antennae serve as air-flow sensors that complement a set of wind-sensitive hairs on the head (Gewecke, 1970). At normal flight speeds, active adjustments of the antennal posture in response to air current keep deflections at the PF-junction below 11 (Gewecke and Heinzel, 1980). Mechanoreceptors monitoring this deflection drive postural adjustment of the antenna itself, but also a change in the wing beat pattern. The crucial mechanoreceptors involved in this behaviour are a ring of campaniform sensilla around the distal pedicel, which codes deflection of the flagellum at the PF-junction up to 1.31 (Heinzel and Gewecke, 1979) (see Sections 2.2.2 and 3.2.2). Reminiscent of the locust flight behaviour, flying crickets (G. bimaculatus) also hold their antennae in a forward-directed posture, although the distal flagellum is bent posteriorly, because it is not stiff enough to maintain a straight posture against an air current. To what extent cricket antennae are involved in steering is unknown, but lesion experiments show that ascending pathways from the suboesophageal and thoracic ganglia are necessary to maintain this posture during flight (Horseman et al., 1997). In flying locusts, antennal sensory cues are also involved in postural adjustments of the hind legs. Tethered flying Schistocerca gregaria attempt to fly an upwind curve, by abducting the leg on the side facing the air current, thus increasing the drag on the inner side of the curve (Arbas, 1986). This ‘upwind response’ is elicited by wind-sensitive hairs on the head. Following occlusion of these hairs, locusts show a ‘downwind response’, in which they abduct the hind leg on the downwind side. Immobilisation of the antennal joints or antennectomy abolishes leg abduction in response to changes in air-current direction. The behavioural significance of the antagonistic steering tendencies triggered by antennae on the one hand, and wind-sensitive hair fields on the other remains unknown. Because the directionality of the response may be reversed in some animals, a fact that Arbas (1986) attributed to ‘‘changes in mood’’, the


137

neural pathway from antennae to hind legs appears to be subject to considerable plasticity. As in locusts, cricket behaviour reveals a relationship between antennal movements and steering movements of the legs. Flying crickets (Teleogryllus oceanicus) show an acoustic startle response to ultrasound stimuli that involves abduction and levation of the contralateral hind leg and bilateral abduction of the antennae (Miles et al., 1992). Yet, whereas in locusts the antennal displacement induces the leg movement, it does not do so in Teleogryllus because activation of the scapal muscles occurs with similar or even longer latency than activation of leg muscles. Hence, the functional significance of antennal movements in flying crickets is less clear than in flying locusts. Nevertheless, the same neural pathway may be recruited as in Arbas’ downwind response. 6.1.3.2 Course control during walking and running. Much as in flying, many insects perform wind-related steering actions during terrestrial locomotion. Linsenmair (1973) lists 38 insect species that show anemomenotactic behaviour, most of which are Coleoptera and Heteroptera. In walking beetles, amputation of a single antenna changes the preferred orientation angle with respect to wind direction (Linsenmair, 1970). A comparative study on the cockroach species Blatella germanica, Periplaneta americana and Blaberus craniifer shows that spontaneous locomotor behaviour varies considerably in terms of looping and other aspects of path curvature. Yet, the walking paths of all species get increasingly straighter with increasing wind speed (Bell and Kramer, 1979). Nevertheless, the preferred angle with respect to wind direction is not the same for all species. Bilateral antennectomy abolishes anemotactic orientation in Blaberus, but unilateral antennectomy or removal of the cerci has no detrimental effect, indicating that the information of at least one intact antenna is sufficient and necessary to elicit this orientation behaviour (Fig. 16). Furthermore, immobilisation of the SP-joint, or of the antennal base beyond the PF-junction revealed that mechanoreceptors of the pedicel (either at the SP-joint or at the PF-junction) are necessary to detect wind direction. It is worth noting, however, that the measures of path straightness and preferred orientation angle, as applied by Bell and Kramer (1979), are insensitive to more subtle changes in locomotor activity. For instance, the trajectories of unilaterally antennectomised B. craniifer turned out to be as straight as in intact animals, but the size of the turns and the duration of straight periods in-between was rather different (Fig. 16B, C). Thus, antennal mechanoreceptive information is likely to be part of a course control circuit that involves both antennae. Further evidence for such a control circuit stems from experiments on running cockroaches of the species P. americana. When running on the ground adjacent to vertical walls, cockroaches often keep one antenna in touch with the wall and use the contact information from their antennae for fast course corrections to regulate the distance to the wall (Camhi and Johnson, 1999).

138


FIG. 16 Antennae are involved in wind orientation of the cockroach Blaberus craniifer. (A) Representative walking track of a cockroach (1000 s) in the absence of visual and wind stimuli. The track is curved and interspersed by repetitive loops. Scale bar is 1 m. (Bi–Biii) Representative walking tracks of intact cockroaches (200 s) at different wind speeds (Bi: 0.03 m s1; Bii: 0.12 m s1; Biii: 0.24 m s1). Upwind direction is indicated by the arrow. Scale bar is 1 m. The preferred upwind direction of locomotion is indicated by a dot at the end of the path. Walking tracks become straightened in the presence of wind. Conspicuous kinks occur regularly. (Ci–Ciii) Same details as in Bi–Biii, but after unilateral antennectomy. Kinks are less pronounced than in B, and the path is more irregular. (Di–Diii) Same details as in Bi–Biii, but after removal of both antennae. The looping behaviour is similar to the no-wind condition shown in A. Adapted from Bell and Kramer (1979) with permission from Elsevier Publishing.


139

This effect is dependent on the speed of locomotion, and decreases as the average distance relative to the wall decreases with increasing speed (Fig. 17). When a cockroach runs along a pleated wall, orientation of the body axis can change by 201 (peak to peak) or more at frequencies around 10 cycles s1 (Fig. 17D–G), which is close to one turn per step cycle. Camhi and Johnson (1999) also found that bend position and contact point along the flagellum may be a good indicator for the distance to the wall, partly because the antennal angle near the base is held rather stable by the cockroach. Accordingly, immobilisation of the antennal base and of the neck does not cause significant changes in contact-mediated turns, whereas loss of flagellar information abolishes turning altogether. Although it is still unknown how a cockroach manages to make the appropriate course corrections on a step-tostep basis, the data of Camhi and Johnson (1999) indicate that the tripod leg coordination pattern is not disturbed during rapid turning. Apparently, antennal contact information causes changes in leg kinematics and/or dynamics, rather than in timing of entire step cycles. In a similar wall-following paradigm, Cowan et al. (2004) measured the time course of steering actions of running cockroaches in response to a step change in sensed heading angle with respect to the wall. Cockroaches alter their heading angle faster in response to a large step in sensed angle than in response to a small step. Because a change in heading angle is always accompanied by a change in lateral distance to the wall, Cowan et al. (2004) suggest that cockroaches control a combination of heading angle and lateral distance, a variable they termed ‘tactile flow’ (see also Section 7.1). Apart from the stunningly fast steering abilities of a running cockroach, the question about the biological significance of a wall-following behaviour in insects remains. Possibly, vertical surfaces in a typical cockroach environment tend to lead to shelters or holes. Other than the mentioned fast course corrections in response to an outward bend of the flagellum, walking cockroaches (P. americana) turn towards an object that they touch with their antenna. If the object is presented to a tethered walking animal, it may continue to turn for several seconds. In this case, angular velocity of turning increases with abduction angle of the antenna (Okada and Toh, 2000). Because the turning tendency is strongly reduced after ablation of the hair plates of the HS-joint, these proprioreceptors appear to serve as angular sensors in tactile orientation behaviour. A further role of the antennae in steering, the ability to follow conspecifics, becomes apparent from descriptions of queuing aggregations in spiny lobsters, but also in the so-called tandem runs of ants, a form of recruiting behaviour. For example, Ho¨lldobler (1985) shows a picture of a tandem run by an Australian Hypoponera species, where the antennae of the follower touch the sides of the leader’s abdomen. Tandem runs are interpreted as a primitive form of communication among nestmates (see Section 6.2.4.2). The antennal sensory information necessary for establishing a tandem run have been studied in some detail for Bothroponera tesserinoda, where it involves both mechanical and

140


FIG. 17 Tactile wall-following in the cockroach: (A) Cockroaches of the species P. americana tend to maintain antennal contact when running along a wall. Observation of body orientation and distance to a pleated wall reveals fast steering responses. (B) Superimposed image sequence of a flagellum as a cockroach runs past a kink of the wall, beginning with first contact with the outward wall projection until 12 ms after it lost contact. Exposures are separated by 4 ms. (C) The distance of the head to the wall depends on the running speed (compare ‘Intact’) and on the length of the flagellum. Unilateral shortening or complete removal of an antenna results in significant differences between runs, in which the shortened antenna faces the wall, and control runs, in which the intact antenna faces the wall (compare ‘Half Ant’ with ‘Control Half’ and ‘No Ant’ with ‘Control None’. All pairs significant except Half Ant/Control for speeds 411 steps/s). Distance is significantly correlated with flagellum length. (D–G) Example run along a pleated wall. (D) Arrows show body axis every 40 ms. (E) If body orientation had been unaltered after the two indicated instances, the cockroach would have touched the wall with a leg (long line segment with *) and its body (short line segment with *). (F) Arrows show body axis 80 ms after the corresponding arrow in E. Owing to fast course corrections, the cockroach only touches the wall once, and only with a leg (*). (G) The time course of body axis angle reveals fast turning after both antennal contacts with the outward projecting wall (white arrows). The black arrow marks the time of leg contact. Adapted from Camhi and Johnson (1999), with permission from the Company of Biologists Ltd.


141

contact-chemosensory cues (Maschwitz et al., 1974), and in Camponotus sericeus, where chemosensory signals are obligatory (Ho¨lldobler et al., 1974). Whereas it seems to be clear that the actual steering, i.e. the close following of the leader’s course, involves tactile cues sensed by the follower’s antennae, the required sensory structures are still unknown. In a peculiar form of migration behaviour involving up to thousands of individuals, the spiny lobster Panulirus argus forms long queues that rely on tactile contact. As the queue moves, each animal keeps physical contact with the animal in front (Herrnkind, 1969). Contact occurs most often with the inner branch of the antennule (which is homologous to the insect antenna) and with the first pair of pereiopods, but also with the antennae. In blinded animals, or in the dark, queue formation is most often initiated by antennal contact. Experiments by Bill and Herrnkind (1976) show that spiny lobsters can greatly reduce the drag opposing their locomotion effort by joining queues (see Section 2.2.2.1).

6.1.3.3 Tactile elicited emergency behaviour. Escape behaviour involves steering, provided that the direction of the escape is correlated to the direction of the triggering stimulus. A prominent example of a steered escape is the reaction of a cockroach to a wind puff, where the air current is sensed by the cerci, and a short latency reaction is mediated by the ascending giant interneuron (GI) pathway. The escape reaction consists of a turn away from the wind stimulus (contraversive turn) and an escape run. First experimental evidence that a descending neural pathway, mediating information from the antennae, i.e. one that does not involve GI, is part of the escape system was presented by Comer et al. (1988) and Stierle et al. (1994). Ablation of the GI in the cockroach P. americana causes a decrease in responsiveness to wind stimuli (97–30%) and an increase in response latency (53–104 ms by Comer et al., 1988; 34–74 ms by Stierle et al., 1994), but does not affect the direction of escape. Bilateral antennectomy further reduces the responsiveness to wind stimuli to 4%, without impairing the motor ability to turn or to escape from tactile stimuli to the body (Comer et al., 1988). However, unilateral sectioning of the neck connective, i.e. introducing an asymmetry in the descending pathway of otherwise intact animals, alters the direction of wind-evoked escape responses without changing responsiveness or latency of escape (Keegan and Comer, 1993). Tactile stimulation of the antenna triggers escape behaviour more reliably when touching the lesioned side, though at longer latency (Comer et al., 1994). Wind-evoked escape of lesioned animals are shorter in distance and duration. Also, lesioned animals show a bias to turn away from the lesioned side, during spontaneous walking as well as during wind- or touch-evoked escape. Nevertheless, lesions do not preclude turning in either direction, indicating that they affect the steering command rather than the neural circuitry underlying the execution of the turn.

142


Indeed, nerve recordings suggest that the descending information modulates the GI-mediated escape responses: whereas, the lesion does not affect the typical asymmetry of wind-evoked activity in the GI (more activity in one connective than in the other), it does so in flexor motoneurons of the middle leg (Keegan and Comer, 1993) which are known to be important for the initiation of an escape turn (Nye and Ritzmann, 1992). Chronic recordings from descending neurons in the neck connective of stationarily running cockroaches show stronger excitation in the connective contralateral to the touched antenna (Ye and Comer, 1996), i.e. the direction of the escape. This neuronal activity is at least partially due to the identified descending neurons DMIa-1 and DMIb-1 (see Section 5.2). Selective wind stimulation of antennae or cerci corroborates the findings of ablation studies: ascending cercal information elicits escape more reliably and at shorter latency than does descending antennal information (Stierle et al., 1994). Consistent with this result, removal of the cerci strongly reduces responsiveness to strong wind stimuli and nearly abolishes escape responses to low wind stimuli. Also, response latencies below 40 ms are hardly ever observed after GI ablation. Bilateral antennectomy in animals with intact GI and cerci has no effect on responsiveness or latency. In a realistic prey–predator encounter, antennal tactile cues prove to be of great behavioural relevance for the appropriate escape of the cockroach (Comer et al., 1994). In a behavioural assay with four different predators (toad, mouse, preying mantis and wolf spider), 96% of escapes from spiders and 18–55% of escapes from other predators commenced only after tactile contact. In spider strikes, 93% of escapes were directed away from the side of first contact, even though this meant that many were directed towards the attacking spider (33% were directed towards the lunge of the spider). Response latency to touch of the predator is about 20 ms, and 25 ms in response to the touch of a mechanical probe. After antennectomy, significantly fewer escape reactions are elicited in response to spider strikes, but not to toad strikes. The reverse is true after cercectomy, indicating that the behavioural relevance of the two ‘emergency warning systems’ – antennae and cerci – differs according to predatorspecific stimulus profiles. Touching mechanoreceptors on the pronotum or on the hind leg with an artificial probe is more effective in eliciting escape than is touch of the antennae. On the other hand, stimulation of the antennae but not of the pronotum or hind leg, with probes having different surfaces (e.g. a glass rod and a pipe cleaner), leads to a differential responses (Comer et al., 1994; see Section 6.2.3 on related work by Comer et al., 2003). This suggests that antennal contact triggers escape only if a certain stimulus quality is detected. This is also supported by the observation that a glass rod may as well elicit a simple antennal avoidance reflex (Frings and Frings, 1949; Section 6.1.2.1). In contrast to surface properties, antennal posture does not affect responsiveness, but escape direction. Since the direction of the tactile stimulus, i.e. imposing a


143

small medial or lateral deflection of the flagellum, has no effect on escape direction (Ye et al., 2003), triggering and steering of escape must involve different neural mechanisms. Immobilisation of the HS- and SP-joint greatly reduces the likelihood of a tactile stimulus to elicit escape, whereas replacement of the flagellum by a prosthesis causes much less reduction of responsiveness (Comer et al., 2003). This result is paralleled by electrophysiological findings, which show that the typical escape-related activity of descending interneurons is altered significantly by immobilisation of the basal joints but not after replacement of the flagellum with a prosthesis. The fact that escape behaviour can be impaired significantly in some individuals with flagellar prosthesis, and that neither treatment completely abolishes escapes, indicates that deflection of the basal joints is neither sufficient nor necessary to trigger escape. In any case, some information processing of both proprioreceptive and exteroreceptive information must be postulated. Comer et al. (2003) suggested that the escapetriggering cue is deflection of the antennal joints, whereas the likelihood of eliciting escape depends on flagellar information. In another kind of emergency reaction, crickets use their antennae in detection of predators and early initiation of defensive behaviour. When attacked by a digger wasp, crickets (Acheta domesticus) often react by turning their abdomen to the offender, and by raising their body axis to assume postures called head-stand and stilt-stand. The latter is the posture that a cricket assumes for a defensive kick with its hind legs. Although the main difference between head- and stilt-stands is only the degree of abdominal elevation, Gnatzy and HeuXlein (1986) found that head stands are often elicited in crickets without antennae but rarely in animals without cerci, whereas stilt-stands are often elicited in crickets without cerci but rarely in animals without antennae. Thus, similar to the escape system of the cockroach, the predator defence system of the cricket uses synergistic action of descending antennal and ascending cercal information. Also, both ‘emergency behaviours’ are initiated by a fast-turning reaction. According to Gnatzy and Hustert (1989), the stiltstand of the cricket is probably elicited by campaniform and scolopidial sensilla of the pedicel. They report ‘‘weak stimulation of this area [in the pedicel] triggered stilt-stand, whereas stronger stimuli often led to a 1801 turn followed by stilt-stand with the cerci and hind legs toward the predator’’ (p. 225). 6.2

ACTIVE SENSING

Some of the behaviours discussed previously involve active antennal movements as part of reflex reactions (e.g. assistance reflexes), or are a facultative result of preceding active exploration (e.g. escape). In contrast, the active antennal movements discussed in the following sections are integral parts of active sensing behaviour, i.e. the animals act to increase the likelihood of tactile stimulation, or to change the quality of stimulus reception. Accordingly, the behavioural functions of active antennal movements and mechanoreception

144


include active exploration of the animal’s environment (Section 6.2.1); aiding complex locomotor actions such as climbing over obstacles (Section 6.2.2); aiding recognition and learning of surface structures or movement sequences (Section 6.2.3); or acting as obligatory components in complex behavioural sequences such as fighting or mating (Section 6.2.4). 6.2.1

Exploratory behaviour and tracking of objects

When regarding an antenna as an articulated, elongated, multi-sensory probe, it is trivial to observe that the likelihood with which any of its sensors will be stimulated depends on both the distribution of stimulants within the probe’s workspace and on its posture. Whereas the antennae of some insects can assume very complicated postures, for example with multiple kinks along the flagellum of Collembola (Simon, 1979), the flagellum of many other insects does not change shape during active movement, simplifying the analysis of active movement strategies. In the case of solid objects and, to some extent, of volatiles too, active movement of the antenna increases the likelihood of contacting a stimulant. Hence, active movement patterns of the antennae must be interpreted as explorative strategies that serve to sample the workspace of the flagellum for contact surfaces, odorants, concentration gradients, and the like. Moreover, not only the arrangement of the antennal joint axes and segment lengths, but also the distribution of the exteroreceptors on the flagellum are morphological parameters that determine the limits of spatial resolution (see Sections 2.2.1 and 3.2.4). These parameters set the frame for sampling efficiency and, therefore, the behavioural significance of any movement strategy. For example, the length of the antenna of the stick insect Carausius morosus is virtually the same as that of an outstretched leg of the same animal (Du¨rr et al., 2001) and a large fraction of the antennal workspace overlaps with the workspace of the leg. Movements of the antenna are therefore likely to be relevant to locomotion. If the stick insect moves its antenna without touching anything, the ipsilateral front leg ‘knows’ a trajectory where obstacles are unlikely to be met. If the antenna does touch something, the front leg must ‘beware of’ an obstacle. Accordingly, the fact that the cockroach P. americana preferentially points its antenna some 301 above the horizon and within a sector ranging from 601 to 901 lateral (Ye et al., 2003), means that this species ‘pays most attention’ to this part of its antennal workspace. Antennal movements that are part of the searching behaviour of walking cockroaches (cf. Okada and Toh, 2000) explore the space ahead for obstacles. Stick insects (C. morosus) not only continuously move their antennae during walking (Fig. 18A), but also coordinate the cyclic movement pattern of each antenna with the step cycle of the ipsilateral front leg (Du¨rr et al., 2001). Each antenna typically moves through one abduction/adduction cycle and two levation/depression cycles per step. The timing of largest abduction and subsequent adduction follows the start of stance movement in the ipsilateral front


145

leg (Fig. 18B). Thus, the pattern of fast antennal abduction and slower adduction becomes part of the gait of the legs. Moreover, the antennal movement pattern is also spatially coordinated with the leg movements. If the antennal searching efficiency was considered the likelihood of the antenna to touch an obstacle prior to leg contact, the typical antennal movement pattern of walking stick insects predict the animal’s behaviour: searching efficiency increases sigmoidally with obstacle height. The strongest increase occurs at obstacle heights of about half the clearance of the body (Du¨rr et al., 2001), which is equivalent to the height range in which model simulations have shown that it becomes necessary to change the normal locomotor pattern in order to climb the obstacle (Cruse et al., 1998; Kindermann, 2002). Furthermore, antennectomised stick insects detect obstacles later and fail to raise their body axis as smoothly as intact animals (Fig. 18C; Du¨rr et al., 2003), showing that stick insects indeed exploit antennal tactile information during walking. Like stick insects, crickets (G. bimaculatus, Horseman et al., 1997) and locusts (Schistocerca gregaria, Saager and Gewecke, 1989) move their antennae during walking. Although no obvious, simple cycle-to-cycle coupling between antennal and leg movements has been documented in either species, ascending neural pathways are crucially involved in this behaviour in crickets (Horseman et al., 1997): transection of a circumesophageal connective completely abolishes ipsilateral antennal movements during walking. Transection of the neck connective can reduce the angular velocity in both antennae during walking. When walking, the crayfish C. destructor typically moves its antennae independently of each other (Sandeman, 1985). During voluntary movements of the lobster antenna (H. americanus), the MC- and CF-joint either move in synchrony, or the CF-joint moves alone (Sigvardt, 1977). This is similar in the crayfish. Sandeman and Wilkens (1983) describe antennal movements of the freshwater crayfish E. armatus that occur in a number of behavioural situations. This species typically points its antennae backward during forward walking and forward during backward walking, possibly to reduce drag. Coupling between antennal and leg movements during walking occurs, for instance, after obstruction of the stance movement. The crayfish responds to such disturbance by moving its antennae in the opposite direction: flexion in response to a forward pull, extension in response to a rearward pull of the walking substrate. Although such neural coupling between leg movements and antennal movements is likely to influence active searching behaviour, it is reminiscent of the equilibrium responses of Palinurus sp. to substrate shift or tilt (Scho¨ne et al., 1976; Barnes and Neil, 1982; Neil et al., 1982, 1984). Occasional rhythmic antennal whipping movements of the walking Euastacus show no coupling to the stepping movement made by the legs (Sandeman and Wilkens, 1983). Thus, a clear relation between leg movements and antennal exploration movements remains to be demonstrated in crayfish. There is evidence that both insects and crayfish adapt their antennal posture and/or movement pattern to the locomotor state of the animal. For example, in

146


FIG. 18 Active exploration and use of antennae in stick insect locomotor behaviour: (A) Antennal movements of the stick insect C. morosus during walking are rhythmical and of fairly regular appearance. Trajectories of antennal tips are drawn to a sphere that is centred on the head and oriented as shown by the inset. The trajectory of the left antenna (right sphere) is drawn as a mirror image, to match the orientation of the right antenna. (B) Abduction phases of the antennae (first and fourth row of black bars) are


147

stick insects, in case of lack of foothold, antennal movements back-up the searching effort of the front leg (Du¨rr, 2001). In stick insects walking a curve, the average abduction angle for the antennae increases for the antenna on the inside of the curve, while it decreases for that on the outside of the curve (Du¨rr and Ebeling, 2005). Similarly, the crayfish C. destructor moves both of its antennae into the walking direction before initiation of walking and before turning (Zeil et al., 1985). Exploratory movements of antennae also occur in response to external stimuli. For example, honeybees move their antennae in response to an air puff containing an odorant (Suzuki, 1975). Both antennal joints contribute to this orienting movement, and fixation of one antennal joint does not preclude movement in the other. Similar stimulus-induced searching movements of the antennae have been described for several insect species, not only in response to odours but also to thermal gradients (e.g. in the bug Triatoma infestans; Flores and Lazzari, 1996). Spatial mapping of external stimulus direction and antennal posture occurs in two forms: first, as directed, open-loop responses, the magnitude of which correlates with the stimulus direction, and second, as antennal tracking of the object, probably in a closed loop. An example of an open-loop response is the reflex-like posture adjustment of the antennae of the spiny lobster P. argus. During rest, the spiny lobster holds its antennae laterally, approximately perpendicular to the body long axis. Upon hydrodynamic stimulation with water jets, or tactile stimuli to the body, the ipsilateral antenna is moved anteriorly or posteriorly, always appropriately directed to reduce the angle between stimulus direction and flagellum (Wilkens et al., 1996). Although response amplitudes

continued

coupled to the pattern of swing movements of the ipsilateral walking legs. Diagonal dotted lines indicate the back-to-front sequence (a metachronal wave) of leg swing movements and antennal abduction. LA and RA: left and right antenna, respectively; L1–L3: left front, middle and hind leg; R1–R3: right legs. (C) The tactile sense is exploited for efficient locomotion. Head trajectories of stick insects walking towards a rectangular obstacle (hatched area, side view) show how the insect raises to climb the edge (walking direction: left to right). Stick insects with intact antennae (left) detect the obstacle earlier (triangles ¼ mean position at first contact) and climb the obstacle with more clearance (circles ¼ position when reaching twice the average walking height) than antennectomised animals (right). Sighted (top) and blind (bottom) animals behave the same. (D) Tactile antennal cues can cause rapid re-targeting of an on-going front leg swing movement. Side view (left) and top view (right) of a stick insect walking sequence towards a vertical pole. Three stick figures show body axis, right front leg and tarsus position (circles) at times of lift-off, first antennal contact with the pole (open circle) and leg contact with the pole (solid line: tarsus trajectory; dotted line: trajectory of the antennal tip). A normal swing movement would have been continued as indicated by the dashed arrow. Instead, no more than 60 ms after antennal contact, the swing movement is re-directed to grasp the pole. Data taken from Du¨rr et al. (2001) (A, B); Du¨rr et al. (2003) (C) and Du¨rr and Krause (2001) (D).

148


are generally low and rarely exceed 301, it shows that Panulirus has a sense of spatial direction, allowing it to transform stimulus angle into a proportional antennal angle. Similarly, the crayfish Procambarus clarkii responds to hydrodynamic stimuli to mechanoreceptors on the telson by backward sweeps of both antennae. As in Panulirus, the angle of the sweep is small (30–601) but correlated with the angle of stimulation (Schmitz, 1992). Evidence for spatial mapping of exteroreceptive information to proprioreceptive cues from the antenna stems from observations on antennal tracking behaviour. For example, the crayfish Procambarus alleni points its antennae at moving objects and tracks them (Bovbjerg, 1956). C. destructor points its antennae towards novel visual stimuli (Sandeman, 1989), and Orconectes limosus directs its antennae towards the source of water vibration (Tautz, 1987). To date, the best-studied antennal tracking behaviour is that of crickets (Gryllus sp.). In experimental trials with visual objects moving horizontally through an arc of 1801, these insects track the object with their ipsilateral antenna through a range of 1101 lateral towards the midline (Honegger, 1981). When the object crosses the midline, typically the antenna stops moving and the contralateral antenna continues tracking. Occasionally, both antennae move in parallel until the adducted antenna points 301 contralaterally. Various visual cues influence antennal tracking in crickets, including the direction of movement (determining the tracked edge), object size and orientation (summarised by Honegger and Campan, 1989). The antennal tracking response of crickets is strongly habituating (Honegger, 1981). Habituation depends on the size of the tracked object, being stronger for small objects (e.g. 131 diameter) than for large objects (e.g. 251 diameter). When tracking small disks the gain of the response is small, i.e. the antenna lags behind, and the response may even subside completely (Honegger et al., 1985). However, a 30 s interval between trials is sufficient for dishabituation of the tracking response. Vibration of a tethered cricket reduces habituation significantly, increasing the correlation between the antennal angle and the target angle. Vibrating the stimulus has no dishabituating effect, indicating that dishabituation is caused mechanically, not visually. There is no frequency dependency of dishabituation by vibration within the tested range of 25–450 Hz. Walking crickets do not track visual objects, indicating that the visual-antennal connection is gated by signals related to thoracic motor activity (Honegger, 1981). Much like active antennal movements during walking, antennal tracking also requires functioning ascending pathways, as transection of a circumoesophageal connective abolishes tracking by the ipsilateral antenna (Horseman et al., 1997). So far, all of the published data refer to horizontal tracking that relies on movement of the SP-joint. To what extent the HS-joint allows tracking of vertically moving objects is unknown. The cockroach P. americana points its antennae towards passing conspecifics and tracks them (Ye et al., 2003), but tracking performance has not been quantified. A visual cue presented at an azimuth beyond the preferred antennal


149

orientation angle causes abduction of the ipsilateral antenna, a movement that often leads to object contact. Whether or not the visually mediated pointing and tracking behaviours of crustaceans and insects are genuine closed-loop behaviours that reduce the visual angle between the image of the target and the image of the antenna is unknown. In fact, it is unclear whether crickets or crayfish can see their own antenna well enough to measure its posture, as visual spatial resolution is fairly low. It is unlikely that these animals can see their antennal tips, but possibly very good lighting conditions could provide sufficient contrast to crudely judge the orientation of the antennal base. An alternative mechanism would be that the animal matches the visual angle of the target with the output of a neural forward model of its own motor action, or by mapping of proprioreceptive cues to external angles (see Section 7.2). 6.2.2

Guidance of locomotion and tactile localisation

As has been indicated in the discussion of stick insect exploratory behaviour (Section 6.2.1), active antennal movements during walking effectively detect obstacles and permit the insect to adjust their gait or trajectories of single leg movements appropriately. Like the stick insect, the potato beetle Leptinotarsa decemlineata raises its body in response to contact with obstacles (Pelletier and McLeod, 1994). While antennectomised beetles bump into obstacles that are higher than 2 mm, blind beetles with intact antennae increase the angle between the body long axis and the substrate with increasing obstacle height. In response to touching very high obstacles (7 mm), some beetles lift both front legs off the ground, elevating the prothorax by a distance that exceeds the length of the front legs. Levation of the prothorax also occurs in stick insect climbing (Du¨rr, 2000; Fig. 18C). Similarly, Watson et al. (2002) show that the cockroach Blaberus discoidalis often stops walking to begin a rearing movement, i.e. body axis inclination, before climbing an obstacle. The kinematic basis underlying this rearing movement is a tibial rotation of the middle legs, caused by the coxatrochanter joint and the trochanter-femur joint (which is not fused in this species). Because this rotation can occur prior to leg contact with the obstacle, Watson et al. (2002) conclude that either visual or antennal tactile cues trigger rearing and subsequent climbing. Antennal movements are part of the searching behaviour of walking cockroaches (P. americana). Upon contact with a vertical rod, unrestrained cockroaches decrease walking speed, orient towards the rod, repeatedly contact the rod with the antennae and climb up it (Okada and Toh, 2000), much like the stick insect C. morosus does in similar situations (Du¨rr, 2000; Du¨rr and Krause, 2001). In stick insects, antennal contact with a vertical rod may lead to short-latency re-targeting of an ongoing swing movement of the ipsilateral front leg (Du¨rr and Krause, 2001; Fig. 18D). This suggests that antennal

150


mechanosensory information is passed rapidly to leg motoneurons and that swing movements are modulated by descending neural input. In the cockroach, shaving off the scapal hair plates affects the antennal contact interval of the repeated contacts with the rod, but not the overall cockroach behaviour, i.e. orienting and climbing (Okada and Toh, 2000). In free walking stick insects of the species Aretaon asperrimus, antennal contact with the far edge of a gap is accompanied by a rearward shift of the liftoff position of the front leg (Bla¨sing and Cruse, 2004). This suggests that the antennae signal the presence of a ‘support area within reach’, leading to an increased forward shift of the thorax which precedes the step across the gap. Tethered walking cockroaches persistently attempt to orient themselves towards a touched object, a behaviour that requires the hair plates of the HSjoint (Okada and Toh, 2000). The ability to orient towards a touched object indicates the ability to code the contact angle. Provided that the contact distance is also coded, a neural polar coordinate system could subserve tactile localisation. However, the fact that cockroaches without HS-joint hair plates are not impaired in their ability to locate and climb obstacles (Okada and Toh, 2000) suggests that these proprioreceptors are not necessary for tactile localisation. To date, not a single study unambiguously demonstrates that insects can indeed code the spatial location of a touched object, i.e. both distance and angle, using solely antennal tactile cues. This will require careful evaluation of a targeted behaviour that is directed by tactile information. In Crustacea, Zeil et al. (1985) were the first to present quantitative evidence that crayfish use their antennae for tactile localisation of external objects. Studying active antennal movements of hungry freshwater crayfish C. destructor, they observed how the animals attacked objects which they had previously touched with their antenna. Vision is of no importance to this behaviour, as blinded crayfish perform as well as sighted ones. Touching a crayfish flagellum with a brush causes retraction of the antenna, turning of the animal towards the contact point, and an attack (Fig. 19A, B). Zeil et al. (1985) show that the turning angle correlates with the angle of the antenna at contact time, and that attack distance correlates with the distance of the contact point on the flagellum. Furthermore, turning angle varies systematically with contact distance (Fig. 19C, D), accounting for the spatial separation of the antennal joints and the centre of rotation of the body. Because the centre of rotation in C. destructor is located some 5 cm behind the antennal base, a distant object, that is contacted at the same antennal angle as a proximal object, requires a larger turning angle than the latter. For the same reason, objects that are touched by the same point on the flagellum but at a different antennal angle, require different attack distances following an appropriate turn. Indeed, crayfish take their own body geometry into account, suggesting that they make appropriate use of both polar coordinates, i.e. contact distance and angle. In a follow-on study, Sandeman and Varju (1988) recorded the average movement trajectory of the crayfish with normal and with manipulated


151

antennae. During an attack, the crayfish rotates so that the touched object is aligned with its long axis, in the region of the mouthparts (Fig. 19E). The turning direction of the crayfish is always correct, and turning angles correlate with the angle that is expected if the crayfish was to turn first and then move towards the object. For distant contact points, however, the turning angle is closer to the antennal angle than to the expected angle. Furthermore, Sandeman and Varju (1988) show that the crayfish turns during the forward movement, i.e. that rotation and translation of the body are not sequential but simultaneous (Fig. 19F). Thus, deviation between the turning angle and the expected angle may be explained by the ability of the crayfish to take the time course of its own movement into account. After immobilisation of the basal antennal joints with a splint, i.e. without impairing proprioreception, attack accuracy remains unaffected. However, bending of the flagellum by a splint biases the attack towards the location of the phantom contact point, i.e. the point at which antennal contact would be expected if the bend of the flagellum had been ignored (Fig. 19E). This bias is stronger with a medial bend than with a lateral bend, suggesting that some proprioreceptive information from the flagellum is integrated to localise the contact point. 6.2.3

Pattern recognition and learning

The ability to discriminate mechanical surface properties on the basis of tactile information is difficult to show unequivocally, at least in animals that cannot be conditioned to associate a tactile cue with a reward. It is probably for this reason that most knowledge about tactile pattern recognition in arthropods has been gained from experiments with honeybees. Moreover, most experimental results on tactile recognition are linked to learning performance. This must be kept in mind because in some learning paradigms it is not evident whether the animal’s ability to succeed is of particular relevance to its natural behaviour. We are aware of only two studies that demonstrate the use of the antennal tactile sense in recognition or evaluation of surface properties in a behavioural experiment without rewarding appropriate performance. One of them concerns the differential assessment of surface structure in situations that are potentially dangerous for a cockroach (Comer et al., 2003), and the other concerns the fine-tuning of honeycomb building activity of the honeybee (Martin and Lindauer, 1966). In the cockroach P. americana, flagellar information qualifies the contact with information about the surface properties, thus modulating the likelihood of an escape. For example, touching a wolf spider is more effective in triggering escape than touching a conspecific, provided there is sufficient time for palpation, i.e. repeated brief contacts (3–5 s used by Comer et al., 2003). Extraction of surface chemicals does not alter the differential response to predator and conspecific, and spider- and cockroach-scented cardboard models do not cause different responsiveness. Therefore, contact-chemoreception is not

152


FIG. 19 Tactile localisation in the crayfish C. destructor: (A) Hungry crayfish attack an object (solid circles) by moving towards the location touched by the antenna. aSENSE: contact angle of flagellum; dSENSE: contact distance along flagellum; CoR: centre of rotation; T: Target position at contact time; V: virtual target position when disregarding


153

necessary to raise or lower the likelihood of an escape response. Comer et al. (2003) took a lot of care to standardise the stimulus presentation in that palpation time, stimulus movement and ‘unscentedness’ were well controlled. Nevertheless, the different body shapes of a spider and cockroach provide many more cues than just cuticular surface structure, and further studies will be necessary to conclude whether it is genuine surface structure information or a mix of structural and positional cues, e.g. the distance or posture of the spider’s legs rather than their hairiness. Honeybees are well known for their ability to build highly ordered and evenly spaced honeycombs. Apart from the geometrical regularity, honeybees keep the thickness of the wax walls within a narrow range, and appropriately manufacture different wall thicknesses for worker cells and drone cells. The thickness of the wall is set by a pressing and scraping the mandibles against the wall, while continuously probing the deformation around the pressure point with the antennal tips. Various impairments of the antennae of entire honeybee colonies, such as unilateral antennectomy or bilateral clipping the distal annuli of the flagellum, have no detrimental effect on average cell diameters or angular arrangement of a newly built honeycomb (Martin and Lindauer, 1966). However, loss or numbing of the antennal tips (by clipping or by dipping into phosphoric acid, respectively) does impair the bees’ fine-tuning ability in controlling the thickness and evenness of the wax walls. On average, walls become much thicker, and occasional holes can occur. Martin and Lindauer (1966) attribute the mechanosensory significance of the antennal tip to a special

continued

the bend; X, Y: Axes of the body coordinate system. (B) Schematic illustrations of four example attacks towards objects located in different directions and distances (dashed lines with open circles. Numbers indicate time in s after first contact). Attacks consist of a turn and forward translation. (C) Turning angle depends on contact angle and contact distance. Attacks to near objects (triangles; d SENSE ¼ 627 cm) have smaller turning angles than attacks to distant objects (squares; d SENSE 413 cm). Grey areas mark predicted angles for exact localisation, if turning preceded translation. (D) Attack distance depends on contact distance and contact angle. Attack paths to lateral objects (triangles; d SENSE 4901) are shorter than paths to rostro-lateral objects (squares; d SENSE o451). Grey areas mark predicted path lengths for exact localisation, if turning preceded translation. (E) Average relative movement of target, T, in body coordinates. Both X- (open symbols) and Y-coordinates (solid symbols) decrease to a value near zero, corresponding to an ideal position near the mouth. If the flagellum is bent laterally, average attack paths are intermediate to what would be the ideal localisation of the real (T, circles) or the virtual target (V, squares). If the flagellum is bent medially, attacks aim at the virtual target. (F) In a real attack, the animal rotates and translates simultaneously. In attacks to a distant target, the turn angle is larger than the sensed angle (left). If the turn angle was equal to the sensed angle, the Y-error would be larger. Adapted from Zeil et al. (1985) (B–D) and from Sandeman and Varju (1988) (A, E, F) with permission from Springer Verlag.

154


arrangement of sensilla on the last annulus, which they call the ‘‘sensory plate’’ (Sinnesplatte of Martin and Lindauer, 1966). Following measurements of the required force to deform the wax wall and the time constant of the damped return of the wall upon release of the pressure point, Martin and Lindauer (1966) suggest that the bees sense these parameters by the ‘sensory plate’. Compared to the experimental evidence on tactile pattern recognition in insects, there is not much known about this ability in Crustacea. Although there are many studies that discuss the role of antennal movements in crayfish behaviour, we are unaware of any study that reports active tactile exploration of an object similar to that seen in insects. Rather, Sandeman (1989) explicitly states ‘‘crayfish [Cherax destructor] have not been seen to explore the contours of a novel object with their flagella in a way to provide them with information about its shape or size’’ (p. 216). Honeybees can be trained to tactually discriminate surface structures, both in unrestrained (Martin, 1965; Kevan and Lane, 1985; Erber et al., 1998) and restrained training situations (e.g. Kevan and Lane, 1985; Erber et al., 1998; Scheiner et al., 1999). Martin (1965) used a system of four horizontal, odourless, narrow corridors for simultaneous presentation of four alternative patterns, only one of which was rewarded with food. Each pattern had a different surface structure (ca. 1-mm wide longitudinal or transverse furrows, 1-mm wide holes, and a plane surface), and lined all four surfaces of a corridor, immediately adjacent to the entrance. Bees were able to associate any of the four surface structures with a reward, though much longer training sessions were necessary than for odour learning (1.5–3 days, compared to 5–6 h). Martin (1965) further showed that clipping the distal one to three annuli of the flagellum abolishes tactile pattern recognition. So does partial fixation of the scape that does not preclude antennal contact with the test stimuli, but prevents the normal contact pattern with the antennal tip. This led Martin (1965) to suggest that appropriate contacts of the ‘sensory plate’ at the tip of the flagellum are crucial for tactile pattern recognition in honeybees. Honeybees are able to discriminate microtextural patterns of flower petals, where the components of the epidermal structures of the petal are of the same size range (ca. 10 mm) as the spacing of the sensilla at the antennal tip (Kevan and Lane, 1985). In a two-alternative learning paradigm, unrestrained bee workers can be trained to associate the petal of the flower Helianthus annuus with a sucrose reward. Thereafter, bees successfully discriminate the trained surface structure against the petals of other flower species, against the reverse petal surface of the same flower, and even against the same petal surface with reversed orientation. In an operant conditioning paradigm, restrained bees learn to extend their proboscis upon presentation of a tactile stimulus (Kevan and Lane, 1985). When conditioned to the petal surface of H. annuus, bees show proboscis extension in 88–100% of presentations of H. annuus (median: 100% correct responses), but only in 0–22% of presentations of petals of another plant species (Xylorhiza wrightii, median: 83% correct responses).


155

When presented with feeder holes with adjacent tactile stimulus plates (on which they can land on) free-flying bees can learn to associate grain size of abrasive paper or the orientation of a coarse grating (1.2 mm spatial wavelength) with a sucrose reward (Erber et al., 1998). However, in the paradigm used in this study, it is difficult to exclude the possibility that bees use other sensory structures than those of their antennae. For instance, free-flying bees fail to learn the same task with finer gratings, despite the fact that they do learn very similar stimuli in a more controlled operant conditioning paradigm that only allows for antennal input. Restrained bees can be conditioned to extend their proboscis in response to tactile contact with an object (Erber et al., 1997), and conditioning can be refined to particular aspects of the touched surface. For example, bees conditioned to a vertical grating of 450 mm spatial wavelength, can differentially respond to various surface properties, such as orientation angle (e.g. detect a 22.51 difference), or spatial wavelength (e.g. detect a 170 mm difference; Erber et al., 1998). Applying the same training paradigm, Erber et al. (1998) also show that bees can learn to distinguish shape, orientation and size of a smooth metal plate. Nevertheless, it should be noted that bees always also respond to the unrewarded alternative in at least 30% of trials. Compared to other sensory modalities, texture learning in free-flying bees is slower than olfactory or colour learning, but faster than visual form learning. Applying the same experimental paradigm as Erber et al. (1998), but only testing vertical and horizontal gratings, Scheiner et al. (1999) show that reversal learning is possible at a similar acquisition rate as initial learning (Fig. 20). Learning performance is different between pollen foragers and nectar foragers, but this difference can be attributed to differences in response threshold to reward concentration, and the resulting ‘perceived value’ of the reward.3 Furthermore, antennal tactile sensing behaviour can be sensitised by a sucrose stimulus and modulated experimentally by injection of the biogenic amines serotonin and octopamine (Pribbenow and Erber, 1996). Although these aspects of neuromodulation are unlikely to be specific to tactile learning (e.g. visually induced antennal movements can be modulated by the same substances; Erber and Kloppenburg, 1995), they are noteworthy when considering response variability within the same and between different behavioural experiments. Apart from discriminating the surface structure, honeybees can also be conditioned to touch a metal plate with their antenna or, in the presence of two plates, to prefer touching one plate over touching the other (Kisch and Erber, 1999). If an object is placed within the workspace of the antennae of a restrained bee, the bee repeatedly antennates the object. This tendency can be enhanced by operant conditioning to a sucrose reward (Fig. 21A, B). Following differential 3 According to Scheiner et al. (1999), the ‘perceived value’ is a function of the true value, i.e. the sucrose concentration, but also of a subjective value that depends on the internal, motivational state of the bee.

156


FIG. 20 Tactile antennal learning and discrimination in pollen and nectar foragers of the honey bee (Apis mellifera). Restrained bees, whose eyes were occluded with paint, were trained to associate the proboscis extension response (PER) with a tactile pattern. The conditioned stimulus consisted of a 3 4 mm copper plate with grooves which were 150–190 mm wide and 30–40 mm deep. In each training trial, bees were free to make antennal contacts with the training stimulus (vertical grating) for 3 s. Subsequently, the PER was elicited by applying a droplet of sucrose to the antennae. When the bee extended its proboscis, it was rewarded with a droplet of sucrose. If the bee responded with a conditioned PER, a drop of sucrose was also offered to the proboscis. The inter-trial interval was 5 min. Learning performance is presented as percentage of bees showing conditioned PER. (A) Learning curve of 75 pollen foragers (circles) and 75 nectar foragers (squares) over a sequence of 5 trials. (B, C) In the subsequent extinction series, bees were presented with either a vertical or a horizontal grating, but no reward was offered. In the first trial, performance is much better in response to the vertical grating compared to the horizontal grating, revealing orientation specificity of learning. Extinction after three unrewarded trials differs between types of foragers (*). (D) Reversal learning over a series of eight trials with the same procedure as in A, but with horizontal grating. Both forager types show successful reversal learning, but differ in performance (*). (E, F) Extinction sequence as in B, C. In the first trial, orientation specificity is reversed compared to B, and differences between forager types are more pronounced. Adapted from Scheiner et al. (1999) permission from Springer Verlag.

conditioning to one out of two objects, the learnt preference can be reversed by a second training period (Kisch and Erber, 1999). Very similar antennal movements can be conditioned by offering a reward only if the muscle activity of the SP-joint flexor muscle reaches a certain threshold (Erber et al., 2000), indicating that tactile contact is not necessary for operant conditioning of antennal movements (Fig. 21A, C). Assuming that the conditioned movement pattern is identical to the movement pattern that occurs in tactile scanning of an object, associative learning of some object properties, e.g. the shape, could be done by learning a movement pattern rather than a contact pattern. The latter interpretation gains support


157

by the fact that the training procedure of Erber et al. (2000) rewarded the activity of a single fast flexor motoneuron. Moreover, this motoneuron is known to govern antennal tactile scanning behaviour (Pribbenow and Erber, 1996; Erber and Pribbenow, 2000). Since the honeybee can recognise object properties by antennal movements, and important features of the movement pattern are dominated by a single motoneuron, the activity of which can be conditioned experimentally, one might speculate that the activity pattern of this motoneuron is indeed sufficient for the honeybee to learn a tactile stimulus. However, because fixation of the antennal joints abolishes associative plasticity of flexor activity, and fixation of only the HS-joint reduces the performance (Erber et al., 2000; Fig. 21C), feedback about the resulting antennal movement appears to be crucial. Nevertheless, it is likely that at least some aspects of tactile learning involve learning of the antennal movement pattern rather than learning the actual contact information. Erber et al. (1997) demonstrate that non-associative, i.e. non-rewarded, learning of antennal movement patterns occurs in restrained honeybees that antennate a presented object for periods of at least 10 min or more. When a metal plate is placed into the workspace of the antennae, bees spontaneously and repetitively touch the object at contact frequencies between 3 and 300 contacts min1, and a median contact duration below 10 ms (Erber et al., 1993). This antennation behaviour is called ‘antennal scanning’ by Erber et al. (1997). The pattern of antennal scanning after removal of the object significantly differs from the spontaneous movement pattern before object presentation: the antennal tip more frequently moves through the area where the object was located and, after 30-min presentation, also avoids areas where the edge of the object was. Total contact time is not the decisive parameter governing this form of motor plasticity, because the required number of subsequent presentation periods depends on the duration of the single presentations (e.g. 1 10 min is enough, but not 2 5 min; Erber et al., 1997). Mechanosensory information about contacts with the object is not necessary either, because cutting the antennal nerve in the flagellum or replacing the flagellum by a prosthesis does not affect it (cf. Erber and Pribbenow, 2000). However, non-associative plasticity cannot be induced when presenting a wire rather than a plate, although the wire is repeatedly touched by the antenna (Erber et al., 1997). Also, fixing the SP-joint precludes the object-related change. This suggests that the observed motor plasticity is not simply a matter of entraining a movement sequence, but also involves some information related to the touched object, e.g. it’s size or shape. Although Erber and Pribbenow (2000) demonstrate that sensory cues have no impact on the ‘scanning parameters’, i.e. contact duration and frequency, the possibility that antennal mechanoreception plays a role in motor plasticity should not be excluded completely. Significant sensory impact on kinematic parameters of the antennal movements may well prove to be important to the natural honeybee behaviour, without showing up in the mere contact sequence.

158


FIG. 21 Operant conditioning of antennal movements in the honeybee works both with and without tactile contact. (A) Restrained honeybees, A. mellifera, were rewarded either if they touched a small metal plate with their antenna with sufficient contact frequency (tactile variant), or if the EMG of the flexor muscle of the SP-joint contained sufficient spikes of the fast flexor motoneuron (EMG variant). In both variants,


6.2.4

159

Communication

The list of reports about the use of antennal movements or tactile contact in intraspecific encounters is long, and by no means can a review of these reports be complete. Many of the published observations are of anecdotal character in that they do not investigate the kind of information that is emitted or received, nor which cues are the decisive ones. The following sections provide an overview of the range of behaviours in which intraspecific communication is known to involve antennal mechanoreception. Emphasis is given where ablation or training procedures have specified the significance of antennal contact. In intraspecific communication, tactile and vibration cues that are sensed by antennal mechanoreceptors can be of great importance. For example, Martin and Lindauer (1966) immobilised the antennae of each insect in an entire honeybee colony. The effect on the colony’s social behaviour was summarised in a footnote by Martin (1965, p. 284):4 ‘‘Bees of an experimental colony (consisting of some continued

electronic signals were amplified (Amp, connected to either of the signal sources), stored on a DAT recorder, digitised by a window discriminator (WD) and monitored on an oscilloscope (OSC). Spontaneous occurrence of the measured signal was determined by its mean frequency, m, and standard deviation s. A computer showed a signal if a sucrose reward was to be offered. (Bi–Biiii) Tactile variant, with experimental protocol shown in Bi. The reward threshold Th 4 ¼ , m þ s was calculated from the contact frequency during a 10 min pretest period (Bii). After a first conditioning session with five rewards (* in Biii), Th was increased to m þ 2s and another five rewards were given in a second conditioning sequence (Biii). Each reward was followed by a 2 min period without evaluation (horizontal lines). Finally, training success was evaluated during a 10 min post-test interval (Biiii). Biiii shows twice the average contact frequency of Bii (8.673.2 compared to 4.372.3 per 10 s). (Ci–Ciiii) EMG variant. Experimental protocol (Ci), pre-test (Cii), conditioning with 10 rewards (Ciii) and post-test (Ciiii) equivalent to the tactile variant, except that only one conditioning threshold Th was used. Graph details as in Bi–Biiii. Conditioning caused an increase of flexor spike frequency from 36.0732.7 in Cii to 99.4738.5 potentials per 10 s in Ciiii. (D) Comparison of tactile variant and three versions of the EMG variant by means of a conditioning index, CI, that quantifies the difference of pre- and post-test signal frequencies, divided by the sum of pre- and post-test signal frequencies. In both, the tactile variant (tactile, nExp ¼ 25, nControl ¼ 10) and the standard EMG variant (EMG, nExp ¼ 10, nControl ¼ 10), differences in CI are statistically different between the conditioned group and a yoke control group (bees that receive the same reward sequences as conditioned group, but independent of their behaviour). If the scape is fixed (EMG HS fixed, nExp ¼ 10, nControl ¼ 9), the conditioning of the EMG variant is still successful. A non-associative increase in CI occurs in the EMG variant, if the antenna is completely fixed (EMG, all fixed, nExp ¼ 5, nControl ¼ 5). Adapted from Kisch and Erber (1999) (B) and Erber et al. (2000) (A, C, D) with permission from Springer Verlag and Elsevier Publishing. 4

Translated from German. The ‘operation done’ was, as one of the experimenters recalls, immobilisation of both antennal joints (Martin Lindauer, personal communication).

160


500 bees) were refrained from mutual antennation by an operation done to their antennae. Bees no longer aggregated to form a building cohort. Social behaviour ceased completely. The animals dispersed uniformly in the small observation hive. Even the queen was disregarded! Moreover, mutual food exchange no longer took place. Control experiments showed that the sense of smell was completely intact’’. Apart from qualitative descriptions, at least three behavioural contexts in arthropods have been shown experimentally to involve communication via antennal contact: agonistic behaviours that serve to establish dominance order or territoriality (Section 6.2.4.1), transfer of information about resources (Section 6.2.4.2), and courtship and mating (Section 6.2.4.3). 6.2.4.1 Agonistic behaviour. Some species of crustaceans and insects establish dominance orders and/or conquer and defend territories, and in the corresponding agonistic behaviours, antennae can be used as both emitters and receivers of information. Many insect species fight over territories and resources, and aggressive encounters in social species may involve whole colonies. For example, colonies of the honeypot ant Myrmecocystus mimicus carry out ritualised tournaments to settle territorial fights. During these tournaments, pairs of ants engage in a peculiar lateral display during which they raise their abdomen and employ their antennae to drum on each other’s behind with great intensity (Ho¨lldobler, 1976). The kind of information transmitted by antennal drumming is yet unknown. It is likely to signal the opponent’s strength, but may also involve additional contact-chemosensory information about the opponent’s chemistry. Agonistic encounters among male crickets (G. bimaculatus) are an ideal behavioural model to study the role of sensory cues in a complex behavioural chain. This is because males express different levels of aggression in an obligatory sequence of behaviours, beginning with antennal fencing, continuing with unilateral and bilateral mandible spreading, escalating to mandible engagement and, finally, wrestling (Fig. 22A). Body weight, or externally visible parameters that are correlated to weight (e.g. size, head width), are not good predictors of the outcome of a fight, or even of the aggression level reached. Even in pairs with a large size difference, larger males tend to win only some 70% of encounters (Hofmann and Schildberger, 2001). Antennal fencing appears to serve the assessment of the opponent’s strength, possibly by sensing the vibrations of the flagella. Males without antennae do not fight, even if two previously dominant males are paired (Fig. 22B). During fencing, the two antennae are moved independently. If the opponent’s flagella hit each other, they often vibrate for periods of 1 s or longer. Hofmann and Schildberger (2001) show that the frequency spectrum of the induced flagellar vibration contains information that correlates with the level of aggression that will be reached subsequently. In particular, the ratio of the main frequency modes of the opponent’s spectra appears to be a good predictor (main mode for animal 1: main mode for animal 2; see Fig. 22D, E): fights between opponents with similar vibration frequencies


161

tend to escalate, whereas opponents with large differences in vibration frequency settle the fight after fencing or mandible spreading. Thus, crickets appear to assess the relative stiffness and/or hit strength, both of which would affect the vibration frequency of the flagellum. This is interesting because it is obvious that a sensor such as Johnston’s organ could sense the animal’s own flagellar vibrations, but sensing the opponent’s flagellar vibration must be rather complicated. To date, it is unclear how this could be done. Antennal fencing also occurs in agonistic encounters of cockroaches. When excised male antennae are used to fence artificially with a conspecific male, P. americana typically responds by backing away from the stimulus, or by elevating its body axis to assume a stilt posture (Bell, 1978). The same experimental situation causes rather different responses in the more aggressive species Nauphotea cinerea. Depending on the social status, dominant males tend to lunge towards the stimulus, whereas the sub-ordinate males rather assume a submissive posture and may even retreat. The stilt posture can be assumed by both groups, showing that the transferred information is sufficiently graded to allow for a range of responses. American lobsters (H. americanus) use their antennae to whip the opponent on the claws, thus emitting information about their aggressive state. The whipped lobster senses the offensive signal with mechanosensory hairs on its claws (Solon and Cobb, 1980). The range of possible responses to the received signal (e.g. escape, attack, etc.) suggests that lobsters transmit graded information about their size or strength, a cue that is likely to be relevant in the establishment of a dominance order among conspecifics. Huber and Kravitz (1995) analysed components of the behavioural sequence of agonistic encounters among juvenile H. americanus, including two kinds of whipping and lashing movements, and a distinct antennal posture where the flagella are pointed upward. The latter ‘‘antenna up’’ posture is part of a threat pattern, and is frequently accompanied by raised claws and an inclined body axis. Single antennal whips occur in a behavioural pattern termed ‘‘do-si-do’’, in which the animal tends to approach its opponent with its claws held downward. Repeated antennal whipping and lashing characterises a further behavioural pattern. The statistical analysis of Huber and Kravitz (1995) indicates that these patterns occur in a fixed sequence that, like the levels of aggression described for the cricket (see above), indicates the intensity of aggression. As encounters reach different intensity, the components of the behavioural patterns appear to serve an exchange of information that determines the course of the encounter. In the crayfish species Procambarus alleni, antennal cues provide sufficient sensory cues to establish a dominance order (Bovbjerg, 1956). Ablation experiments show that blind crayfish still know which end of their opponent to attack, an ability that is abolished by additional antennectomy. Lack of antennae and eyesight precludes the establishment of a dominance order. Whether or not antennal tactile cues are augmented by contact-chemosensory cues in this behaviour is unknown. Because either intact vision or intact antennal sensing are

162


FIG. 22 Antennal fencing behaviour is an indicator for intensity of cricket fights: (A) Agonistic encounters in the cricket G. bimaculatus follow a behavioural sequence that can be described by six distinct levels of aggression. Antennal fencing (level 2) is the first fighting stage. (B) Antennae are crucial for the initiation of fighting. Pairs of crickets either fight (black bars), show pre-established dominant or sub-ordinate behaviour


163

sufficient but not necessary to establish a dominance order, it is likely that normal crayfish use cues received by both of these senses in aggressive encounters. Crayfish of the related species Orconectes virilis respond strongly to the antennal posture of an approaching conspecific, and antennal posture is a good predictor for the intensity of the encounter in this species (Heckenlively, 1970). Orconectes also performs antennal waving displays during agonistic encounters, a behaviour that is also executed in the dark (Bruski and Dunham, 1990) and even with antennal stumps (Ameyaw-Akumfi, 1979). In the latter situation, the display lasts longer and the performing animal is held longer by its opponent, leading Ameyaw-Akumfi (1979) to interpret antennal waving as an ‘appeasement display’ that transmits mechanical cues.5 In support of this view, Tierney et al. (2000) describe antennal waving to be performed by the subordinate animal, and in one species of Orconectes, it is sex-specific to males. The display results in release of chela contact by the dominant animal. Visual cues affect various aspects of antennal waving displays (Bruski and Dunham, 1990), but whether visual cues transmit information to the opponent is still a matter of debate. 6.2.4.2 Transferring information about resources. Possibly the most famous form of insect communication is the waggle dance of the honeybee (Apis mellifera), a locomotor display by which a worker bee transmits information about the location of a food source to its colony companions. Although it has been known for many decades that the directional information is coded by the dance direction relative to gravity, the mode by which this information was transmitted was unknown for a long time. Owing to the detailed analyses of

continued

(white bars), or court each other (hatched bars). Shortening of antennae to half of the flagellum length does not change a pair’s behaviour. If both flagella are shortened to 1/8th, encounters without fights are more frequent compared to intact animals, and courtship may occur. Bilateral antennectomy completely abolishes fighting behaviour and further increases the frequency of courtship behaviour. (C) Duration of antennal fencing in cricket encounters with same treatments as in (B). The duration increases significantly in crickets with antennae shortened to 1/8th of the normal length. (D) Frequency spectra of flagellar vibrations after mutual antennal contact during fencing. In pair 1 the modes with highest frequency components were similar (ratio 23:23). This fight escalated to level 5. In pair 2 the frequency spectra differ more strongly (ratio 35:22). This fight stopped at level 3. (E) Among six fights where flagellar frequency spectra were obtained, the final level of aggression correlated significantly with the frequency ratio of the opponents’ high-frequency modes (rS ¼ 0:88, po0:02). Adapted from Hofmann and Schildberger (2001) with permission from Elsevier Publishing. 5 Although Ameyaw-Akumfi (1979) uses the word ‘antennule’ in one table and in a few places of the text, the remaining text and images suggest that the appendage referred to really is the 2nd antenna, not the antennule (1st antenna).

164


Michelsen et al. (reviewed in Michelsen, 1993), it is now established that the vibrations produced by the waggle dancer are airborne sounds, and cannot be detected via substrate vibrations. Rather, the cues produced by the dancer are characterised by an acoustic near field with 94 dB sound pressure near the wings and a steep pressure gradient that declines within a few millimetres. Moreover, peak to peak particle velocities are large (ca. 1 m s1), and the sound spectrum varies at different locations around the dancer’s abdomen, potentially allowing a follower bee to detect the location behind the dancer as it moves from the side of the dancer towards her rear end (which is the course a follower normally takes). All of these cues could be picked up by an appropriate vibration sensor on the antenna. Indeed, honeybees can be trained to associate a sound of a given frequency with an electric shock, and to avoid the sound thereafter in a frequency-specific manner (Towne and Kirchner, 1989). The conditioned sound is also avoided during flight behaviour, even if the conditioned frequency (265 Hz) is close to the bee’s own wing beat frequency. When placed in a standing-wave tube, bees perform characteristic antennal lifting movements at locations where the air particle velocity is high (1 m s1, at 18 Pa rms sound pressure), but not if air pressure oscillations are strong and air particle velocity is low (0.07 m s1, 60 Pa rms sound pressure). This indicates that bees perceive air particle movements rather than sound pressure. Unilateral antennectomy reduces the likelihood that a bee gets recruited to a new food source, and increases the time needed to find the feeder (Dreller and Kirchner, 1993b). In contrast, clipping the antennal tips or shaving all hairs on the head is not detrimental to recruitment success. Because controls showed that bees with only one antenna can fly sufficiently well, attend waggle displays and are motivated to leave the hive for foraging, Dreller and Kirchner (1993b) conclude that tactile cues are not necessary for information transmission in waggle displays, but a ‘binaural’, or rather bi-antennal mechanism must be postulated. The sensory organ involved was shown to be located in the pedicel (Dreller and Kirchner, 1993a). In an operant conditioning paradigm developed by Kirchner et al. (1991), honeybees can learn to associate a sound with sucrose reward in a two-alternative paradigm (Dreller and Kirchner, 1993a). Subsequent testing of bees with a variety of ablations reveals that a single antenna is sufficient to maintain discrimination, albeit with reduced performance. Removal of both antennae further reduces the performance to chance level. Immobilisation of the SP-joint and the PF-junction also causes significant impairments, but fixing only the SP-joint has no effect. Also, shaving the hairs off the bee’s head is ineffective. In conclusion, the bee’s flagellum is well suited to transmit the sound frequencies emitted by the dancer (see Section 2.2.2.2) to the PF-junction, where Johnston’s organ can transduce the information into a neural activity. There is little or no contralateral transfer of auditory information from the antennae, since unilateral antennectomy impairs choice performance only if the sound is


165

presented on the manipulated side, but not on the intact side (Dreller and Kirchner, 1993a). Within the honeybee genus Apis, not only the Western species A. mellifera can be operantly conditioned to sound stimuli, but also the Asian species A. dorsata and A. florea (Dreller and Kirchner, 1994). Because all three species recruit other foragers by a waggle dance, but A. florea does not emit any sound while dancing, it is likely that the ability to detect sound preceded the use of sound production in the evolution of the waggle dance. According to this scenario, the sensory ability to detect near-field sound using antennal mechanoreceptors allowed the introduction of an additional cue into intraspecific communication. Apart from receiving the information of the waggle dancer, antennae appear to play a role in attracting bees to follow the waggle dance, i.e. in initiating the behaviour that will lead to information transmission about food sources. Tautz and Rohrseitz (1998) traced back the paths of follower bees to determine when and where surrounding bees became interested in the waggle dance. Their results show that the fraction of bees that become follower bees only after they make antennal contact with the dancer or another follower bee, is dependent on the lighting conditions and on the properties of the dance floor (e.g. open vs. sealed cells). The worse the transmission of vibration of the floor, and the darker the ‘room’, the more bees need physical contact with others before following the waggle dance. A closer look at the antennal contact pattern of the follower bees with the dancer reveals that follower bees make regular brief contact with the body of the dancer, during which the flagellum is bent laterally (Rohrseitz and Tautz, 1999). The duration of coincident antennal contact of both antennae decreases as the follower bee gradually moves from a position beside the dancer to one behind the dancer. Given that a follower bee knows its own orientation relative to gravity, the correlation between bilateral contact coincidence and orientation relative to the dancer may allow the follower to either ‘calculate’ the orientation of the dancer with respect to gravity, or help the follower to align herself with the dancer, i.e. find the location where it is easiest for her to determine the orientation of the waggle dance. Antennation is frequently observed in other hymenopterans too, and often has been particularly emphasised in reports on social behaviour in ants (e.g. see reviews by Lenoir and Jaisson, 1982; Ho¨lldobler, 1999). For example, in a study on liquid food exchange in the ponerine ant Pachycondyla villosa, Ho¨lldobler (1985) describes how antennation precedes food transmission between the forager and the solicitor. Food transmission in Pachycondyla is entirely external, i.e. food is not carried in the crop and then regurgitated. As regurgitation of food from the crop occurs in species that are considered to be more ‘advanced’, this behaviour may represent an evolutionarily intermediate stage between social food transmission with and without regurgitation. Because other ponerine ant species that do not transmit food, use antennation cues to ‘invite’ nest mates to follow them in a tandem run back to the nest,

166


Ho¨lldobler (1985) supposes that antennation, being a widely used behavioural component in ant communication, became a ritualised signal for food solicitation, allowing ant colonies to form an external ‘social bucket’ at first, and later even a distributed internal ‘social stomach’. Similar evolutionary scenarios are discussed by Lenoir and Jaisson (1982). Solicitation behaviours in different ant species may also serve as a tool to determine the relative significance and benefit of the two sensory modalities that are involved: mechano- and chemoreception. For example, the ant species Bothroponera tesserinoda and Camponotus sericeus both recruit conspecifics by tandem runs, a behaviour during which both participating ants expect regular physical contact with the other. Whereas both of these species use contactchemosensory cues to identify a leader as ‘being worth-while following’, B. tesserinoda can be convinced to follow an unscented dummy, provided she has searched for a lost leader for long enough to raise her motivation (Maschwitz et al., 1974). C. sericeus on the other hand, does not follow unscented dummies (Ho¨lldobler et al., 1974). Thus, C. sericeus is more restrained and more specific in its requirements to complete successful recruitment. If a leader is lost, only an appropriately scented substitute will do. B. tesserinoda appears to weigh the importance of mechano- and chemosensory cues, depending on its behavioural state. Whether or not such differences are linked to the likelihood of an individual of a given species to lose its leader before reaching the nest remains to be shown. To our knowledge, antennation between ant species or between ants and myrmecophile insects is to date the only use of antennae in interspecific communication. For example, the staphylinid beetle Atemeles pubicollis imitates antennation signals used by colonies of the Formicidae for food solicitation, allowing the beetle to parasitise the intraspecific food flow (Ho¨lldobler, 1970). 6.2.4.3 Courtship and mating. In a study on the behavioural sequence in the mating cycle of the cricket Teleogryllus commodus, Loher and Rence (1978) repeatedly emphasise the significance of antennal movements at various times of the mating cycle. Pair formation begins with the male calling-song that attracts the female. Upon the first encounter, the male seeks antennal contact with the approaching cricket. In a male–female encounter, the pair engages either in agitated antennal contact for several seconds, or the male antennates the female’s elytra, abdomen and hind legs, depending on whether the female is approached from the front or from the rear, respectively. Following this antennation, the male withdraws and vibrates its antennae. Blinded and deafened crickets begin courtship upon contact with a female antenna, indicating that mutual antennal contact is important for sex recognition. Antennectomised females respond appropriately to the male calling sound but are ‘‘reluctant to mount the male and experienced difficulties in assuming the right position on the back of the male’’ (Loher and Rence, 1978, p. 231). During courtship, female antennation of the male elytra and head releases a range of courtship-specific


167

behaviours in the male, such as searching movements of the cerci. Once the female has mounted the male, the male maintains antennal contact by reaching back and flagellating her antennae, keeping her in place until copula. After copulation, the male stays close to the female for some 50–60 min and lays his antennae on her back, as if to guard her. If, during this ‘guarding period’, antennal contact is lost due to restlessness of the female, the male reacts with antennal vibration and flagellation, apparently trying to calm her down. Guarding is important, because unguarded females get rid of the spermatophore within a few minutes. Loher and Rence (1978) claim that lack of antennation is ‘‘the single most disruptive influence on spermatophore formation and mating frequency’’ (p. 240). The specific role of antennal tactile cues in parts of the cricket mating cycle, particularly in sex recognition, has been studied to some detail. Whereas Huber (1955) suggested that male Gryllus campestris use tactile cues about the strength of antennal contacts to tell the opponent’s sex, species of the genus Teleogryllus appear to use contact-chemosensory cues. In T. commodus, there are no structural differences between male and female flagella, nevertheless male crickets respond to contact with the cut-off antenna of a conspecific with sex-specific behaviours (courtship song vs. aggressive song; Rence and Loher, 1977). This differential response is abolished if the cuticular wax layer on the antenna is removed with chloroform. Also, olfactory cues can be ruled out, because neither behavioural choice of appropriately scented containers, nor electroantennograms reveal any sex-specificity. In the closely related Teleogryllus oceanicus, selective blockage of antennal chemoreceptors by zinc sulphate abolishes initiation of courtship in males, without impairing tactile avoidance responses (Balakrishnan and Pollack, 1997). Males of G. bimaculatus frequently respond to encounters with dead females with courtship song, but not if cuticular hydrocarbons are removed by a solvent. However, re-painting the dead female’s flagellum with the hydrocarbon extract is sufficient to re-establish courtship response in some males (Tregenza and Wedell, 1997). The latter study also shows that male and female hydrocarbon extracts differ strongly in concentrations of individual components, but not in the presence of a single pheromone. Note, however, that the extracts were not obtained from antennae only, and that pronounced sex differences in hydrocarbon composition could be restricted to particular parts of the cuticle, for instance to the flagellum. Apart from sex recognition, a second component of the mating cycle in T. oceanicus, the female mounting the male, is significantly impaired by antennectomy of the females (Balakrishnan and Pollack, 1997). Because immobilisation of the female’s pedicel has no effect on mounting frequency, it is not crucial for the female to sense the wing vibration of the courting male with her Johnston’s organ, but other cues seem to aid mounting. Possibly, these are cues related to those involved in rearing or climbing obstacles (see Section 6.2.2).

168


Similar to crickets, antennal contact of the male cockroach Blatella with a female is necessary for sex recognition, and a non-volatile compound produced by the female’s integument is the decisive cue (Roth and Willis, 1952; cf. Cornwell, 1968, p. 179ff). As soon as the contacted cockroach is recognised as a female, the couple repeatedly stroke each other’s antenna, whereby a sexually stimulating substance is transferred to the male antenna. The male then twitches his abdomen and turns it towards the female, lifting his fore and hind wings, thereby exposing a gland on the dorsal side of its eighth abdominal segment. ‘Sparring’ the antennae between the couple, and active movement by the female, are both described as important factors in getting the female to take the secretion of the dorsal gland. Sparring is not essential though, as males will also court antennectomised females. Antennectomy of the male strongly reduces his response. Antennal movements also play a conspicuous role in various components of the mating cycle in cockroaches of the genus Blaberus (Wendelken and Barth, 1987). A male B. discoidalis approaches the female upon reception of a volatile sex pheromone, waves his antennae and, upon contact with her, strokes her dorsal surface with the antennae. The female indicates her receptivity by turning towards the male, holding her antennae in front of her, sometimes followed by a brief period of antennal fencing. Following female antennation, the male almost always begins his courtship behaviour by wing pumping. The direction of the subsequent 1801 turn of the male appears to be away from the side that receives more tactile contact from the female. B. giganteus males maintain antennal contact with the female as they perform a wing pumping display with clockwise circling movements. The females of B. giganteus can assume an unreceptive posture, in which the antennae are held posteriorly, below the body and between the legs. Thus, females actively avoid antennal contact with the male. If receptive, she antennates the male. Unreceptive females of B. craniifer also hide their antennae, but beneath the tegmina. B. craniifer males typically maintain antennal contact during wing pumping behaviour, and even during intermittent turns. In conclusion, antennal waving and antennation of the male are fixed components in the courtship behaviour of all Blaberus species observed by Wendelken and Barth (1987). So is female antennation to indicate her receptivity: a prerequisite for the behavioural sequence that leads to copulation. Very likely, the described behaviours in Blaberus include both contact-chemosensory and tactile cues.

7

Biomimetics and ‘antennal engineering’

In recent years, more and more experiments have been performed to steer insects by implanting electrodes into their nervous systems, adjacent structures or muscles, or to build robots that, in one way or another, mimic insect function (e.g. Espenschied et al., 1993; Frik and Amendt, 1995; Pfeiffer et al., 1995;


169

Ritzmann et al., 2000). Naturally, such attempts have stirred public interest beyond the scientific community (Suplee, 1997; Yam, 1998; Trachtman, 2000). The following sections will review some attempts to use insect antennae to steer robots, or to remote control an insect via its antennal system (Section 7.1), as well as various engineering issues concerning the design and fabrication of artificial tactile antennae (Section 7.2). 7.1

STEERING INSECTS AND ROBOTS

In two cases, attempts to steer a cockroach by electrical stimulation of antennal afferents and the connecting descending pathways have been successful. In the first study, electrical stimulation of 10–100 mA via electrodes, inserted into antennal stumps, was used to steer cockroaches (Periplaneta), which were held stationary on top of an air-supported ball (Holzer and Shimoyama, 1997). The animals responded by turning away from the stimulated side, but the variance of the response angles was rather large. In the second study, electrodes were inserted at the base of the antennae of Gromphadorhina portentosa (Moore et al., 1998), by penetrating the soft cuticle of the SP-joint. The indifferent electrode was located in the dorsal abdomen. The effects of the electrical stimulation mimic escape responses described for Periplaneta americana (e.g. Comer et al., 1994; Stierle et al., 1994; see Section 6.1.3.3), but it is unknown whether the underlying neural escape circuit (Burdohan and Comer, 1990; Ye and Comer, 1996) was actually recruited. The rather large variations of the behavioural responses suggest that the electrical stimulation recruited a multitude of circuits. Passive tactile sensors have been used on mobile robots, including legged robots, for detecting contact with walls or with large obstacles, but typically they send only an unspecific binary stop signal to the controller (e.g. hexapod robot ‘Ghengis’; Brooks, 1989). A mobile robot that uses tactile antennae for steering in a more sophisticated way is the small hexapod robot ‘Sprawlette’, which is equipped with a flexible probe with several strain gauges (Cowan et al., 2004). The sensor reading is used to estimate the lateral distance to the contacted wall, and simple proportional–differential (PD) control of the distance is sufficient to achieve reasonably stable course control. Intuitively, the PD control loosely matches the idea of keeping stable both the lateral distance to the wall and the angle between wall and heading direction. As a mismatch in the angle relative to the wall soon leads to a change in lateral distance to the wall, controlling both lateral distance and its derivative stabilises a variable that Cowan et al. (2004) call ‘‘tactile flow’’. This was inspired by experimental results on the cockroach P. americana (Johnson and Camhi, 1999; see Fig. 17 and relating text). Instead of building an artificial antenna, Kuwana et al. (1999) use the antennae of the silk moth Bombyx mori as a chemosensory device to steer a small wheel-driven robot. Pheromone-induced electrical activity in the antennal

170


nerve can be picked up by electrodes, amplified, drift-cancelled, band-pass filtered and compared to an offset voltage. After a period of about 25 min, the baseline drift has decreased sufficiently, and the sensor produces reliable signalto-noise ratios for a period of half an hour. During this test interval, the robot can track a realistic odour plume to proceed towards the pheromone source, although its track is considerably different to that of a moth.

7.2

ENGINEERING OF ACTIVE TACTILE SENSORS

Similar to the approach of Kuwana et al. (1999), insect antennae have been coupled repeatedly to stationary electronic devices, e.g. via field-effect transistors, to detect volatile chemicals at very low concentrations (e.g. Scho¨ning et al., 1998, 2000; Schu¨tz et al., 1999, 2000; Schroth et al., 2001; Park et al., 2002; Park and Baker, 2002). In principle, similar arrangements could be employed to record mechanoreceptor activity for tactile contact sensing, although the mechanical interaction with the contact surface would probably make a stable recording even more difficult than it is already for chemical sensors. To our knowledge, no attempts have been made to use an insect antenna as a natural contact sensor. The mechatronics of tactile sensors is a strongly growing field of engineering. Research areas concerning transduction methods, incorporation of sensors into the contact material, integration of information about tactile contacts with information about posture and movement (haptics), and data processing schemes have witnessed considerable improvement over the past decade (see Lee and Nicholls, 1999, for review). Artificial fabrication of an antenna-like tactile sensor still poses many problems, however, some of which concern the mechanics of the sensor-carrying probe and others, which concern the actual sensing process. In case of the probe, problems mainly concern conflicting requirements. For instance, an ideal probe would be reasonably stiff to simplify determination of the contact point, but also to avoid strong movement-induced vibrations. On the other hand, the probe would need to be sufficiently light to allow for fast active sampling strategies, and fairly compliant for passive bending upon object contact. Moreover, it has to be durable. In case sensors are to be mounted to the probe, mechanical robustness is required, but also very basic issues of transduction and coding are still far from the so-called standard solutions. Lee and Nicholls (1999) highlight the three main engineering difficulties of tactile sensing as being (1) the lack of a localised sensory organ, i.e. transduction involves signals from a much wider area than the sensor area alone, (2) the signal complexity, i.e. the mixed and differing impact of shape, texture, friction, force, etc. on the transduction process and (3) the ‘‘difficulty to imitate’’ the decisive sensory quality, i.e. to define the physical properties that are most appropriate to be quantified.


171

The importance of the right choice of probe is emphasised by studies on bending and resonance properties. Recently, Barnes et al. (2001) showed that the tapered shape of a mechanical probe allows distinction of drag-induced bending, caused by flow of the surrounding medium, from contact-induced bending, caused by a touched obstacle. In a biological context this is interesting, as the antennae of crayfish and lobsters, but also those of locusts, are tapered (see Section 2.3.2). In a series of papers, Ueno and co-workers analysed the possibility of using contact-induced oscillations of an insensitive beam to determine the contact distance (Ueno and Kaneko, 1995; Ueno et al., 1996, 1998). Their experiments show that the fundamental and second-order natural frequencies of a beam are sufficient to give a good estimate of the contact point. The major difficulty in this application, however, is to measure the frequencies, because repeated object contacts preclude sufficiently long oscillations to apply standard routines such as a Fourier transformation. Ueno et al. (1996) also demonstrate that an appropriate mass distribution along the beam is a prerequisite for a unique mapping of oscillation frequencies to contact point. In a biological context, these results imply that the afferent activity of Johnston’s organ, a vibration sensor common to all higher insects, could in principle contain information about the contact point during tactile sensing. Exploiting force and position measurements at the base of an insensitive compliant probe, Tsujimura and Yabuta (1992) were able to reconstruct the shape of a touched object. Additionally to such measurements, Kaneko et al. (1995, 1998) detect lateral slip on the contact surface and use an active movement strategy to determine the touch direction that avoids lateral slip. Provided that the contacted object is sufficiently static, an insensitive probe can, thus, determine the contour of a touched object. Nevertheless, equipping a probe with a set of mechanosensors would certainly boost the applicability of an artificial tactile antenna. Recently, microelectric mechanical systems (MEMS) of various types have been fabricated and mounted to a biomimetic lobster antenna (McGruer et al., 2002), the aim being to simultaneously measure water flow, bending of the probe and contact point with an external object. The resulting MEMS antenna is part of an underwater ambulatory robot that is inspired by an eight-legged lobster (Ayers, 2002). The antennae can be positioned at one out of four postures, but active movement is not used in the sensing process. In the control of legged locomotion, Du¨rr and Krause (2001) argue that it can be particularly efficient to use a leg-like sensor to guide a leg, much as insects use a pair of highly specialised sensor legs – their antennae – to keep locomotion efficient, even on rough terrain. A great potential in active tactile sensing certainly lies in the context-dependent choice of sampling strategy and its adaptive control. It will pay off to analyse further the tactile sampling strategies of insect antennae, that appear well-adapted to tactile detection of obstacles (see Sections 6.2.1 and 6.2.2). For adaptive control, Du¨rr et al. (2003) have advocated the use of forward models for the control of a tactile sensor. In

172


a simulation study, they used a parallel set of extended self-organising maps to track an external target with an antenna, a behaviour known from insects and Crustacea (see Section 6.2.1). Unless it is performed under closed-loop control, tracking requires the solution of the inverse kinematics problem. Whereas technical systems typically calculate the solution to this problem, animals cannot do so, because they do not know how to mathematically invert the analytical description of their forward kinematics. Yet, animals can learn their forward kinematics by comparing the motor signal to the corresponding sensory information. Moreover, the obtained forward model of a limb has to be adaptive, because the forward kinematics may change during growth or due to injury. Du¨rr and Krause (2002) implement a forward model by means of extended self-organising maps, i.e. extended Kohonen maps. The adaptive forward model is then applied to iteratively solve the inverse kinematics problem during object tracking. Both in technical and biological systems, forward models are a very powerful concept for prediction of the system’s state, including its proprioreceptive information. In an active sensor, forward models are an appropriate means for comparing the expected input to the real input, thus permitting the separation of active and passive movement. Biologists have repeatedly claimed the necessity of a special case of forward model, the efference copy or corollary discharge (Holst and Mittelstaedt, 1950; Sperry, 1950), where a scaled copy of the motor command is used to cancel the sensory input that is induced by active movement. In the nervous system, many variants of such forward models are likely to be implemented. In technical systems, they are likely to be indispensable for the separation of actively and passively induced tactile inputs.

8

Conclusions

This review focuses on the biology of the antennal tactile sense of insects and crustaceans. It covers the biomechanics, kinematics and behavioural biology of antennal movements, the peripheral and central neurobiology of antennal mechanoreception and interdisciplinary aspects of biology and engineering of tactile sensing. The discussion of these topics is centred around a set of insect model organisms, all of which employ their antennae in active tactile sensing. Work on other species, particularly on decapod crustaceans, is used to complete and contrast the insights from the insect model organisms. In conclusion, we will discuss three aspects of the antennal tactile sense that draw on information from various sections and, therefore, require separate treatment. Firstly, we will address the behavioural similarities between insect and crustacean antennal tactile sensing (Section 8.1), which are particularly remarkable when bearing in mind the morphological differences between these taxa. Secondly, we will contrast in hindsight the categories of passive and active sensing, as used in Section 6, by relating them to a behaviour-based


173

classification of tactile sensing (Section 8.2). Finally, we will discuss this classification in an evolutionary context (Section 8.3). 8.1

THE ANTENNAL TACTILE SENSE OF INSECTS AND CRUSTACEA

The antennae of insects are sensory organs that carry sensilla of at least three modalities: mechano-, chemo- and thermoreception. Accordingly, different insect taxa have evolved specialised antennae that are more suitable for one of these modalities than for another. For example, the pectinate antennae of some moths (e.g. Manduca sexta), are well-adapted for olfaction, but are less suitable for tactile sensing. This is because an increase of surface area also increases the delicateness of the structure, which can be detrimental where physical contacts are necessary for sensing. Generally, highly specialised antennae, i.e. those that strongly deviate from a linear morphology, are abundant in holometabolous insect orders (particularly in Coleoptera, Lepidoptera and Diptera; e.g. the Hymenoptera being an exception) but less common in hemimetabolous insect orders (e.g. Orthoptera, Phasmatodea, Blattodea; e.g. the Odonata and Hemiptera being exceptions). Crickets, stick insects and cockroaches have long, straight flagella and move their long antennae constantly during walking, thus raising the likelihood of tactile contacts. In species of these orders, objects are rarely touched with the antennal tip. Rather, contacts typically occur along the side of the flagellum. Sensory hairs along the cricket and cockroach antennae, are arranged in rings around the distal ends of the annuli, and their number is highest at the tip of their antenna. Honeybees, on the other hand, use their antennae in a very different way (e.g. Section 6.2.3). The geniculate morphology of the Hymenopteran antenna can be considered as a kinematic specialisation to targeted probing with the antennal tip, requiring a highly dexterous antenna. In the honeybee, the highest number of sensilla is also found at the tip of the antenna, but a peculiarity to the honeybee antenna is a specialised arrangement of sensory hairs, the sensory plate, on the terminal annulus (cf. Section 3.2.4.1). The differences in kinematics, sensing strategies and the arrangement of receptors on the antenna are paralleled by neuroanatomical differences between the model systems of cricket and cockroach on the one side, and of the honeybee on the other. In crickets and cockroaches, antennal afferents project to three distinct neuropils, the AL, the DL and the VFA (Section 4.1.1). In the highly structured VFA, afferents connect to interneurons, some of which show conspicuously regular arborisation patterns. This indicates a topological organisation of this neuropil (Section 4.2.3). In comparison, although honeybees do have ALs and DLs, they appear to lack the VFA. Quite possibly, the mentioned differences in movement strategy and receptor arrangement on the flagellum imply differences in neural processing of antennal tactile information, reflected by a different layout of the deutocerebral neuroanatomy.

174


From a behavioural perspective, decapod Crustacea appear to use their antennae in similar ways as insects with long linear antennae. At the level of antennal reflexes, there seems to be little difference between the palinurid rock lobsters on the one hand, and locusts on the other (Section 6.1.2). Similarly, speed-dependent abduction, antennal-tracking behaviour and the use of antennae in agonistic encounters, bear many similarities in crustacea and insects. Yet, the underlying physiological mechanisms may be different. These differences may have their basis in differences in antennal form, size, ‘construction’ of the joints and the viscosity of the surrounding medium, but also in the involvement of different brain regions. Differences in antennal movement strategies must be related to the kinematic peculiarity of the crustacean antenna, where two redundant joints (MC- and CF-joint) move the flagellum within a plane that is rotated by the complex BIM-joint. In comparison, the basic construction of an insect antenna is that of a Cardan joint (Section 2.2). The presence of redundant joints in crustaceans is mirrored in the existence of common excitatory motoneurons, whereas the antennal musculature in insects is innervated by ‘dedicated’ excitatory and common inhibitory motoneurons (Section 2.3). Also, the crayfish antenna has a bi-articular chordotonal organ, whereas functionally related proprioreceptors in insects only span a single joint. Similarly, environmental and biomechanical aspects probably had a bearing on the differences between some insect and crustacean mechanosensory structures and, possibly, their transduction mechanisms. For example, transduction may be affected by differences in cuticular characteristics, but also by the external medium. Finally, unlike insects, crayfish and lobsters have two pairs of specialised antennae, the first pair of which are used primarily for olfaction, while the second antennae serve as tactile probes. Afferents from first and second antennae project to separate neuropils in the deuto- and tritocerebrum, respectively. The degree of structure of the tritocerebral antenna II neuropil correlates with the length of the flagellum of the different species. The length of the flagellum can be considered indicative for the tactile usage of the antenna. Thus, it is intriguing to hypothesise that the necessity for topological coding of contact location along the flagellum is related to the finding of a structured neuropil in crustaceans and insects with long tactile antennae (cf. Section 4.1.2). Certainly, the behavioural significance of the antennal tactile sense in insects and Crustacea suggests that these taxa have indeed evolved serially homologous structures into an analogous set of active tactile sensors. 8.2

ACTIVE AND PASSIVE MECHANICAL SENSING

In the present review, behavioural functions have been loosely grouped into passive and active sensing. This distinction may appear somewhat artificial,


175

particularly as certain behavioural contexts presented in Section 6 may involve both passive and active aspects. The reason for this choice of classification is the attempt to arrange behavioural contexts into a sequence of increasing complexity. For example, assistance reflexes (Section 6.1.2) are treated within the passive sensing context, although they occur during voluntary, hence ‘active’ behaviour. Similarly, tactile wall-following in cockroaches is classified as passive sensing (Section 6.1.3.2), although cockroaches must be expected to actively control their antennal posture. Accordingly, the passive sensing section (Section 6.1) comprises behavioural contexts in which any active component solely concerns postural reflexes. Active components, such as posture stabilisation in tactile wall-following or during flight speed control, may be necessary to keep passive sensing simple, e.g. by keeping the SP-joint angle constant or the average torque acting on the PFjunction in a linear range. Thus, local postural reflexes may be prerequisites for simple control schemes, which, in turn, underlie orientation behaviours. Exploration movements are the most basic form of active sensing, in that they serve to expand the workspace of an otherwise passive sensor. Similarly, tracking of visual targets adjusts the workspace to a region with likely sensory input. All further behavioural contexts discussed in Section 6 are variations of this scheme: antennal movements act either to raise the likelihood of gathering potentially important information (as during locomotion), or of seeking physical contact at a given location (as in pattern recognition) and possibly in a given sequence (as in communication). Honeybee hearing does not fit this scheme, because it does not involve active antennal movement. Nevertheless, the small sensing range of honeybee hearing, e.g. a few millimetres in waggle dance communication, requires active movements of the follower bee and, therefore, active positioning of the antennal sensory structures. Moreover, the vibration cues heard by a honeybee bear much similarity to vibration cues associated with other examples of insect communication, such as antennal fencing in crickets and, possibly, temporal sequences of antennation in ant communication systems. Therefore, honeybee hearing may be considered as active sensing – although it is a special case. Active movement contributes to the efficiency of tactile sensing by increasing the likelihood of the antennal contacts with an external object. On the other hand, active movement of a movement-sensitive sensor requires the ability to discriminate self-induced stimulation from external stimulation. Hence, active sensing produces ambiguities with passive sensing. The use of forward models can effectively resolve this ambiguity as discussed in Section 7.2 from an engineering point of view. The motor-dependent suppression of antennal mechanosensory inputs to a cricket descending interneuron probably represents an example where a feed-forward motor signal modifies a sensory signal (Section 5.2). Thus, this may be indicative for the biological implementation of a forward model, for instance an ‘efference copy’. Moreover, biomechanical

176


properties may also help to discriminate active from passive sensory cues (Section 2.2.2.1). One could even hypothesise that, depending on the specific quality of the antenna–object contact differential sets of antennal mechanoreceptors are activated by self-induced and external contacts, thus enabling an active organism to discriminate between these sources of stimulation. Another aspect of active sensing, e.g. tactile scanning of external objects or tracking of a visual stimulus is dependent on the spatial and temporal alignment of the sensor and its target. Firstly, spatial information from different sensors, e.g. the compound eye and the antenna, must be brought into congruence, either by transforming one of the coordinate frames into the other, or by mapping afferent inputs of both sensors to a common frame of reference. As a consequence on the neuronal level, interneurons must exist, which integrate the information from different sensory modalities to generate appropriate motor commands to adjust the position of the antenna. Accordingly, interneurons have been found which are the likely components of a multimodal integration stage (Sections 4.2.2 and 5.1). Assessing their behavioural significance, e.g. for a co-ordinate transformation between the eye and the antennal base, however, requires profound knowledge of their sensory physiology. This knowledge, however, is currently not available. Secondly, spatial acuity of the tactile sense depends crucially on the acuity of the elements of the underlying neuronal circuitries. Passive and active sensing may be very similar in terms of acuity, assuming that acuity depends on sensor spacing and its regularity alone. This, however, and how sensor spacing is represented in the brain remains so far unknown. The data that are available for antennal sensory neurons (Sections 3.2 and 3.3) provide a reference that is necessary to judge the acuity of the coding of stationary and dynamic parameters of antennal posture and movement. Moreover, local interneurons in the deutocerebrum appear to process the primary sensory information to extract specific features of movement parameters (Section 5.1). From a comparative point of view, the great extent to which passive and active sensing behaviours resemble each other in insects and crustaceans, emphasises the similarities rather than differences between both arthropod taxa. Except in the context of pattern recognition (Section 6.2.3), insect sensing movements find their counterparts in crustaceans, hinting at the evolutionary benefit that a sensor gains from adopting active properties. 8.3

LEVELS OF BEHAVIOURAL COMPLEXITY

Following the common assumption that antennae have evolved from serially homologous locomotor appendages, it follows that they have probably never passed an immobile stage in their evolution. Rather, they are likely to represent a body appendage that has always been mobile and has always been equipped with the sensory infrastructure that is common to arthropod legs. Therefore, it


177

may be hypothesised that antennae have always been actively moveable, complex sensors, and that their sensory potential initially subserved the behavioural function of posture stabilisation and propulsion. Consequently, antennal searching behaviour during locomotion, or antennal tracking (Sections 6.2.1 and 6.2.2) may be fairly close to the origins of antennal motor programs, i.e. rhythmic stepping movements, whereas apparently ‘simple’ behaviours like the ‘static’ antennal flight posture, may well represent a rather derived condition. On the basis of the distinction of passive and active sensing, however, behavioural contexts may be arranged in a sequence of increasing complexity of motor action and of neural processing of sensory information. Basically, orientation behaviours like geotaxis (Section 6.1.1) and anemotaxis (Section 6.1.3.2) are crude forms of course control, as it occurs in the control of flight speed (Section 6.1.3.1) and wall-following (Section 6.1.3.2). Moreover, the ability to sense small deflections is the prerequisite for appropriate ‘turn-away’ reactions, as is the case in antenna-mediated escape in cockroaches and the defensive stilt posture and corresponding kick reaction of the cricket (Section 6.1.3.3). It is reasonable to assume that the sensory abilities required for these behaviours were already present in an articulated leg. Furthermore, assuming the gradual conversion of a front leg into a dedicated sensor, it is plausible that antennae served as sensors to signal the presence of an obstacle within the action range of the legs, thus aiding efficient locomotion on unpredictable terrain (Section 6.2.2). The involvement of antennae in climbing of obstacles and gaps suggests that lack of antennal contact signals ‘space to move’ to the legs, allowing the animal to stick with the current locomotor behaviour, whereas antennal contact elicits a fast change in locomotor strategy, e.g. by recruiting the thoracic joints to incline the body axis. The presence of dedicated, large calibre descending antennal mechanosensory interneurons that project to the thoracic ganglia, possibly mediating fast reactions to detected obstacles ahead of the animal would support this behavioural consideration (Sections 4.2.3 and 5.2). Pattern recognition and communication are yet more complex behaviours that require the coding of spatial and/or of temporal information. In contrast to mere spatial coding of the location of a single contact point, pattern recognition requires sequential antennal contacts, and, therefore, the ability to either map each contact information into a common 3D representation, or to code a four-dimensional pattern by including time as a variable. Only in cases where the recognised microtextures are of the spatial range as the mechanoreceptor spacing (e.g. Kevan and Lane, 1985), may sequential contacts be unnecessary. The ability of honeybees to learn movement patterns (e.g. Erber et al., 1997, 2000), which is reflected in the activity of a single motoneuron; and the similarity of tactile learning with movement learning (see Fig. 21), hints at the possibility that tactile pattern recognition in insects may in part be associated with the serial coding of a movement pattern and, thus,

178


temporal coding of a sequence, in addition to spatial coding of location. So far, the literature suggests that both spatial and temporal information can be learned. As a consequence, further studies will have to reveal to what extent these two aspects are linked. Finally, antenna-mediated communication in insects requires the ability to control a ‘voluntary’ action sequence by the emitter, as well as the ability to decode the transmitted temporal information and/or the stimulus intensity. Most reported forms of communication that involve the antennae rely either on vibration cues (e.g. antennal fencing in crickets, waggle dance in honeybees) or repetitive antennation (e.g. the tournaments of Myrmecocystus). A special case may be courtship behaviour, as contact-chemoreception appears to be more important than in other forms of communication. In conclusion, the utilisation of active properties appears to be a major theme in the design of tactile antennal behaviours. This may be an ‘a priori’ condition of arthropod antennae, which is expressed in both insects and crustaceans. The various organisational levels of antennal movements range from simple reflexes to complex sequences. Some of the aspects of the biology of arthropod antennae relate to core interests of behavioural biology, neuroethology and neurophysiology. Accordingly, we are confident that continued comparative research on the antennal tactile sense and its underlying neural substrates will provide insight into the evolution and organisation of sensorimotor aspects of arthropod behaviour in general.

Acronyms and abbreviations

ACH ACHE AL AMMC BE-joint BIM-joint CB-joint CF-joint CI CNS DBN DC DL

acetylcholine acetylcholine esterase antennal lobe of the insect deutocerebrum antennal mechanosensory and motor centre basipodite–exopodite joint basipodite–ischiopodite–meropodite (3rd) joint complex of the crustacean 2nd antenna coxopodite–basipodite (2nd) joint of the crustacean 2nd antenna carpopodite–flagellum (5th) joint of the crustacean 2nd antenna common inhibitor neuron central nervous system descending brain neuron deutocerebrum dorsal lobe of the insect deutocerebrum


DUM EJP EMG GABA GI HC-joint HS-joint IJP IM-joint IR LAN MAN MC-joint MCF-CO MEMS N1 NO OSC PF-junction sCO SOG SP-joint T1-7 TC TCG VFA VUM WD

179

dorsal unpaired median neuron excitatory junction potential Electromyogram g-amino butyric acid giant interneurons and corresponding ascending neural pathway of the cockroach escape system ‘head’–coxopodite (1st) joint of the crustacean 2nd antenna head–scape (or 1st) joint of the insect antenna inhibitory junction potential ischiopodite–meropodite joint immunoreactive lateral antennular neuropil medial antennular neuropil meropodite–carpopodite (4th) joint of the crustacean 2nd antenna biarticular chordotonal organ of the crustacean antenna, spanning the MC- and CF-joint. micro-electric-mechanical systems main antennal nerve, Nervus antennalis, that branches inside the antenna into a lateral N1l and a medial N1m. nitric oxide oscilloscope not actively moveable joint between the pedicell and the flagellum of the insect antenna scapal chordotonal organ suboesophageal ganglion scape–pedicel (or 2nd) joint of the insect antenna tract number one to seven tritocerebrum tritocerebral commissure giant interneuron ventral area of flagellar afferents, a neuropil of the insect deutocerebrum ventral unpaired median neuron window discriminator

Acknowledgements V.D. is supported by the Deutsche Forschungsgemeinschaft grant DU-380/1. V.D. thanks J. Camhi (Jerusalem), J. Erber (Berlin), H. Hofmann (Cambridge/ MA), D. Sandeman (Wellesley) and J. Zeil (Canberra) for consenting and/or commenting the presentation of their original figures and corresponding legends, J. Tautz and M. Lindauer (Wu¨rzburg) for information related to the translated citation in Section 6.2.4, as well as S. Ku¨hn and H. Cruse (Bielefeld) for

180


constructive feedback on large portions of the manuscript. M.G. thanks P. Braünig (Aachen), for providing access to his literature database and for commenting on Fig. 14. M.G. is also grateful to H. W. Honegger (Nashville), for granting permission to re-use Fig. 3. E.M.S. thanks G. A. Jacobs (Bozeman), and P. Braünig (Aachen) for helpful comments on the manuscript and P. Braünig (Aachen), for some discussions. The authors thank T. Matheson for helpful suggestions on the manuscript.

Note added in proof Du¨rr, V. (2005). Biomechanics of an active tactile sensor: the flagellum of Carausius morosus. Proc. German Zool. Soc., in press. Du¨rr, V. and Ebeling, W. (2005). The behavioural transition from in straight to curve walking: kinetics of leg movement parameters and the initiation of turning. J. Exp. Biol. 208, 2237–2252. Fields, D. M. and Weissburg, M. J. (2004). Rapid firing rates from mechanosensory neurons in copepod antennules. J. Comp. Physiol. A 190, 877–882. Han, Q., Hansson, B. S., and Anton, S. (2005). Interactions of mechanical stimuli and sex pheromone information in antennal lobe neurons of a male moth, Spodoptera littoralis. J. Comp. Physiol. A 191, 521–528. Krause, A.F. and Du¨rr, V. (2005). Proprioreceptive influence of sensory hair fields on antennal movements of Carausius morosus. Proc. German Zool. Soc., in press. Kung, C. (2005). A possible unifying principle for mechanosensation. Nature 436, 647–654. Kwon, H.-W., Lent, D. D., and Strausfeld, N. J. (2004). Spatial learning in the restrained American cockroach Periplaneta americana. J. Exp. Biol. 207, 377–383. Lambin, M., Deglise, P., and Gauthier, M. (2005). Antennal movements as indicators of odor detection by worker honeybees. Apidologie 36, 119–126. Lange, O., Reimann, B., Saenz, J., Du¨rr, V., and Elkmann, N. (2005). Insectoid obstacle detection based on an active tactile approach. Proc. Int. Symp. Adapt. Motion Anim. Mach. 5 (AMAM5), in press. Lewinger, W. A., Harley, C. M., Ritzmann, R. E., Branicky, M. S., and Quinn, R. D. (2005). Insect-like antennal sensing for climbing and tunneling behavior in a biologically-inspired mobile robot. In: Proc. IEEE Int. Conf. Robotics Automation (ICRA’05), Barcelona, April 18–22. Lent, D. D. and Kwon, H.-W. (2004). Antennal movements reveal associative learning in the American cockroach Periplaneta americana. J. Exp. Biol. 207, 369–375. Okada, J. and Toh, Y. (2004). Spatio-temporal patterns of antennal movements in the searching cockroach. J. Exp. Biol. 207, 3693–3706.

References Abzhanov, A. and Kaufman, T. C. (1999). Homeotic genes and the arthropod head: expression patterns of the labial, proboscipedia, and deformed genes in crustaceans and insects. Proc. Nat. Acad. Sci. USA 96, 10224–10229. Allgaüer, C. (1989). Analyse der Antennenbewegungen bei Grillen (Gryllus campestris und Gryllus bimaculatus): motorische Kontrolle, sensorische Bewegungskodierung, integrative Interneurone. Doctoral dissertation, TU Mu¨nchen.


181

Allgaüer, C. and Honegger, H.-W. (1993). The antennal motor system of crickets: modulation of muscle contractions by a common inhibitor, DUM neurons, and proctolin. J. Comp. Physiol. A 173, 485–494. Altner, H. (1977a). Insect sensilla: principles of structure and function. Verh. Dtsch. Zool. Ges. 1977, 139–153. Altner, H. (1977b). Insect sensillum specificity and structure: an approach to a new typology. In: 6th International Symposium on Olfaction and Taste, (eds Magnen, J. and MacLeod, P.), pp. 295–303. Paris, London: Gif-sur-Yvette Information Retrieval Ltd. Altner, H. and Prillinger, L. (1980). Ultrastructure of invertebrate chemo-, thermo-, and hygroreceptors and its functional significance. Int. Rev. Cytol. 67, 69–139. Altner, H., Routil, C. and Loftus, R. (1981). The structure of bimodal chemoreceptive, thermoreceptive, and hygroreceptive sensilla on the antenna of Locusta migratoria. Cell Tiss. Res. 215, 289–308. Altner, H., Sass, H. and Altner, I. (1977). Relationship between structure and function of antennal chemo-, hygro- and thermoreceptive sensilla in Periplaneta americana. Cell Tiss. Res. 176, 389–405. Altner, H., Schaller-Selzer, L., Stetter, H. and Wohlrab, I. (1983). Poreless sensilla with inflexible sockets. A comparative study of a fundamental type of insect sensilla probably comprising thermo- and hygroreceptors. Cell Tiss. Res. 234, 279–307. Ameyaw-Akumfi, C. (1979). Appeasement displays in cambarid crayfish (Decapoda, Astacoidea). Crustaceana Suppl. 5, 135–141. Arbas, E. A. (1986). Control of hindlimb posture by wind-sensitive hairs and antennae during locust flight. J. Comp. Physiol. A 159, 849–857. Arnold, G., Masson, C. and Budharugsa, S. (1985). Comparative study of the antennal lobes and their afferent pathway in the worker bee and the drone (Apis mellifera). Cell Tiss. Res. 242, 593–605. Ayers, J. (2002). A conservative biomimetic control architecture for autonomous underwater robots. In: Neurotechnology for biomimetic robots, (eds Ayers, J., Davis, J. L. and Rudolph, A.), pp., 241–259. Cambridge/MA, London:MIT Press. Baba, Y. (2000). New methods of dye application for staining motor neurons in an insect. J. Neurosci. Meth. 98, 165–169. Baba, Y. N. and Comer, C. M. (2003). Antennal mechanosensory neurons in the cockroach their location and projection sites. Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience. CD-ROM. 2003, 269.12. Bacon, J. (1980). A homologous interneurone in a locust, cricket and a mantid. Verh. Dtsch. Zool. Ges. 85, 300. Bacon, J. and Mo¨hl, B. (1983). The tritocerebral commissure giant (TCG) wind-sensitive interneuron in the locust. I. Its activity in straight flight. J. Comp. Physiol. 150, 439–452. Bacon, J. and Tyrer, M. (1978). The tritocerebral commissure giant (TCG): a bimodal interneurone in the locust Schistocerca gregaria. J. Comp. Physiol. 126, 317–325. Balakrishnan, R. and Pollack, G. S. (1997). The role of antennal sensory cues in females responses to courting males in the cricket Teleogryllus oceanicus. J. Exp. Biol. 200, 511–522. Ball, E. and Young, D. (1974). Structure and development of the auditory system in the prothoracic leg of the cricket Teleogryllus commodus (Walker) II. Postembryonic development. Z. Zellforsch. 147, 313–324. Balss, H. (1940). Decapoda (Zehnfu¨Xer). In: Bronn’s Klassen und Ordnungen des Tierreichs 5(1), 116–123.

182


Barnes, T. G., Truong, T. Q., Adams, G. G. and McGruer, N. E. (2001). Large deflection analysis of a biomimetic lobster antenna due to contact and flow. ASME J. Appl. Mech. 68, 948–951. Barnes, W. J. P. and Neil, D. M. (1982). Reflex antennal movements in the spiny lobster, Palinurus elephas. II. Feedback and motor control. J. Comp. Physiol. 147, 269–280. Barth, F. G. (2002). A Spider’s World: Senses and Behavior. Heidelberg, Berlin: Springer. Bartos, M., Allgaüer, C., Eckert, M. and Honegger, H.-W. (1994). The antennal motor system of crickets: proctolin in slow and fast motoneurons as revealed by double labelling. Eur. J. Neurosci. 6, 825–836. Bartos, M. and Honegger, H.-W. (1997). Impact of motor activity and antennal mechanosensory input on the intensity of proctolin-like immunoreactivity in antennal motoneurons of crickets (Gryllus bimaculatus). J. Comp. Physiol. A 181, 59–70. Bartos, M., Lauer, B. and Honegger, H. W. (1995). Colocalisation of glutamate and proctolin in antennal motoneurons: an immunocytochemical double-labelling study in crickets. Proc. Go¨ttingen. Neurobiol. Conf. 23, 600. Bauer, C. (1991). Modulatory action of proctolin in the locust (Locusta migratoria) antennal motor system. J. Insect Physiol. 37, 519–532. Bauer, C. and Gewecke, M. (1991). Motoneuronal control of antennal muscles in Locusta migratoria. J. Insect Physiol. 37, 551–562. Ba¨ssler, U. (1971). Zur Bedeutung der Antennen fu¨r die Wahrnehmung der Schwerkraftrichtung bei der Stabheuschrecke Carausius morosus. Kybernetik 9, 31–34. Ba¨ssler, U. (1993). The walking- (and searching-) pattern generator of stick insects, a modular system composed of reflex chains and endogenous oscillators. Biol. Cybern. 69, 305–317. Ba¨ssler, U. and Bu¨schges, A. (1998). Pattern generation for stick insect walking movements – multisensory control of a locomotor program. Brain Res. Rev. 27, 65–88. Bell, W. J. (1978). Directional cues in tactile stimuli involved in agonistic encounters in cockroaches. Physiol. Entomol. 3, 1–6. Bell, W. J. and Kramer, E. (1979). Search and anemotactic orientation of cockroaches. J. Insect Physiol. 25, 631–640. Bender, M., Gnatzy, W. and Tautz, J. (1984). The antennal feathered hairs in the crayfish: a non-innervated stimulus-transmitting system. J. Comp. Physiol. A 154, 45–47. Bentley, D. (1977). Control of cricket song patterns by descending interneurons. J. Comp. Physiol. 116, 19–38. Bicker, G. (1992). Taurine in the insect nervous system. Comp. Biochem. Physiol. C 103, 423–428. Bicker, G., Schaefer, S., Ottersen, O. P. and Storm-Mathisen, J. (1988). Glutamate-like immunoreactivity in identified neuronal populations of insect nervous systems. J. Neurosci. 8, 2108–2122. Bicker, G., Schafer, S. and Kingan, T. G. (1985). Mushroom body feedback interneurones in the honeybee show GABA-like immunoreactivity. Brain Res. 360, 394–397. Bill, R. G. and Herrnkind, W. F. (1976). Drag reduction by formation movement in spiny lobsters. Science 193, 1146–1148. Bla¨sing, B. and Cruse, H. (2004). Mechanisms of stick insect locomotion in a gapcrossing paradigm. J. Comp. Physiol. A 190, 173–183. Blochlinger, K., Jan, L. Y. and Jan, Y. N. (1991). Transformation of sensory organ identity by ectopic expression of cut in Drosophila. Genes Dev. 5, 1124–1135. Bloechl, R. and Selzer, R. (1988). Embryogenesis of the connective chordotonal organ in the pedicel of the American cockroach: cell lineage and morphological differentiation. Cell Tiss. Res. 252, 669–678.


183

Bodmer, R., Barbel, S., Sheperd, S., Jack, J. W., Jan, L. Y. and Jan, Y. N. (1987). Transformation of sensory organs by mutations of the cut locus of Drosophila melanogaster. Cell 51, 293–307. Boeckh, J., Ernst, K. D. and Polz, H. (1983). Convergence of inputs from different sense organs upon neurons of the central antennal pathway in insects. Fortschr. Zool. 28, 143–148. Boeckh, J., Ernst, K. D., Sass, H. and Waldow, U. (1984). Anatomical and physiological characteristics of individual neurones in the central antennal pathway of insects. J. Insect Physiol. 30, 15–26. Boeckh, J. and Tolbert, L. P. (1993). Synaptic organization and development of the antennal lobe in insects. Micr. Res. Tech. 24, 260–280. Bornhauser, B. C. and Meyer, E. P. (1997). Histamine-like immunoreactivity in the visual system and brain of an orthopteran and a hymenopteran insect. Cell Tiss. Res. 287, 211–221. Bovbjerg, R. V. (1956). Some factors affecting aggressive behavior in crayfish. Physiol. Zool. 29, 127–136. Boyan, G., Williams, L. and Meier, T. (1993). Organization of the commissural fibers in the adult brain of the locust. J. Comp. Neurol. 332, 358–377. Boyan, G. S. and Williams, J. L. D. (1981). Descending interneurones in the brain of the cricket. Naturwissenschaften 68, 486–487. Boyan, G. S., Williams, J. L. D., Posser, S. and Braünig, P. (2002). Morphological and molecular data argue for the labrum being non-apical, articulated, and the appendage of the intercalary segment in the locust. Arthropod Struct. Dev. 31, 65–76. Boyan, G. S. and Williams, L. (2002). A single cell analysis of engrailed expression in the early embryonic brain of the grasshopper Schistocerca gregaria: ontogeny and identity of the secondary headspot cells. Arthropod Struct. Dev. 30, 207–218. Bo¨hm, H. and Schildberger, K. (1992). Brain neurons involved in the control of walking in the cricket Gryllus bimaculatus. J. Exp. Biol. 166, 113–130. Bo¨hm, L. (1911). Die antennalen Sinnesorgane der Lepidopteren. Arbeiten aus dem Zoologischen Institut der Universita¨t Wien und der Zoologischen Station in Triest 19, 219–246. Braünig, P. (1985). Mechanoreceptive neurons in an insect brain. J. Comp. Neurol. 236, 234–240. Braünig, P. (1990). The morphology of suboesophageal ganglion cells innervating the nervus corporis cardiaci III of the locust. Cell Tiss. Res. 260, 95–108. Braünig, P. (1991). Subesophageal DUM neurons innervate the principal neuropils of the locust brain. Philos. Trans. R. Soc. Lond. B 332, 221–240. Braünig, P., Allgaüer, C. and Honegger, H.-W. (1990). Suboesophageal DUM neurons are part of the antennal motor system of locusts and crickets. Experientia 46, 259–261. Braünig, P. and Hustert, R. (1980). Proprioceptors with central cell bodies in insects. Nature 283, 768–770. Braünig, P., Pflu¨ger, H.-J. and Hustert, R. (1983). The specificity of central nervous projections of locust mechanoreceptors. J. Comp. Neurol. 218, 197–207. Braünig, P. and Pflu¨ger, H. J. (2001). The unpaired median neurons of insects. Adv. Insect Physiol. 28, 185–266. Breer, H. (1981). Comparative studies on cholinergic activities in the central nervous system of Locusta migratoria. J. Comp. Physiol. 141, 271–275. Breer, H. and Sattelle, D. B. (1987). Molecular properties and functions of insect acetylcholine receptors. J. Insect Physiol. 33, 771–790. Brewster, R. and Bodmer, R. (1995). Origin and specification of type-II sensory neurons in drosophila. Development 121, 2923–2936.

184


Brewster, R., Hardiman, K., Deo, M., Khan, S. and Bodmer, R. (2001). The selector gene cut represses a neural cell fate that is specified independently of the AchaeteScute-Complex and atonal. Mech. Dev. 105, 57–68. Brodfuehrer, P. D. and Hoy, R. R. (1990). Ultrasound sensitive neurons in the cricket brain. J. Comp. Physiol. 166, 651–662. Brooks, R. A. (1989). A robot that walks; emergent behaviors from a carefully evolved network. Neural Computat. 1, 253–262. Brusca, R. C. and Brusca, G. J. (1990). Invertebrates. Sunderland, MA: Sinauer Associates. Bruski, C. A. and Dunham, D. W. (1990). Antennal waving in the crayfish Orconectes rusticus (Girard, 1852) (Decapoda, Astacidea). Crustaceana 58, 83–87. Burdohan, J. A. and Comer, C. M. (1990). An antennal-derived mechanosensory pathway in the cockroach: descending interneurons as a substrate of evasive behaviour. Brain Res. 535, 347–352. Burdohan, J. A. and Comer, C. M. (1996). Cellular organization of an antennal mechanosensory pathway in the cockroach, Periplaneta americana. J. Neurosci. 16, 5830–5843. Burrows, M. (1992). Local circuits for the control of leg movements in an insect. TINS 15, 226–232. Burrows, M. (1996). The Neurobiology of an Insect Brain. Oxford: Oxford University Press. Burrows, M. and Newland, P. L. (1993). Correlation between the receptive fields of locust interneurons, their dendritic morphology, and the central projections of mechanosensory neurons. J. Comp. Neurol. 329, 412–426. Bu¨schges, A. and El Manira, A. (1998). Sensory pathways and their modulation in the control of locomotion. Curr. Opin. Neurobiol. 8, 733–739. Camhi, J. M. and Johnson, E. N. (1999). High-frequency steering maneuvers mediated by tactile cues: antennal wall-following in the cockroach. J. Exp. Biol. 202, 631–643. Campbell, F. L. (1972). A new antennal sensillum of Blatella germanica (Dictyoptera: Blattellidae) and its presence in other Blattaria. Ann. Entomol. Soc. Am. 65, 888–892. Chapman, K. M., Duckrow, R. B. and Moran, D. T. (1973). Form and role of deformation in excitation of an insect mechanoreceptor. Nature 244, 453–454. Chapman, R. F. (2003). Contact chemoreception in feeding by phytophagous insects. Ann. Rev. Entomol. 48, 455–484. Chapman, R. F. and Lee, J. C. (1991). Environmental effects on numbers of peripheral chemoreceptors on the antennae of a grasshopper. Chem. Senses 16, 607–616. Clarac, F., Neil, D. M. and Vedel, J.-P. (1976). The control of antennal movements by leg proprioceptors in the rock lobster, Palinurus vulgaris. J. Comp. Physiol. 107, 275–292. Clarac, F. and Vedel, J.-P. (1971). Etude des relations fonctionelles entre le muscle flećhisseur accessoire et les organes sensoriels chordotonaux des appendices locomoteurs de la langouste Palinurus vulgaris. Z. Vergl. Physiol. 72, 386–410. Clarac, F. and Vedel, J. P. (1975). Neurophysiological study of the antennal motor patterns in the rock lobster. I. Reflex modulation of extensor and flexor motoneurone activity. J. Comp. Physiol. 102, 201–221. Comer, C. M., Dowd, J. P. and Stubblefield, T. G. (1988). Escape responses following elimination of the giant interneuron pathway in the cockroach, Periplaneta americana. Brain Res. 445, 370–375. Comer, C. M., Mara, E., Murphy, K. A., Getman, M. and Mungy, M. C. (1994). Multisensory control of escape in the cockroach Periplaneta americana. II. Patterns of touch-evoked behavior. J. Comp. Physiol. A 174, 13–26.


185

Comer, C. M., Parks, L., Halvorsen, M. B. and Breese-Terteling, A. (2003). The antennal system and cockroach evasive behavior. II. Stimulus identification and localization are separable antennal functions. J. Comp. Physiol. A 189, 97–103. Cornwell, P. B. (1968). The Cockroach. I. A laboratory Insect and an Industrial Pest. London: Hutchinson. Cowan, N. J., Ma, E. J., Cutkosky, M. and Full, R. J. (2004). A biologically inspired passive antenna for steering control of a running robot. Proceedings of 11th International Symposium on Robotics Research (ISRR 2004), in press. Crouau, Y. (1997). Comparison of crustacean and insect mechanoreceptive setae. Int. J. Ins. Morphol. Embryol. 26, 181–190. Cruse, H., Kindermann, T., Schumm, M., Dean, J. and Schmitz, J. (1998). Walknet – a biologically inspired network to control six-legged walking. Neural Networks 11, 1435–1447. Dambly-Chaudiere, C., Jamet, E., Burri, M., Bopp, D., Basler, K., Hafen, E., Dumont, N., Spielmann, P., Ghysen, A. and Noll, M. (1992). The paired box gene pox-neuro – a determinant of poly innervated sense-organs in drosophila. Cell 69, 159–172. Damen, W. G. M., Hausdorf, M., Seyfarth, E. A. and Tautz, D. (1998). A conserved mode of head segmentation in arthropods revealed by the expression pattern of hox genes in a spider. Proc. Nat. Acad. Sci. USA 95, 10665–10670. Dietz, A. and Humphreys, W. J. (1971). Scanning electron microscopic studies of antennal receptors of the worker honeybee, including sensilla campaniformia. Ann. Entomol. Soc. Am. 64, 919–925. Distler, P. G. and Boeckh, J. (1996). Synaptic connection between olfactory receptor cells and uniglomerular projection neurons in the antennal lobe of the American cockroach, Periplaneta americana. J. Comp. Neurol. 370, 35–46. Dostal, B. (1958). Riechfa¨higkeit und Zahl der Riechsinneselemente bei der Honigbiene. Z. Vergl. Physiol. 41, 179–203. Dreller, C. and Kirchner, W. H. (1993a). Hearing in honeybees: localization of the auditory sense organ. J. Comp. Physiol. A 173, 275–279. Dreller, C. and Kirchner, W. H. (1993b). How bees perceive the information of the dance language. Naturwissenschaften 80, 319–321. Dreller, C. and Kirchner, W. H. (1994). Hearing in the Asian honeybees Apis dorsata and Apis florea. Insectes Sociaux 41, 291–299. Du¨rr, V. (2000). Grasping movements in an insect: effects of tactile cues and step cycle phase. Eur. J. Neurosci. 12 (Suppl. 11), 150. Du¨rr, V. (2001). Stereotypic leg searching-movements in the stick insect: kinematic analysis, behavioural context and simulation. J. Exp. Biol. 204, 1589–1604. Du¨rr, V., Ko¨nig, Y. and Kittmann, R. (2001). The antennal motor system of the stick insect Carausius morosus: anatomy and antennal movement pattern during walking. J. Comp. Physiol. A 187, 131–144. Du¨rr, V. and Krause, A. (2001). The stick insect antenna as a biological paragon for an actively moved tactile probe for obstacle detection. In: Climbing and Walking Robots – From Biology to Industrial Applications. Proc. 4th Int. Conf. Climbing and Walking Robots (CLAWAR 2001, Karlsruhe), (eds Berns, K. and Dillmann, R.), pp. 87–96. Bury St. Edmunds, London: Professional Engineering Publishing. Du¨rr, V. and Krause, A. (2002). Design of a biomimetic active tactile sensor for legged locomotion. In: Climbing and Walking Robots, and the Support Techologies for Mobile Machines. Proceedings of 5th International Conference on Climbing and Walking Robots, CLAWAR 2002, Paris, (eds Bidaud, P. and Ben Amar, F.), pp. 255–262. Bury St. Edmunds, London: Professional Engineering Publishing. Du¨rr, V., Krause, A., Schmitz, J. and Cruse, H. (2003). Neuroethological concepts and their transfer to walking machines. Int. J. Robot. Res. 22, 151–167.

186


Du¨rr, V. and Ebeling, W. (2005). The behavioural transition from in straight to curve walking: kinetics of leg movement parameters and the initiation of turning. J. Exp. Biol. 208, 2237–2252. Ehmer, B. and Gronenberg, W. (1997a). Antennal muscles and fast antennal movements in ants. J. Comp. Physiol. B 167, 187–296. Ehmer, B. and Gronenberg, W. (1997b). Proprioceptors and fast antennal reflexes in the ant Odontomachus (Formicidae, Ponerinae). Cell Tiss. Res. 290, 153–165. Eichmu¨ller, S. and Scha¨fer, S. (1995). Sensory neuron development revealed by taurine immunocytochemistry in the honeybee. J. Comp. Neurol. 352, 297–307. Ellerkmann, E. (1994). Neuronale Kontrolle und Modulation der Antennenbewegung bei der Honigbiene. Doctoral dissertation, Technische Universita¨t Berlin. Erber, J., Kierzek, S., Sander, E. and Grandy, K. (1998). Tactile learning in the honeybee. J. Comp. Physiol. A 183, 737–744. Erber, J. and Kloppenburg, P. (1995). The modulatory effects of serotonin and octopamine in the visual system of the honeybee (Apis mellifera L.). Behavioral analysis of the motion-sensitive antennal reflex. J. Comp. Physiol. A 176, 111–118. Erber, J. and Pribbenow, B. (2000). Antennal movements in the honeybee: how complex tasks are solved by a simple neuronal system. In: Prerational Intelligence. Adaptive Behaviour and Intelligent Systems without Symbols and Logic, (eds Cruse, H., Dean, J. and Ritter, H.), pp. 109–121. Dordrecht: Kluwer. Erber, J., Pribbenow, B., Bauer, A. and Kloppenburg, P. (1993). Antennal reflexes in the honeybee – tools for studying the nervous system. Apidologie 24, 283–296. Erber, J., Pribbenow, B., Grandy, K. and Kierzek, S. (1997). Tactile motor learning in the antennal system of the honeybee (Apis mellifera L.). J. Comp. Physiol. A 181, 355–365. Erber, J., Pribbenow, B., Kisch, J. and Faensen, D. (2000). Operant conditioning of antennal muscle activity in the honeybee (Apis mellifera L.). J. Comp. Physiol. A 186, 557–565. Ernst, K. D. and Boeckh, J. (1983). A neuroanatomical study on the organisation of the central antennal pathways in insects. III. Neuroanatomical characterization of physiologically defined response types of deutocerebral neurons in Periplaneta americana. Cell Tiss. Res. 229, 1–22. Espenschied, K. S., Quinn, R. D., Chiel, H. J. and Beer, R. D. (1993). Leg coordination mechanisms in stick insect applied to hexapod robot locomotion. Adapt. Behav. 1, 455–468. Esslen, J. and Kaissling, K. E. (1976). Number and distribution of sensilla on antennal flagellum of honeybee (Apis mellifera L.). Zoomorphology 83, 227–251. Field, L. H. and Matheson, T. (1998). Chordotonal organs of insects. Adv. Insect Physiol. 27, 1–228. Flanagan, D. and Mercer, A. R. (1989). Morphology and response characteristics of neurons in the deutocerebrum of the brain in the honeybee Apis mellifera. J. Comp. Physiol. A 164, 483–494. Fleming, R. C. (1968). Head musculature of sphinx moths (Lepidoptera: Sphingidae). Contrib. Am. Entomol. Inst. 3, 1–32. Flores, G. B. and Lazzari, C. R. (1996). The role of the antennae in Triatoma infestans: orientation towards thermal sources. J. Insect Physiol. 42, 433–440. Foelix, R. F. (1996). Biology of Spiders. 2nd ed. Oxford, New York: Oxford University Press. French, A. S. and Sanders, E. J. (1979). The mechanism of sensory transduction in the sensilla of the trochanteral hair plate of the cockroach, Periplaneta americana. Cell Tiss. Res. 198, 159–174. French, A. S. and Sanders, E. J. (1981). The mechanosensory apparatus of the femoral tactile spine of the cockroach, Periplaneta americana. Cell Tiss. Res. 219, 53–68.


187

Frik, M. and Amendt, O. (1995). Neural control of a walking robot in variable terrain. IFTOMM 1995, Mailand, pp. 2297–2301. Frings, H. and Frings, M. (1949). The loci of contact chemoreceptors in insects. Am. Mid. Nat. 41, 602–658. Fudalewicz-Niemczyk, W. and Rosciszewska, M. (1973). The peripheric nervous system of the larva of Gryllus domesticus L. (Orthoptera). Part I. Antenna. Acta Biol. Cracov. Ser. Zool. XVI, 209–217. Gaffal, K. P., Tichy, H., Theiss, J. and Seelinger, G. (1975). Structural polarities in mechanosensitive sensilla and their influence on stimulus transmission (Arthropoda). Zoomorphology 82, 79–103. Gebhardt, M. (1996). Intersegmentale und lokale Informationsverarbeitung im neuronalen Netzwerk zur Kontrolle von Antennenbewegungen bei Grillen (Gryllus campestris (L.), Gryllus bimaculatus De Geer). Doctoral dissertation, Technische Universita¨t Mu¨nchen. Gebhardt, M. (2004). In vivo-application of anti-proctolin-antiserum affects antennal flight posture in crickets. J. Comp. Physiol. A 190, 359–364. Gebhardt, M. and Honegger, H.-W. (2001). Physiological characterisation of antennal mechanosensory descending interneuons in an insect (Gryllus bimaculatus, Gryllus campestris) brain. J. Exp. Biol. 204, 2265–2275. Gewecke, M. (1970). Antennae: another wind sensitive receptor in locusts. Nature 225, 1263–1264. Gewecke, M. (1972a). Antennen und Stirn-Scheitelhaare von Locusta migratoria L. als Luftstro¨mungs-Sinnesorgane bei der Flugsteuerung. J. Comp. Physiol. 80, 57–94. Gewecke, M. (1972b). Bewegungsmechanismus und Gelenkrezeptoren der Antennen von Locusta migratoria. Z. Morphol. Tiere 71, 128–149. Gewecke, M. (1975). The influence of air current sense organs on the flight behaviour of Locusta migratoria. J. Comp. Physiol. A 103, 79–95. Gewecke, M. (1979). Central projection of antennal afferents for the flight motor in Locusta migratoria (Orthoptera: Acrididae). Entomol. Gen. 5, 317–320. Gewecke, M. and Heinzel, H.-G. (1980). Aerodynamic and mechanical properties of the antennae as air-current sense-organs in Locusta migratoria. I. Static characteristics. J. Comp. Physiol. A 139, 357–366. Ghysen, A. and Dambly-Chaudiere, C. (1993). The specification of sensory neuron identity in drosophila. Bioessays 15, 293–298. Gillespie, P. G. and Walker, R. G. (2001). Molecular basis of mechanosensory transduction. Nature 413, 194–202. Gnatzy, W. (1984). Campaniform structures on lobster antennae are dermal glands. Cell Tiss. Res. 236, 729–731. Gnatzy, W. and HeuXlein, R. (1986). Digger wasp against crickets. I. Receptors involved in the antipredator strategies of the prey. Naturwissenschaften 73, 212–215. Gnatzy, W. and Hustert, R. (1989). Mechanoreceptors in behavior. In: Cricket Behavior and Neurobiology, (eds Huber, F., Moore, T. E. and Loher, W.), pp. 198–226. Ithaka, London: Cornell University Press. Gnatzy, W., Schmidt, M. and Rombke, J. (1984). Are the funnel-canal organs the campaniform sensilla of the shore crab Carcinus maenas (Crustacea, Decapoda). 1. Topography, external structure and basic organization. Zoomorphology 104, 11–20. Goodman, L. J. (1981). Organisation and physiology of the insect dorsal ocellar system. In: Comparative Physiology and Evolution of Vision in Invertebrates. C: Invertebrate Visual Centers and Behavior II C, (eds Autrum, H., Goodman, L. J., Messenger, J. B. and Wehner, R.), pp. 201–286. Springer: Berlin. Goulding, S. E., zur Lage, P. and Jarman, A. P. (2000). Amos, a proneural gene for Drosophila olfactory sense organs that is regulated by lozenge. Neuron 25, 69–78.

188


Go¨pfert, M. C., Briegel, H. and Robert, D. (1999). Mosquito hearing: sound-induced antennal vibrations in male and female Aedes aegypti. J. Exp. Biol. 202, 2727–2738. Go¨pfert, M. C. and Robert, D. (2001a). Active auditory mechanics in mosquitoes. Proc. R. Soc. Lond. B 268, 333–339. Go¨pfert, M. C. and Robert, D. (2001b). Turning the key on Drosophila audition. Nature 411, 908. Gras, H. and Ho¨rner, M. (1992). Wind-evoked escape running of the cricket Gryllus bimaculatus. J. Exp. Biol. 171, 189–214. Greenwood, M. and Chapman, R. F. (1984). Differences in numbers of sensilla on the antennae of solitarious and gregarious Locusta migratoria L. (Orthoptera: Acrididae). Int. J. Insect Morphol. Embryol. 13, 295–301. Griss, C. and Rowell, H. F. (1986). Three descending interneurons reporting deviation from course in the locust. I. Anatomy. J. Comp. Physiol. A 158, 765–774. Gronenberg, W. and Strausfeld, N. J. (1990). Descending neurons supplying the neck and flight motor of diptera: physiological and anatomical characteristics. J. Comp. Neurol. 302, 973–991. Guharay, F. and Sachs, F. (1984). Stretch-activated single ion channel currents in tissuecultured embryonic chick skeletal muscle. J. Physiol. Lond. 352, 685–701. Gundel, M. and Penzlin, H. (1995). Anatomy of neurons crossing the tritocerebral commissures of the cockroach Periplaneta americana (Blattaria). J. Morphol. 223, 225–242. Gupta, B. P. and Rodrigues, V. (1997). Atonal is a proneural gene for a subset of olfactory sense organs in Drosophila. Genes Cells 2, 225–233. Haas, M. S., Brown, S. J. and Beeman, R. W. (2001). Homeotic evidence for the appendicular origin of the labrum in Tribolium castaneum. Dev. Genes Evol. 211, 96–102. Habig, C. and Taylor, R. C. (1982a). The crayfish second antennae. I. Muscles, innervation patterns and movements of antennal segments. Comp. Biochem. Physiol. 72A, 339–347. Habig, C. and Taylor, R. C. (1982b). The crayfish second antennae. II. Motoneuron structure as revealed by cobalt chloride backfilling. Comp. Biochem. Physiol. A 72, 349–358. Hallberg, E. and Hansson, B. (1999). Arthropod sensilla: morphology and phylogenetic considerations. Microsc. Res. Tech. 47, 428–439. Hansson, B. S. (ed.). (1999). Insect Olfaction. Berlin: Springer. Heckenlively, D. B. (1970). Intensity of aggression in the crayfish Orconectes virilis (Hagen). Nature 225, 180–181. Hedwig, B. (1994). A cephalothoracic command system controls stridulation in the acridid grasshopper Omocestus viridulus L. J. Neurophysiol. 72, 2015–2025. Hedwig, B. (1996). A descending brain neuron elicits stridulation in the cricket Gryllus bimaculatus (de Geer). Naturwissenschaften 83, 428–429. Heinrich, R. (2002). Impact of descending brain neurons on the control of stridulation, walking, and flight in Orthoptera. Micros. Res. Tech. 56, 301. Heinzel, H.-G. and Gewecke, M. (1979). Directional sensitivity of the antennal campaniform sensilla in locusts. Naturwissenschaften 66, 212–213. Heinzel, H.-G. and Gewecke, M. (1987). Aerodynamic and mechanical properties of the antennae as air-current sense-organs in Locusta migratoria. II. Dynamic characteristics. J. Comp. Physiol. A 161, 671–680. Hennig, W. (1986). Taschenbuch der speziellen Zoologie. WirbelloseII:Gliedertiere. 4th ed. Frankfurt/M., Thun: Verlag Harry Deutsch. Hensler, K. (1988). The pars intercerebralis neurone PI(2)5 of locusts: convergent processing of inputs reporting head movements and deviations from straight flight. J. Exp. Biol. 140, 511–533.


189

Hensler, K. (1992). Neuronal co-processing of course deviation and head movements in locusts. I. Descending deviation detectors. J. Comp. Physiol. A 171, 257–271. Hensler, K. and Rowell, C. H. F. (1990). Control of opromotor responses by descending deviation detector neurons in intact flying locusts. J. Exp. Biol. 149, 191–205. Heran, H. (1957). Die Bienenantenne als MeXorgan der Flugeigengeschwindigkeit. Naturwissenschaften 44, 475. Heran, H. (1959). Wahrnehmung der Flugeigengeschwindigkeit bei Apis mellifica L. Z. Vergl. Physiol. 42, 103–163. Herrnkind, W. (1969). Queuing behavior of spiny lobsters. Science 164, 1425–1427. Hertel, H. and Maronde, U. (1987). The physiology and morphology of centrally projecting visual interneurons in the honeybee brain. J. Exp. Biol. 133, 301–315. Hicks, J. B. (1857). On a new organ in insects. J. Proc. Linn. Soc. (London) 1, 136–140. Higashijima, S., Michiue, T., Emori, Y. and Saigo, K. (1992). Subtype determination of Drosophila embryonic external sensory organs by redundant homeo box genes BarH1 and BarH2. Genes Dev. 6, 1005–1018. Hildebrand, J. G. (1996). Olfactory control of behavior in moths: central processing of odor information and the functional significance of olfactory glomeruli. J. Comp. Phys. A 178, 5–19. Hill, E. S., Iwano, M., Gatellier, L. and Kanzaki, R. (2002). Morphology and physiology of the serotonin-immunoreactive putative antennal lobe feedback neuron in the male silkmoth Bombyx mori. Chem. Senses 27, 475–483. Hofmann, H. A. and Schildberger, K. (2001). Assessment of strength and willingness to fight during aggressive encounters in crickets. Anim. Behav. 62, 337–348. Hofmann, T., Koch, U. T. and Ba¨ssler, U. (1985). Physiology of the femoral chordotonal organ in the stick insect, Cuniculina impigra. J. Exp. Biol. 114, 207–223. Holst, E. V. and Mittelstaedt, H. (1950). Das Reafferenzprinzip. Wechselwirkungen zwischen Zentralnervensystem und Peripherie. Naturwissenschaften 37, 464–476. Holzer, R. and Shimoyama, I. (1997). Locomotion control of a bio-robotic system via electric stimulation. Proc. Int. Conf. Intell. Robots Systems 1514–1519. Homberg, U. (1984). Processing of antennal information in extrinsic mushroom body neurons of the bee brain. J. Comp. Physiol. 154, 825–836. Homberg, U. (1994). Flight-correlated activity changes in neurons of the lateral accessory lobes in the brain of the locust Schistocerca gregaria. J. Comp. Physiol. A 175, 597–610. Homberg, U. (2002). Neurotransmitters and neuropeptides in the brain of the locust. Microsc. Res. Tech. 56, 189–209. Homberg, U., Christensen, T. A. and Hildebrand, J. G. (1989). Structure and function of the deutocerebrum in insects. Ann. Rev. Entomol. 34, 477–501. Homberg, U. and Hildebrand, J. G. (1989). Serotonin-immunoreactive neurons in the median protocerebrum and suboesophageal ganglion of the sphinx moth Manduca sexta. Cell Tiss. Res. 258, 1–24. Homberg, U., Kingan, T. G. and Hildebrand, J. G. (1987). Immunocytochemistry of GABA in the brain and suboesophageal ganglion of Manduca sexta. Cell Tiss. Res. 248, 1–24. Homberg, U., Kingan, T. G. and Hildebrand, J. G. (1990). Distribution of FMRFamide-like immunoreactivity in the brain and suboesophageal ganglion of the sphinx moth Manduca sexta and colocalization with SCPB-, BPP-, and GABA-like immunoreactivity. Cell Tiss. Res. 259, 401–419. Homberg, U., Wurden, S., Dircksen, H. and Rao, K. R. (1991). Comparative anatomy of pigment-dispersing hormone-immunoreactive neurons in the brain of orthopteroid insects. Cell Tiss. Res. 266, 343–357.

190


Honegger, H.-W. (1981). A preliminary note on a new optomotor response in crickets: antennal tracking of moving targets. J. Comp. Physiol. A 142, 419–421. Honegger, H.-W., Allgaüer, C., Klepsch, U. and Welker, J. (1990a). Morphology of antennal motoneurons in the brains of two crickets, Gryllus bimaculatus and Gryllus campestris. J. Comp. Neurol. 291, 256–268. Honegger, H.-W., Bartos, M., Gramm, T. and Gebhardt, M. J. (1995). Peripheral modulation and plasticity of antennal movements in crickets. Verh. Dt. Zool. Ges. 88, 129–137. Honegger, H.-W., Brunninger, B., Braünig, P. and Elekes, K. (1990b). GABA-like immunoreactivity in a common inhibitory neuron of the antennal motor system of crickets. Cell Tiss. Res. 260, 349–354. Honegger, H.-W. and Campan, R. (1989). Vision and visually guided behavior. In: Cricket Behavior and Neurobiology, (eds Huber, F., Moore, R. and Loher, W.), pp. 147–177. Ithaka, London: Cornell University Press. Honegger, H. W., Garcia-Scheible, I., Kammerer, R. and Laurent, G. (1985). Vibration blocks habituation of an optical released antennal response in crickets. In: Acoustic and Vibrational Communication in Insects, (eds Kalmring, K. and Elsner, N.), pp. 223–230. Berlin, Hamburg: Paul Parey. Horn, E. (1983). Gravity. In: Comprehensive Insect Physiology, Biochemistry and Pharmacology, Vol. 6 (eds Kerkut, G. A. and Gilbert, L. I.), pp. 223–230. Oxford, NY: Pergamon Press. Horn, E. and Bischof, H.-J. (1983). Gravity reception in crickets: the influence of cercal and antennal afferents on the head position. J. Comp. Physiol. A 150, 93–98. Horn, E. and Fo¨ller, W. (1985). Tonic and modulatory subsystems of the complex gravity receptor system in crickets, Gryllus bimaculatus. J. Insect Physiol. 31, 937–946. Horseman, B. G., Gebhardt, M. J. and Honegger, H.-W. (1997). Involvement of the suboesophageal and thoracic ganglia in the control of antennal movements in crickets. J. Comp. Physiol. A 181, 195–204. Hoskins, S. G., Homberg, U., Kingan, T. G., Christensen, T. A. and Hildebrand, J. G. (1986). Immunocytochemistry of GABA in the antennal lobes of the sphinx moth Manduca sexta. Cell Tiss. Res. 244, 243–252. Howse, P. E. (1965). The structure of the subgenual organ and certain other mechanoreceptors of the termite Zootermopsis angusticollis (Hagen). Proc. R. Ent. Soc. Lond. A 40, 137–146. Howse, P. E. (1968). The fine structure and functional organization of chordotonal organs. Symp. Zool. Soc. Lond. 23, 167–198. Howse, P. E. and Claridge, M. F. (1970). Fine structure of Johnston’s organ of leaf hopper, Oncopsis flavicollis. J. Insect Physiol. 16, 1665. Hoyle, G. (1978). Distribution of nerve and muscle fibre types in locust jumping muscle. J. Exp. Biol. 73, 203–233. Ho¨lldobler, B. (1970). Zur Physiologie der Gast-Wirt-Beziehungen (Myrmecophilie) bei Ameisen: II. Das Gastverha¨ltnis des imaginalen Atemeles pubicollis Bris. (Col. Staphylinidae) zu Myrmica und Formica (Hym. Formicidae). Z. Vergl. Physiol. 66, 215–250. Ho¨lldobler, B. (1976). Tournaments and slavery in a desert ant. Science 192, 912–914. Ho¨lldobler, B. (1985). Liquid food transmission and antennation signals in ponerine ants. Israel J. Entomol. XIX, 89–99. Ho¨lldobler, B. (1999). Multimodal signals in ant communication. J. Comp. Physiol. A 184, 129–141. Ho¨lldobler, B., Mo¨glich, M. and Maschwitz, M. (1974). Communication by tandem running in the ant Camponotus sericeus. J. Comp. Physiol. 90, 105–127.


191

Ho¨rner, M. (1992). Wind-evoked escape running of the cricket Gryllus bimaculatus. II. Neurophysiological analysis. J. Exp. Biol. 171, 215–245. Ho¨rner, M. and Gras, H. (1985). Physiological properties of some descending neurons in the cricket brain. Naturwissenschaften 72, 603–604. Huber, F. (1955). Sitz und Bedeutung nervo¨ser Zentren fu¨r Instinkthandlungen beim Ma¨nnchen von Gryllus campestris L. Z. Tierpsychol. 12, 12–48. Huber, R. and Kravitz, E. A. (1995). A quantitative analysis of agonistic behavior in juvenile American lobsters (Homarus americanus L.). Brain Behav. Evol. 46, 72–83. Hughes, G. M. and Wiersma, A. G. (1960). Neuronal pathways and synaptic connexions in the abdominal cord of the crayfish. J. Exp. Biol. 37, 291–307. Imms, A. D. (1939). On the antennal musculature in insects and other arthropods. Quart. J. Microsc. Sci. 81, 273–320. Jacobs, G. A. and Theunissen, F. E. (1996). Functional organization of a neural map in the cricket cercal sensory system. J. Neurosci. 16, 769–784. Janet, C. (1911). Sur l’existance du’un organe chordotonal et d’une vesicle pulsatile antennaires chez l’abeille et sur la morphologie de la tete de cet. Comptes Rendus hebdomadaires Seances l’Acadiemie Sci. 152, 110–113. Jhaveri, D., Sen, A., Reddy, G. V. and Rodrigues, V. (2000). Sense organ identity in the Drosophila antenna is specified by the expression of the proneural gene atonal. Mech. Dev. 99, 101–111. Ka¨mper, G. (1992). Development of cricket sensory hairs changes of dynamic mechanical properties. J. Comp. Physiol. A 170, 49–55. Kammerer, R. and Honegger, H. W. (1988). The role of mechanoreceptors in the control of antennal tracking movements of crickets. Proc. Go¨ttingen. Neurobiol. Conf. 201 . Kaneko, M., Kanayma, N. and Tsuji, T. (1995). 3-D active antenna for contact sensing. Proc. IEEE Int. Conf. Robot. Automat. 1995, 1113–1119. Kaneko, M., Kanayma, N. and Tsuji, T. (1998). Active antenna for contact sensing. IIEEE Trans. Robot. Automat. 14, 278–291. Kanzaki, R., Arbas, E. A. and Hildebrand, J. G. (1991). Physiology and morphology of descending neurons in the pheromone-processing olfactory pathways in the male moth Manduca sexta. J. Comp. Physiol. A 169, 1–14. Kanzaki, R., Arbas, E. A., Strausfeld, N. J. and Hildebrand, J. G. (1989). Physiology and morphology of projection neurons in the antennal lobe of the male moth Manduca sexta. J. Comp. Physiol. A 165, 427–453. Keegan, A. P. and Comer, C. M. (1993). The wind-elicited escape response of cockroaches (Periplaneta americana) is influenced by lesions rostral to the escape circuit. Brain Res. 620, 310–316. Keil, T. A. (1982). Contacts of pore tubules and sensory dendrites in antennal chemosensilla of a silkmoth – demonstration of a possible pathway for olfactory molecules. Tissue Cell 14, 451–462. Keil, T. A. (1989). Fine structure of the pheromone-sensitive sensilla on the antenna of the hawkmoth, Manduca sexta. Tissue Cell 21, 139–151. Kevan, P. G. and Lane, M. A. (1985). Flower petal microtexture is a tactile cue for bees. Proc. Natl. Acad. Sci. USA 82, 4750–4752. Kim, J. L. and Yamasaki, T. (1996). Sensilla of Carabus (Isiocarabus) fiduciarius saishutoicus Csiki (Coleoptera: Carabidae). Int. J. Insect Morphol. Embryol. 25, 153–172. Kim, Y., Lee, H. and Kim, K. (1990). Projection of antennal receptor cells into the dorsal lobe of brain of Pieris rapae (Insecta, Lepidoptera). Korean J. Zool. 33, 28–34. Kindermann, T. (2002). Behavior and adaptability of a six-legged walking system with highly distributed control. Adapt. Behav. 9, 16–41.

192


Kirchner, W. H. (1994). Hearing in honeybees: The mechanical response of the bee’s antenna to near-field sound. J. Comp. Physiol. A 175, 261–265. Kirchner, W. H., Dreller, C. and Towne, W. F. (1991). Hearing in honeybees: operant conditioning and spontaneous reactions to airborne sound. J. Comp. Physiol. A 168, 85–89. Kisch, J. and Erber, J. (1999). Operant conditioning of antennal movements in the honeybee. Behav. Brain Res. 99, 93–102. Klein, U., Bock, C., Kafka, W. A. and Moore, T. E. (1988). Antennal sensilla of Magicicada cassini Fisher (Homoptera, Cicadidae). Fine structure and electrophysiological evidence for olfaction. Int. J. Insect Morphol. Embryol. 17, 153–167. Kloppenburg, P. (1995). Anatomy of the antennal motoneurons in the brain of the honeybee (Apis mellifera). J. Comp. Neurol. 363, 333–343. Kloppenburg, P., Camazine, S. M., Sun, X. J., Randolph, P. and Hildebrand, J. G. (1997). Organization of the antennal motor system in the sphinx moth Manduca sexta. Cell Tiss. Res. 287, 425–433. Kloppenburg, P. and Erber, J. (1995). The modulatory effects of serotonin and octopamine in the visual system of the honeybee (Apis mellifera L.): II. Electrophysiological analysis of motion-sensitive neurons in the lobula. J. Comp. Physiol. A 176, 119–129. Kloppenburg, P., Kirchhof, B. S. and Mercer, A. R. (1999). Voltage-activated currents from adult honeybee (Apis mellifera) antennal motor neurons recorded in vitro and in situ. J. Neurophysiol. 81, 39–48. Klose, M. (1996). Development of leg chordotonal sensory organs in normal and heat shocked embryos of the cricket Teleogryllus (Walker). Roux’s. Arch. Dev. Biol. 205, 344–355. Knyazeva, N. I. (1974). Campaniform sensilla of Locusta migratoria L. (Orthoptera, Acrididae). Entom. Rev. 53, 62–66. Knyazeva, N. I., Fudalewicz-Niemczyk, W. and Rosciszewska, M. (1975). Proprioceptors of the house-cricket (Gryllus domesticus L.) (Orthoptera). Acta Biol. Cracov. Ser. Zool. 18, 33–44. Knyazeva, N. I., Fudalewicz-Niemczyk, W. and Rosciszewska, M. (1978). Proprioreceptors in the nymph of the house cricket Gryllus domesticus (Orthoptera, Gryllidae). Entom. Rev. 58, 16–19. Koch, M. (2000). The cuticular cephalic endoskeleton of primarily wingless hexapods: ancestral state and evolutionary changes. Pedobiologia. 44, 374–385. Kondoh, Y., Okuma, J. and Newland, P. L. (1995). Dynamics of neurons controlling movements of a locust hind leg – wiener kernel analysis of the responses of proprioceptive afferents. J. Neurophysiol. 73, 1829–1842. Konings, N. M., Vullings, G. B., Geffard, M., Buijs, R. M., Diederen, H. B. and Jansen, W. F. (1988). Immunocytochemical demonstration of octopamine-immunoreactive cells in the nervous system of Locusta migratoria and Schistocerca gregaria. Cell Tiss. Res. 251, 371–379. Krause, A. F. and Du¨rr, V. (2004). Tactile efficiency of insect antennae with two hinge joints. Biol. Cybern. 91, 168–181. Kreissl, S. and Bicker, G. (1989). Histochemistry of acetylcholinesterase and immunocytochemistry of an acetylcholine receptor-like antigen in the brain of the honeybee. J. Comp. Neurol. 286, 71–84. Kreissl, S., Eichmu¨ller, S., Bicker, G., Rapus, J. and Eckert, M. (1994). Octopamine-like immunoreactivity in the brain and suboesophageal ganglion of the honeybee. J. Comp. Neurol. 348, 583–595. Kros, C. J., Rusch, A. and Richardson, G. P. (1992). Mechano-electrical transducer currents in hair cells of the cultured neonatal mouse cochlea. Proc. R. Soc. Lond. B 249, 185–193.


193

Kuster, J. E., French, A. S. and Sanders, E. J. (1983). The effects of microtubule dissociating agents on the physiology and cytology of the sensory neuron in the femoral tactile spine of the cockroach, Periplaneta americana L. Proc. R. Soc. Lond. B 219, 397–412. Kuwana, Y., Nagasawa, S., Shimoyama, I. and Kanzaki, R. (1999). Synthesis of the pheromone-oriented behaviour of silkworm moths by a mobile robot with moth antennae as pheromone sensors. Biosensors Bioelectron. 14, 195–202. Lambin, M. (1973). Le sensilles de l’antenne chez quelques blattes et en particulier chez Blaberus cranifer (Burm.). Z. Zellforsch. 143, 183–206. Laurent, G. (1996). Dynamical representation of odors by oscillating and evolving neural assemblies. TINS 19, 489–496. Lee, J. K. and Strausfeld, N. J. (1990). Structure, distribution and number of surface sensilla and their receptor cells on the olfactory appendage of the male moth Manduca sexta. J. Neurocytol. 19, 519–538. Lee, M. H. and Nicholls, H. R. (1999). Tactile sensing for mechatronics – a state of the art survey. Mechatronics 9, 1–31. Lenoir, A. and Jaisson, P. (1982). Evolution et roˆle des communications antennaires chez les insects sociaux. In: Social Insects in the Tropics, (ed. Jaisson, P.), pp. 157–180. Paris: Presses de l’Universite´ de Paris XIII. Linsenmair, K. E. (1970). Die Interaktion der paarigen antennalen Sinnesorgane bei der Windorintierung laufender Ka¨fer (Insecta, Coleoptera). Z. Vergl. Physiol. 70, 247–277. Linsenmair, K. E. (1973). Die Windorientierung laufender Insekten. Fortschr. Zool. 21, 59–79. Loftus, R. (1966). Cold receptor on the antenna of Periplaneta americana. Z. Vergl. Physiol. 52, 380–385. Loftus, R. (1968). The response of the antennal cold receptor of Periplaneta americana to rapid temperature changes and to steady temperature. Z. Vergl. Physiol. 59, 413–455. Loftus, R. (1969). Differential thermal components in the response of the antennal cold receptor of Periplaneta americana to slowly changing temperature. Z. Vergl. Physiol. 63, 415–433. Loftus, R. (1976). Temperature-dependent dry receptor on antenna of Periplaneta – Tonic response. J. Comp. Physiol. 111, 153–170. Loher, W. and Rence, B. (1978). The mating behavior of Teleogryllus commodus (Walker) and its central and peripheral control. Z. Tierpsychol. 46, 225–259. Lutz, E. M. and Tyrer, N. M. (1988). Immunohistochemical localization of serotonin and choline-acetyltransferase in sensory neurons of the locust. J. Comp. Neurol. 267, 335–342. Malun, D. (1991). Inventory and distribution of synapses of identified uniglomerular projection neurons in the antennal lobe of Periplaneta americana. J. Comp. Neurol. 305, 348–360. Malun, D., Waldow, U., Kraus, D. and Boeckh, J. (1993). Connections between the deutocerebrum and the protocerebrum, and neuroanatomy of several classes of deutocerebral projection neurons in the brain of male Periplaneta americana. J. Comp. Neurol. 329, 143–162. Manning, A. (1967). Antennae and sexual receptivity in Drosophila melanogaster females. Science 158, 136–137. Markl, H. (1962). Borstenfelder an den Gelenken als Schweresinnesorgane bei Ameisen und anderen Hymenopteren. Z. Vergl. Physiol. 45, 475–569. Maronde, U. (1991). Common projection areas of antennal and visual pathways in the honeybee brain Apis mellifera. J. Comp. Neurol. 309, 328–340.

194


Martin, H. (1965). Leistungen des topochemschien Sinnes bei der Honigbiene. Z. Vergl. Physiol. 50, 254–292. Martin, H. and Lindauer, M. (1966). Sinnesphysiologische Leistungen beim Wabenbau der Honigbiene. Z. Vergl. Physiol. 53, 372–404. Maschwitz, U., Ho¨lldobler, B. and Mo¨glich, M. (1974). Tandemlaufen als Rekrutierungsverhalten bei Bothroponera tesserinoda Forel (Formicidae: Ponerinae). Z. Tierpsychol. 35, 113–123. Mathenia, N. R., Leung, V., Baba, Y. and Comer, C. M. (2003). Comparison of a specific descending mechanosensory interneuron in cockroaches and crickets. Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience. CD-ROM. 2003, 269.9. Matheson, T. (1990). Responses and locations of neurons in the locust metathoracic femoral chordotonal organ. J. Comp. Physiol. A 166, 915–927. Matheson, T. (1992). Range fractionation in the locust metathoracic femoral chordotonal organ. J. Comp. Physiol. A 170, 509–520. Matsuda, R. (1965). Morphology and Evolution of the Insect Head. Ann Arbor, MI: American Entomological Institute. McFarlane, J. E. (1953). The morphology of the chordotonal organs of the antenna, mouthparts and legs of the lesser migratory grasshopper, Melanoplus mexicanus mexicanus (Saussure). Can. Entomol. LXXXV, 81–103. McGruer, N. E., Adams, G. G., Truong, T. Q., Barnes, T. G., Lu, X. and Aceros, J. C. (2002). Biomimetic flow and contact/bending MEMs sensors. In: Neurotechnology for Biomimetic Robots, (eds Ayers, J., Davis, J. L. and Rudolph, A.), pp. 11–30. Cambridge/MA, London: MIT Press. McIndoo, N. E. (1914). The olfactory sense of the honey-bee. J. Exp. Biol. 16, 265–346. McIndoo, N. E. (1922). The auditory sense of the honey-bee. J. Comp. Neurol. 34, 173–199. McIver, S. B. (1975). Structure of cuticular mechanoreceptors of arthropods. Ann. Rev. Entomol. 20, 381–397. McIver, S. B. (1985). Mechanoreception. In: Comprehensive Insect Physiology, Biochemistry and Pharmacology, (eds Kerkut, G. A. and Gilbert, L. I.), pp. 71–132. Oxford, NY: Pergamon Press. Menzel, R. and Mu¨ller, U. (1996). Learning and memory in honeybees: from behavior to neural substrates. Ann. Rev. Neurosci. 19, 379–404. Mercer, A., Mobbs, P. G., Evans, P. D. and Davenport, A. (1983). Biogenic amines in the brain of the honeybee, Apis mellifera. Cell Tiss. Res. 234, 655–677. Merritt, D. J. (1997). Transformation of external sensilla to chordotonal sensilla in the cut mutant of Drosophila assessed by single-cell marking in the embryo and larva. Microsc. Res. Tech. 39, 492–505. Michelsen, A. (1993). The transfer of information in the dance language of honeybees: progress and problems. J. Comp. Physiol. A 173, 135–141. Milde, J. J. and Strausfeld, N. J. (1990). Cluster organization and response characteristics of the giant fiber pathway of the blowfly Calliphora erythrocephala. J. Comp. Neurol. 294, 59–75. Miles, C. I., May, M. L., Holbrook, E. H. and Hoy, R. R. (1992). Multisegmental analysis of acoustic startle in the flying cricket (Teleogryllus oceanicus): kinematics and electromyography. J. Exp. Biol. 169, 19–36. Mill, P. J. (1976). Chordotonal organs of crustacean appendages. In: Structure and Function of Proprioceptors in the Invertebrates, (ed. Mill, P. J.), London: Chapman & Hall. Monteforti, G., Angeli, S., Petacchi, R. and Minnocci, A. (2002). Ultrastructural characterization of antennal sensilla and immunocytochemical localization of a


195

chemosensory protein in Carausius morosus Brunner (Phasmida : Phasmatidae). Arthropod Struct. Dev. 30, 195–205. Moore, T. E., Crary, S. B., Koditschek, D. E. and Conklin, T. A. (1998). Directed locomotion in cockroaches: ‘‘Biobots’’. Acta Entomol. Slovenica 6, 71–78. Moran, D. T. and Varela, F. G. (1971). Microtubules and sensory transduction. Proc. Natl. Acad. Sci. USA 68, 757–760. Moran, D. T., Varela, F. J. and Rowley, J. C. (1977). Evidence for active role of cilia in sensory transduction. Proc. Natl. Acad. Sci. USA 74, 793–797. Mo¨hl, B. and Bacon, J. (1983). The tritocerebral commissure giant (TCG) wind-sensitive interneuron in the locust. II. Directional sensitivity and role in flight stabilisation. J. Comp. Physiol. 150, 453–465. Murphey, R. K., Possidente, D., Pollack, P. and Ghysen, A. (1989). Compartments and topography of leg afferent projections in Drosophila. J. Neurosci. 9, 3209–3217. Mu¨cke, A. and Lakes-Harlan, R. (1995). Central projections of sensory cells of the midleg of the locust, Schistocerca gregaria. Cell Tiss. Res. 280, 391–400. Myers, C. M. and Evans, P. D. (1987). An FMRFamide antiserum differentiates between populations of antigens in the brain and retrocerebral complex of the locust, Schistocerca gregaria. Cell Tiss. Res. 250, 93–99. Naessel, D. R., Cantera, R. and Karlsson, A. (1992). Neurons in the cockroach nervous system reacting with antisera to the neuropeptide Leukokinin I. J. Comp. Neurol. 322, 45–67. Neil, D. M., Barnes, W. J. P. and Burns, M. D. (1982). Reflex antennal movements in the spiny lobster Palinurus elephas. I. Properties of reflexes and their interaction. J. Comp. Physiol. 147, 259–268. Neil, D. M., Priest, T. D., Miyan, J. A., Wotherspoon, R. M. and Schoene, H. (1984). Co-ordinated equilibrium responses at two joints in the spiny lobster antenna in relation to the pattern of movements imposed upon the legs. J. Comp. Physiol. 155, 351–363. Neuhaüser, T., Sorge, D., Stay, B. and Hoffmann, K. H. (1994). Responsiveness of the adult cricket (Gryllus bimaculatus and Acheta domesticus) retrocerebral complex to allatostatin-1 from a cockroach, Diploptera punctata. J. Comp. Physiol. B 164, 23–31. Newland, P. L. (1990). Morphology of a population of mechanosensory ascending interneurones in the metathoracic ganglion of the locust. J. Comp. Neurol. 299, 242–260. Newland, P. L., Rogers, S. M., Gaaboub, I. and Matheson, T. (2000). Parallel somatotopic maps of gustatory and mechanosensory neurons in the central nervous system of an insect. J. Comp. Neurol. 425, 82–96. Nottebohm, E., Dambly-Chaudiere, C. and Ghysen, A. (1992). Connectivity of chemosensory neurons is controlled by the gene poxn in Drosophila. Nature 359, 829–832. Nye, S. W. and Ritzmann, R. E. (1992). Motion analysis with leg joints associated with escape turns of the cockroach, Periplaneta americana. J. Comp. Physiol. 171, 183–194. Ochieng, S., Hallberg, E. and Hansson, B. (1998). Fine structure and distribution of antennal sensilla of the desert locust, Schistocerca gregaria (Orthoptera: Acrididae). Cell Tiss. Res. 291, 525–536. Okada, J., Kanamaru, Y. and Toh, Y. (2002). Mechanosensory control of antennal movement by the scapal hair plate in the American cockroach. Zool. Sci. 19, 1201–1210. Okada, J. and Toh, Y. (2000). The role of antennal hair plates in object-guided tactile orientation of the cockroach (Periplaneta americana). J. Comp. Physiol. A 186, 849–857.

196


Okada, J. and Toh, Y. (2001). Peripheral representation of antennal orientation by the scapal hair plate of the cockroach Periplaneta americana. J. Exp. Biol. 204, 4301–4309. Okada, R., Ikeda, J. and Mizunami, M. (1999). Sensory responses and movementrelated activities in extrinsic neurons of the cockroach mushroom bodies. J. Comp. Physiol. A 185, 115–129. Okada, R., Sakura, M. and Mizunami, M. (2003). Distribution of dendrites of descending neurons and its implications for the basic organisation of the cockroach brain. J. Comp. Neurol. 458, 158–174. Ott, S. R. and Elphick, M. R. (2002). Nitric oxide synthase histochemistry in insect nervous systems: methanol/formalin fixation reveals the neuroarchitecture of formaldehyde-sensitive NADPH diaphorase in the cockroach Periplaneta americana. J. Comp. Neurol. 448, 165–185. Ott, S. R., Jones, I. W., Burrows, M. and Elphick, M. R. (2000). Sensory afferents and motor neurons as targets for nitric oxide in the locust. J. Comp. Neurol. 422, 521–532. Pareto, A. (1972). Die zentrale Verteilung der Fu¨hlerafferenz bei Arbeiterinnen der Honigbiene, Apis mellifera L. Z. Zellforsch. 131, 109–140. Park, K. C. and Baker, T. C. (2002). Improvement of signal-to-noise ratio in electroantennogram responses using multiple insect antennae. J. Insect Physiol. 48, 1139–1145. Park, K. C., Ochieng, S. A., Zhu, J. W. and Baker, T. C. (2002). Odor discrimination using insect electroantennogram responses from an insect antennal array. Chem. Senses 27, 343–352. Paydar, S., Doan, C. A. and Jacobs, G. A. (1999). Neural mapping of direction and frequency in the cricket cercal sensory system. J. Neurosci. 19, 1771–1781. Pelletier, Y. and McLeod, C. D. (1994). Obstacle perception by insect antennae during terrestrial locomotion. Physiol. Entomol. 19, 360–362. Petit, L. (1958). Le de´terminisme geńe´tique et psychophysiologique de la compe´tition sexuelle chez Drosophila melanogaster. Bull. Biol. France Belg. 92, 248–327. Petryszak, A. (1975). Sensory peripheric nervous system of Periplaneta americana L. (Blattoidea). II. Antenna. Acta Biol. Cracov. Ser. Zool. 18, 257–263. Pfaffmann, C. (1955). Gustatory nerve impulses in rat, cat and rabbit. J. Neurophysiol. 18, 429–440. Pfeiffer, F., Eltze, J. and Weidemann, H. J. (1995). Six-legged technical walking considering biological principles. Robot. Autonom. Systems 14, 223–232. Poulet, J. F. A. and Hedwig, B. (2002). A corollary discharge maintains auditory sensitivity during sound production. Nature 418, 872–876. Poulet, J. F. A. and Hedwig, B. (2003). A corollary discharge mechanism modulates central auditory processing in singing crickets. J. Neurophysiol. 89, 1528–1540. Pribbenow, B. and Erber, J. (1996). Modulation of antennal scanning in the honeybee by sucrose stimuli, serotonin, and octopamine: behavior and electrophysiology. Neurobiol. Learn. Memory 66, 109–120. Pringle, W. S. (1938a). Proprioception in insects. I. A new type of mechanical receptor from the palps of the cockroach. J. Exp. Biol. 15, 100–113. Pringle, W. S. (1938b). Proprioception in insects. II. The action of the campaniform sensilla on the legs. J. Exp. Biol. 15, 114–131. Pringle, W. S. (1948). The gyroscopic mechanism of the halteres of diptera. Philos. Trans. R. Soc. Lond. B 233, 347–384. Rehder, V., Bicker, G. and Hammer, M. (1987). Serotonin-immunoreactive neurons in the antennal lobes and suboesophageal ganglion of the honeybee. Cell Tiss. Res. 247, 59–66.


197

Reichert, H. (1993). Sensory inputs and flight orientation in locusts. Comp. Biochem. Physiol. A 104, 647–657. Reichert, H. and Rowell, C. H. F. (1986). Neural circuits controlling flight in the locust: how sensory information is processed for motor control. TINS 9, 281–283. Remy, C., Guy, J., Pelletier, G. and Boer, H. H. (1988). Immunohistological demonstration of a substance related to neuropeptide-Y and FMRFamide in the cephalic and thoracic nervous systems of the locust Locusta migratoria. Cell Tiss. Res. 254, 189–195. Rence, B. and Loher, W. (1977). Contact chemoreceptive sex recognition in the male cricket, Teleogryllus commodus. Physiol. Entomol. 2, 225–236. Richard, D., Preteur, V., Campan, R., Beugnon, G. and Williams, L. (1985). Visual interneurones of the neck connectives in Gryllus bimaculatus. J. Insect Physiol. 31, 407–417. Rind, F. C. and Leitinger, G. (2000). Immunocytochemical evidence that collision sensing neurons in the locust visual system contain acetylcholine. J. Comp. Neurol. 423, 389–401. Risler, H. (1977). The construction of the auditory organ in male mosquitoes. Fortschr. Zool. 24, 143–147. Ritzmann, R. E., Quinn, R. D., Watson, J. T. and Zill, S. N. (2000). Insect walking and biorobotics: a relationship with mutual benefits. Bioscience 50, 23–33. Rogers, S. M. and Simpson, S. J. (1997). Experience-dependent changes in the number of chemosensory sensilla on the mouthparts and antennae of Locusta migratoria. J. Exp. Biol. 200, 2313–2321. Rogers, S., Matheson, T., Despland, E., Dodgson, T., Burrows, M. and Simpson, S. (2003). Mechanosensory-induced behavioural gregarization in the desert locust Schistocerca gregaria. J. Exp. Biol. 206, 3991–4002. Rohrseitz, K. and Tautz, J. (1999). Honeybee dance communication: waggle run direction coded in antennal contacts?. J. Comp. Physiol. A 184, 463–470. Rosciszewska, M. and Petryszak, A. (1986). The sense organs of the antennae of Periplaneta americana L. (Blattoidea), Carausius morosus B. (Cheleutoptera), Gryllus domesticus L. (Orthoptera). Folia histochemica et cytobiologica 24, 318–319. Rospars, J. P. (1988). Structure and development of the insect antennodeutocerebral system. Int. J. Insect Morphol. Embryol. 17, 243–294. Rospars, J. P. and Hildebrand, J. G. (1992). Anatomical identification of glomeruli in the antennal lobes of the male sphinx moth Manduca sexta. Cell Tiss. Res. 270, 205–227. Rossi-Durand, C. and Vedel, J. P. (1981). Antennal proprioception in the rock lobster Palinurus vulgaris: anatomy and physiology of a bi-articular chordotonal organ. J. Comp. Physiol. 145, 505–516. Rossi, E. L. and Vedel, J. P. (1980). Encoding of the angular position of the two distal joints of the spiny lobster antennae by a single mechanoreceptor. Neurosci. Lett. 18, 301–308. Roth, L. M. and Willis, E. R. (1952). A study of cockroach behaviour. Amer. Mid. Nat. 47, 65–129. Rowell, C. H. F. and O’Shea, M. (1976a). Neuronal basis of a sensory analyser, acridid movement detector system. 1. Effects of simple incremental and decremental stimuli in light and dark adapted animals. J. Exp. Biol. 65, 273–288. Rowell, C. H. F. and O’Shea, M. (1976b). Neuronal basis of a sensory analyser, acridid movement detector system. 3. Control of response amplitude by tonic lateral inhibition. J. Exp. Biol. 65, 617–625. Rowell, H. F. and Reichert, H. (1986). Three descending interneurons reporting deviation from course in the locust. II. Physiology. J. Comp. Physiol. A 158, 775–794.

198


Ro¨mer, H. (1983). Tonotopic organization of the auditory neuropile in the bushcricket Tettigonia viridissima. Nature 306, 60–62. Ro¨mer, H., Marquart, V. and Hardt, M. (1988). Organization of a sensory neuropile in the auditory pathway of two groups of Orthoptera. J. Comp. Neurol. 275, 201–215. Ruppert, E. E. and Barnes, R. D. (1994). Invertebrate Zoology. 6th ed. New York: Saunders College Publishing. Ru¨th, E. (1976). Electrophysiology of sensilla chaetica on antennae of Periplaneta americana. J. Comp. Physiol. 105, 55–64. Saager, F. and Gewecke, M. (1989). Antennal reflexes in the desert locust Schistocerca gregaria. J. Exp. Biol. 147, 519–532. Salecker, I. and Distler, P. (1990). Serotonin-immunoreactive neurons in the antennal lobes of the American cockroach Periplaneta americana – light microscopic and electron-microscopic observations. Histochemistry 94, 463–473. Sandeman, D., Sandeman, R., Derby, C. and Schmidt, M. (1992). Morphology of the brain of crayfish, crabs, and spiny lobsters: a common nomenclature for homologous structures. Biol. Bull. 183, 304–326. Sandeman, D. C. (1985). Crayfish antennae as tactile organs: their mobility and the responses of their proprioceptors to displacement. J. Comp. Physiol. A 157, 363–374. Sandeman, D. C. (1989). Physical properties sensory receptors and tactile reflexes of the antenna of the Australian freshwater crayfish Cherax destructor. J. Exp. Biol. 141, 197–218. Sandeman, D. C., Scholtz, G. and Sandeman, R. E. (1993). Brain evolution in decapod crustacea. J. Exp. Zool. 265, 112–133. Sandeman, D. C. and Varju, D. (1988). A behavioral study of tactile localization in the crayfish Cherax destructor. J. Comp. Physiol. A 163, 525–536. Sandeman, D. C. and Wilkens, L. (1983). Motor control of movements of the antennal flagellum in the Australian crayfish, Euastacus armatus. J. Exp. Biol. 105, 253–273. Schafer, R. and Reagan, P. D. (1981). Colchicine reversibly inhibits electrical activity in arthropod mechanoreceptors. J. Neurobiol. 12, 155–166. Schafer, R. and Sanchez, T. V. (1973). Antennal sensory system of cockroach, Periplaneta americana: postembryonic development and morphology of sense organs. J. Comp. Neurol. 149, 335–353. Schafer, R. and Sanchez, T. V. (1976). Nature and development of sex attractant specificity in cockroaches of genus Periplaneta. I. Sexual dimorphism in distribution of antennal sense organs in 5 species. J. Morphol. 149, 139–157. Schaller, D. (1978). Antennal sensory system of Periplaneta americana L. Cell Tiss. Res. 191, 121–139. Scha¨fer, S. and Bicker, G. (1986a). Common projection areas of 5-HT-like and GABAlike immunoreactive fibers in the visual system of the honeybee. Brain Res. 380, 368–370. Scha¨fer, S. and Bicker, G. (1986b). Distribution of GABA-like immunoreactivity in the brain of the honeybee. J. Comp. Neurol. 246, 287–300. Scha¨fer, S., Bicker, G., Ottersen, O. P. and Storm-Mathisen, J. (1988). Taurine-like immunoreactivity in the brain of the honeybee. J. Comp. Neurol. 268, 60–70. Scha¨fer, S. and Rehder, V. (1989). Dopamine-like immunoreactivity in the brain and suboesophageal ganglion of the honeybee. J. Comp. Neurol. 280, 43–58. Scheidler, A., Kaulen, P., Bruning, G. and Erber, J. (1990). Quantitative autoradiographic localization of I-125 alpha-bungarotoxin binding sites in the honeybee brain. Brain Res. 534, 332–335. Scheiner, R., Erber, J. and Page, R. E. (1999). Tactile learning and the individual evaluation of the reward in honeybees (Apis mellifera L.). J. Comp. Physiol. A 185, 1–10.


199

Schildberger, K. (1984). Multimodal interneurons in the cricket brain – properties of identified extrinsic mushroom body cells. J. Comp. Physiol. 154, 71–79. Schildberger, K. and Agricola, H. (1992). Allatostatin-like immunoreactivity in the brains of crickets and cockroaches. Proc. Go¨ttingen. Neurobiol. Conf. 20, 494. Schmidt-Ott, U., Gonzales-Gaitan, M., Jaeckle, H. and Technau, G. M. (1994). Number, identity, and sequence of the Drosophila head segments as revealed by neural elements and their deletion patterns in mutants. Proc. Natl. Acad. Sci. USA 91, 8363–8367. Schmidt-Ott, U. and Technau, G. M. (1992). Expression of en and wg in the embryonic head and brain of Drosophila indicates a refolded band of seven segment remnants. Development 116, 111–125. Schmidt, K. (1973). Comparative morphology of insect mechanoreceptors. Verh. Dtsch. Zool. Ges. 1972, 15–25. Schmidt, M. (1990). Ultrastructure of a possible new type of crustacean cuticular strain receptor in Carcinus maenas (Crustacea, Decapoda). J. Morphol. 204, 335–344. Schmidt, M. and Ache, B. W. (1992). Antennular projections to the midbrain of the spiny lobster. II. Sensory innervation of the olfactory lobe. J. Comp. Neurol. 318, 291–303. Schmidt, M. and Ache, B. W. (1993). Antennular projections to the midbrain of the spiny lobster. III. Central arborisations of motoneurons. J. Comp. Neurol. 336, 583–594. Schmidt, M. and Ache, B. W. (1996). Processing of antennular input in the brain of the spiny lobster, Panulirus argus: I. Non-olfactory chemosensory and mechanosensory pathway of the lateral and median antennular neuropils. J. Comp. Physiol. A 178, 579–604. Schmidt, M., Ekeris, L. V. and Ache, B. W. (1992). Antennular projections to the midbrain of the spiny lobster. I. Sensory innervation of the lateral and medial antennular neuropils. J. Comp. Neurol. 318, 277–290. Schmidt, M. and Gnatzy, W. (1984). Are the funnel-canal organs the campaniform sensilla of the shore crab Carcinus maenas (Crustacea, Decapoda). 2. Ultrastructure. Cell and Tiss. Res. 237, 81–93. Schmitz, B. (1992). Directionality of antennal sweeps elicited by water jet stimulation of the tailfan in the crayfish Procambarus clarkii. J. Comp. Physiol. A 171, 617–627. Schneider, D. (1964). Insect antennae. Ann. Rev. Entomol. 9, 103–122. Schneider, D. and Kaissling, E. (1957). Der Bau der Antenne des Seidenspinners Bombyx mori L. II. Sensillen, cuticula¨re Bildungen und innerer Bau. Zool. Jahrb. Abt. Anat. Ontol. Tiere. 76, 223–250. Schoch, K., Rillich, J., Stevenson, P. A. and Schildberger, K. (2002). Aggression in crickets IV: sensory and motor systems – functional anatomy. Proc. Int. Congr. Neuroethol. 6, 248. Schoch, K., Stevenson, P. A. and Schildberger, K. (2001). Neurochemical architecture of a putative mechanosensory neuropil in the deutocerebrum of the cricket Gryllus bimaculatus. Proc. Go¨ttingen. Neurobiol. Conf. 28, 324. Scho¨ne, H., Neil, D. M., Stein, A. and Carlstead, M. K. (1976). Reactions of the spiny lobster, Palinurus vulgaris, to substrate tilt (I.). J. Comp. Physiol. 107, 113–128. Scho¨ning, M. J., Schroth, P. and Schu¨tz, S. (2000). The use of insect chemoreceptors for the assembly of biosensors on semiconductor field-effect transistors. Electroanal. 12, 645–652. Scho¨ning, M. J., Schu¨tz, S., Schroth, P., WeiXbecker, B., Steffen, A., Kordos, P., Hummel, H. E. and Lu¨th, H. (1998). A BioFET on the basis of intact insect antennae. Sensors Actuat. B 47, 235–238. Schram, F. R. (1986). Crustacea. New York: Oxford University Press.

200


Schroth, P., Scho¨ning, M. J., Lu¨th, H., WeiXbecker, B., Hummel, H. E. and Schu¨tz, S. (2001). Extending the capabilities of an antenna/chip biosensor by employing various insect species. Sensors Actuat. B 78, 1–5. Schu¨rmann, F. W., Ottersen, O. P. and Honegger, H. W. (2000). Glutamate-like immunoreactivity marks compartments of the mushroom bodies in the brain of the cricket. J. Comp. Neurol. 418, 227–239. Schu¨tz, S., Scho¨ning, M. J., Schroth, P., WeiXbecker, B., Schmitz, H. and Bleckmann, H. (2000). An insect-based BioFET as a bioelectronic nose. Sensors Actuat. B 65, 291–295. Schu¨tz, S., Weissbecker, B., Hummel, H. E., Apel, K. H. et al (1999). Insect antenna as a smoke detector. Nature 398, 298–299. Sen, A., Reddy, G. V. and Rodrigues, V. (2003). Combinatorial expression of prospero, seven-up, and elav identifies progenitor cell types during sense-organ differentiation in the Drosophila antenna. Dev. Biol. 254, 79–92. Shepherd, D. and Murphey, R. K. (1986). Competition regulates the efficacy of an identified synapse in crickets. J. Neurosci. 6, 3152–3160. Shields, D. C. and Hildebrand, J. G. (1999a). Fine structure of antennal sensilla of the female sphinx moth, Manduca sexta (Lepidoptera: Sphingidae). I. Trichoid and basiconic sensilla. Can. J. Zool. 77, 290–301. Shields, D. C. and Hildebrand, J. G. (1999b). Fine structure of antennal sensilla of the female sphinx moth, Manduca sexta (Lepidoptera: Sphingidae). II. Auriculate, coeloconic, styliform complex sensilla. Can. J. Zool. 77, 302–313. Shields, V. D. C. and Hildebrand, J. G. (2001). Recent advances in insect olfaction, specifically regarding the morphology and sensory physiology of antennal sensilla of the female sphinx moth Manducasexta. Micros. Res. Tech. 55, 307–329. Siegler, M. V. S. and Pousman, C. A. (1990a). Distribution of motor neurons into anatomical groups in the grasshopper metathoracic ganglion. J. Comp. Neurol. 297, 313–327. Siegler, M. V. S. and Pousman, C. A. (1990b). Motor neurons of grasshopper metathoracic ganglion occur in stereotypic anatomical groups. J. Comp. Neurol. 297, 298–312. Sigvardt, K. A. (1977). Sensory-motor interactions in antennal reflexes of the American lobster. J. Comp. Physiol. A 118, 195–214. Simon, H. R. (1979). Raümliches Orientierungsverhalten durch Antennenbewegungen von epigeisch-hemiedaphischen Collembolen. Dtsch. Ent. Z. 26, 89–100. Slifer, E. H. (1966). Sense organs on the antennal flagellum of a walkingstick Carausius morosus Bru¨nner (Phasmida). J. Morphol. 120, 189–202. Smith, D. (1969). The fine structure of haltere sensilla in the blowfly, Calliphora erythrocephala (Meig.), with scanning electron microscopic observations on the haltere surface. Tissue and Cell 1, 443–484. Snodgrass, R. E. (1935). Principles of Insect Morphology. New York, London: McGraw-Hill Book Company. Snodgrass, R. E. (1956). Anatomy of the Honeybee. Ithaka, London: Cornell University Press. Solon, M. H. and Cobb, J. S. (1980). Antennae-whipping behavior in the American lobster, Homarus americanus. J. Exp. Mar. Biol. Ecol. 48, 217–224. Sperry, R. W. (1950). Neural basis of the spontaneous optokinetic response produced by visual inversion. J. Comp. Physiol. Psychol. 43, 482–489. Spo¨rhase-Eichmann, U., Vullings, H. G. B., Buijs, R. M., Ho¨rner, M. and Schu¨rmann, F. W. (1992). Octopamine-immunoreactive neurons in the central nervous system of the cricket, Gryllus bimaculatus. Cell Tiss. Res. 268, 287–304.


201

Staudacher, E. (1998a). Confocal imaging of antennal afferents and their overlap with arborisations of descending neurons in the cricket. Proc. Go¨ttingen Neurobiol. Conf. 26, 551. Staudacher, E. (1998b). Distribution and morphology of descending brain neurons in the cricket Gryllus bimaculatus. Cell Tiss. Res. 294, 187–202. Staudacher, E. M. (2001). Sensory responses of descending brain neurons in the walking cricket, Gryllus bimaculatus. J. Comp. Physiol. A 187, 1–17. Staudacher, E. and Schildberger, K. (1998). Gating of sensory responses of descending brain neurons during walking in crickets. J. Exp. Biol. 201, 559–572. Staudacher, E. and Schildberger, K. (1999). A newly described neuropile in the deuterocerebrum of the cricket: antennal afferents and descending interneurons. Zoology: Anal. Complex Systems, 102, 212–226. Stay, B., Tobe, S. S. and Bendena, W. G. (1994). Allatostatins – identification, primary structures, functions and distribution. Adv. Ins. Physiol. 25, 267–337. Steinbrecht, R. A. (1997). Pore structures in insect olfactory sensilla: a review of data and concepts. Int. J. Insect Morphol. Embryol. 26, 229–245. Stengl, M., Homberg, U. and Hildebrand, J. G. (1990). Acetylcholinesterase activity in antennal receptor neurons of the sphinx moth Manduca sexta. Cell Tiss. Res. 262, 245–252. Stern, M. (1999). Octopamine in the locust brain: cellular distribution and functional significance in an arousal mechanism. Microsc. Res. Tech. 45, 135–141. Stern, M., Thompson, K. S. J., Zhou, P., Watson, D. G., Midgley, J. M., Gewecke, M. and Bacon, J. P. (1995). Octopaminergic neurons in the locust brain – morphological, biochemical and electrophysiological characterization of potential modulators of the visual-system. J. Comp. Physiol. A 177, 611–625. Stevenson, P. A. (1999). Colocalisation of taurine with transmitter-immunoreactivities in the nervous system of the migratory locust. J. Comp. Neurol. 404, 86–96. Stevenson, P. A. and Spo¨rhase-Eichmann, U. (1995). Localization of octopaminergic neurones in insects. Comp. Biochem. Physiol. B 110, 203–215. Stierle, I. E., Getman, M. and Comer, C. M. (1994). Multisensory control of escape in the cockroach Periplaneta americana. I. Initial evidence from patterns of windevoked behavior. J. Comp. Physiol. A 174, 1–11. Stocker, R. F. (2001). Drosophila as a focus in olfactory research: mapping of olfactory sensilla by fine structure, odor specificity, odorant receptor expression, and central connectivity. Micros. Res. Tech. 55, 284–296. Strausfeld, N. J. (1976). Atlas of an Insect Brain. Berlin: Springer. Strausfeld, N. J. (1998). Crustacean-insect relationships: the use of brain characters to derive phylogeny amongst segmented invertebrates. Brain Behav. Evol. 52, 186–206. Strausfeld, N. J., Bassemir, U., Singh, R. N. and Bacon, J. P. (1984). Organizational principles of outputs from dipteran brains. J. Insect. Physiol. 30, 73–93. Strausfeld, N. J., Hansen, L., Li, Y., Robert, S. G. and Ito, K. (1998). Evolution, discovery, and interpretations of arthropod mushroom bodies. Learning and Memory 5, 11–37. Sun, X. J., Tolbert, L. P. and Hildebrand, J. G. (1993). Ramification pattern and ultrastructural characteristics of the serotonin-immunoreactive neuron in the antennal lobe of the moth Manduca sexta. A laser-scanning confocal and electron-microscopic study. J. Comp. Neurol. 338, 5–16. Suplee, C. (1997). Robot revolution. Natl. Geogr. 192, 81–85. Suzuki, H. (1975). Antennal movements induced by odour and central projection of the antennal neurones in the honey-bee. J. Insect Physiol. 21, 831–847.

202


Tamarelle, M., Belles, X., Piulachs, D. and Girardie, J. (1990). Gastrin-cholecystokininlike and neuroparsin-like immunoreactivities in the brain and retrocerebral neuroendocrine complex of the cockroach Blatella germanica. Histochemistry 93, 433–438. Tamarelle, M., Romeuf, M. and Vanderhaeghen, J. J. (1988). Immunocytochemical localization of gasterin-cholecystokinin-like immunoreactivity in the central nervous system of the migratory locust. Histochemistry 89, 201–207. Tautz, J. (1987). Water vibration elicits active antennal movements in the crayfish, Orconectes limosus. Anim. Behav. 35, 748–754. Tautz, J., Masters, W. M., Aicher, B. and Markl, H. (1981). A new type of water vibration receptor on the crayfish antenna. I. Sensory physiology. J. Comp. Physiol. 144, 533–541. Tautz, J. and Mu¨ller-Tautz, R. (1983). Antennal neuropile in the brain of the crayfish: morphology of neurons. J. Comp. Neurol. 218, 415–425. Tautz, J. and Rohrseitz, K. (1998). What attracts honeybees to a waggle dancer?. J. Comp. Physiol. A 183, 661–667. Taylor, R. C. (1975). Physical and physiological properties of the crayfish antennal flagellum. J. Neurobiol. 6, 501–519. Telford, M. J. and Thomas, R. H. (1998). Expression of homeobox genes shows chelicerate arthropods retain their deutocerebral segment. Proc. Natl. Acad. Sci. USA 95, 10671–10675. Thany, S. H., Lenaers, G., Crozatier, M., Armengaud, C. and Gauthier, M. (2003). Identification and localization of the nicotinic acetylcholine receptor alpha3 mRNA in the brain of the honeybee, Apis mellifera. Insect Mol. Biol. 12, 255–262. Thompson, K. S. and Bacon, J. P. (1991). The vasopressin-like immunoreactive (VPLI) neurons of the locust, Locusta migratoria. II. Physiology. J. Comp. Physiol. A 168, 619–630. Thompson, K. S., Tyrer, N. M., May, S. T. and Bacon, J. P. (1991). The vasopressinlike immunoreactive (VPLI) neurons of the locust, Locusta migratoria. I. Anatomy. J. Comp. Physiol. A 168, 605–617. Thurm, U. (1963). Die Beziehungen zwischen mechanischen Reizgro¨Xen und stationa¨ren Erregungszusta¨nden bei Borstenfeld-Sensillen von Bienen. Z. Vergl. Physiol. 46, 351–382. Thurm, U. (1964). Mechanoreceptors in the cuticle of the honeybee: fine structure and stimulus mechanism. Science 145, 1063–1065. Thurm, U. (1965). An insect mechanoreceptor. Part I: fine structure and adaequate stimulus. Cold Spring Harbor Symp. Quant. Biol. 30, 75–82. Thurm, U. (1983). Biophysics of Mechanoreception. In: Biophysics, (eds Hoppe, W. B. and Bauer, R. D.), pp. 666–671. Berlin, N. Y.: Springer. Thurm, U., Stedtler, A. and Foelix, R. (1975). Stimulus-effective deformations of the terminal structures of a mechanoreceptor. Verh. Dtsch. Zool. Ges. 67, 37–41. Tichy, H. (1979). Hygro- and thermoreceptive triad in the antennal sensillum of the stick insect, Carausius morosus. J. Comp. Physiol. A 132, 149–152. Tichy, H. and Loftus, R. (1983). Relative excitability of antennal olfactory receptors in the stick insect, Carausius morosus L.: in search of a simple, concentrationindependent odor-coding parameter. J. Comp. Physiol. A 152, 459–473. Tichy, H. and Loftus, R. (1987). Response characteristics of a cold receptor in the stick insect Carausius morosus. J. Comp. Physiol. A 160, 33–42. Tichy, H. and Loftus, R. (1990). Response of moist-air receptor on antenna of the stick insect Carausius morosus to step changes in temperature. J. Comp. Physiol. A 166, 507–516. Tichy, H. and Loftus, R. (1996). Hygroreceptors in insects and a spider: humidity transduction models. Naturwissenschaften 83, 255–263.


203

Tierney, A. J., Godleski, M. S. and Massanari, J. R. (2000). Comparative analysis of agonistic behavior in four crayfish species. J. Crust. Biol. 20, 54–66. Toh, Y. (1977). Fine structure of antennal sense organs of the male cockroach, Periplaneta americana. J. Ultrastruct. Res. 60, 373–394. Toh, Y. (1981). Fine-structure of sense organs on the antennal pedicel and scape of the male cockroach, Periplaneta americana. J. Ultrastruct. Res. 77, 119–132. Tominaga, Y. and Yokohari, F. (1981). External structure of the sensillum capitulum, a hygro- and thermoreceptive sensillum of the cockroach, Periplaneta americana. Cell Tiss. Res. 226, 309–318. Torkkeli, P. H. and French, A. S. (1999a). Primary culture of antennal mechanoreceptor neurons of Manduca sexta. Cell Tiss. Res. 297, 301–309. Torkkeli, P. H. and French, A. S. (1999b). Stretch activated ion channels in cultured mechanosensory neurons of Manduca sexta. Biophys. J. 76, A203. Towne, W. F. and Kirchner, W. H. (1989). Hearing in honeybees: detection of airparticle oscillations. Science 244, 686–688. Trachtman, P. (2000). Redefining robots. Smithsonian 30, 97–112. Tregenza, T. and Wedell, N. (1997). Definitive evidence for cuticular pheromones in a cricket. Anim. Behav. 54, 979–984. Tsujimura, T. and Yabuta, T. (1992). A tactile sensing method employing force/torque information through insensitive probes. Proc. IEEE Int. Conf. Robot. Autom. 1992, 1315–1320. Tyrer, N. M., Pozza, M. F., Humbel, U., Peters, B. H. and Bacon, J. P. (1988). The tritocerebral commissure ‘‘dwarf’’ (TCD): a major GABA-immunoreactive descending interneuron in the locust. J. Comp. Physiol. A 164, 141–150. Tyrer, N. M., Turner, J. D. and Altman, J. S. (1984). Identifiable neurons in the locust central nervous system that react with antibodies to serotonin. J. Comp. Neurol. 227, 313–330. Ueno, N. and Kaneko, M. (1995). On a new contact sensing strategy for dynamic active antenna. Proc. IEEE Int. Conf. Robot. Autom. 1995, 1120–1125. Ueno, N., Kaneko, M. and Svinin, M. (1996). Theoretical and experimental investigation on dynamic active antenna. Proc. IEEE Int. Conf. Robot. Autom. 1996, 3557–3563. Ueno, N., Svinin, M. M. and Kaneko, M. (1998). Dynamic contact sensing by flexible beam. IEEE-ASME Trans. Mechatron. 3, 254–264. Urvoy, J., Fudalewicz-Niemczyk, W., Petryszak, A. and Rosciszewska, M. (1984). Etude des organs sensoriels externes de l’antenne de Phasme Carausius morosus B. (Cheleutoptera). Acta Biol. Cracoviensia Ser.: Zool. XXVI, 57–68. vande Berg, J. S. (1971). Fine structural studies of Johnston’s organ in the Tobacco Hornworm moth, Manduca sexta (Johannson). J. Morphol. 133, 439–456. Vedel, J. P. (1980). The antennal motor system of the rock lobster: competitive occurrence of resistance and assistance reflex patterns originating from the same proprioceptor. J. Exp. Biol. 87, 1–22. Vedel, J. P. (1985). Cuticular mechanoreception in the antennal flagellum of the rock lobster Palinurus vulgaris. Comp. Biochem. Physiol. A 80, 151–158. Vedel, J. P. and Monnier, S. (1983). A new muscle receptor organ in the antenna of the rock lobster Palinurus vulgaris: mechanical, muscular and proprioceptive organization of the two proximal joints JO and J1. Proc. R. Soc. Lond. B 218, 95–110. Verhaert, P., Grimmelikhuijzen, J. P. and Deloof, A. (1985). Distinct localization of FMRFamide-like and bovine pancreatic polypeptide-like material in the brain, retrocerebral complex and suboesophageal ganglion of the cockroach Periplaneta americana L. Brain Res. 348, 331–338.

204


Vitzthum, H., Homberg, U. and Agricola, H. (1996). Distribution of dip-allatostatin I-like immunoreactivity in the brain of the locust Schistocerca gregaria with detailed analysis of immunostaining in the central complex. J. Comp. Neurol. 369, 419–437. Vitzthum, H., Homberg, U., Wu¨rden, S. and Agricola, H. (1995). Allatostatin-like immunoreactivity in the central complex of the locust Schistocerca gregaria. Proc. Go¨ttingen. Neurobiol. Conf. 23, 598. Waldow, U. (1975). Multimodal neurons in deutocerebrum of Periplaneta americana. J. Comp. Physiol. 101, 329–341. Waldow, U. (1977). CNS units in cockroach (Periplaneta americana) – specificity of response to pheromones and other odor stimuli. J. Comp. Physiol. 116, 1–17. Watson, A. H. and Burrows, M. (1983). The morphology, ultrastructure, and distribution of synapses on an intersegmental interneuron of the locust. J. Comp. Neurol. 214, 154–169. Watson, A. H. and Pflu¨ger, H. J. (1989). Regional specialisation for synaptic input and output on a locust intersegmental interneurone with multiple spike-initiating zones. J. Comp. Neurol. 279, 515–527. Watson, J. T., Ritzmann, R. E., Zill, S. N. and Pollack, A. J. (2002). Control of obstacle climbing in the cockroach, Blaberus discoidalis. I. Kinematics. J. Comp. Physiol. A 188, 39–53. Weber, H. and Weidner, H. (1974). Grundriss der Insektenkunde. Stuttgart: Gustav Fischer. Weide, W. (1960). Einige Bemerkungen u¨ber die antennalen Sensillen sowie u¨ber das Fu¨hlerwachstum der Stabheuschrecke Carausius (Dixippus) morosus Br. (Insecta: Phasmida). Wiss. Z. Univ. Halle Math.-Nat. IX/2, 247–250. Wendelken, P. W. and Barth, R. H. (1987). The evolution of courtship phenomena in neotropical cockroaches of the genus Blaberus and related genera. Adv. Ethol. 27, 1–98. Wendler, G. (1965). U¨ber den Anteil der Antennen an der Schwererezeption der Stabheuschrecke Carausius morosus Br. Z. Vergl. Physiol. 51, 60–66. Wendt, B. and Homberg, U. (1992). Immunocytochemistry of dopamine in the brain of the locust Schistocerca gregaria. J. Comp. Neurol. 321, 387–403. Wilkens, L. A., Schmitz, B. and Herrnkind, W. F. (1996). Antennal responses to hydrodynamic and tactile stimuli in the spiny lobster Panulirus argus. Biol.Bull. 191, 187–198. Williams, L. D. (1975). Anatomical studies of the insect central nervous system: a ground-plan of the midbrain and an introduction to the central complex in the locust Schistocerca gregaria (Orthoptera). J. Zool. Lond. 176, 67–86. Wolf, H. and Burrows, M. (1995). Proprioceptive sensory neurons of a locust leg receive rhythmic presynaptic inhibition during walking. J. Neurosci. 15, 5623–5636. Wolf, H. and Laurent, G. (1994). Rhythmic modulation of the responsiveness of locust sensory local interneurons by walking patter generating networks. J. Neurophysiol. 71, 110–118. Yagodin, S. V. (1980). Role of antenna in flight maintenance of the cockroach Periplaneta americana. J. Evol. Biochem. Physiol. 16, 30–36. Yam, P. (1998). Roaches at the wheel. Sci. Am. 278, 45. Ye, S. and Comer, C. M. (1996). Correspondence of escape-turning behavior with activity of descending mechanosensory interneurons in the cockroach, Periplaneta americana. J. Neurosci. 16, 5844–5853. Ye, S., Leung, V., Khan, A., Baba, Y. and Comer, C. M. (2003). The antennal system and cockroach evasive behavior. I. Roles for visual and mechanosensory cues in the response. J. Comp. Physiol. A 189, 89–96.


205

Yildiz, F. and Gebhardt, M. (2003). Complex innervation of the antennal base of crickets. Proc. Go¨ttingen. Neurobiol. Conf. 29, 238. Yokohari, F. (1981). The sensillum capitulum, an antennal hygro- and thermoreceptive sensillum of the cockroach, Periplaneta americana L. Cell Tiss. Res. 216, 525–543. Young, D. (1970). The structure and function of a connective chordotonal organ in the cockroach leg. Philos. Trans. R. Soc. Lond. B, 256, 401–426. Younossi-Hartenstein, A. and Hartenstein, V. (1997). Pattern, time of birth, and morphogenesis of sensillum progenitors in Drosophila. Microsc. Res. Tech. 39, 479–491. Zacharuk, R. Y. (1985). Antennae and Sensilla. In: Comprehensive insect physiology, biochemistry and pharmacology, (eds Kerkut, G. A. and Gilbert, L. I.), pp. 1–69. Oxford, NY: Pergamon Press. Zeil, J., Sandeman, R. and Sandeman, D. C. (1985). Tactile localisation: the function of active antennal movements in the crayfish Cherax destructor. J. Comp. Physiol. A 157, 607–617. Zeiner, R. and Tichy, H. (1998). Combined effects of olfactory and mechanical inputs in antennal lobe neurons of the cockroach. J. Comp. Physiol. A 182, 467–476.

Eupyrene and Apyrene Sperm: Dichotomous Spermatogenesis in Lepidoptera Michael Friedla¨ndera, Rakesh K. Sethb and Stuart E. Reynoldsc a

Department of Life Sciences, Ben Gurion University of the Negev, Beer Sheva, Israel Department of Zoology, University of Delhi, Delhi, India c Department of Biology & Biochemistry, University of Bath, Bath, UK b

1 2

3

4

5

6

7

8 9

Introduction: dichotomous spermatogenesis 207 1.1 What is dichotomous spermatogenesis? 207 1.2 Eupyrene and apyrene sperm in Lepidoptera 209 Spermatogenesis in Lepidoptera 214 2.1 The testes 214 2.2 Spermatogonia 218 2.3 Spermatocytes 219 Dichotomous spermatogenesis in Lepidoptera 220 3.1 How eupyrene and apyrene spermatogenesis differ 220 3.2 The cellular mechanism of apyrene spermatogenesis 226 3.3 Duration of eupyrene and apyrene divisions 230 Spermiogenesis 231 4.1 Formation of eupyrene and apyrene spermatozoa 231 4.2 Peristaltic squeezing 234 4.3 Surface structures of lepidopteran sperm 235 Regulation of dichotomous spermatogenesis 236 5.1 Control of mitosis and meiosis 236 5.2 The switch from eupyrene to apyrene spermatogenesis 248 5.3 Control of spermiogenesis 251 5.4 Control of spermatogenesis during diapause 253 Sperm movement and transfer 255 6.1 Release of eupyrene and apyrene sperm from the testis into the male tract 255 6.2 Descent of eupyrene and apyrene sperm along the male tract 259 6.3 Transfer of eupyrene and apyrene sperm to the female 262 6.4 Sperm retention by female moths and effects on female sexual behaviour 268 Behaviour of eupyrene and apyrene sperm in the female 269 7.1 Sperm migration from bursa to spermatheca 269 7.2 Sperm sorting in the spermatheca 271 7.3 Loss of sperm from the spermatheca 273 Sperm maturation 274 8.1 Sperm activation 274 The evolutionary rationale of dichotomous spermatogenesis 283

ADVANCES IN INSECT PHYSIOLOGY VOL. 32 ISBN 0-12-024232-X DOI: 10.1016/S0065-2806(05)32003-0


EUPYRENE AND APYRENE SPERM

9.1 Evolution of apyreny 283 9.2 Possible functions of apyrene sperm 10 Conclusion 291 Acknowledgements 292 References 293

1 1.1

207

284

Introduction: dichotomous spermatogenesis WHAT IS DICHOTOMOUS SPERMATOGENESIS?

Dichotomous spermatogenesis is a process by which all the males of a given species produce two kinds of concomitant spermatozoa that differ markedly in their pathways of differentiation, structure, and DNA content. One kind is the typical haploid sperm that fertilises the eggs and is similar, in principle, to the generalised type of spermatozoa found in most animal species (e.g. Baccetti, 1991). The other kind, sometimes called parasperm (Jamieson, 1987), is atypical in both form and function, either being anucleated or having an unbalanced set of chromosomes; this type of spermatozoon does not fertilise the egg and its function(s) in different systematic groups is still unclear. Dichotomous spermatogenesis produces only these two kinds of spermatozoa and there are no intermediate morphs between them. It is hardly possible to imagine a trait that would at first sight appear to be more costly to fitness than the production of anucleate sperm. Consequently, it is not surprising that the evolution of dichotomous spermatogenesis and the selective forces that have driven this are topics of considerable interest to evolutionary biologists (e.g. Swallow and Wilkinson, 2002). Notwithstanding that dichotomous spermatogenesis produces these two concomitant morphs of spermatozoa, it must be stressed that it is a normal process, intrinsic to the species concerned. This type of spermatogenesis that leads to two kinds of spermatozoa, one the regular haploid and the other with reduced content of DNA, or even no DNA at all, is genetically controlled and occurs in a regular and predictable way. Dichotomous spermatogenesis is therefore essentially different from abnormal spermatogeneses that produce aberrant spermatozoa of unpredictable and variable nature. Such abnormal spermatogeneses occur sporadically in isolated individuals of a given species, either as a result of infectious disease, abnormal genes, or under irregular physiological, developmental or environmental conditions (e.g. Sorour and Larink, 2001; Murugavel et al., 2002; Chang et al., 2004). The normal occurrence of dichotomous spermatogenesis is well documented for a large number of invertebrate species belonging to phylogenetically distant taxa, such as insects (Tazima, 1967), lumbricids (Boi et al., 2001), rotifers (Koehler and Birky, 1966), snails (Hodgson, 1997), opilionids (Juberthie et al., 1976), and sea urchins (Au et al., 1998), among others (Fain-Maurel, 1966). Dichotomous spermatogenesis has even been reported in some vertebrates,

208

¨ NDER ET AL M. FRIEDLA

fish, for example (Hayakawa et al., 2002). Moreover, we think it would not be unreasonable to assume that other cases of dichotomous spermatogenesis have gone unreported, and that the phenomenon is actually even more widespread than has been documented in the literature. Some investigators may be unaware of the existence of dichotomous spermatogenesis and simply dismiss it as ‘‘abnormal’’. Indeed, one of the purposes of writing this review is to increase awareness of dichotomous spermatogenesis as a normally occurring phenomenon. On the other hand, there are also examples of animal species in which normal spermatogenesis produces more than one sperm morph, but which do not qualify as ‘‘dichotomous spermatogenesis’’, in the sense that we use the term here. In these cases, such as that reported for Drosophila species of the obscura group (Joly and Lachaise, 1994), the sperm morphs are classified mainly by their size and they are functionally less extremely different than those resulting from the ‘‘true’’ dichotomous spermatogenesis. Although in these cases, the spermatozoa of a given species show two or more coexisting size classes, the spermatozoa belonging to any of those classes have a common structure, organelle distribution, and DNA content, regardless of whether they are larger or smaller (Takamori and Kurokawa, 1986; Pasini et al., 1996). Nevertheless, it seems that despite this apparent similarity, the size classes do differ in their ability to fertilise the eggs (Swallow and Wilkinson, 2002). Thus, in Drosophila subobscura and several other drosophilid species, it is the long spermatozoa that preferentially fertilise the eggs and the short spermatozoa of these species undertake fertilisation only after the long ones have been depleted (Snook et al., 1994; Bressac and Hauschteck-Jungen, 1996; Snook, 1997; Snook and Karr, 1998). Rhabditid nematodes also produce two size classes of spermatozoa. In this case too, large spermatozoa outcompete small spermatozoa within the female genital tract during fertilisation. However, in these nematodes, each distinct type of sperm is actually produced by a different kind of individual of the same species; the small spermatozoa are produced by hermaphrodite individuals, while the larger ones are produced by males (LaMunyon and Ward, 1998). This system is therefore essentially different from dichotomous spermatogenesis as we have defined it, and in which both of the two types of spermatozoa are produced by the same testes but only one of them is able to fertilize the eggs. The distinctive features of dichotomous spermatogenesis are characteristic of the species, varying not only among systematic groups, but even between species of the same group. Following Holmes (1979), the typical haploid spermatozoa are here called eupyrene (eu ¼ true; pyren ¼ fruit stone or nucleus), while the atypical ones ð¼ paraspermÞ are classified as (a) apyrene ða ¼ notÞ, completely lacking a nucleus, as found in lepidopterans and rotifers, (b) oligopyrene ðoligo ¼ fewÞ, possessing a nucleus but having fewer chromosomes than the haploid spermatozoon, as found in snails and some pentatomid


209

bugs and (c) hyperpyrene ðhyper ¼ aboveÞ, possessing a nucleus but having more chromosomes than the haploid spermatozoon, as found in some other bugs and carabid beetles (Fain-Maurel, 1966). 1.2

EUPYRENE AND APYRENE SPERM IN LEPIDOPTERA

The best-known case of dichotomous spermatogenesis is that of the Lepidoptera. The fact that moths and butterflies produce two kinds of sperm was first reported by Meves (1903), who in a beautiful paper published at the very beginning of the twentieth century accurately described the process at the cytological level in considerable detail (Figs. 1 and 2). Although Katsuno (1977a) has asserted that Toyama’s (1894) study of spermatogenesis in the commercial silkmoth Bombyx mori was the first publication to note the dichotomous spermatogenesis of Lepidoptera, careful examination of both text and figures of Toyama’s paper reveals no sign that he recognised either the distinction between eupyrene and apyrene meiosis, or the subsequent presence in the testis of two kinds of sperm. In the same year as Meves’s paper appeared, Voinov (1903) published a brief and apparently independent note in which he observed that a number of butterflies produced spermatocytes of two different kinds, and that one of these (the smaller) underwent an aberrant form of meiosis. However, Voinov did not report that the smaller spermatozoa ultimately discarded their irregularly formed nuclei and became anucleate. Both authors struggled to understand the function of apyrene sperm. Meves recognised that if an apyrene sperm lacking a nucleus were to fertilise an egg, the resulting progeny could express only maternal characteristics. Voinov considered the possibility that apyrene sperm might be incapable of fertilisation, but rejected as ‘‘peu probable’’, the idea that a large part of the insect’s sperm would play no part in fertility. Instead, Voinov suggested, wrongly, that the FIG. 1 A reproduction of the original figure (Plate VI), from Meves (1903) showing apyrene spermatogenesis in the testis of the notodontidid, Pygaera ð¼ PhaleraÞ bucephala (the buff tip). Note particularly the early elaboration of the flagella of the prospective spermatids, all four of which can already be seen at a stage that considerably precedes the dissolution of the nuclear membrane of the first meiotic prophase (121). Meves clearly recognised the lack of a real metaphase plate in apyrene spermatocytes, and illustrated the irregular distribution of the chromosomes in anaphase (125). Note that in this species, the apyrene chromosomes form micronuclei at both the first telophase (128, 129) and the second telophase (134), and remain in this condition in the spermatid (135–142). The apyrene spermatid contains a dark-staining cone, which appears in the sperm head in place of the nucleus that would be present there in its eupyrene counterpart. Note also that the mitochondria assemble together before the first division (126–127), surround the spindle (128), and separate again towards the second meiotic division (131–133); they now coalesce to form the nebenkern of the spermatid (135–137), finally elongating to form the mitochondrial derivatives that are found alongside the axoneme of the mature sperm apyrene sperm.

210



211

FIG. 2 A reproduction of the original figure (Plate VII), from Meves (1903) showing the intratesticular development of apyrene spermatozoa (143–151) and the developed intratesticular form of a single eupyrene bundle in P. bucephala. In this figure Meves described the progress of enucleation as it occurs during the development of apyrene sperm. In (147) and (148), the micronuclei move down the apyrene bundle towards the tail, and in (150) they are shown being ejected from the cells at the tail end of the cyst. Each bundle is shown to be enveloped by a small number of flattened cyst cells with large nuclei. In the eupyrene cyst (152), note the nucleus of the head cyst cell. The head of each eupyrene spermatozoon (containing its nucleus) is individually located within a deep tubular invagination of the head cell.

two different sperm morphs might determine the sex of the fertilised egg, males being produced by one kind of sperm, and females by the other. The difficulty of understanding why male moths and butterflies would produce sperm without nuclei nevertheless meant that the phenomenon of lepidopteran sperm dichotomy remained an obscure curiosity for more than half a century, until being ‘‘discovered’’ again in the 1960s. Indeed, during this intervening period some investigators (e.g. Munson, 1906) continued to deny the existence of apyrene sperm, while others (e.g. Machida, 1929; Oˆmura, 1936b; Sado, 1963) although recognising that dichotomous spermatogenesis occurred in Lepidoptera, erroneously asserted that the resulting apyrene sperm did not ordinarily leave the testis. Today, however, the phenomenon of lepidopteran apyrene spermatogenesis and the fact that both eupyrene and apyrene sperm are transferred to the female are firmly established. Dichotomous spermatogenesis has been found in

212


almost every lepidopteran species that has been studied, including members of basal systematic groups of the order (Friedla¨nder, 1983; Kristensen, 1984; Hamon and Chauvin, 1992). The only known exceptions are two species of the genus Micropterix, belonging to the most primitive extant lepidopteran suborder, Zeugloptera (Sonnenschein and Haüser, 1990). It therefore appears that the dichotomous spermatogenesis that characterises Lepidoptera is an evolutionary novelty of the order, which has developed independently from analogous systems found in other systematic groups of the animal kingdom. The following observations support this assertion: (1) eupyrene–apyrene dichotomous spermatogenesis already exists in the most primitive extant lepidopteran suborder, the Zeugloptera (Friedla¨nder, 1983); while (2) only regular eupyrene spermatozoa are produced in the closely related orders of Diptera and Siphonaptera (Jamieson, 1987). But most significantly, (3) apyrene spermatogenesis is absent throughout the Trichoptera (Friedla¨nder, 1983), the sister order of the Lepidoptera, together with which it forms the superorder Amphiesmenoptera (Hennig, 1981). Swallow and Wilkinson (2002) have proposed that the presence of both dichotomous and non-dichotomous spermatogenetic species within the family Micropterigidae means that dichotomous spermatogenesis must have evolved more than once within the Lepidoptera. However, they did not apparently consider the possibility that this character might have been secondarily lost in some members of this basal family. To help solve this question, it will be necessary to study the spermatogenesis of additional zeuglopteran species. How did dichotomous spermatogenesis evolve? We suggest that a less extreme type of sperm dimorphism, such as those described above for drosophilids and nematodes, was ancestral to the evolution of ‘‘true’’ sperm dichotomy in the Lepidoptera. In this scenario, an ancestral lepidopteran species with intraspecific sperm classes that had the same DNA content, but which differed in both their dimensions and their competitive ability to fertilise the eggs, is supposed to have undergone successive genetic events that widened the dissimilarities between the corresponding characters of the two sperm size classes, the final result being two classes of spermatozoa that now have completely different morphology and behaviour, the two kinds being either fertile or sterile, and the sterile class completely lacking nuclear DNA. We will suggest below that a single genetic event affecting the control of chromosome pairing underlay the transition to true sperm dichotomy. We suggest that such extreme dichotomy of sperm types could not have evolved unless it conferred some adaptive benefit. Wolf (1994) has previously suggested that there is no such benefit, describing the origin of lepidopteran sperm dichotomy as being ‘‘degenerative evolution at the cellular level’’ resulting from the previous loss of function, and consequent non-use of the apparatus of chromosome separation in meiosis. Wolf does not deny the possibility that the original emergence of sperm dichotomy may have been adaptive, but does suggest that its present form in Lepidoptera confers no


213

additional advantage, and draws the analogy with the loss of eyes and body pigmentation that occurs in cave-dwelling animals. He contrasts the failure of chromosome separation in apyrene spermatogenesis with cytokinesis, which continues to be performed normally in the same cells, implying that the latter function is still required and therefore maintained by natural selection in order to supply the sterile spermatozoa needed by the moths to ensure fertility. We must point out, however, that if apyreny is indeed due to the loss of function of some gene essential for meiosis, then this loss of function is specific to the apyrene line since eupyrene divisions continue to occur normally in the same individual. The implication is not that a simple loss of function mutation has occurred, but rather that the mutation concerns a discriminative change of regulation of that gene’s expression through the span of life of the individual. Further, the proportions of eupyrene and apyrene spermatozoa differ remarkably between lepidopteran species, varying from 11% apyrene in the hepialid Hepialus behrensi, to 99% apyrene in the sphingid Pachysphinx modesta (Swallow and Wilkinson, 2002). This extreme variability suggests that dichotomous spermatogenesis continues to be under active selection. We will return later to the questions of the possible function of apyrene spermatogenesis and how it arose. In reviewing the relevant literature, it is hard to escape the conclusion that the existence of a heterogeneous eupyrene–apyrene sperm population has sometimes been overlooked when dealing with basic or applied questions of lepidopteran reproduction. This omission is of crucial importance because it has now been indisputably shown (Sahara and Kawamura, 2002) that in the commercial silkmoth B. mori, the transfer of apyrene spermatozoa to the genital tract of the inseminated female is essential for the fertilization of the eggs by the eupyrene spermatozoa. Unfortunately, such experiments are difficult to accomplish, but we suggest that this essential role for apyrene sperm will also be found to be true throughout the order. Despite this demonstration of the crucial reproductive importance of apyrene spermatozoa, however, it is still unknown how these anucleate sperm perform their reproductive function. Moreover, the two kinds of spermatozoa differ profoundly in their (1) relative counts in the ejaculate, (2) hormonal control and timing of differentiation, (3) resistance to chemical, endocrine, physical, and genetic manipulations, (4) behaviour within the male and female genital tracts, and (5) responses to sperm competition, etc. (Friedla¨nder, 1991). Disregarding these differences would necessarily lead to erroneous conclusions in both basic and applied experimentation. Thus, our second major purpose in writing this review is to make clear the importance for lepidopteran reproductive biology of the dichotomous spermatogenesis universally found in this order. We deal here with (a) structural differences between the two kinds of spermatozoa and their distinctive modes of differentiation, (b) developmental mechanisms underlying, and hormonal systems regulating dichotomous spermatogenesis, (c) transformations undergone by the two spermatozoa types during their passage through the male and


214

female genital tracts and (d) evaluation of the diverse functions that have been proposed for the apyrene spermatozoa.

2

Spermatogenesis in Lepidoptera

2.1 2.1.1

THE TESTES

Anatomy of the testes

Lepidopterans have two testes showing, in principle, the architecture that characterises this organ throughout the Insecta and other arthropod groups. In the imago, the paired testes may either remain separate, as in Hepialidae and Zeugloptera (Matsuda, 1976) or fuse together into a single structure as has been described for the sphingid, Manduca sexta (Reinecke et al., 1983) and the noctuid Heliothis virescens (LaChance and Olstad, 1988). Each testis is kidney- or bean-shaped and contains four pyriform lobuli or follicles (Fig. 3). Each follicle has a large, ‘‘free’’ distal, blind end situated on the convex side of the testis and a smaller, ‘‘open’’ proximal end located at the concave side of the testis. The proximal end is prolonged into a separate vas efferens and the four vasa efferentia of each testis fuse to form one of the two paired vasa deferentia. The germarium, located at the distal end of each follicle, produces spermatogonia throughout the span of the insect’s life, even during diapause during which other spermatogenetic stages may disappear, as will be detailed later. 2.1.2

The Verson’s cell

The germarium contains a giant apical cell, first described by Verson (1889), and commonly designated as Verson’s cell (Toyama, 1894; Carson, 1945). This cell is surrounded by, and is in close contact with, a number of primary spermatogonia. There are, however, no cytoplasmatic bridges between the apical cell and these primary spermatogonia; the possible significance of this will emerge below. The apical Verson’s cell is already present at the time of hatching and persists throughout the larval stages of development. In some species, the apical cell disappears in the late pupa or early imago (Witalis and Godula, 1993), while in other species it persists in the adult, although becoming smaller than at the earlier ontogenetic stages (Sonoli and Hooli, 1992). Verson (1889) had originally suggested that the apical cell was a germ cell, but Toyama (1894) showed that it is derived from the follicular epithelium, and is not from the germline. He instead suggested that it was a ‘‘supporting cell, connecting all the younger genital elements with the wall of the testicular follicle and probably nourishing them’’. The apical Verson’s cell accumulates packages of glycogen particles (Wolf, 1991) and appears to be actively involved in phagocytosis and RNA synthesis.


215

FIG. 3 Male reproductive system of Bombyx mori. (A) Schematic figure of male reproductive tract of a pupa 1 or 2 days before emergence. add, ampulla ductus deferentis; dd, ductus deferens; ga, glandula alba; gl, glandula lacteola; gp, glandula prostatica; gpl, glandula pellucida; gs, glandula spermatophorae; p, penis; t, testis; vs, vesicula seminalis. Approximately 2. From Oˆmura (1938). (B) Diagrammatic representation of the structure of the testis of a full-grown B. mori larva. cl, capsula lobuli; dd, ductus deferens; de, ductulus efferens; l, lobulus ð¼ follicleÞ; mc, membrana communis; mb, membrana basilaris; te, tunica externa; ti, tunica interna. Approximately 30. From Oˆmura (1936b).

However, the significance of these cellular activities is unknown, and the few published results of experiments designed specially to clarify this question (Leclercq-Smekens, 1978), have been considered inconclusive (Szo¨llo¨si, 1982). So, the function of the lepidopteran testis apical cell still remains an open question. Wolf (1991) suggested that the function of the Verson’s cell is to maintain the primary spermatogonia, which he called ‘‘prespermatogonia’’ as germline stem cells (GSC). This suggestion is based on the observation that the Verson’s cell maintains close contact with the primary spermatogonia, but not with other cells and that numerous projections of the primary spermatogonia are located in invaginations of the Verson’s cell. This suggests the transfer of


216

material, or at least signals, between them, although no direct coupling structures are evident. When a primary spermatogonium divides, one daughter cell remains in contact with the apical cell and continues to function as a stem line primary spermatogonium. The other daughter cell or gonioblast is eventually surrounded by somatic cells that form the wall of a closed cyst ð¼ spermatocystÞ. Accordingly, now deprived of contact with the Verson’s cell, this newly divided spermatogonial cell loses its stem line status and proceeds to differentiate. Wolf (1991) suggested that the ‘‘Verson’s cells may create an extracellular microenvironment controlling the further development’’ of the primary spermatogonia, conferring ‘‘stemness’’ on the latter through this close contact. Consistent with this, once away from the Verson’s cell and within the cyst, the spermatogonium now undergoes a predetermined number of synchronous mitotic divisions, eventually producing a clone of 2n cells. This interpretation is consistent with the finding that in the testis of Drosophila melanogaster specialised somatic cells provide a signal that maintains the adjacent GSC by causing their mitotic divisions to be self-renewing (Kiger et al., 2001; Tulina and Matunis, 2001). In Drosophila, the maintenance of GSC depends on the secretory activity of a number of non-mitotic somatic support cells (the ‘‘hub’’) located at the apex of this insect’s tubular testis. The hub cells produce a protein signal, Unpaired (Upd), which interacts with a receptor, Domeless (Dom) that is expressed specifically in the neighbouring GSC. In turn, this activates a Janus Kinase–Signal Transducer and Activator of Transcription (JAK–STAT) signalling pathway, involving the gene products Hopscotch (Hop) and Stat92E, the latter being a nuclear transcription factor that presumably promotes the expression of target genes involved in the GSC self-renewal response. Mutants of genes in this pathway lead to male sterility consequent upon the failure of GSC to self-renew. It seems probable that a similar signalling pathway operates in the lepidopteran testis, the fly’s hub cells being functional homologues of the moth’s Verson’s cells. 2.1.3

Cysts

The division of a primary spermatogonial cell, in addition to maintaining the GSC status of one of the progenies, produces a gonioblast that initiates a series of divisions, giving rise to a clone of gonial cells. The clone is contained within a hollow cyst made of flat, thin somatic cells. In the testicular follicles of most insects, the cysts are spatially organised in a polarised way, reflecting the temporal progression of spermatogenesis. This is also true of Lepidoptera, in which those cysts that contain the earlier stages of spermatogenesis are also the closest to the apical cell, while those containing mature spermatozoa are the closest to the vas efferens, into which they will eventually be released. As we shall see below, divisions of primary spermatogonial cells occur continuously from late embryo to imago and there is therefore an uninterrupted production of new cysts throughout life (Friedla¨nder, 1982).


217

In lepidopterans, the cyst wall consists of several similar large, thin cells and a single ‘‘head cyst cell’’, that is located at the anterior end of those cysts that contain elongated bundles of either differentiating spermatids or spermatozoa. The head cyst cell has a highly differentiated, bulky configuration and the area of its surface that faces the spermatid heads displays deep tubular invaginations. Within the cyst, the head of each elongating spermatid or spermatozoon is individually tightly held within one of these invaginations. Thus, the head cyst cell is analogous to the Sertoli cell of the vertebrate testis, in that it maintains close contact with the maturing spermatozoa embedded in it. In mammals, the interaction between the maturing germ cells and the Sertoli cell has been shown to be dependent on the presence of complex carbohydrates (asparagine-linked glycans) on the germ cell surface. Their synthesis requires a testis-specific mannosidase that is found only in spermatocytes and early spermatids. Mice in which the gene encoding this enzyme (Man2a2) had been knocked out by targeted mutation and homologous recombination showed greatly reduced fertility due to the premature release of spermatogenic cells from the testis (Akawa et al., 2002). It is not known whether the interactions of maturing lepidopteran spermatids with the head cyst cell are similarly mediated by cell surface interactions but this could be so, since a late stage of intratesticular sperm maturation (peristaltic squeezing – see below) depends in Lepidoptera on the ability of the head cyst cell to hold the bundle of maturing spermatozoa together. In cysts containing cells of the eupyrene line, the head cyst cell shows an increasing amount of cytoplasmic lysine-rich proteins during the nucleoprotein transition undergone by the spermatid. No such phenomenon has been observed in the comparable apyrene cysts and, consequently, it has been proposed that this differential behaviour may be related in some way to the regulation of dichotomous spermatogenesis (Friedla¨nder and HauschteckJungen, 1982b). 2.1.4

The blood– germ cell barrier

The four follicles of each of the paired testes are enwrapped together by a sheath consisting of two layers, the internal tunic and the external tunic (Fig. 3). The internal tunic folds inwards forming three septa, each composed of two adjacent, symmetrical layers that are oppositely oriented, back-to-back and located between two contiguous follicles. The external tunic covers the entire outer surface of the testes, without penetrating between the follicles (Szo¨llo¨si et al., 1980). There is a long-standing suggestion (Kambysellis and Williams, 1971a,b) that the permeability properties of the testicular sheath are important in regulating spermatogenesis during diapause in saturniid and other moths. These authors suggested that during pupal diapause (and only during diapause), the sheath becomes impermeable to an unknown haemolymph trophic factor


218

named ‘‘Macromolecular Factor’’ (MF), which promotes spermatogenesis. The supposed impermeability of the sheath to MF was suggested to be the cause of the apparent cessation of spermatogenesis that characterises saturniid pupal diapause. As we discuss below, in Lepidoptera this ‘‘cessation’’ during diapause does not in fact represent a complete halting of spermatogenesis, but rather a block to the progression of the spermatogenetic cells beyond a species-specific stage of their differentiation. This block has been documented to be at meiotic prophase for the tortricid Cydia pomonella, while in the sphingid M. sexta it is at the stage of the elongating spermatid. The supposed trophic factor would thus act to remove this block. The key experimental result of Kambysellis and Williams (1971a,b) was that explanted diapausing testes renewed production of sperm when ecdysteroid moulting hormones were added to the culture medium. They interpreted these results to mean that that the ecdysteroids had rendered the sheath permeable again to the proposed MF. However, the question of the permeability of the sheath to MF has never been directly examined and the nature of MF remains obscure. In any case, it has since been shown that addition of ecdysteroid to the culture medium of diapausing testes promotes renewal of spermatogenesis most probably by sustaining the functional integrity of the sheath cells and by enhancing their glycogen metabolism (Friedla¨nder, 1989), rather than by increasing sheath permeability to the postulated MF. In addition to this possible separation by its sheath of the whole testis from putative trophic factors in the haemolymph, however, there is in Lepidoptera a real ‘‘blood–germ cell barrier’’ that selectively isolates the germ cells from haemolymph factors, as in other insects (Jones, 1978). This barrier functions in a way similar to that of the ‘‘blood–testis barrier’’ found in mammalian testes (Lui et al., 2003). In lepidopterans, the barrier is located beyond the internal tunic, at the level of the cyst wall, as has been shown by using electron-opaque tracers such as horseradish peroxidase, injected into the haemolymph. The injected tracer, a globular protein of 40 kDa molecular weight and 60 A˚ molecular diameter, penetrates the testicular sheath and the follicular envelope but not the cyst wall, showing that large molecules can enter certain parts of the testes but are excluded from, or enjoy only restricted access to other parts of the organ. This is at least in part due to the presence of septate and other tight junctions between the somatic cells forming the cyst wall (Szo¨llo¨si, 1982). 2.2

SPERMATOGONIA

The dichotomous spermatogenesis of Lepidoptera follows a generalised pattern of spermatogenetic proliferation and meiosis throughout the order. Spermatogonia begin to divide even before the larva hatches from the egg. Late embryonic testes of the bombycid B. mori already show the four developing follicles containing mitotically dividing spermatogonia (Takeuchi, 1969). In this and other lepidopterans, the testes of first instar larvae already


219

display encysted spermatogonia and spermatogonial proliferation continues uninterruptedly throughout the rest of the insect’s life. Within a cyst, all the differentiating germ cells remain interconnected by cytoplasmic bridges and they divide and develop synchronously throughout spermatogenesis. The structure of these cytoplasmic bridges (‘‘polyfusosomes’’) has been studied by a microspreading technique and transmission electron microscopy (Marec et al., 1993). In Lepidoptera and other insects, as in other invertebrates and vertebrates (Roosen-Runge, 1977), the number of mitotic divisions, n, occurring during spermatogonial development is a characteristic of the species. The fixed number of these divisions undergone by encysted spermatogonia before they are transformed into spermatocytes, indicates that the succeeding generations of these cells represent successive stages of their differentiation. In man and other mammals, in which classification of the spermatogenous cells is of cardinal importance in studies of male fertility, successive generations of spermatogonia have been morphologically recognised and characterised (Kerr and de Kretser, 1981). Unfortunately, studies of this kind are scarce in insects but the meagre related data indicate that also in these taxa, there is a progressive differentiation of the spermatogonia following the successive spermatogonial division. Thus, in spermatogonia of several Orthoptera species, the X chromosome is heterochromatic, being negatively heteropycnotic in the early cysts but becoming progressively positively heteropycnotic through the successive spermatogonia generations, towards their transformation into spermatocytes (Hanna-Alava, 1965). Concerning lepidopteran spermatogonia, Lai-Fook (1982a) found that in the hesperiid Calpodes ethlius, the ultrastructure of the bridges interconnecting the spermatogonia of the cystic clone differs among successive generations, indicating that there is a progressive differentiation of the spermatogonia from one generation to the next towards their ultimate transformation into spermatocytes. In lepidopteran species, the number of generations of spermatogonia (n) has generally been found to be six although there are exceptions to this ‘‘rule’’ (Phillips, 1970). This results in the presence of a clone of 64 of the last generation of spermatogonia within each cyst of most lepidopteran species (e.g. Garvey et al., 2000). These clones of 64 mitotic cells now undergo a last phase of DNA replication and subsequently enter meiotic prophase, thus being transformed into primary spermatocytes. 2.3

SPERMATOCYTES

As is characteristic of meiosis, the late primary spermatocytes now undertake two successive divisions without an intervening DNA replication. It is before the first of these divisions that the differing eupyrene and apyrene fates of the presumptive sperm first become evident (see below). From this time on, each developing cyst contains exclusively either eupyrene or apyrene spermatocytes,


220

indicating that the decision to develop along the eupyrene or apyrene pathway is shared by and communicated among all the cells of the cyst at the same time. The message is presumably mediated by the intercellular bridges that interconnect all the spermatogonia within a cyst. But communication among the spermatogonia could alternatively be mediated by some factor present in the cyst extracellular fluid that bathes them all, or by cell surface contacts (such as the glycans discussed above), or possibly by all three mechanisms. Following the two divisions of meiosis, in most lepidopteran species the cysts now each contain 256 spermatids, which subsequently elongate and differentiate into bundles of 256 mature spermatozoa (Phillips, 1971). In Lepidoptera, after pachytene, primary spermatocytes enter a diffuse stage that encompasses both diplotene and most of diakinesis. At the diffuse stage, the nuclei resemble those of interkinetic cells, as they show no recognisable chromosomes but display diffuse, decondensed chromatin (Kawaguchi, 1928). As during the meiotic prophase of other insects (Fox et al., 1974; Henning and Kremer, 1990), the interkinetic-like nucleus is actively involved in RNA synthesis throughout the diffuse stage, as shown by autoradiography using [3H]uridine incorporation (Friedla¨nder, M., unpublished observations). This RNA synthesis appears to be sufficient for the complete, normal development of both the eupyrene and apyrene spermatids since no nuclear incorporation of labelled uridine can be detected from diakinesis onwards (Friedla¨nder, M., unpublished observations). This is consistent with the fact that apyrene spermatids continue to develop even after their nuclei have been extruded from the cells (Friedla¨nder and Miesel, 1977). The importance for developmental biology of this observation concerning the development of apyrene spermatids should be emphasised. We are aware of no other case, where similarly extensive structural and functional developmental changes are known to occur over an extended period in metazoan cells that have no nucleus. By comparison, the changes that occur in mammalian erythropoiesis after nuclear expulsion are extremely limited (Ogawa, 1993). Since apyrene spermatids show a clearly defined sequence of ontogenetic changes, apyrene spermatogenesis offers a very good model in which the posttranscriptional control of gene expression without any possibility of interfering mRNA synthesis can be studied.

3 3.1

Dichotomous spermatogenesis in Lepidoptera HOW EUPYRENE AND APYRENE SPERMATOGENESIS DIFFER

The differences between the eupyrene and apyrene pathways of development are clearly visible using conventional light microscope techniques and most of the pertinent features were described in detail by Meves (1903) as shown in Fig. 1. Here, we also reproduce images of eupyrene and apyrene spermatocytes


221

FIG. 4 Eupyrene and apyrene meiotic divisions in the carob moth Ectomyelois ceratoniae. The eupyrene metaphase (a), anaphase (b), and telophase (c) are regular, and the resulting spermatids are mononucleated (g). The chromosomes of the apyrene metaphase do not form an equatorial plate (d) and are irregularly distributed at anaphase (e). During apyrene telophase, only part of the chromosomes clusters at the poles (f), and the resulting spermatids (h) contain numerous micronuclei of different sizes. Testes squashed in aceto-orcein and observed with Nomarski interference contrast optics. From Leviatan and Friedla¨nder (1979).

from testis squashes viewed under interference contrast (Fig. 4), a rapid technique that is useful for experimental approaches. Primary spermatocytes of Lepidoptera already display indications of the eupyrene–apyrene dichotomy as early as zygotene. At this time, in comparison

222


with the early corresponding eupyrene spermatocytes, apyrene spermatocytes show (1) smaller cytoplasm volume, (2) fewer, smaller, and less electronopaque mitochondria, (3) lack of the chromosomal bouquet configuration, (4) premature condensed chromatin and shorter pachytene chromosomes, (5) absence of chiasmata, and (6) two separated nucleoli instead of the single one found in eupyrene cells (Kawamura et al., 1998; Garvey et al., 2000; Reinholdt et al., 2002). After the diffuse stage, the chromosomes of both eupyrene and apyrene spermatocytes reappear in their condensed, orthodox form in late diakinesis. But now the apyrene chromosomes become even more different from those of the eupyrene line than before. Eupyrene diakinesis is characterised by a regular and orderly array of paired bivalents that is found at this stage in typical meiotic cells throughout the animal kingdom. By contrast, apyrene diakinesis does not follow this pattern, instead variable numbers of both unpaired chromosomes and heteromorphic masses of chromatin are seen. The two consecutive meiotic metaphases of eupyrene cells are of orthodox type. The first metaphase shows a regularly formed equatorial plate, containing pairs of homologous chromosomes; the second one shows an equally regular equatorial plate that contains the haploid number of chromosomes. The subsequent eupyrene anaphases–telophases are likewise orthodox, leading eventually to the formation of four regular and equivalent haploid spermatids. Lukhtanov and Dantchenko (2002) have drawn attention to the fact that in some but not all Lepidoptera, notably in species of the lycaenid genus Agrodiaetus, the arrangement of bivalents in the metaphase plates of eupyrene spermatocytes is both unusual and exceptionally highly ordered, with large chromosomes being centrally located within the array. They speculate that this is a consequence of the polycentric/holocentric nature of lepidopteran chromosomes, providing a model to explain how the larger chromosomes would be pulled to the middle of the metaphase plate. In stark contrast, during the apyrene equivalent of the first meiotic metaphase, lepidopteran apyrene spermatocytes never form a real equatorial plate and their condensed chromosomes appear either to be unpaired or to form heteromorphic chromatin bodies, which are located at undetermined distances from the polar spindles (Leviatan and Friedla¨nder, 1979; Reinholdt et al., 2002). The subsequent apyrene anaphases–telophases are likewise highly irregular, as the chromatin bodies and univalents split, separate, and move asymmetrically and asynchronously towards the spindle poles. In some species, the univalents and chromatin bodies form discrete, variable numbers of micronuclei (Friedla¨nder and Miesel, 1977). The resulting nuclei or micronuclei have unequal DNA contents and are eventually extruded from the cells during spermatid differentiation (Friedla¨nder and Miesel, 1977; Kawamura et al., 2000). Wolf and Bastmeyer (1991) and Wolf (1994) recognised that two different strains of Ephestia ku¨hniella showed two different types of apyrene divisions


223

FIG. 5 Types of apyrene meiosis. In type Sbr meiosis (a), large chromatin clumps are initially located at the equator of the spindle and separate in a highly unequal fashion. In type L meiosis (b), small chromatin clumps are initially scattered throughout the spindle area and separate in a roughly even fashion. From Wolf (1994).

(Fig. 5), distinguishing them as ‘‘L’’ and ‘‘Sbr’’, the names being derived from the strains in question. In the former type of division, the chromatin is seen during anaphase–telophase to separate as an unequal number of amorphous masses, while in the latter type, the separating chromatin forms two approximately equal collections of large numbers of smaller masses that later become micronuclei. They classified a number of other Lepidoptera as undertaking apyrene spermatogenesis according to one or other of these ‘‘types’’. The significance of this distinction is not clear, however, since a single species may be observed to follow both ‘‘types’’ and it is possible that there is in reality a continuum of intermediate forms. As in other lepidopteran cells, the chromosomes of eupyrene spermatocytes have either polycentric (Wolf et al., 1987), or holocentric (Wolf, 1994) centromeres, a circumstance that probably contributes to the well-documented high resistance of male moths, and indeed of lepidopteran cells in general, to the adverse effects of ionising radiation (e.g. North and Holt, 1968; Koval, 1983; Seth and Reynolds, 1993; Seth et al., 2003). In contrast, in the dividing apyrene spermatocytes, the ‘‘chromosomes’’ show neither centromeres nor any other kind of special structural differentiation by which they could attach to the spindle microtubules. Instead, the microtubules reach the ‘‘chromosomes’’ at apparently random locations, which may be situated either confronting the poles or at the lateral faces of the chromatin masses (Friedla¨nder and Wahrman, 1970). The spindles of eupyrene spermatocytes are larger and contain more tubulin microtubules than those of the corresponding apyrene cells (Wolf, 1992, 1995). The eupyrene spindles display a typical appearance in that they contain microtubules attached both to the spindle poles and to the kinetochores of the chromosomes (but see below). By contrast, however, apyrene spindles have

224


only polar microtubules and the kinetochore set of microtubules appear to be completely missing (Reinholdt et al., 2002), a feature that is consistent with the irregular, asynchronous and asymmetric distribution of the chromosomes that occurs during apyrene meiotic divisions. The above account has emphasised the non-standard character of apyrene meiosis in Lepidoptera. Wolf (1994), however, has pointed out that lepidopteran eupyrene meiosis also has a number of unusual features. In particular, eupyrene meiotic spindles differ from most others in that the majority of the spindle microtubules appear to be attached neither to the chromosomes nor to the spindle poles. During anaphase, the separating chromosomes move past these microtubules towards the poles, and at telophase these non-kinetochore microtubules can be seen to remain between the polar clusters of chromosomes. Presumably, they serve as cytoskeletal guides for the movement of the separating chromosomes. These observations are puzzling, because microtubules are conventionally considered to be highly dynamic structures that need to be stabilised by association with a microtubule organising center (MTOC). What is the nature of the MTOC for the unattached spindle fibres of lepidopteran eupyrene spermatocytes? Wolf (1994) speculated that the nonkinetochore microtubules might be stabilised (a) by non-conventional MTOCs associated with intraspindle membranes, (b) by high concentrations of monomeric tubulin contained within the membrane that surrounds this type of spindle (see below), and (c) by stabilising cellular factors (such as microtubuleassociated proteins). Actually, this issue has recently become even more interesting with the discovery that the spindles of the ‘‘normal’’ spermatocytes of Drosophila also contain non-kinetochore fibres, although they are fewer in number than in Lepidoptera (Rebollo et al., 2004). These non-kinetochore spindle fibres appear to be nucleated by membranes present in the spindle region of the cell. A further difference between meiosis in Lepidoptera and in most other organisms is that during the division of lepidopteran spermatocytes from late prometaphase until early anaphase, the spindle is contained within a complex of tightly wrapped membranes derived from the endoplasmic reticulum. A similar system has been reported for drosophilids (Fuller, 1993). Wolf (1990a) termed this condition ‘‘sheathed nuclear division’’. This could explain option (b) above, since the sheath could retain a high concentration of tubulin monomers within the nuclear compartment. There is some evidence for the presence of stabilising factors. Immunostaining of eupyrene spermatocytes of the arctiid Phragmatobia fuliginosa with specific antibodies (Wolf and Joshi, 1996) showed that, as expected, b-tubulin was present throughout the meiotic spindle microtubules. But g-tubulin, which is known to be associated with MTOCs, was present not only in the region of the centrosomes, from which radiate the asters of the polar microtubules, but was also present within spindle microtubule arrays located on either side of the metaphase plate, known to be largely made up of unattached microtubules.


225

This g-tubulin staining persisted during anaphase when the kinetochore microtubules depolymerised, indicating that g-tubulin was present in the nonkinetochore spindle microtubules. While exposure to cold shock depolymerised the aster microtubules, the non-kinetochore microtubules were not depolymerised. Wolf and Joshi (1996) speculated that the presence of g-tubulin might stabilise them. Wolf (1996) also studied the acetylation state of tubulin in meiotic and mitotic spindles of the testis in E. ku¨hniella and Pieris brassicae using immunofluorescent staining. Acetylated tubulin is usually supposed to be present only in stable microtubules, while the non-acetylated protein is found in dynamic microtubules, meaning those that assemble and disassemble continuously. The mitotic spindles of gonial cells in both species were found to stain with an antibody directed against acetylated tubulin, indicating that these microtubules are relatively stable. The picture for eupyrene meiotic spindles differed between species. In E. ku¨hniella, no spindle fibres at all were labelled in early spermatocytes, while in P. brassicae, kinetochore microtubules and the small microtubule arrays around the centrosomes were found to be acetylated. Following chromosome separation, during late telophase, the spindle microtubules of both species became highly acetylated. Wolf (1996) considered that these observations were consistent with the suggestion that in both species all spindle fibres are dynamic from prometaphase to anaphase, and become more stable in telophase. The greater degree of acetylation seen in the early stages of division in P. brassicae was supposed to be because the spindle fibres in this species are longer, allowing tubulin subunits to persist longer even under dynamic conditions. Unfortunately, these studies reported only observations of eupyrene, and not apyrene spermatocytes. In addition to the characters described above, which are clearly related directly to the behaviour of the chromosomes during meiosis, the eupyrene and apyrene cells of Lepidoptera also differ markedly from the diffuse stage onward in the organisation of the cellular membranes associated with the nucleus. Thus, as mentioned above, in eupyrene spermatocytes the nucleus is surrounded by several concentric layers of endoplasmic reticulum (ER) embedded in a material that is relatively electron opaque, and from late prometaphase these membranes ensheath the metaphase–telophase spindles of eupyrene cells. In contrast, in apyrene cells, only a few loose and discontinuous layers of ER are found close to the nucleus and the apyrene metaphase– telophase spindles are surrounded by only a few similarly loose and discontinuous layers of ER (Wolf, 1992). ER is also found among the spindle microtubules in the dividing spermatocytes of both eupyrene and apyrene lines. Lukhtanov and Dantchenko (2002) have speculated that the presence of this membraneous material may account for the fact that the bivalents of the eupyrene metaphase plate always appear well separated, in contrast to the chromosomes of mitotically dividing spermatogonia, where such membranes are absent (Wolf, 1990b).

226


In the spindles of eupyrene spermatocytes, ER is abundant and mostly vesicular, while in the spindles of apyrene cells it is scarce and forms elongated strips (Friedla¨nder and Wahrman, 1970). The significance of these structural differences between the spindle-related membranous systems of eupyrene and apyrene spermatocytes is uncertain. It has been suggested (Wolf, 1995) that they might be related to the differential sequestration and release of Ca2+ ions, the ER being supposed here to have a role similar to that played by the sarcoplasmic reticulum of muscles, and that changes in intracellular Ca2+ might be important in regulating chromosome behaviour. However, by the time that the spindle is forming there are already considerable differences between the chromosomal organisation of eupyrene and apyrene cells, so that in our opinion such differences in perinuclear and spindle ER are unlikely to be the primary cause in determining the fate of the dividing spermatocyte (see Section 3.2). 3.2 3.2.1

THE CELLULAR MECHANISM OF APYRENE SPERMATOGENESIS

Failure of chromosome pairing in apyrene spermatogenesis

The earliest and key difference between eupyrene and apyrene spermatogenesis is the failure of the homologous chromosomes in the apyrene cells to pair properly during the zygotene stage of meiosis. Although a considerable number of other differences exists between apyrene and eupyrene spermatocytes, as detailed above, these differences all follow the initial lack of normal chromosome pairing. It has long been known, in fact since the very first observations of Meves (1903), that in apyrene spermatocytes the chromosomes fail to form bivalents. Where it does occur, pairing is inefficient and leads to the synapsis of nonhomologous chromosomes and non-homologous recombination ensues. Lack of pairing means that most chromosomes remain as univalents as the first meiotic division enters what ought to be metaphase, and without the double attachment to the spindle of the bivalents, no organised metaphase plate can form. With the onset of the apyrene anaphase, the lack of metaphase organisation inevitably leads to mis-segregation of the chromosomes and aneuploidy of the daughter cells. Where non-homologous recombination has occurred, chromatids are unable to separate properly, leading to characteristic chromatin bridges in the anaphase–telophase profile. Why do bivalents fail to form? Their formation necessarily involves two processes: the recognition by each chromosome of its homologous partner, and the cohesion of the chromosome pairs thus formed. The lack of bivalents by the time of the first metaphase could result from the earlier failure of either or both of these processes. Two papers by Garvey et al. (2000) and Reinholdt et al. (2002) reporting a microscopic study of spermatocytes of the gypsy moth, Lymantria dispar, have


227

reinforced this conclusion that chromosome pairing is indeed abnormal in apyrene spermatocytes at stages much earlier than metaphase. One line of evidence came from light microscopic observation of silver nitrate-stained nucleoli. Nucleolar fusion, which results from the pairing of the two nucleolar organisers, may be regarded as a visible indicator at the light microscope level of more general chromosome pairing. Reinholdt et al. (2002) found that nucleolar fusion was clearly evident in eupyrene meiotic divisions by pachytene and was maintained thereafter in eupyrene cells throughout the diffuse stage. By contrast, in apyrene spermatocytes nucleolar pairing was incomplete during pachytene, only being observed in about half of the apyrene spermatocytes at this stage. Moreover, by diplotene no apyrene spermatocytes at all continued to show nucleolar pairing. One interpretation of this observation is that not all of the chromosomes in apyrene spermatocytes undergo pairing, and that by diplotene any paired homologues that might have previously formed have separated. Thus, in this view, chromosome pairing is initiated in apyrene spermatocytes but the chromosomes then fail to cohere. A second, stronger interpretation is that pairing of homologous chromosomes in apyrene cells never occurs at all. In such a view, the transient nucleolar fusion seen in about 50% of apyrene cells during pachytene could result simply from the tendency of the heterochromatic regions, containing the strongly amplified rRNA genes of the nucleolar organiser, to associate with each other, as has been noted to occur in Drosophila (McKee and Karpen, 1990). 3.2.2

The synaptonemal complex

The same investigators (Garvey et al., 2000) also obtained further evidence that the normal pairing of homologous chromosomes never occurs in apyrene spermatocytes from studies of the synaptonemal complex (SC), a ladder-like structure that spans the space between the two homologous chromosomes of the bivalent, holding them together during the first meiotic division. Such cohesion is essential; without it, the bivalents do not form and the univalents do not line up on the metaphase plate. In eupyrene spermatocytes, which follow the typical meiotic programme, chromosome pairing during zygotene is accompanied by the formation of an SC during pachytene, as has been documented in an electron microscopic study of spermatocytes in E. ku¨hniella (Marec and Traut, 1993). Garvey et al. (2000) confirmed that this was also true for eupyrene divisions in the gypsy moth L. dispar, but found that apyrene spermatocytes showed no evidence of the presence of an SC at any stage. Moreover, immunostaining of apyrene spermatocytes (Reinholdt et al., 2002) showed that their chromosomes completely lacked the protein SCP3, a key component of the SC, while this protein was clearly present in the chromosomes of the corresponding eupyrene cells.


228

All of this implies that failure of synapsis underlies the apyrene condition in lepidopteran sperm. However, the possibility cannot be excluded that the chromosomes of apyrene spermatocytes do indeed initially come together but that synapsis cannot be maintained because of the absence of the SC. There are several well-known cases of germ cells in other organisms in which normal meiosis occurs essentially according to the usual pattern, despite the complete absence of any recognisable SC, such as in male D. melanogaster, the fission yeast Schizosaccharomyces pombe, and the filamentous fungus Aspergillus nidulans (Hawley, 2002; Page and Hawley, 2003). This means that in these organisms, and presumably also in others, mechanisms have evolved to ensure that pairing of meiotic chromosomes can be both initiated and maintained without SC formation. Thus, it is difficult to argue that absence of the SC alone is necessarily responsible for apyreny in the lepidopteran testis. Nevertheless, the facts that in the apyrene spermatocytes of male Lepidoptera bivalents do not form, that nucleolar fusion does not occur in many apyrene cells, that when it does it is transient, and that SCs are absent in apyrene divisions, are all strong indications that synapsis does not occur normally. 3.2.3

Interaction of telomeres with the nuclear envelope

Why does chromosome pairing fail to occur normally in lepidopteran apyrene meiosis? We suggest that it is because in apyrene spermatocytes either the terminal heterochromatic regions of the chromosomes, called telomeres, are either abnormal or unable to interact normally with the nuclear envelope. Support for this proposal comes from observations of the spatial arrangement of spermatocyte chromosomes during meiotic prophase. In most cases of typical meiosis, during leptotene the chromosome telomeres are dispersed, but are associated with the nuclear envelope. Subsequently, the telomeres cluster together at the nuclear envelope, at an intranuclear location close to the extranuclear location of the centrosomes. Since their ends are now all held close together, the orientation of the condensing chromosomes is constrained and they adopt a characteristic arrangement in space called the ‘‘bouquet’’ (Bass, 2003). Bouquet formation is clearly evident in lepidopteran eupyrene spermatocytes, but is not observable in the apyrene cells (Meves, 1903; Reinholdt et al., 2002). The absence of bouquet formation in apyrene meiosis suggests that the chromosomes are unable to associate with their specific binding site on the nuclear membrane. Pairing will therefore be expected to be at least hindered, and possibly prevented, since it is also known from studies in fission yeast (Niwa et al., 2000) that bouquet formation facilitates the pairing of homologous chromosomes during prophase. Moreover, in the achiasmate meiosis that occurs during Drosophila spermatogenesis, pairing of homologous chromosomes occurs without the benefit of an SC. Instead, the chromosome pairs


229

are each sequestered into discrete domains within the nucleus, associated with a particular region of the nuclear envelope (Vazquez et al., 2002). Hawley (2002) has suggested that this sequestration is sufficient to ensure the proper segregation of the chromosomes in the ensuing first meiotic division. Since it has been shown in maize (Carlton and Cande, 2002) that chromosome telomeres are involved in organising the bouquet, one hypothesis to explain the lack of bouquet formation in lepidopteran apyrene spermatocytes, is that telomere structure and/or behaviour is abnormal in these cells. An alternative hypothesis is that the chromosome telomeres of apyrene spermatocyte are normal, but that the nuclear periphery of these cells lacks some component with which the telomeres would otherwise associate during clustering. Possibilities that fit this description would be the meiosis-specific nuclear lamins C2 and B3, or an associated SC-attachment protein (Alsheimer et al., 1999). 3.2.4

Genetic control of chromosome pairing in meiosis

On the basis of the above analysis of events during lepidopteran apyrene meiosis, we suggest that a possible explanation for the failure of homologous chromosomes in these cells to pair normally is a change in the expression of a gene essential for telomere clustering. This gene would be expressed normally in eupyrene spermatocytes, but its product would not be functional in apyrene cells, either due to selective lack of transcription or some other post-transcriptional deficiency. This would probably be due to a change in the signalling system that controls the timing of the gene’s expression. The gene in question is unknown, but we can point to a number of mutations that phenocopy some important features of apyrene spermatogenesis that have been described in various organisms. These mutations interfere with both bouquet formation and chromosome pairing. Thus, in fission yeast, the product of the kms1 gene is required for normal bouquet formation, and in its absence irregularities in chromosome pairing and recombination occur (Niwa et al., 2000). In budding yeast, the protein Ndj1p is localised to telomeres, which are in turn localised in premeiotic aggregates at the nuclear periphery. When the gene ndj1 is disrupted, however, the telomeres are scattered throughout the nucleus, and not associated with the periphery, indicating that Ndj1p may be required for this interphase interaction with the nuclear envelope. During meiosis, ndj1 mutants fail to form telomere clusters and bouquet formation does not occur. As a consequence, homologue pairing and the progress of meiosis are delayed (Trelles-Sticken et al., 2000). In maize, pam1 mutants suffer abnormal telomere clustering, and defects in chromosomal synapsis ensue (Golubovskaya et al., 2002). We consider it improbable that apyreny originally arose through a mutation that led to the loss of function of an expressed protein that is essential to telomere clustering. This would be most unlikely if not impossible, since meiosis


230

proceeds normally in the eupyrene spermatocytes of the same insect. But an altered pattern of expression that led to the selective absence of the gene product from apyrene cells would lead to a phenotype very like that expressed in apyrene spermatocytes. An alternative (though in our view less likely) scenario would involve a mutation that did indeed lead to the loss of function of a telomere clustering protein, coupled with functional rescue by the selective expression of some other gene(s) in eupyrene, but not apyrene spermatocytes. 3.3

DURATION OF EUPYRENE AND APYRENE DIVISIONS

The duration of meiosis differs between the two types of spermatocytes, a fact that has important consequences for the content of the two types of sperm present in the testis at any one time (see below). Apyrene meiosis lasts for a markedly shorter time than eupyrene meiosis, as has been shown directly by autoradiography in C. pomonella, using incorporation of radioactive thymidine at the premeiotic S-phase as a marker for the timing of DNA synthesis (Friedla¨nder and Hauschteck-Jungen, 1986). Conspecific larval testis transplantation experiments have also led to the conclusion that apyrene spermatogenesis is compressed compared to the eupyrene process. In these experiments, testes containing spermatocytes that had developed only up until the diffuse stage were transplanted into pupae. This procedure caused the induction of precocious apyrene development and as a result apyrene metaphases were seen as early as 2 days after the transplant. This lapse of only 2 days from induction of apyreny to the first appearance of apyrene metaphases indicates that in this species, the meiotic prophase leading to the production of apyrene spermatozoa must be much shorter than the corresponding meiotic prophase that leads to eupyreny (Friedla¨nder and Benz, 1981). A particularly marked difference, noticed in the very first descriptions of apyreny (Meves, 1903), is that during the first meiotic prophase of apyrene nuclei the chromosomes are prematurely condensed compared with the eupyrene condition, in which the chromatin is more diffuse. The premature condensation of apyrene chromosomes is consistent with the marked temporal compression of the apyrene developmental programme (Friedla¨nder and Benz, 1981). Reinholdt et al. (2002) hypothesised that the premature condensation of apyrene chromosomes might be due to a difference in histone H3 phosphorylation between apyrene and eupyrene spermatocytes in prophase. However, immunocytochemistry revealed no apparent difference, suggesting that there is no fundamental difference in chromosome packing at this time. The shorter duration of apyrene meiosis is accompanied by the failure to synthesise a lysine-rich cytoplasmic protein fraction that is present in the comparable eupyrene cells (Friedla¨nder and Hauschteck-Jungen, 1982b). In turn, it is possible that the lack of this protein may lead to the atypical chromosome behaviour of the apyrene division.


231

Ordinarily, progression through meiosis is subject to a checkpoint in pachytene that arrests the first meiotic division when disruptions such as failure of synapsis occur (Roeder and Bailis, 2000). Intriguingly, the pachytene checkpoint is a regulatory step that appears to be particularly important in spermatogenesis, perhaps operating a quality control device for spermatozoa (Hunt and Hassold, 2002). Reinholdt et al. (2002) have speculated that a key event in allowing the evolutionary innovation of lepidopteran apyrene spermatogenesis to occur may have been the overriding of the pachytene checkpoint. Without the removal of this checkpoint, it would be expected that the apyrene spermatocytes would be permanently arrested, and subsequently die. Reinholdt et al. (2002) considered the possibility that the pachytene checkpoint does not operate in lepidopteran meiosis at all. Although there is no direct evidence, we point out here that the arrest of spermatogenesis that occurs in the diffuse stage spermatocytes of larval Lepidoptera (see above) points strongly toward the existence of a functional pachytene checkpoint in the immature testes. A further checkpoint that operates in cell division ensures accurate chromosome segregation by monitoring spindle microtubule tension and attachment to kinetochores, chromosome separation being prevented until these parameters have been satisfactorily achieved (Howell et al., 2004; Logarhinho et al., 2004; Taylor et al., 2004). It is evident that this checkpoint must be disabled during apyrene divisions. Using the same arguments as above, however, it seems unlikely that the nature of the genetic event that underlies the eupyrene–apyrene dichotomy is a structural change in the checkpoint proteins; much more probable is that they are regulated differently.

4 4.1

Spermiogenesis FORMATION OF EUPYRENE AND APYRENE SPERMATOZOA

Spermiogenesis is the morphogenetic process that changes spermatids (the pyriform, unripe, immature haploid cells that come into existence following the second division of meiosis) into fully differentiated and highly specialised spermatozoa. Eupyrene and apyrene spermiogenesis are illustrated diagrammatically in Fig. 6. In both eupyrene and apyrene cells, spermatid differentiation begins well ahead of the end of meiosis and before any division of the nucleus or cytoplasm has occurred. The first structural indication of spermatid differentiation of the two spermatogenetic lines is the appearance of four precociously developing flagella in the primary spermatocyte, at the beginning of the diffuse stage, or even earlier. The axonemes (flagellar axes) of these flagella originate from four basal bodies ð¼ centriolesÞ (Friedla¨nder and Wahrman, 1970) displaying apparently identical structural characteristics in both eupyrene and apyrene meiotic lines (Yamashiki and Kawamura, 1998). The four flagella are then

232


equally distributed among the four spermatids that result from the two subsequent divisions of meiosis (Friedla¨nder and Wahrman, 1970). In the early eupyrene spermatid, the proximal end of the basal body contacts the nucleus, which is still spherical and located in the anterior part of the elongating cell, close to the head cyst cell. At this stage, the spermatids within a cyst are interconnected by cytoplasmic bridges. Also, numerous nuclear membrane pores are present (LaChance and Olstad, 1988). The implication is that signals travel from the nucleus to the cytoplasm, and that these signals are exchanged among spermatids, presumably coordinating and synchronising their developmental progress. Subsequently, the nucleus of each spermatid begins to elongate and the basal body and proximal end of the axoneme appear attached laterally to it. Subsequently, the nucleus undergoes a change in its content of nucleoprotein in which a lysine-rich nucleoprotein fraction is replaced by an arginine-rich one. The nucleoprotein shift is evidently not causally related to elongation as it occurs after the spermatid begins to elongate (Friedla¨nder and HauschteckJungen, 1982a,b). A vesicle having an electron-dense wall occurs in front of the basal body of the developing eupyrene spermatid. This vesicle is frequently called an ‘‘acrosome’’ in the literature (e.g. Jamieson, 1987), presumably by analogy with the acrosomes of other animals, which are located at the same position within the spermatozoon. However, despite this topographical and structural similarity, there is as yet no report, as far as we are aware, that confirms that the vesicle found in insect spermatozoa is functionally a lysosome containing multiple hydrolytic enzymes, the essential qualification for its designation as an acrosome (Baccetti, 1979). This deserves further study. During spermiogenesis, the mitochondria of the developing eupyrene spermatid fuse into a structure called a ‘‘Nebenkern’’ that subsequently divides into two mitochondrial derivatives flanking the axoneme (Andre´, 1959; Phillips, 1970). In some species, both derivatives have similar dimensions (Friedla¨nder and Gitay, 1972), while in others one is always much larger than the other (Phillips, 1970; LaChance and Olstad, 1988). The derivatives display clear parallel cristae in longitudinal sections (Friedla¨nder and Gitay, 1972). Throughout spermiogenesis, the developing germ cells are enclosed by the somatic cells of the cyst wall. As development progresses, however, the cyst wall cells extend pseudopodia into the spaces between the maturing spermatozoa (Smith, 1968; Phillips, 1970, 1971). The function of this invasion of the sperm bundle by supporting cells is unknown, but it seems likely that it is required for the normal progress of spermiogenesis. The invasion coincides with continued change within the flagella of the spermatozoa (which may be supposed to respond to signals from the enveloping cellular processes) and also to the elaboration of the surface appendages of the sperm (see below). It is not known whether the cyst wall cells have a role in the synthesis of these appendages, but it seems probable.


233

FIG. 6 Stages of spermiogenesis in Bombyx mori. (A) Eupyrene spermiogenesis. (a) In the early eupyrene spermatid, the proximal end of the centriole is in a small depression in the posterior part of the nucleus. (b) A spatial redistribution occurs, which results in the attachment of the centriole, along its entire length, lateral to the nucleus. In front of the centriole, a prominent protruberance of the nucleus is present. The Nebenkern


234

Apyrene spermiogenesis differs from the above account in a number of ways. In the apyrene spermatid, the basal body is always separated from the micronuclei but attaches to the cell membrane at the anterior end of the elongating cell, close to the head cyst cell. The cell contains no acrosome-like vesicle and, in its place, there is an electron-opaque truncated cone of extracellular origin and unknown function (Friedla¨nder and Wahrman, 1971; Yamashiki and Kawamura, 1997). During elongation, the micronuclei (nuclear fragments) are initially located at undetermined positions along the flagellum. They are successively partially degraded by intranuclear DNase II, and extruded from the cell (Friedla¨nder and Miesel, 1977). In the normally developing apyrene spermatid, the micronuclei are extruded and do not undergo the nucleoprotein shift that occurs in the corresponding eupyrene cells. In apyrene spermatids, as in eupyrene ones, a Nebenkern is present and divides into two mitochondrial derivatives. However, the derivatives differ from those of the eupyrene type by showing no clear cristae in longitudinal sections (Friedla¨nder and Gitay, 1972) and by containing a rod of paracrystalline material. Metabolic significance has been attributed to this difference in mitochondrial appearance (see below). 4.2

PERISTALTIC SQUEEZING

LaChance and Olstad (1988) observed that ‘‘maturation of the sperm bundle involves the disappearance of a large volume of cytoplasm’’. These authors were at that time unable to suggest a mechanism for the removal of the unwanted cytoplasm, and although they drew attention to the presence in the developing spermatid of multivesicular bodies, which disappear along with the cytoplasm, they admitted that these structures might have nothing to do with the disappearance of excess cytoplasm.

continued

elongates behind the nucleus. A vesicle with thick, electron-dense walls appears in front of the nucleus. (c) Later, the vesicle comes in touch with the proximal end of the centriole. (d) The nucleus and the vesicle elongate simultaneously leaving the centriole behind the vesicle, in a latero-posterior position in relation to the nucleus. (B) Apyrene spermiogenesis. (e) In the early apyrene spermatid, the nucleus Nebenkern and centriole are close together in the anterior part of the cell. The centriole is not attached to the nucleus. (f) The centriole then moves to a forward position but the nucleus remains posterior to the centriole. The Nebenkern becomes situated between the nucleus and the centriole. At the same time the cell elongates. The proximal end of the centriole attaches itself anteriorly to the cell membrane. (g) The proximal end of the apyrene spermatid is engulfed in a deep invagination of a supporting cell. In the space between the spermatid and the supporting cell, an osmiophilic substance of unknown origin appears close to the proximal end of the centriole. (h) This substance appears to penetrate the anterior end of the spermatid forming a truncated cone in front of the centriole. From Friedla¨nder and Wahrman (1971).


235

In fact, Meves (1903), Zylberberg (1969) and Katsuno (1977a) had all observed that material was ejected from the developing spermatids. It is now evident that surplus cytoplasm is actively ejected from the cell as it is transformed into a mature spermatozoon. During the final steps of spermiogenesis the bundles of both eupyrene and apyrene sperm undergo a process of ‘‘peristaltic squeezing’’ (Fig. 7) that begins at the sperm’s anterior tip and advances towards the posterior tip, as reported for B. mori (Kawamura et al., 2000, 2001). This squeezing is effected by a network of actin filaments in the cyst cells that enclose the sperm bundle (Sahara and Kawamura, 2004). As squeezing proceeds, circlets of actin filaments develop within the bundle, surrounding individual sperm. It is not clear if these circlets are still physically attached to the cyst cells. Squeezing appears to cause only a slight elongation of the developing spermatozoa, but they become obviously thinner at this time. The function of squeezing appears to be the ejection of superfluous cytoplasm and organelles from the developing sperm. In apyrene cysts, squeezing is involved in the elimination of both cytoplasmic debris and also the nuclear fragments left over at the conclusion of the aberrant anaphase–telophase. In the strain of B. mori studied by Kawamura et al. (2000), these took the form of micronuclei. The nuclear fragments are initially located roughly in the middle of the sperm, and are swept by the passage of peristaltic furrow of contraction from there towards the rear of the sperm, from which they can be seen to be ejected. In eupyrene haploid cysts, produced by normal diploid males, squeezing achieves the elimination of cytoplasmic debris only. Thus, eupyrene and apyrene spermiogenesis differ in the consequences of squeezing. However, the basic nature of the squeezing process probably does not differ between the two sperm morphs, as has been shown by observing spermiogenesis in triploid and tetraploid Bombyx males. These insects are sterile, presumptive eupyrene and apyrene spermatocytes both displaying aberrant meiotic divisions during spermatogenesis that result in the formation of irregular non-haploid nuclei. The abnormal sperm that contain these irregular nuclei behave in a manner similar to the normally occurring apyrene spermatids. The irregular nuclei remain unattached to the basal body of the axoneme, and like true apyrene nuclear fragments, are also discarded from the tail end of the sperm by peristaltic squeezing (Kawamura et al., 2001). This is an important observation, because it illustrates the point that the later differences between eupyrene and apyrene developmental pathways may simply be consequences of earlier determining events, and that these later differences do not depend on essential differences between the two sperm types. 4.3

SURFACE STRUCTURES OF LEPIDOPTERAN SPERM

Spermiogenesis in Lepidoptera also involves the intratesticular elaboration of the surface of the eupyrene spermatozoa to a remarkable degree (Phillips, 1971). The completion of this process leaves the two kinds of sperm with very different appearances: while the surface of the eupyrene sperm is entirely


236

covered with extraordinary decorations, that of the apyrene sperm remains uncovered and has no particular surface structures. The surface structures of eupyrene sperm are here termed lacinate and reticular appendages (Fig. 8). The lacinate appendages are stacks of parallel laminate structures that in transverse sections resemble rays of a rising sun projecting from the cell body. They are present principally on the surface of the sperm that encloses the mitochondrial derivative and are less pronounced on that part of the surface that overlies the flagellar motor. Despite their extracellular location, the lacinate rays appear to be tubulin derivatives since they are disassembled by the antimitotic agent vinblastine (Friedla¨nder, 1976; Friedla¨nder and Gershon, 1978). These structures have also been termed the ‘‘radial mantle’’ (LaChance and Olstad, 1988). The reticular appendage is a rod of spherical subunits embedded in electronopaque material, which is attached to the length of the cell by radial thin laminae (Phillips, 1971; LaChance and Olstad, 1988). It has additionally been termed the ‘‘clear band’’ (Yasuzumi and Oura, 1965), and also the ‘‘satellite body’’ (LaChance and Olstad, 1988). The exact location of the reticular appendage on the sperm surface is precisely controlled relative to both the lacinate appendages and also the internal structures of the sperm, and is the same in every mature sperm. In H. virescens (LaChance and Olstad, 1988), the reticular appendage is initially located apparently randomly on the surface of the developing spermatid, but as maturation proceeds, it becomes closely associated with the underlying axoneme. The surface appendages of eupyrene spermatozoa are apparently a relatively late phylogenetic development since they are absent from the corresponding cells in ‘‘lower’’ systematic groups of Lepidoptera, such as Zeugloptera and Hepialoidea (Friedla¨nder, 1983). Their functions are unknown, but since eupyrene and apyrene sperm differ in their surface ornamentation, it is reasonable to suppose that differences in the way that these sperm morphs differ in their behaviour in the male and female tracts may be something to do with this. Some suggestions as to the function of eupyrene sperm surface structures will be made in Section 7.2. The nature of the processes that led to the different development of surface structures in eupyrene and apyrene sperm is unknown.

5

Regulation of dichotomous spermatogenesis

5.1 5.1.1

CONTROL OF MITOSIS AND MEIOSIS

The timetable of spermatogenesis

The timetable of spermatogenesis is characteristic of the relevant lepidopteran species, and under controlled conditions is strictly adhered to, being tightly


237

238


FIG. 7 (A) Peristaltic squeezing during eupyrene spermiogenesis in Bombyx mori. Eupyrene sperm bundles displayed by confocal images with immunofluorescence staining for tubulin (green), with rhodaminephalloidin for actin (red), and with Hoechst 33258 for nuclei (blue). (a) Early eupyrene sperm bundle bearing round nuclei. (b) The nuclei assume a spearhead shape. Green colour for microtubules is eliminated to show actin filaments more clearly. (c) Eupyrene sperm bundle shortly before the initiation of peristaltic squeezing. (d) Eupyrene sperm bundle just after the start of squeezing action. (e) Squeezing proceeds to the halfway point in a eupyrene sperm bundle. Note the transformation of actin bands and particles. (f) Completion of peristaltic squeezing. Arrows (a, c–e) represent actin particles moving posteriorly as peristaltic squeezing proceeds. CCN, cyst cell nucleus; HCN, head cyst cell nucleus; SN, sperm nucleus. Scale bars ¼ 80 mm. (B) Peristaltic squeezing during apyrene spermiogenesis in Bombyx mori. Staining as in 9(A). (a) Young apyrene sperm bundle. The nuclei are scattered in the bundle. (b) Apyrene sperm bundle with the nuclei gathering in the middle region. (c) Apyrene sperm bundle shortly before the initiation of peristaltic squeezing. (d) Peristaltic squeezing initiates from the anterior of the bundle. (e) Squeezing proceeds to the halfway point. (f) Completion of peristaltic squeezing. Nuclei are eliminated from the posterior end. Arrows (a–c) represent actin particles. CCN, cyst cell nucleus; HCN, head cyst cell nucleus; SN, sperm nucleus. Scale bar ¼ 40 mm. Both (A) and (B) from Sahara and Kawamura (2004).

correlated with the ontogenetic status of the individual (Fig. 9). Spermatogenesis is a discontinuous process, being punctuated by checkpoints at predetermined stations. Progress from one station to the next is under hormonal control. If a particular hormonal cue fails to occur at the right time, then


239

FIG. 8 Surface structures of eupyrene and apyrene spermatozoa of Bombyx mori as they appear within the testis. (a) Eupyrene spermatozoa. Transverse sections through flagella in an intratesticular cyst. Extracellular rays of alternating electron-dense, electron-lucid areas form the lacinate appendages. They overlay two apposed mitochondrial derivatives (M) with very poorly defined cristae. The reticular appendages (arrows) are composed of small spheres and electron-dense material and are typically located at one edge of the zone covered with the other appendages. (b) Apyrene spermatozoa. Transverse sections through flagella in an intratesticular cyst. No extracellular appendages are present. The two mitochondrial derivatives (M) are separated from one another forming a V configuration. They are more elliptical than those of the eupyrene spermatozoa and have conspicuous electron-lucid cristae. Both (a) and (b) are shown 42 500. From Friedla¨nder and Gitay (1971).

further development is blocked. The blocked cell will eventually lyse, should the necessary cue for progression fail to materialise. 5.1.2

Spermatogonial proliferation

Proliferating encysted spermatogonia are already present in early first instar male larvae (Leviatan and Friedla¨nder, 1979; Gelbic and Metwally, 1981). Proliferation continues uninterruptedly, and apparently automatically, throughout the rest of the insect’s life. The progress and timetable of spermatogonial divisions can neither be changed nor stopped by experimental manipulations that are compatible with cell viability (Dumser, 1980). Thus, spermatogonia are apparently unaffected by procedures using doses of exogenous hormones that effectively stop the meiotic divisions in the same

240


FIG. 9 Ecdysteroids and spermatogenesis. (A) Timetable of spermatogenesis in the testes of developing Manduca sexta. The stages at which particular meiotic events are observable in the testis are indicated by the open bars. Primary spermatocytes are already present in the 4th instar larvae and continue to appear throughout the rest of the insect’s life. Eupyrene metaphases occur from day 3 wandering larvae until day 13 pupae, while apyrene metaphases appear from day 3 pupae throughout the pupal and adult stages. The timing of the first observed (eupyrene) meiotic metaphases suggests that the resumption of meiosis may be triggered by the pre-pupation surge of ecdysteroids. (B) Appearance of meiotic metaphases in thorax-ligated day 0 wandering male larvae of M. sexta, which were injected immediately after ligation with varying amounts of 20-hydroxyecdysone (arrows). Spermatocytes at the diffuse stage were continuously present in each isolated abdomen. The length of time during which meiotic metaphases were observed (indicated by the bars) was dose dependent. After cessation of meiotic metaphases induced by 10 mg 20-hydroxyecdysone, a second injection of the same dose of the hormone (second arrow) caused the resumption of meiotic metaphases. These experiments confirm that ecdysteroids are required for the lifting of the block on meiosis. W, wandering larvae; P, pupae. From Friedla¨nder and Reynolds (1988).


241

testes (Friedla¨nder, 1982; see below). Although treatments with unphysiologically large doses of juvenile hormone (JH) and juvenoids have been noted to stop spermatogonial divisions, this effect is caused by inducing necrosis of the relevant cells and not by interfering with the normal course of the cell cycle (Gelbic and Metwally, 1981). Further, the proliferation of spermatogonia persists throughout diapause in lepidopteran insects that enter this condition. This happens despite profound changes in developmental hormone titres, and the concurrent death of the spermatogenous cells at later stages of spermatogenesis that occurs during diapause (Friedla¨nder and Benz, 1982). Although it has been reported that inactive, pre-existing spermatogonia of non-lepidopteran insects may be induced to initiate mitotic activity by exogenous ecdysteroid treatments, resulting in an apparent increase in the rate at which spermatogenesis proceeds (Dumser and Davey, 1975; Dumser, 1980), there is no indication that this is true in Lepidoptera. 5.1.3

Initiation of meiosis

The factors inducing lepidopteran spermatogonia to enter meiosis remain unknown. As noted above, the transformation to meiotic development occurs after a fixed number (n) of spermatogonial mitoses has occurred. In most species, n ¼ 6, so that the number of primary spermatocytes in the cyst is then 64. Primary spermatocytes at the pachytene stage are observed to occur as early as the second instar larva in B. mori (Takeuchi, 1969) and Ectomyelois ceratoniae (Leviatan and Friedla¨nder, 1979), but they appear at later instars in other species, such as Trichoplusia ni (Holt and North, 1970). Like spermatogonia, once their production has begun, early spermatocytes are subsequently always present in the testis throughout the rest of the insect’s life, a permanent pool of early developing spermatocytes being continuously restocked by new cells resulting from gonial proliferation. These early spermatocytes are able to endure drastic experimental treatments that are compatible with cell viability and they continue to be produced throughout diapause. Thus, during both larval life and diapause, spermatocytes develop uninterruptedly, progressing as far as the diffuse stage of meiotic prophase, but at this point development halts. Subsequently, unless the block to progression is removed, the arrested spermatocytes die, apparently by apoptosis (Kubo-Irie et al., 1999a). Only after a species-specific stage of ontogenetic development has been passed, or when diapause has been lifted, are the spermatocytes able to progress beyond a spermatogenetic checkpoint that is located in the diffuse stage. As has been discussed above, this seems to correspond with a stage at which the meiotic divisions of many organisms may be halted, the pachytene checkpoint (Roeder and Bailis, 2000). It has not been determined how long a primary spermatocyte can persist in this arrested condition. Probably this varies between species.


242

5.1.4

Control of meiosis by ecdysteroids

Since most aspects of insect postembryonic development are controlled by the endocrine system, it seems reasonable to relate the ability of lepidopteran spermatogenesis to progress beyond the diffuse stage to the presence or absence of one or more developmental hormones. An important clue as the nature of the hormone(s) involved was given by experiments with the sphingid M. sexta. In this insect, further progress of spermatocytes towards the first meiotic metaphase occurs for the first time close to the time of pupation. Experimental interference with the time of pupation also interferes with the time at which the metaphases are first seen. Accordingly, it was shown by a series of experiments (Friedla¨nder and Reynolds, 1988) combining ligatures, implantations of prothoracic glands and injections of 20-hydroxyecdysone, that the cue for the continuation in this insect of the meiotic advance towards the first meiotic metaphase is given by the post-wandering surge of 20-hydroxyecdysone experienced by the insect in the last larval instar. This is the same ecdysteroid peak that also triggers pupal cuticle deposition (Truman and Riddiford, 1974; Bollenbacher et al., 1981). Thus, in abdomens isolated from last instar larvae of Manduca, just before the rise of the post-wandering ecdysteroid peak, spermatocytes cease developing at the diffuse stage and eventually lyse without entering the first meiotic metaphase. The stoppage can be removed and spermatogenesis re-initiated by either implanting active prothoracic glands or injecting exogenous 20-hydroxyecdysone into the isolated abdomens (Fig. 9). In contrast, earlier surges of ecdysteroids preceding the post-wandering peak are unable to promote progress beyond the meiotic diffuse stage. Neither the premoult peaks that trigger the previous moults, nor the pre-wandering peak that changes the developmental commitment of the fifth instar epidermis, are related to the progress of meiosis. Because we have a particularly complete understanding of the roles of developmental hormones in controlling moulting and metamorphosis in Manduca (Truman and Riddiford, 1974; Riddiford et al., 2003), we can attempt to explain these experimental results in terms of the roles of these hormones in controlling meiosis in spermatocytes. The positive effect of the post-wandering peak in re-initiating meiosis, in contrast with the lack of a similar effect of the earlier peaks, may reflect at least two dissimilarities between the situations in which the respective peaks function. First, the last instar’s post-wandering surge of ecdysteroid is the first moult-initiating peak that occurs in the absence of JH, whereas the latter hormone has always previously been present at moult-initiating peaks prior to this. This suggests that ecdysteroids promote lifting of the meiotic block, while JH maintains it. The pre-wandering ecdysteroid peak, which does not initiate a moult, also occurs without the presence of JH but the concentration


243

of ecdysteroid experienced during this event is much lower (Fig. 9), and is possibly inadequate to trigger the removal of the meiotic block. The effects of 20-hydroxyecdysone are both concentration and time dependent, suggesting that the lifting of the meiotic block depends on an effect that is accumulated in the hormone’s target cells (Friedla¨nder and Reynolds, 1988). This is consistent with what we know of the actions of ecdysteroids in promoting moulting. Second, the spermatocytes evidently need to acquire the ability to respond to ecdysteroid by lifting the prophase checkpoint. In vivo, this competence is acquired only following exposure to the pre-wandering peak of ecdysteroid. This is shown by the observation that abdomens of Manduca larvae, which were isolated and immediately injected with 20-hydroxyecdysone 1 day before wandering began, failed to show meiotic metaphases. However, when similarly treated abdomens were injected 4 days later with an additional, similar dose of 20-hydroxyecdysone, they subsequently displayed meiotic metaphases (Friedla¨nder and Reynolds, 1988). This suggests that the initial exposure to ecdysteroid had the effect of ‘‘priming’’ the spermatocytes to respond to the second ecdysteroid treatment, precisely as would be expected if this priming was normally effected by the pre-wandering ecdysteroid peak. Again, this is consonant with what is known of the action of ecdysteroids in triggering pupation in Manduca. The experiments with Manduca provide conflicting evidence as to whether ecdysteroid is continuously required for progress through meiosis to continue, once the block has been lifted. Injecting a small amount of 20-hydroxyecdysone into isolated abdomens caused only a transient response, and meiosis stopped again after only 2 days. However, a second injection led to a renewed meiotic response. This implies that ecdysteroid is continuously required. By contrast, Friedla¨nder and Reynolds (1992) showed that during pupal diapause in this insect, eupyrene metaphases continued unchecked despite undetectably low levels of ecdysteroid (this result is discussed further below). This shows that ecdysteroids (at least those detected by the radioimmunoassay) are unnecessary for the maintenance of eupyrene meiosis. A possible reconciliation of these two sets of results would follow from the proposal that the effects of ecdysteroids on spermatocytes are initially reversible, but that one of the effects that is accumulated is progression towards a threshold, beyond which further ecdysteroid is not required. Again, this is much like what happens during the other actions of ecdysteroids (Riddiford et al., 2001). The situation in C. pomonella is quite similar, although less completely researched. In diapausing last larval instar insects, primary spermatocytes fail to progress beyond the diffuse stage and eventually lyse. As is the case in Manduca, exposing isolated abdomens of these diapausing larvae of C. pomonella to an ecdysteroid signal (in this case, the non-steroidal ecdysone agonist tebufenozide), revokes the stoppage and renews the progression of meiosis towards metaphase (Friedla¨nder and Brown, 1995).

244


In other Lepidoptera, however, the mechanism of control of meiosis appears to be somewhat different. In spermatocytes of B. mori, the lifting of the prophase checkpoint evidently does not depend on the prepupal surge of ecdysteroid. This conclusion rests on several different lines of evidence, as follows. Kawamura and Sahara (2002) devised a culture system that allows isolated spermatocysts explanted from Bombyx testes to develop into fully matured sperm bundles, including the completion of peristaltic squeezing. Such in vitro development, a remarkable technical achievement, requires that the cysts are isolated and the culture is continually shaken; the addition of Bombyx haemolymph to the culture medium is essential. However, just as is the case for Manduca and Cydia, this culture system evidently cannot support the further development of spermatocytes that are arrested in the first meiotic prophase. The first explants in the reported experiments were made from silkworms immediately after ecdysis to the fourth larval stage. With donors of this age, almost no sperm formation occurred, but the number of eventually maturing sperm bundles increased progressively with the age of the donor insect during the fourth and early fifth stages, reaching a plateau after about 4 days. This suggests that the Bombyx culture system allowed those spermatocytes to develop, which had already experienced some endogenous permissive condition that removes the block on meiosis, but that the in vitro conditions did not permit development in those cysts that had not experienced this event. Unlike the situation in Manduca, the permissive condition in Bombyx evidently does not coincide with an identifiable ecdysteroid surge. It persists over several days, and different cysts evidently take different lengths of time to respond to the permissive condition. The apparent timing of this endogenous licensing event is nevertheless in accord of what is known of the timing of spermatogenesis in Bombyx. According to Sado (1961), spermatocytes first appear in the second larval stage, 6 days after hatching, and by the time of entry into the fourth stage, a majority of cells in the testis are already recognisably spermatocytes in prophase. Metaphases are first seen early in the late fourth and early fifth larval stage. This suggests that declines in JH during the fourth and fifth larval stages may be the crucial endocrine event, but the requisite experiments to show this have not been performed, and it is possible that endogenous ecdysteroids are also required. A similar conclusion can be drawn from the transplantation experiments of Takeuchi (1969), who showed that the formation of spermatocytes (i.e. the initiation of meiotic prophase) probably did not require developmental hormones (spermatocytes were formed in isolated pupal abdomens). But because development of the transplanted germ cells did not progress further than this, except in the environment of the fifth instar larva, in which ecdysteroids are unaccompanied by JH, Takeuchi suggested that the known pre-pupation decline in JH played the most important role in permitting spermatogenesis to continue towards the formation of elongating spermatids; as we have discussed


245

above, this means removing the block on progression of meiosis beyond the first prophase. The situation in the noctuid H. virescens is again different. In a study by Loeb et al. (1985), thorax ligatures failed to prevent subsequent progression of ‘‘premeiotic’’ spermatocytes through meiosis. This was originally taken to imply that ecdysteroids synthesised by the prothoracic glands are not required for meiosis, although it is possible that the stage at which the ligature was applied corresponded to a point after the meiotic block had already been lifted, but before the consequences of this were evident (in these experiments continued spermatogenesis was judged to occur only when elongating spermatids were seen). Another possibility, however, is that the ecdysteroids necessary for removal of the meiotic block were supplied by another tissue in the abdomen. And indeed, when spermatocytes from the same stage were cultured in vitro after the removal of the testis sheath, they eventually stopped developing and lysed. This stoppage could be revoked and re-initiation of spermatogenesis induced by adding ecdysteroids to the medium. Similar results were obtained by replacing the addition of ecdysteroids with the inclusion of testis sheath tissue in the medium (Giebultowicz et al., 1987). It has subsequently been shown that the sheath cells of Heliothis are able to synthesise ecdysteroids, including 20-hydroxyecdysone among others (Loeb et al., 2001). It is still unclear, however, whether the synthesis of testicular ecdysteroids is related in any way to the discontinuity of spermatogenesis that is found in larval testis, or in the naturally occurring diapause of this insect. The situation for Galleria mellonella is also not clear. Lender and DuveauHagege (1963) reported that when testes of last stage larvae of this insect were explanted into tissue culture medium, the germ cells were able to continue developing as far as elongating spermatids without the benefit of added haemolymph. This was interpreted at the time to mean that endogenous hormones were not required for the progress of spermatogenesis. However, the wandering larvae that were used as donors would already have experienced a pre-wandering surge of ecdysteroid. If these insects behave like Manduca, this previous exposure would have been enough to allow the testes to reinitiate development and continue it in vitro in the absence of the hormone. On the other hand, if in this insect the prepupal decline of JH is the crucial endocrine factor in permitting the progress of spermatogenesis beyond meiotic prophase, then that decline would also have preceded the explantation of the testes. 5.1.5

Indirect role of ecdysteroids

The above evidence shows that at least in M. sexta and C. pomonella, lifting of the prophase checkpoint in male meiosis in vivo is due to ecdysteroid hormones. However, the effect of the ecdysteroid in these insects appears either to be indirect or to require unknown additional factors. This conclusion


246

follows from the failure to induce in vitro progression of meiosis in isolated larval testes of either non-diapausing Manduca (Friedla¨nder and Reynolds, 1988) or diapausing Cydia (Friedla¨nder and Benz, 1982) by adding ecdysteroids to the culture media. These experiments were unsuccessful despite exhaustive attempts to optimise the culture conditions for the response. Moreover, exactly the same culture conditions were adequate to allow continued meiotic metaphases in testes of the same age that had already experienced different hormonal conditions prior to their explantation. The results of Kawamura and Sahara (2002) are consistent with a similar situation in B. mori, although their experiments were not designed to test whether ecdysteroids can act alone to relieve the meiotic block in vitro. In experiments with C. pomonella, partial in vitro differentiation of the spermatids did indeed occur in response to ecdysteroid, but the resulting cells were unreduced since the renewed spermatogenesis skipped over the meiotic metaphases (Friedla¨nder and Benz, 1982). It is difficult to avoid the conclusion that some unknown hormonal factor promoting meiosis is missing from the culture medium. 5.1.6

Role of juvenile hormone

Just as development of the insect as a whole is controlled by the interplay between ecdysteroids and JH, fluctuations of the JH titre are also involved in the regulation of spermatogenesis by interfering with the progress of meiosis. However, much less information is available about the role of JH than for ecdysteroids. Although the post-wandering, prepupal surge of ecdysteroid has clearly been identified by Friedla¨nder and Reynolds (1988) as the immediate cause of removal of the meiotic block in Manduca, it is notable that the ecdysteroid surges that precede larval–larval moults in this insect do not remove the block. Presumably this is because, as indicated before, these larval premoult surges all occur in the presence of JH, and because JH prevents the lifting of the block. The absence of JH at the time of the prepupal surge permits the ecdysteroid to exert its prospermatogenetic effect. The best evidence for the antispermatogenetic effect of JH is the work of Takeuchi (1969) with B. mori. Precocious meiotic development towards metaphases and beyond was found to occur following allatectomy of early 3rd instar Bombyx larvae. Further, testes transplanted into pupal hosts formed spermatids, but this was repressed by cotransplanting adult corpora allata (CA). The CA hormone that restrains spermatogenesis is presumably JH. Takeuchi’s (1969) experiment depends on the fact that in Lepidoptera, the adult CA resume secretion of JH. The CA from both male and female donors were equally effective. It is also known that CA in both male and female moths secrete JH. These experiments indicate that the CA secretes a hormone that acts in young larvae to restrain the removal of the block on meiosis, but say little about how it does so.


247

Kajiura et al. (1993) used flow cytometry to analyse the ploidy of nuclei isolated from Bombyx testes. The presence of 1C cells was taken to indicate the presence of spermatids and spermatozoa. Repeated injections of the JH analogue methoprene arrested development to produce dauer larvae, and also blocked the development of 1C cells. However, methoprene treatment of larvae in which the production of 1C cells had already begun (1C cells first appear from day 0 of the fifth instar) was ineffective in preventing the appearance of further 1C cells. Instead, the juvenoid treatment prevented 2C cells progressing to 4C (i.e. prevented DNA replication, presumably in gonial cells). But the same treatment repeated 2 days later had no effect on the proportions of 1C, 2C and 4C cells, and the authors suggested that by this time the testes were no longer able to respond to JH. The situation appears to be similar in C. pomonella (Friedla¨nder, 1982). Allatectomy of early last instar larvae, predetermined by their environment to enter diapause and to arrest meiosis at the diffuse stage, prevented the stoppage and enabled the spermatogenic cells to advance towards metaphase. Conversely, application of the synthetic juvenoid altosid to late penultimate larvae, predetermined by their environment to skip diapause, induced stoppage of meiosis at the diffuse stage in the last larval instar. It thus seems clear that the action of JH, at least in these two species, is to oppose the pro-spermatogenic effects of ecdysteroids. This makes sense, since it explains why the regular exposure of the testis to a surge of ecdysteroid that occurs before each moult in premetamorphic larvae does not lead to the premature lifting of the block on meiosis. The opposing effects of JH and ecdysteroid on spermatogenesis probably do not represent a simple antagonism. Although the mechanism by which JH affects its target cells is still not understood, the effects of JH and ecdysteroids on moulting and metamorphosis are not simply antagonistic in this way (Riddiford et al., 2003; Dubrovsky et al., 2004). Rather, JH may be seen as modulating the response of the target tissue to ecdysteroid, so that in the presence of JH, ecdysteroids cause one kind of response, while in the absence of JH the response to ecdysteroid is different. This may be because in the absence of JH, different ecdysteroid receptor isoforms are produced in target cells (Riddiford et al., 2001; Sullivan and Thummel, 2003). More experiments on the actions of JH on lepidopteran spermatogenesis are required. 5.1.7

Similarity to oogenesis and other male reproductive systems

The failure of meiosis to progress beyond mid-prophase, due to the absence of the prospective cue for its re-initiation, is a phenomenon that is common to the gametogenesis of both male and female insects. The mechanism regulating the phenomenon appears to be essentially similar in females and males since in both spermatocytes (Friedla¨nder and Reynolds, 1988) and oocytes (Lanot et al., 1988, 1990): (a) meiosis discontinues at diplotene–diakinesis; (b) the


248

chromosomes in the arrested cells then show interphase-like configurations, becoming either diffuse or lampbrush-shaped; (c) these chromosomes are highly active in RNA synthesis; and (d) meiosis re-initiation appears to be controlled by a rise of the ecdysteroid level. Indeed, regulation of progression through meiosis at this stage is very common throughout the animal kingdom, including humans (McLaren, 1995). It will be recalled that reinitiation of arrested meiosis at puberty in both male and female mammals has long been known to be subject to the dual control system of two circulating protein hormones, luteinising hormone (LH) and follicle-stimulating hormone (FSH). In the case of the male, LH exerts its effects by stimulating the synthesis of the steroid testosterone in the Leydig cells (somatic supporting cells within the testis). Although testosterone has many independent effects on sexual development, both FSH and testosterone are required for the resumption of spermatogenesis (Zirkin, 1998). This has clear parallels with the control of spermatogenesis in Lepidoptera, in which more than one hormone plays a role. 5.2 5.2.1

THE SWITCH FROM EUPYRENE TO APYRENE SPERMATOGENESIS

The apyrene spermatogenesis inducing factor

The decision of the spermatocytes of a particular cyst to follow either the eupyrene or the apyrene developmental pathway is unrelated to the location of the cyst within the testis, as eupyrene and apyrene cysts appear to be distributed at random. Apyrene cysts occupy no particular positions within the testis that might indicate either topographical diversification or the presence of morphogenetic gradients inducing apyreny. Instead, both kinds of cells appear to derive from the same kind of bipotential early primary spermatocytes (Leviatan and Friedla¨nder, 1979; Katsuno, 1987; Reinholdt et al., 2002). However, timing and the developmental stage of the whole insect are clearly important factors in the eupyrene–apyrene shift decision. According to species, eupyrene meiotic metaphases are first seen in the testes of either the penultimate or the last larval instars, and later disappear at a species-characteristic stage of pupal development. In contrast, the earliest apyrene metaphases are seen close to the time of pupation, just before or after, according to the species, and they may then continue throughout the rest of life (Tazima, 1967; Friedla¨nder and Reynolds, 1988). Fluctuations of the circulating levels of either JH (Leviatan and Friedla¨nder, 1979) or ecdysteroids (Friedla¨nder and Benz, 1982) appear to be related only indirectly to this shift of spermatogenesis from eupyrene to apyrene development. Instead, the changing developmental fate of the bipotential primary spermatocyte from the eupyrene to the apyrene pathway is apparently induced by a still undetermined haemolymph Apyrene Spermatogenesis Inducing Factor (ASIF) (Friedla¨nder and Benz, 1981) that


249

becomes active close to pupation, before or after, according to the species (Jans et al., 1984). The causal relation between pupation and ASIF activity has been shown by experimental manipulation of the timetable of development. In one set of experiments, penultimate instar larvae of the saturniid Actias selene were allatectomised and responded by pupating directly, skipping the last larval instar. Apyrene metaphases appeared on time, 2 days after pupation, in both allatectomised individuals and intact controls, despite the fact that the allatectomised insects were 12 days younger than the corresponding controls (Friedla¨nder et al., 1981). Similarly, C. pomonella, pseudoparasitised in the egg stage by the hymenopteran parasitoid Ascogaster quadridentata, developed apparently normally, but then moulted directly into precocious last instar larvae by skipping the penultimate larval stage. Again, apyrene metaphases appeared on time in the pseudoparasitised larvae, shortly before pupation, just as they did in the controls, despite the difference in age of these insects (Brown and Friedla¨nder, 1995). Accordingly too, testes of penultimate larval instar of C. pomonella, showing diffuse stage spermatocytes as the most advanced meiotic stage, precociously undertook apyrene spermatogenesis when transplanted into pupae (Jans et al., 1984). In C. pomonella, it was shown experimentally that ASIF becomes active concurrently with the commitment of other tissues to pupal development, 2 days before appearance of the first apyrene metaphase. In such experiments, there was a lapse of only 2 days between induction of apyreny and appearance of the earliest apyrene metaphases (Friedla¨nder and Benz, 1981). During this relatively very short period, the apyrene cells completed prophase, a process that takes the corresponding eupyrene cell more than a week (Friedla¨nder and Benz, 1981). It has been noted above (Section 3.3) that apyrene meiosis is completed more quickly than eupyrene meiosis. It is not clear whether ASIF induces apyreny by shortening the meiotic prophase, and that this, in turn, would interfere with the expression in the spermatocytes of a gene that is essential to the normal progression of meiosis, or whether ASIF interferes directly with the expression of this gene, which in turn would shorten the apyrene prophase (Friedla¨nder and HauschteckJungen, 1982a, 1986). The nature of ASIF is still obscure. There is at present no evidence showing whether ASIF is a single chemical entity or a particular set of different factors acting coordinately. ASIF is certainly present in the haemolymph, since testes transplanted from larvae to pupae respond to it, despite being inserted at sites (e.g. in the thorax) far from their original location. It is not sex-specific, since transplantation into both male and female pupae induces apyrene divisions to occur. Although it has not been definitively ruled out that ASIF is actually the lack of some factor present in the larva but not the pupa, and which is essential for the continuance of eupyreny, this seems unlikely, since apyrene divisions are not induced by explanting into tissue culture conditions testes that


250

have already begun to display eupyrene divisions, but which have not yet experienced exposure to pupal conditions. Such testes continue to show only eupyrene divisions. 5.2.2

Factors within the testis that may determine apyreny

The nature of the changes within the testes that are responsible for the switch from eupyrene to apyrene spermatogenesis are unknown. Kawamura et al. (2002) have shown in B. mori that addition of exogenous ecdysteroid hormones or glucose both increase the number of apyrene cysts developing when incubated in vitro. They suggested that this may be relevant to the switchover from eupyreny to apyreny that occurs in vivo. They pointed out that ecdysteroid hormones are present in the insect at the time that the switch is taking place, and also that the level of glycogen in the testis declines (presumably due to its conversion to glucose) at the same time. The significance of this experiment is, however, not simple to interpret. It is important to understand that, at the time that they were explanted, at least some of the cultured testicular cysts were already determined to produce apyrene sperm. It is possible that the effect of the added ecdysteroid and glucose may have been to enhance the survival of those apyrene cysts, rather than to promote the changeover from eupyrene to apyrene meiosis. This would be consistent with what is observed in diapausing M. sexta, in which both eupyrene and apyrene cysts are deprived of ecdysteroid, and in which apyrene cysts die at an earlier stage of development than the eupyrene ones (Friedla¨nder and Reynolds, 1992) (see Section 5.4). Enhanced supply of nutrients might easily be envisaged to have an effect of this kind too. 5.2.3

Co-existence of eupyrene and apyrene cysts

The difference in the duration of meiotic prophase between the eupyrene and apyrene pathways has an important consequence. The appearance of ASIF induces all cysts that now enter prophase to switch to the apyrene pathway. However, in every case examined, the earliest cysts containing apyrene metaphases are found to coexist, during a short period of time, with other cysts containing eupyrene metaphases, despite the presence of ASIF in the haemolymph (e.g. Friedla¨nder and Reynolds, 1988). This indicates that the primary spermatocytes must have a relatively short period of sensitivity, during which they are able to respond to the presence of ASIF activity. When ASIF first appears, those spermatocytes that are exposed to it will fall into two populations. Some will already have passed through the restricted sensitive phase during which they can respond. So, these more ‘‘mature’’ spermatocytes continue to follow uninterruptedly the path of eupyrene development that they have already effectively begun. Other slightly less advanced cells will still be sensitive to ASIF and will respond by changing their commitment, so as to


251

enter the apyrene path of development. These spermatocytes will all develop to form apyrene sperm, and will do so, as indicated before, more quickly than the cells following the eupyrene pathway. Thus, for a short period, cysts containing developing spermatocytes of both types will coexist within the testis. However, since the duration of prophase in spermatocyctes destined to become eupyrene sperm is longer than in the case of apyrene spermatocytes, this period of coexistence will be longer than the sensitive period during which the switchover occurs. Further, there will be a period of time during which both types of cells will arrive together at metaphase. After the short period of coexistence of eupyrene and apyrene metaphases of the two kinds, all the newer cysts that begin their development in the presence of ASIF will display apyrene metaphases only, as indeed occurs in mature pupae and adults (Friedla¨nder and Hauschteck-Jungen, 1986). 5.3

CONTROL OF SPERMIOGENESIS

Spermiogenesis, or spermatid differentiation, is frequently but erroneously considered to be a sequel and direct consequence of meiosis. A closer scrutiny of data on spermatogenesis shows that meiosis and spermatid differentiation are two different and separable processes, but that in most cases meiosis overlays and at least partially conceals the early preparatory stages of spermiogenesis. In certain cases, however, the two processes appear in nature to be clearly separated and distinguishable. An example is seen in the case of hymenopteran species that produce haploid males. Males of such species produce normal, haploid spermatozoa by direct differentiation of their already haploid primary spermatocytes (Wilson, 1925; White, 1973). In such cases, the primary spermatocytes skip the chromosome reduction that normally occurs during meiosis, a process clearly superfluous for haploid individuals, but which is indispensable for the production of haploid sperm by diploid males. Another example of a different kind is seen in the case of Drosophila males bearing a meiotic mutation ms(1)413. Despite the failure of meiosis to progress, these insects continue spermiogenesis to produce abnormal, non-haploid but nevertheless elongating spermatids (Brick et al., 1979; Fuller, 1993). Moreover, meiosis and spermatid differentiation can be separated experimentally. Thus, testes of fertile male adult antlions treated with the antimitotic agent 2-mercaptoethanol, show direct differentiation of spermatocytes into anomalous spermatids that are either diploid or tetraploid. The appearance of the anomalous spermatids occurs because the chemical causes breakdown of the meiotic metaphase spindles, either the second or the first metaphase, respectively. This experiment shows that completion of meiosis is not a precondition for spermatid differentiation (Friedla¨nder and Wahrman, 1965). A similar phenomenon has been reported in B. mori, where isolated spermatocytes were treated for 24 h in vitro with the antimicrotubule

252


agent colcemid, followed by prolonged culture under conditions that would otherwise have allowed normal development (Kawamura and Yamashiki, 1999). This resulted in the formation of sperm in which either both meiotic divisions, or in some cases only the second meiotic division, were blocked. Nevertheless, following the removal of the colcemid, spermiogenesis ensued despite the fact that meiosis had not taken place. Where both meiotic divisions had been blocked, sperm with four flagella and a giant Nebenkern were formed, while where only the second division had been prevented, sperm with two flagella were formed. Additionally, there are clear indications in the normal progress of lepidopteran spermatogenesis that spermatid differentiation is not a sequel of meiosis. Thus, as noted before, the flagella of the four prospective spermatids resulting from the two subsequent divisions of meiosis have already begun to develop in the primary spermatocyte long before the cell reaches the first meiotic metaphase (Henneguy, 1897; Friedla¨nder and Wahrman, 1970). The same phenomenon has been reported for the sister order Trichoptera (Friedla¨nder, 1993; Klein and Wolf, 1997), other orders of insects (Daub and Hauser, 1988), non-insect invertebrates (Paulus, 1989), as well as for vertebrates (Abe´ et al., 1988). Further, in Lepidoptera (Friedla¨nder and Hauschteck-Jungen, unpublished observations), as in other insects (Fox et al., 1974), the informative RNA involved in spermatid differentiation is transcribed during the diffuse stage of the primary spermatocyte, long before the cell arrives at metaphase. Accordingly, as is the case for other insects (Gould-Somero and Holland, 1974), no RNA synthesis has been observed in lepidopteran spermatids (Friedla¨nder and Hauschteck-Jungen, unpublished results), and at this time the spermatids display compact chromatin with no indication of being involved in any synthetic activity. And we may be certain that, being anucleate, the apyrene spermatid relies on pre-existing informative RNA for its subsequent differentiation. Moreover, as indicated above, it has been shown experimentally that in isolated testes of diapausing C. pomonella larvae, cultured in media containing ecdysteroids, the spermatocytes can be induced to skip metaphases and undergo partial differentiation, producing unreduced eupyrene spermatids (Friedla¨nder and Benz, 1982). This again shows that spermiogenesis is not obligatorily coupled with meiosis. Spermatid elongation consists of at least two phases. During the first, only the flagellum elongates while the nucleus remains spherical; during the second, both the flagellum and nucleus elongate together. Thus, in E. ceratoniae, flagellum elongation begins in fourth instar larvae while the nuclei start elongating only in mid-fifth instar larvae. Nuclear elongation is apparently dependent on a low JH titre. Triggering of nuclear elongation is preceded by a refractory period during which preparatory synthetic processes, related to the nuclear elongation, take place. Thus, elongation of the nucleus, but not that of the already elongating flagellum, can be inhibited by high doses of exogenous


253

juvenoids. But this occurs only when the juvenoid is applied before the spermatid has completed the synthetic preparatory processes for nuclear elongation. After the nucleus starts elongating, the juvenoid is no longer able to prevent elongation. Accordingly too, precocious nuclear elongation could not be induced in in vivo experiments solely by lowering the JH titre, before the preparatory synthetic period ended (Leviatan and Friedla¨nder, 1979). Further, in vitro, nuclear elongation could not be induced solely by culturing testes in JH-free media (Friedla¨nder, M., unpublished observations). Lepidopteran apyrene differentiation includes only those basic features of spermatogenesis needed to produce a highly simplified cell, which is specialised mainly for motility, and which lacks the capacity for fertilisation. Accordingly, basic axoneme development, which starts in the primary spermatocyte, probably even before the shift to apyreny has occurred, proceeds in a similar way in both eupyrene and apyrene cells. But in contrast with the eupyrene situation, during apyrene spermatogenesis other cellular components remain relatively underdeveloped. Thus, in addition to the eventual loss of the nucleus by the end of meiosis, lepidopteran apyrene spermatids display both reduced volume and smaller and fewer mitochondria containing less DNA in relation to their eupyrene counterparts (Kawamura et al., 1998). 5.4

CONTROL OF SPERMATOGENESIS DURING DIAPAUSE

Spermatogenesis, like other morphogenetic processes in insects, discontinues during diapause and resumes afterwards (Chippendale and Alexander, 1973). But in contrast with the morphogenetic standstill of other tissues that occurs during diapause, the apparent stoppage of spermatogenesis conceals the fact that its early stages remain operative and continue without interruption. The diapause ‘‘arrest’’ of spermatogenesis is characterised by (1) an uninterrupted supply of newly produced spermatogonia, (2) that differentiate continuously, leading to the presence of a permanent pool of primary spermatocytes, (3) which begin progressing through spermatogenesis, (4) but which eventually degenerate in an apoptosis-like process, before the end of their differentiation; (5) the stage at which the spermatogenetic cells interrupt their development is species-specific and is correlated with the instar at which the species diapauses, either as a larva or pupa. Thus, in C. pomonella, a species that enters diapause during the last larval instar, the spermatocytes stop developing and die at the diffuse stage. The testes of diapausing C. pomonella lack any indication of apyrene development since the spermatocytes die before ASIF becomes active close to the time of pupation (Friedla¨nder and Benz, 1982). The onset of this spermatogenic arrest is directly and reversibly related to fluctuations in the titre of JH. Accordingly, juvenoid applications to penultimate larvae reared under diapause-averting conditions cause spermatocyte death and discontinuity of spermatogenesis.

254


However, the persistence of this discontinuity in spermatogenesis is unrelated to the maintenance of a high titre of JH, since both diapause and the discontinuity of spermatogenesis can extend for months, while the JH titre drops drastically during the first month of diapause. On the other hand, allatectomy of the last instar larvae that are otherwise predetermined to enter diapause, prevents both diapause and the arrest of spermatogenesis (Friedla¨nder, 1982) indicating that JH is essential for the onset of both diapause and the interruption of sperm production. However, the timing of the death of the spermatogenous cells differs in M. sexta. In this species, and other Lepidoptera that diapause as pupae, diapause is due to the failure of the prothoracic glands to secrete ecdysteroid moulting hormones (Denlinger, 1985). In Manduca, both eupyrene and apyrene spermatogenesis are initiated throughout diapause, but in each case the process is interrupted at a different specific stage before it can be completed. Interruption of eupyrene spermatogenesis in this species occurs quite late; the eupyrene spermatids die only after their axonemes have extended considerably and nuclear elongation has began. In contrast, apyrene spermatogenesis is discontinued relatively early, before the first division of meiosis (Friedla¨nder and Reynolds, 1992). In another lepidopteran species that diapauses as a pupa, the swallowtail moth, Atrophaneura alcinous, spermatogeneous cells of diapausing insects die at the secondary spermatocyte stage, before the beginning of spermiogenesis (Kubo-Irie et al., 1999a). This again shows that death of the spermatogeneous cell during diapause occurs at a genetically predetermined specific stage of spermatogenesis. Cell death during development, including that of germ cells (Print and Loveland, 2000), is now generally acknowledged to occur by apoptosis, with the active participation of an intrinsic cell death programme within the prospective dying cell. The death of spermatogenous cells during lepidopteran diapause is no exception, and indeed Kubo-Irie et al. (1999a) have shown that in A. alcinous both the prospective and actually dying cells clearly display apoptotic characteristics, as shown by their ultrastructure, supravital staining, and the use of the TUNEL technique to show the presence of double-stranded breaks in nuclear DNA. As discussed above, the apoptotic process is closely integrated with the development of the whole insect, and is thus controlled by a complex mechanism in which the two main circulating morphogenetic hormones, ecdysteroid and JH, are both involved. However local hormones could also play a role. For example, it is still unknown whether fluctuations in the titre of intratesticularly synthesised ecdysteroids (Loeb et al., 1988) are involved in the diapauserelated discontinuity of spermatogenesis. Also, as indicated previously, it is not in any case known that the actions of ecdysteroids on spermatogenesis are direct.


6

255

Sperm movement and transfer

6.1

RELEASE OF EUPYRENE AND APYRENE SPERM FROM THE TESTIS INTO THE MALE TRACT

6.1.1

Spermiation

Spermiation is the displacement of spermatozoa from the testes into the male genital duct. The following description of the cellular processes of spermiation in Lepidoptera is based mainly on the works of (B. mori) Oˆmura (1936b) and Katsuno (1977a), (L. dispar) Giebultowicz et al. (1997), and (Spodoptera littoralis) Bebas et al. (2001). The bundles of spermatozoa leave the testes through a one cell-thick layer, which is currently designated as the terminal epithelium, although it lacks a basement membrane (Giebultowicz et al., 1997). In earlier light-microscope studies (Oˆmura, 1936b), the same structure was called the ‘‘membrana basilaris’’, but the term is inappropriate because this structure contains not only extracellular components, but also cells. On the other hand, because it has no basement membrane, the terminal epithelium corresponds only imperfectly to the definition of a real epithelium. Nevertheless, we will continue to use the name terminal epithelium for the sake of continuity. A terminal epithelium of this type separates each follicle singly from the lumen of the vas efferens. The epithelium lies within the plane of a ring of striated muscle cells that surrounds this junction between the follicle and the vas efferens. The muscle does not penetrate between the follicles. The muscular ring undergoes myogenic contractions changing the size of the passage to the vas efferens, despite the fact that the ring lacks innervation. Before spermiation, the cells of the terminal epithelium display cytoplasmic interdigitations that effectively occlude the passage of sperm from the follicle into the vas efferens. During this period, cysts containing either eupyrene or apyrene spermatozoa can be seen to be located at right angles to the terminal epithelium, in a ‘‘head-first’’ position i.e. with their head tips close but unattached to the epithelium. Oˆmura (1936b) described in detail how, subsequently, the entire cyst vibrates and bends in the middle through 1801, thus acquiring a ‘‘U-shaped’’ conformation. Consequently, both the head and flagellar tips of the U-shaped cyst are now close to the epithelium. After bending, the tips of the flagella press on the testicular face of the epithelium. Then, the follicle, like a spring, returns to its unbent–erect configuration. But it is the head ends of the spermatozoa that move, and not the tails. Consequently, the sperm are now in a ‘‘tail-first’’ position, and the flagella will precede the heads of the sperm during their passage through the epithelium. All authors agree that the eupyrene bundles do indeed exit the testis tail first. Katsuno (1977a), however, specifically commented that he did not observe the previous head-first orientation of the eupyrene bundles. This could represent a

256


difference between the strains of silkworm used by these two authors, but was probably simply because Katsuno did not observe the insects at the right time of day. The reorientation of the eupyrene bundles may be responsible for the movement that has been observed in some intratesticular spermatozoa in in vitro preparations since, otherwise, the spermatozoa show no intrinsic motility within the testes. The coordinated movement of all the spermatozoa within the cyst implies the existence of some form of communication between them. It is not known how this is achieved. It may be relevant however that by this time, all of the spermatozoa with any one eupyrene cyst are radially oriented about their longitudinal axes in the same way and with great precision, relative to the position of the flagella, the mitochondrial derivative and the surface appendages (see Fig. 8). This would in principle enable all the sperm to react to an external signal by bending in the same direction. This argument is however rather spoiled by the observation that apyrene sperm are not all radially oriented about their longitudinal axes in the same way within their own cysts. Thus, either the bending and unbending of the apyrene spermatozoa is coordinated differently from that of the eupyrene spermatozoa, or there is a common mechanism of control that does not require a high degree of precision in their radial orientation. At the beginning of spermiation, the sperm bundles, still covered by the somatic cyst cells, protrude towards the vas efferens through newly opened intercellular channels in the terminal epithelium that are framed by the epithelial cells. During their passage through the epithelium, the bundles of spermatozoa are stripped of their cystic envelopes, which remain trapped within the intercellular channels. However, the lysis of the apical head cyst cell is delayed in the case of the apyrene (but not eupyrene) bundles. The mechanism of sperm release through the terminal epithelium is not known. However, there is some evidence that the process of penetration of the terminal epithelium may involve the cytoskeletal protein actin. Gvakharia et al. (2003) have shown that the apyrene sperm bundles of S. littoralis have conspicuous bundles of actin microfilaments located at their proximal (i.e. ‘‘head end’’) tips, and that inhibition of actin polymerisation using the drug cytochalasin inhibits sperm release. Moreover, when double-stranded b-actin RNA was used to produce RNA interference (RNAi) and thus deplete the amount of actin in the testis, the treatment also inhibited sperm release. All this strongly suggests that the bundle-tip actin microfilaments play an important role in spermiation, but it is still possible that both myogenic contractions of the muscular ring and perhaps autonomous movements of the spermatozoa bundles may additionally be involved in the rapid discharge of the spermatozoa from the testes, both during and after they become free from their cystic envelopes. Spermiation does not occur continuously, but only at certain times of the day, according to a circadian rhythm (see Section 6.1.4). Following the exit of the spermatozoa at the end of the release phase of this rhythm, the channels in


257

the terminal epithelium are empty and their diameter now becomes progressively smaller. Subsequently, the channels are occluded as the epithelium is both reduced in dimensions and reconstructed to its original pre-spermiation condition by intense morphogenetic activity of its cells. It is unknown whether on the following day, during the next period of spermiation, the same already used channels re-open again or, alternatively, whether new, different channels are formed. 6.1.2

Differential release of eupyrene and apyrene spermatozoa

During spermiation, apyrene bundles are the first to cross the terminal epithelium and to appear in the upper male tract (e.g. Riemann et al., 1974; Katsuno, 1977a; Bebas et al., 2001; Seth et al., 2002b). Although apyrene cysts are more numerous than eupyrene cysts, and so might be expected to appear in the upper vas deferens (UVD) first simply on statistical grounds, the precedence of apyrene sperm release has been so consistently reported that it seems unlikely to be due simply to chance, and it is consequently tempting to propose that spermiation of apyrene cysts in some way opens the way for the eupyrene cysts. However, it is unknown whether the two types of cysts cross the epithelium by the same channels or not. 6.1.3

Disruption of the apyrene bundles and maintenance of eupyrene bundles

The lysis of their enclosing cyst cells as they cross the terminal epithelium means that the apyrene bundles are no longer held together along their length. However, the delayed lysis of the head cyst cell means that the complete dissolution of the apyrene bundle is also delayed (Riemann and Giebultowicz, 1992). In at least some species (e.g. L. dispar), as they pass through the terminal epithelium, the apyrene sperm are temporarily closely associated with ‘‘annulate bodies’’, which subsequently disappear. These longitudinally extended structures appear to be axially oriented around or at least to lie in parallel with the apyrene sperm. The material of which the annulate bodies are made appears to be secreted by the cells of the terminal epithelium and the walls of the UVD, and to be associated with the formation of the apyrene sperm envelope. The head cyst cell finally lyses in the UVD and the apyrene bundle is disrupted, so liberating their spermatozoa, and by the time the apyrene sperm reach the seminal vesicle, they are completely dispersed (Fig. 10). This suggests that the apyrene spermatozoa were not originally held together within their cysts by anything other than their simple enclosure within the cellular walls of the cyst envelope, and by their common attachment to the head cyst cell. By contrast, the eupyrene spermatozoa remain organised in their original bundles for the entire duration of their passage through and storage within the male genital duct. This occurs despite the fact that at the time of their passage through the terminal epithelium, the eupyrene bundles lose their enclosing

258


FIG. 10 Structure of eupyrene and apyrene spermatozoa within the male tract. (a) Transverse section through seminal vesicle. Eupyrene spermatozoa remain in bundles and are embedded in a matrix of entangled electron-opaque wavy fibres. Note that somatic cyst cells are absent from the bundle periphery. Each spermatozoon is surrounded by inner and outer envelopes. Reticular appendages (arrows) protrude through a longitudinal slit of the outer envelope. Apyrene spermatozoa ðÞ are not in bundles and are dispersed around the eupyrene bundle. Scale bar ¼ 300 nm. (b) Longitudinal sections through flagellae of eupyrene spermatozoa in seminal vesicle. The upper flagellum (a) displays axoneme (A) and inner (I) and outer (O) envelopes. A tangentially oblique sectioned flagellum (b) displays mitochondrial derivatives (D) and a regular pattern of equidistant parallel striations of outer envelope (E). Scale bar ¼ 200 nm. From Friedla¨nder et al. (2001).

cellular cystic envelope. This indicates that there are additional factors that hold eupyrene sperm together in bundles, which are not present in the intratesticular bundles of apyrene sperm. This cohesion of the eupyrene bundles is probably due at least in part to the presence of a matrix of long electron-dense ‘‘wavy fibres’’ (Fig. 10), which forms in the spaces between the individual


259

sperm of the bundle during its passage through the UVD (Friedla¨nder et al., 2001). However, this begs the question of why the eupyrene bundle does not disperse immediately on its passage through the terminal epithelium. It may be that there are other factors that hold the eupyrene sperm together at this time, before the matrix containing the wavy fibres is formed or, alternatively, perhaps the matrix is already beginning to be formed but is not seen in the microscope. 6.1.4

Circadian rhythm of sperm release from the lepidopteran testis

Spermiation has been reported to be subject to a daily rhythm in several species belonging to different lepidopteran families, and which may therefore be universal. Among the species known to be rhythmic in this way are E. ku¨hniella (Riemann et al., 1974), Pectinophora gossypiella (La Chance et al., 1977), L. dispar (Giebultowicz et al., 1988), C. pomonella (Giebultowicz and Brooks, 1998), S. littoralis (Bebas et al., 2001) and S. litura (Seth et al., 2002b). In all these species, the release of spermatozoa from the testis into the UVD occurs only at night. In several of them, rhythmic spermiation has been shown to be controlled by an intrinsic circadian clock, located in the testes–genital duct complex, since the rhythm is maintained in vitro by such complexes, isolated and cultured in absence of any exogenous neuroendocrine contribution (Giebultowicz et al., 1989; Bebas et al., 2001). The circadian rhythm of sperm release is markedly affected in vivo by exposure of male insects to constant light (Riemann et al., 1974; La Chance et al., 1977; Giebultowicz et al., 1988, 1990; Bebas and Cymborowski, 1999; Seth et al., 2002b), a fact that has implications for the design of insect mass release-rearing programmes. The photoreceptor that entrains the rhythm is extraretinal, since the rhythm can be shifted in vitro by changing the photoperiodic regime, and even abolished by exposing in vitro preparations to constant illumination (Giebultowicz et al., 1989). The rhythm of sperm release can be entrained by both light and temperature zeitgebers (timing signals) (Syrova et al., 2003) and, like constant light, constant temperature has a deleterious effect on sperm release and fertility (Proshold, 1997). 6.2 6.2.1

DESCENT OF EUPYRENE AND APYRENE SPERM ALONG THE MALE TRACT

Sperm movement and storage in the male tract

Neither eupyrene bundles nor individual apyrene sperm are motile while they are in the male tract, and their movement along it must therefore be powered by peristaltic movements of the tract walls. Such movements are known to occur in the upper vas deferens, and have been reported to follow a daily rhythm (Giebultowicz et al., 1996). The dynamics of movement of sperm along


260

the tract have been described in some detail for S. litura by Seth et al. (2002b). In this insect, as in others (see above) release of sperm from the testis into the UVD occurs during the early part of the night. But subsequent movement of sperm into the seminal vesicle (SV) appears to take some time and sperm numbers do not peak in the SV until about 8 h later. Rhythmicity of movement from the SV to the duplex is less obvious, because sperm accumulate progressively in this organ, but the time taken for sperm to travel from the SV to the duplex appears to be approximately 12 h. The site of storage of sperm varies among Lepidoptera. In some, the SV is the principal site of storage (e.g. B. mori, Omura, 1938; Hyalophora cecropia and other saturniid silkmoths, Shepherd, 1974a; M. sexta, Reinecke et al., 1983; A. alcinous, Kubo-Irie et al., 1999b). In other species (T. ni, Holt and North, 1970; E. ku¨hniella, Riemann et al., 1974; P. gossypiella, La Chance et al., 1977; H. virescens, Proshold, 1991; Polygonia c-aureum, Hiroyoshi, 1995; P. rapae, Wedell and Cook, 1999a; S. litura, Seth et al., 2002b) sperm accumulate in the duplex and not in the SV. 6.2.2

Morphological changes in sperm in the male tract

In general, eupyrene sperm are modified considerably in appearance as they pass along the male reproductive tract, while the extent of change in apyrene sperm is much less notable (Figs. 8 and 10). The process is essentially the same in all the species that have been studied. In E. ku¨hniella (Riemann and Thorson, 1971), C. ethlius (Lai-Fook, 1982b), L. dispar (Riemann and Giebultowicz, 1991, 1992), A. alcinous (Kubo-Irie et al., 1999b), M. sexta (Friedla¨nder et al., 2001) and Euptoieta hegesia (Mancini and Dolder, 2003), the apyrene sperm acquire an outer sheath or envelope as they pass through the terminal epithelium and enter the lumen of the vas deferens. This envelope is continuous over the entire surface of the sperm, and specifically lacks the axial slit that is present in the eupyrene sperm (see below). Detailed electron microscopic observations of Riemann and Giebultowicz (1991, 1992) on timed specimens of L. dispar strongly suggest that the envelope is formed from the secretions of the epithelial cells of the UVD. Displaying a rhythmic synthetic activity that coincides with the time when sperm are released, these cells elaborate an electron-dense material, which surrounds the lumenal microvilli in apparently stiff cylinders, displaying an annular structure with axial spacings of 18–20 nm. This material separates off from the microvilli by splitting along one side of the cylinder, and is released into the UVD lumen in the form of annulate bodies. These annulate bodies bear a strong similarity to the material of the apyrene sperm envelope, and it is probable that they are made of the same material. In the first few hours following release from the testes, annulate bodies associate closely with the apyrene sperm in the UVD. Those annulate bodies that do so appear to have no lumen. As noted above, the dissolution of the apyrene sperm bundles is somewhat delayed on entry into the UVD, but once


261

the apyrene sperm have separated, the formation of the outer envelope is seen to be complete, and annular bodies no longer associate with them. The surface decorations of eupyrene sperm develop completely differently. They are decorated with lacinate appendages in the testis, but these are seen to be disintegrating even as the eupyrene sperm are crossing the terminal epithelium. As the appendages disappear, new inner and outer envelopes are formed around each individual sperm. Within a few hours, the lacinate appendages have entirely disappeared, and the inner and outer envelopes are complete, except for the axial opening that extends along its length, and which is plugged by the reticular appendage that survives from the structures that were elaborated within the testis. A number of possibilities suggest themselves to explain the differences between the surface decorations of eupyrene and apyrene sperm. First, as suggested by Andre´ (1962), the lacinate appendages may represent a store of material that is subsequently used to construct one or both of the extracellular envelopes that come to surround individual eupyrene sperm as they travel within their bundles down the male tract (see below). Second, it is possible that the material of the lacinate appendages is remodelled to form the network of ‘‘wavy fibres’’ that apparently holds the eupyrene bundle together until it is transferred to the female. Third, the reticulate appendage may be a device to ensure that when the outer envelope forms around the eupyrene sperm a longitudinal slit is left in the envelope. The reticular appendage plugs this slit, while the eupyrene sperm remains within the male tract, but is removed revealing the slit, when the sperm is activated on transfer to the female. In contrast, the apyrene spermatozoon lacks both the reticular appendage and the slit and, unable to hatch, therefore remains within the envelope. The changes in surface decoration of both types of sperm are probably effected at least in part through the changing synthetic activities of the epithelial cells of the UVD. These cells undergo a daily rhythm of secretion, as seen by periodic acid–Schiff (PAS) staining (a marker for glycoproteins) (Riemann and Giebultowicz, 1991). A gel electrophoretic study of sperm associated proteins showed that several proteins present in sperm bundles while they are in the testis are lost as they enter the UVD. Furthermore, additional proteins appear to be transferred from the walls to the sperm bundles as the latter pass down the tract (Giebultowicz et al., 1994). An additional factor that may play a role in the modification of sperm surface structures is the fact that the contents of the UVD lumen are acidic at the time when sperm are being released from the testes (Bebas et al., 2002). Acidification of the lumen is rhythmic and persists in culture; H+ ion transport is probably effected by the activity of a V-ATPase in the UVD epithelial cells. It is not clear whether acidification is required for normal sperm maturation. One may speculate that acidification of the UVD lumen might play a part in the dissolution of the lacinate appendages, and/or that it is required for the synthesis and/or assembly of new surface decorations.


262

6.3 6.3.1

TRANSFER OF EUPYRENE AND APYRENE SPERM TO THE FEMALE

Mating

Both eupyrene and apyrene sperm are transferred during mating from the male to the female moth. The male introduces the aedeagus (penis) into the female’s genital opening and during the ensuing copula, a spermatophore is formed within the female’s bursa copulatrix from the secretions of the lower male tract (see Fig. 11). The spermatophore is filled with semen (sperm and other secretions); in some species, the mating process is finally completed by the addition of secretions that set within the female’s ductus bursae to form a mating plug (spermatophragma or sphragis), a structure that functions to guard recently transferred sperm. Species that form a spermatophragma frequently have highly modified male genitalia to enable the removal of a previous mating plug, and the frequency of removal may be high in natural populations (Orr, 2002). Even in those species that can form a spermatophragma, it will not be utilised by the male on every occasion, implying that its formation is expensive, and that a trade-off exists between such expenditure and the risk of sexual competition. Both eupyrene and apyrene sperm enter the spermatophore and both kinds of sperm are activated to become motile while in it. The sperm leave the spermatophore and the bursa by swimming along the female’s ductus

FIG. 11 The spermatophore. Diagram of female reproductive system of Bombyx mori after mating, showing the location and orientation of the spermatophore. From Osanai et al. (1987a).


263

seminalis to reach the spermatheca, in which they are stored until they are needed. Only at the time of oviposition will the eupyrene (but not the apyrene) sperm leave the spermatheca in order to meet the descending ova and effect fertilisation. The duration of mating varies greatly between species. In S. litura (Seth et al., 2002a), mating lasts about 1 h, and spermatophore formation and transfer proceed during the whole of this time. Likewise, in Papilio xuthus, mating lasts about 50 min (Watanabe et al., 2000), but in B. mori mating lasts for several hours, during which time there are several separate ejaculations into the bursa (Oˆmura, 1938). In M. sexta, the mean duration of mating is about 3 h (Woods and Stevenson, 1996). 6.3.2

Transfer of sperm during mating

As previously described, prior to mating, eupyrene and apyrene sperm are stored together in the male tract, the exact site of storage varying between species. As mating begins, the apyrene sperm are fully dissociated, but the eupyrene sperm are still held together in bundles. Both kinds of sperm are immotile at this time. It must be presumed that the transfer of sperm into the spermatophore during mating is achieved by propulsive movements of the walls of the male tract. Oˆmura (1938) comments that such movements are evident in vitro after immediate dissection of males taken in flagrante delicto. The number of sperm transferred to the female within the spermatophore varies considerably among lepidopteran species. Thus, the small white butterfly Pieris rapae transfers only around 50 eupyrene bundles (equivalent to 12 800 individual eupyrene sperm) (Watanabe et al., 1998) and around 50 000 apyrene sperm (Watanabe et al., 2000) in a single mating, while a 1-day-old virgin male common cutworm S. litura transfers approximately 350 eupyrene bundles (89 600 individual sperm) and an estimated 268 800 apyrene sperm during the first mating (Seth et al., 2002a). Marcotte et al. (2003) determined the numbers of sperm transferred by the two Choristoneura species to be 20 385 eupyrene, 214 644 apyrene in C. fumiferana, and 28 177 eupyrene and 161 206 apyrene in C. rosaceana. Watanabe et al. (2000) give the numbers of eupyrene and apyrene sperm transferred into the spermatophore in P. xuthus as approximately 11 000 (i.e. 43 bundles) and 160 000, respectively. Although eupyrene sperm can be accurately and relatively easily counted in the spermatophore of a newly mated female since the eupyrene bundles take some time to dissociate, it is hard to count apyrene sperm accurately as they are already fully dissociated before they are deposited in the spermatophore. Therefore, counts of apyrene spermatozoa must be made by diluting semen and counting individual sperm in a microscope slide counting chamber. Moreover in some species, apyrene sperm (but not eupyrene) degenerate quite quickly within the spermatheca, leading to further uncertainty in counts of


264

apyrene sperm unless these are done very soon after mating. The often unchallenged supposition is that eupyrene and apyrene sperm are transferred to the female in the same overall proportions in which they are formed in the testis. This is an assumption that may not be true (see Section 6.3.6) and we suggest that additional efforts should be made to count eupyrene and apyrene sperm separately in the spermatheca. 6.3.3

The spermatophore

Spermatophores vary considerably in size and shape among lepidopteran species. The most complete description of spermatophore formation, structure and content is for the commercial silkworm, B. mori (Oˆmura, 1938; Osanai et al., 1987a). In this species, the spermatophore is ovoid in shape (see Fig. 11) with its opening at the base, located in the neck of the bursa close to the opening of the ductus seminalis. The spermatophragma is located in the ductus bursae at the entrance to the bursa and does not occlude the exit of sperm from the spermatophore into the ductus seminalis. According to Oˆmura (1938), in Bombyx the first material to be transferred by the male to the female is a transparent fluid derived from the male’s glandula prostatica. However, most of the secretion from this part of the male tract is reserved until later. Shortly afterwards, the pearly body, a secretion of the glandula alba of the male tract, appears in the bursa. Osanai et al. (1987a) consider that the pearly body is not a part of the spermatophore proper, but it is continuous with the outermost layer of the spermatophore wall. It is therefore supposed that the material of the pearly body speads laterally over the internal surface of the bursa. No specific function has been attributed to the pearly body. As mating continues, the spermatophore proper is now formed. The walls of the spermatophore are derived from secretions of the glandula spermatophorae. At the same time as the walls of the spermatophore are being formed, the contents are also being transferred. In Bombyx, sperm are transferred into the spermatheca quite early, at the same time as other components (Oˆmura, 1938). Similarly in S. litura, semen is transferred early (Seth et al., 2002a). In some other lepidopterans, however, sperm are transferred only during the latter part of mating. This is true for Colias eurytheme (Rutowski and Gilchrist, 1986), P. rapae (Watanabe and Sato, 1993) and P. brassicae (Tschudi-Rein and Benz, 1990). The spermatophore represents a considerable investment on the part of the male moth. The male transfers 6–7% of his body mass to the female at a single mating in Colias philodice and C. eurytheme (Marshall, 1982); about 7% in P. rapae (Watanabe and Sato, 1993) and M. sexta (Woods and Stevenson, 1996), while in P. napi the spermatophore and its contents represent an average of 15% of male mass (Kaitala and Wiklund, 1994). Spermatophores also contain proteins, sugars and other valuable nutrients (Marshall, 1982;


265

Watanabe and Sato, 1993); that of P. napi contains 14% N (Karlsson, 1998). The mass of the spermatophore is in some cases limited by the intake of nutrients such as sugars (P. xuthus; Watanabe and Hirota, 1999). Following mating in some but not all Lepidoptera, the spermatophore is degraded by the female and its components are absorbed (e.g. Boggs and Gilbert, 1979; Oberhauser, 1992; Watanabe and Sato, 1993). Materials transferred in this way include amino acids (Boggs and Gilbert, 1979) and sugars (Watanabe and Sato, 1993), as well as minerals such as zinc (Engebretson and Mason, 1980), sodium (Pivnick and McNeil, 1988) and phosphorus (Lai-Fook, 1991). In the last case, transport of P was convincingly shown to occur directly across the walls of the bursa. The size of the male’s investment in the spermatophore will be expected to vary according to the mating strategy of the moth in question. The resource cost to the male of increasing the size of the spermatophore is clear. The benefit of the investment may be direct in that it increases not only the male’s fitness by providing energy for sperm movement (see below), but also the female’s fitness by providing resources for egg production (this has been shown for Heliconius spp., Boggs and Gilbert, 1979; Danaus plexippus, Oberhauser, 1989; P. napi, Karlsson, 1998). But there are also indirect benefits to the male’s fitness through enhanced competition with other males: for example, a large spermatophore may lead to more effective inhibition of further mating by the inseminated female. This has been shown for P. rapae (Wedell and Cook, 1999a,b); P. xuthus (Watanabe, 1988); D. plexippus (Oberhauser, 1989); Pseudaletia separata (He and Tsubaki, 1991); Papilio machaon (Svard and Wiklund, 1991); and P. napi (Kaitala and Wiklund, 1994). Thus, a high degree of female choice in selecting a mate might be predicted to lead to the maximisation of spermatophore size, probably revealed as a strong relationship between male size and spermatophore size. On the other hand, an increasing degree of male choice might lead to males investing more in high-quality females, probably revealed as a strong relationship between spermatophore size and female size. But these predictions are complicated by the observations that high-quality ð¼ largeÞ females may mate more often, as may those females that have received smaller spermatophores. Both of these eventualities will lead to increased risk of post-copulatory female choice (sperm competition), so that there are additional reasons to suppose that spermatophore size will be maximised. In fact, in almost every case where the matter has been investigated, size (mass) of the spermatophore has been found to be significantly positively correlated with the body size (mass) of the male. This is true for the European cornborer, Ostrinia nubilalis (Royer and McNeil, 1993) and the pierid butterflies P. napi and P. rapae (Bissoondath and Wiklund, 1996). The case of two closely related moths of the genus Choristoneura is however instructive. In the obliquebanded leafroller, C. rosaceana spermatophore size is positively correlated with male mass (Delisle and Bouchard, 1995). The same is true of the spruce budworm, C. fumiferana (Delisle and Hardy, 1997); although the


266

relationship was not quite significant in another study (Marcotte et al., 2003). On the other hand, spermatophore size was found to be highly significantly correlated with female mass in C. rosaceana, while the relationship was only just significant in the related spruce budworm, C. fumiferana (Marcotte et al., 2003). This implies that the mating strategy of C. rosaceana may involve a greater element of male choice than is usual among Lepidoptera. 6.3.4

The number of sperm transferred

On the other hand, the number of sperm transferred to the female has been found to be independent of the spermatophore size in a number of Lepidoptera. This is true for H. virescens (LaMunyon and Huffman, 2001), and both C. rosaceana and C. fumiferana (Marcotte et al., 2003). This implies that the selective forces leading to the evolution of large spermatophores are independent of post-mating sperm competition. This may well be true for these two moth species, since in general many more sperm are transferred than are necessary to fertilise all of the eggs that the female can lay, and since a second mating by the female secures the complete displacement of the previous male’s sperm (e.g. S. litura, Seth et al., 2002a). This means that the main selective force acting to shape the male’s reproductive strategy will be to impose on the female the avoidance of a second mating. The total number of sperm transferred to the female is influenced by the male moth’s age. In S. litura, a 0–1-day-old virgin male transfers an average of 338 eupyrene bundles when mating with a 0–1-day-old virgin female, but this figure rises to 407 bundles for a 1–2-day-old virgin male, and 518 bundles for a 2–3-day-old virgin male (Seth et al., 2002a). These figures represent essentially all of the eupyrene sperm within the duplex that are available for transfer. Comparison with measurements of the sperm content of the duplex (Seth et al., 2002b) suggests that the percentage of sperm stored in the duplex that was transferred to the female at mating was close to 100%. Thus in 0–1-day-old males, 100% of the sperm in the duplex were transferred; in 1–2-day-old males, 96%; and in 2–3-day-old males, 90%. Although some eupyrene bundles were found to be present in the duplex after mating (0–1-day-old males, 54 bundles; 1–2-day-old males, 73 bundles; 2–3-day-old males, 98 bundles), these were probably sperm that had subsequently descended from the upper regions of the tract. Proshold (1991) similarly concluded that all the sperm that were available in the duplex at the time of mating were transferred in H. virescens. Therefore, an older but still virgin S. litura male moth may transfer more sperm at mating, simply because more are available. On the other hand, however, this trend may not be maintained in all cases since when mating of Plodia interpunctella was delayed, the fecundity and fertility of the inseminated females declined continuously with time (Huang and Subramanyam, 2003). When the delay reached 5 days, the males were unable to inseminate the females at all. The reason for this decline is unknown.


6.3.5

267

Sperm allocation

Increasing sperm transfer with increasing male age at first mating is also probably not true for all Lepidoptera. Hiroyoshi and Mitsuhashi (1999) found that in the butterfly P. c-aureum, not all of the sperm stored in the male tract were used in mating. They found that sperm reflux occurred from the duplex into higher regions of the male tract, and that this allowed sperm to be retained for subsequent matings. The implication of this is that some degree of male choice exists in the allocation of sperm, and that the male can assess the quality of the female with whom he mates and/or the probability of finding another female. Evidence that at least some lepidopterans can indeed allocate sperm according to circumstance has been found in P. rapae (Wedell and Cook, 1999a,b). Male butterflies of this species at their second mating were shown to retain sperm within the duplex, and to adjust the number of sperm transferred to the female according to the risk of sperm competition. Interestingly, when mating for the first time, males did not respond to the risk of sperm competition according to the mating status of their female partner, but they did provide more sperm to larger females. Since larger females remate sooner, this behaviour does in fact follow the prediction of sperm competition theory (Pizzari and Birkhead, 2002; Wedell et al., 2002) that more sperm should be transferred where the risk of competition is greater. The prevalence of sperm retention by the male and/or retrograde movement of sperm at mating in Lepidoptera other than butterflies has been little investigated. H. virescens does not alter the number of sperm transferred to the female, even under conditions where the size of the spermatophore is varied (LaMunyon and Huffman, 2001), which implies that in this species the number of sperm transferred during mating is not subject to male control. No evidence for the retention or reflux of either eupyrene or apyrene sperm was found in S. litura (Seth et al., 2002a), and as noted above, it appears that in this species all available sperm are transferred. However, these last studies were on laboratory pairings; the same might not be true under natural conditions. The number of sperm transferred to the female during mating is also affected by the male’s previous mating history. In those moths that transfer all the available stored sperm at each mating, it stands to reason that this must be so. In S. litura, the number of eupyrene bundles transferred at mating by 1–2-dayold males was reduced from 408 to 78 if the moth had mated on the previous night. This corresponds with a dramatic reduction in the number of eupyrene bundles stored in the duplex (Seth et al., 2002b). Male H. virescens also transferred fewer sperm in their second mating (LaMunyon and Huffman, 2001). 6.3.6

Differential transfer of eupyrene and apyrene sperm

Since eupyrene spermatogenesis precedes apyrene spermatogenesis, it is a reasonable question whether the proportions of the two sperm types in the


268

ejaculates of male moths vary with age and/or mating history. In some cases where the question has been examined, it seems that neither eupyrene nor apyrene sperm is preferentially transferred to the female during mating. Watanabe et al. (2000) interrupted mating in P. xuthus in order to follow the process of sperm transfer into the spermatophore and found that transfer of eupyrene and apyrene sperm was roughly parallel. While this does not prove that the insect cannot select either of the two sperm types for transfer, it is at least consistent with the hypothesis that there is no such selection. In H. virescens, the proportion of eupyrene and apyrene sperm transferred to the female did not vary according to the perceived risk of sperm competition (i.e. when mating with previously mated females) (LaMunyon and Huffman, 2001). However, the butterflies P. rapae (Cook and Wedell, 1996) and P. napi (Cook and Wedell, 1999) have been shown to transfer fewer eupyrene at the first than at the second mating, while the number of apyrene sperm transferred did not differ between these two events. In the case of P. napi, the ratio of apyrene to eupyrene sperm fell from 8.7:1 at the first mating to 5.3:1 at the second. This is actually the reverse of what would be expected if the first formed eupyrene sperm were released first. Nevertheless, this observed change in the proportions of eupyrene and apyrene sperm transferred according to the age and sexual experience of the male moth still does not prove that he is able to handle the two types of sperm differentially in the male tract. Sperm might be transferred only on the basis of availability. This is a question that requires study. 6.4

SPERM RETENTION BY FEMALE MOTHS AND EFFECTS ON FEMALE SEXUAL BEHAVIOUR

There are direct consequences of mating on the behaviour and physiology of the female moth. The number of eggs laid is directly influenced by the number of eupyrene sperm transferred by the males of Acrolepiopsis assectella (Thibout, 1979), and S. litura (Seth et al., 2002a). In B. mori, the main factor eliciting oviposition is the presence of mature eupyrene sperm in the spermatheca (Karube and Kobayashi, 1999). On the other hand, in H. virescens (Park et al., 1998) and Helicoverpa armigera (Jin and Gong, 2001), oogenesis and oviposition are stimulated by factors from the male accessory glands. Additionally, as a result of mating, the female moth becomes sexually unreceptive for a characteristic period of time. This too has been shown in some cases to be determined by factors derived from the testis (e.g. H. cecropia, Riddiford and Ashenhurst, 1973; M. sexta, Sasaki and Riddiford, 1984), although this has not been proven to be due to the presence of sperm. In L. dispar, suppression of female sex pheromone production has been shown to be a two-step process involving both the physical stimulus of mating and also the transfer of sperm (Giebultowicz et al., 1991; Raina et al., 1994). In other cases, e.g. Helicoverpa zea, the inhibition of female receptivity is


269

mediated not by sperm, but by components of the spermatophore derived not from the testis but the accessory glands (Raina, 1989; Raina et al., 1994). One such component has been identified as a pheromonostatic peptide (Kingan et al., 1995). It is not known how the number of eupyrene, as opposed to apyrene sperm, in the spermatheca would be assessed by the female. One possibility is that because the two types of sperm are differentially sorted within the spermatheca (see Section 7.2), with eupyrene sperm being located primarily in the utriculus, all that is necessary would be to monitor the state of fullness of the utriculus. Another possibility is that one or more chemicals carried on the surface of eupyrene (but not apyrene) sperm are directly monitored by chemoreceptors in the spermatheca wall. Accordingly, the composition of the surface of the spermatozoa differs considerably within the spermatheca between the two types of spermatozoa as the eupyrene ‘‘hatches’’ from its extracellular envelopes while the apyrene does not, remaining entrapped within its extracellular envelopes (Friedla¨nder et al., 2001). Female H. virescens mate with more than one male, and store sperm from each mating, so that sperm competition ensues. Sperm precedence does not consistently favour the first or second male. LaMunyon (2000) studied sperm retention by females in order to understand why and how the pattern of paternity varied in the multiply mated females. Females stored more sperm from older males, and older males were more likely to gain sperm precedence over younger rivals. The advantage of the older male is likely due to his increased sperm count (LaMunyon and Huffman, 2001). The amount of sperm stored by the female was positively correlated with female size, male size and spermatophore size. Interestingly, in this species, it was found that the amount of eupyrene sperm stored in the spermatheca was never more than the equivalent of the content of one ejaculate, whereas more than one ejaculate’s worth of apyrene sperm was stored. This underlines the point that in this species, apyrene sperm have a relatively long life within the female reproductive tract, and are not quickly destroyed, as occurs in Bombyx for example (see above). LaMunyon’s (2000) finding also implies that eupyrene sperm are selectively ejected (or destroyed) following the second mating. This must be dependent on the sorting of sperm within the spermatheca (see below).

7

Behaviour of eupyrene and apyrene sperm in the female

7.1 7.1.1

SPERM MIGRATION FROM BURSA TO SPERMATHECA

Dynamics of sperm movement within the female

Once they have been transferred to the female, both eupyrene and apyrene sperm migrate from the spermatophore to the spermatheca, the site of sperm

270


storage in the female. It is not certain whether this migration relies only on either sperm motility or contractions of the walls of the ductus seminalis or on both. Since apyrene sperm are pre-dissociated they are able to leave the spermatophore more quickly than eupyrene sperm. Accordingly, it has been noted in several cases that apyrene sperm arrive at the spermatheca earlier than their eupyrene counterparts (C. pomonella, White et al., 1975; Antheraea pernyi and M. sexta, Silberglied et al., 1984; P. brassicae, Tschudi-Rein and Benz, 1990; S. litura, Seth et al., 2002a; Choristoneura fumiferana and C. rosaceana, Marcotte et al., 2003). The difference may be quite dramatic. In the case of the swallowtail butterfly P. xuthus (Watanabe et al., 2000) apyrene sperm begin to arrive in the spermatheca only an hour after mating, but eupyrene sperm do not begin to arrive until about 1 day after mating. On the other hand, it has been reported that in a number of other species, transport of eupyrene and apyrene sperm to the spermatheca is synchronous (T. ni, Holt and North, 1970; H. virescens and H. subflexa, Proshold et al., 1975; B. mori, Katsuno, 1977a; Diatraea saccharalis, Miskimen et al., 1983; L. dispar, Proshold, 1995; P. separata, He et al., 1995). Transfer of sperm to the spermatheca can be quite fast. In the most rapid case so far documented, S. litura, apyrene sperm arrive in the spermatheca as early as 30 min after the mating pair have separated, and eupyrene sperm follow only 15 min later. Although the progressive deflation of the spermatophore over this period suggests that sperm transfer from this structure may continue for as long as 12 h after mating, most sperm arrive in the spermatheca much sooner than this. Visual observations of sperm density suggested that a plateau was reached after only 90 min (Seth et al., 2002a). The dynamics of sperm movement within the female reproductive tract vary considerably among lepidopteran species, but there seems to be little correlation with size or taxonomic affiliation. As in S. litura, sperm movement is also rapid in the much larger moth B. mori (Suzuki et al., 1996). In this species, sperm begin to enter the spermatheca only 1 h after mating ended, and the spermatheca was full 1.5 h later. On the other hand, Proshold et al. (1975) found that in H. virescens and H. subflexa (moths similar in size to S. litura) sperm took at least 3 h to arrive in the spermatheca. In the butterfly P. brassicae, sperm were first observed in the spermatheca 5.5–8 h after mating (Tschudi-Rein and Benz, 1990). In the two Choristoneura species studied by Marcotte et al. (2003), arrival of apyrene sperm in the spermatheca occurred between 2 and 4 h after mating, but eupyrene sperm took much longer (9–11 h). Maximum sperm density within the spermatheca was not achieved until 9–12 h after mating. Marcotte et al. (2003) noted that the first arrival of eupyrene sperm was at roughly the same time that the density of apyrene sperm reached its peak. With the continued passage of time after mating, the numbers of eupyrene sperm in the spermatheca continued to increase, while apyrene sperm numbers decreased, so that the eupyrene:apyrene ratio was reversed. This reversal was closely associated with the onset of oviposition.


7.1.2

271

The spermatophore and the source of energy for sperm migration

Dissociation of eupyrene bundles and the acquisition of motility by both eupyrene and apyrene sperm takes place within the spermatophore. The process has been studied in detail in B. mori (Osanai et al., 1987a). As the ejaculated sperm become motile, characteristic changes take place within the spermatophore’s inner and outer matrices. The inner matrix of the spermatophore contains both eupyrene sperm bundles and individual apyrene sperm. It also contains male secretions that include granules of glycogen, as shown by PAS staining and by digestion with amylase, and also a basophilic (haematoxylin-staining) material. The outer matrix contains no sperm, but abundant granules of both types as well as non-staining vacuoles. Sperm activation occurs during 80 min after mating. During this time, both glycogen and basophilic granules are depleted, and the extent of unstained spaces within the spermatophore increases. Osanai et al. (1987a) interpreted these structural changes to be consistent with a model advanced by them, which proposes that Bombyx sperm obtain energy for movement by degrading arginine-rich proteins to glutamate, while glycogen is simultaneously broken down to pyruvate. These two anaerobic pathways are then coupled to allow the further conversion of the end products to alanine and succinate. The disappearance of the PAS-positive granules represents the breakdown of glycogen by glycolysis, while the disappearance of haematoxylinstaining granules represents the utilisation of arginine-rich storage proteins. The model is additionally supported by the presence in the spermatophore of arginase (Osanai et al., 1986) and glycogen phosphorylase (Osanai et al., 1995), together with the accumulation of alanine (Osanai et al., 1987b) and succinate (Osanai and Isono, 1997) in the spermatophore after mating. The evidence in favour of the linked arginase and glycolytic pathways in Bombyx sperm is thus quite strong. Osanai and Isono (1997) have pointed out that the metabolic pathways used by Bombyx sperm are in fact quite similar to those employed by parasitic helminths and intertidal bivalves, both of which must operate under hypoxic or anaerobic conditions. Whether these metabolic pathways operate in other Lepidoptera is unknown. Also, it is not clear even in Bombyx whether the metabolism of eupyrene and apyrene sperm employs the same pathways. 7.2

SPERM SORTING IN THE SPERMATHECA

It has been noted in a number of cases that eupyrene and apyrene sperm are separately sorted within the spermatheca (e.g. B. mori, Katsuno 1977b; S. litura, Seth et al., 2002a). Although the two types of sperm are separately formed within the testis in distinct eupyrene or apyrene cysts, once they have been released they always occur concomitantly within the male reproductive tract, and are transferred together to the female during mating. Once they

272


arrive at the spermatheca, however, the two types of sperm become separated. The spermatheca has two recognisable subdivisions, a large utriculus and a smaller lagena. Eupyrene sperm are found mainly in the utriculus, while apyrene sperm are found in the lagena. This assortment has been studied in considerable detail for B. mori by Katsuno (1977b). It was found (Fig. 12) that apyrene and eupyrene sperm arrived together in the large lobe ð¼ utriculusÞ of the spermatheca, being present in large numbers from about 105–110 min after the beginning of copulation. At this time the eupyrene sperm were well dispersed, being mixed with a relatively small number of individual apyrene sperm. Within a few minutes after this, however, larger numbers of apyrene sperm were observed to have separated from the eupyrene sperm and to occur within the spermatheca in large ‘‘masses’’. These apyrene masses very quickly disappeared from the utriculus and appeared in the lagena, in which only small numbers of eupyrene sperm were observed at this time. Katsuno described this change in location of the apyrene masses as ‘‘migration’’ but it is far from clear that the sperm sorting he observed is due to active migration on the part of the apyrene sperm. If such active migration were indeed the case, then some form of chemotaxis would have to be considered, but no direct evidence for this exists. Further, Katsuno’s description makes it clear that the accumulation of masses of apyrene sperm in the lagena occurs at a time when additional sperm, both

FIG. 12 Sperm sorting in the spermatheca. Schematic representation of events in Bombyx mori. (A) 105–110 min after commencement of mating. Masses of apyrene spermatozoa are observed in the large lobe of the spermatheca. (B) 110–120 min. Masses of apyrene spermatozoa migrate from large lobe to small lobe. (C) 140–210 min. Masses of apyrene sperm in small lobe are dispersed. (D) 225–300 min. Eupyrene spermatozoa in large lobe migrate to small lobe as apyrene sperm in small lobe degenerate. (E) 330–360 min. A considerable number of apyrene sperm in large and small lobes have degenerated. The female insect was separated from the male 30 min after mating commenced. ag, accessory gland; ll, large lobe; sl, small lobe; mas, mass of apyrene spermatozoa. From Katsuno (1977a).


273

apyrene and eupyrene are entering the spermatheca. Thus, bulk movement of masses of sperm might be a consequence of ‘‘pushing from the back’’ by the crowd of newly entered sperm. The mechanism for the assortment of eupyrene and apyrene sperm remains unknown. We suggest that two (highly speculative) possibilities must be considered. First, maintained beating of the sperm flagella may lead to a mechanical self-sorting of the different-sized sperm morphs. Second, differential adhesion of sperm to each other or to the walls of one division of the spermatheca might lead to an affinity-based separation of the sperm types. Such differential adhesion might arise through mechanical or molecular means but the two possible mechanisms are by no means exclusive. It is known that mammalian sperm adhere to the specific regions of the female reproductive tract by means of a carbohydrate–protein interaction, with mannosyloligosaccharides expressed on the surface of oviductal epithelial cells being bound by sperm-associated lectins (Topfer-Petersen et al., 2002). A similar mechanism of sperm adhesion could explain sperm sorting in the lepidopteran spermatheca. The necessary adhesive molecules might reside in the different surface structures of the eupyrene and apyrene sperm (Section 6.1.4). This could be related to the fact that in the spermatheca the eupyrene spermatozoa discard their extracellular envelopes while the apyrene ones do not. The significance to lepidopteran reproductive biology of this sorting of eupyrene and apyrene sperm within the spermatheca will become evident below. 7.3

LOSS OF SPERM FROM THE SPERMATHECA

Katsuno (1977b) observed that in B. mori, the masses of apyrene sperm in the lagena soon ‘‘degenerate’’ and disappear (225–300 min after copulation begins). By 225 min, eupyrene sperm were beginning to enter the lagena in appreciable numbers. By 330 min, when the number of apyrene sperm in both lobes of the spermatheca was considerably reduced from its peak at 120–310 min, those female moths that had separated from the males began oviposition. Similar rather rapid losses of apyrene sperm from the spermatheca have also been described for P. separata (He et al., 1995), Choristoneura rosaceana and C. fumiferana (Marcotte et al., 2003). In the case of P. xuthus (Watanabe et al., 2000), apyrene sperm persist in the spermatheca for only about 12 h, while eupyrene sperm were still present 7 days after mating. Watanabe did not determine the fate of the lost apyrene sperm. He speculates that they may migrate out of the spermatheca, or alternatively that they may be destroyed. However, no evidence of the selective death or phagocytosis of apyrene sperm within the spermatheca has so far been adduced in these species. Selective loss or degeneration of apyrene sperm does not occur in all Lepidoptera, however. In M. sexta, Friedla¨nder et al. (2001) observed that 2–3 days after mating both eupyrene and apyrene sperm were still present.


274

Apparently healthy sperm of both types could be seen, although sperm of both types that were unable to ‘‘hatch’’ and had degenerated, were also present. In some species there appears to be individual variation among females in the persistence of apyrene sperm within the spermatheca, and this has behavioural consequences. Cook and Wedell (1999) showed that there was a strong inverse correlation between the number of apyrene sperm in the spermatheca and the tendency to remate female P. napi butterflies (more sperm was associated with less remating), but there was no correlation between mating behaviour and the number of eupyrene sperm. Interestingly there was a genetic component to the retention of apyrene sperm by individual females, the tendency to which was inversely correlated with the tendency to multiple mating (Wedell, 2001). This implies that the persistence or otherwise of apyrene sperm in this species is an attribute of the female rather than the male, and thus that apyrene sperm are killed or expelled rather than just age and die.

8

Sperm maturation

8.1 8.1.1

SPERM ACTIVATION

Acquisition of sperm motility

Lepidopteran sperm remain immotile as long as they are stored in the male tract. Motility is acquired only at the time of transfer to the female (B. mori: Oˆmura, 1938; Katsuno, 1977b; Osanai et al., 1987a; saturniid silkmoths: Shepherd, 1974a; A. alcinous: Kubo-Irie et al., 1999b; M. sexta: Friedla¨nder et al., 2001). The process of ‘‘activation’’ by which motility is acquired clearly differs between eupyrene and apyrene sperm, since the former (but not the latter) are stored in the form of sperm bundles in which the ability of individual sperm to move is strongly constrained by their entrapment in a fibrous, extracellular matrix. These matrix fibres disappear during the activation of eupyrene sperm. Moreover, the anatomical changes that occur in eupyrene and apyrene sperm during their activation also differ markedly (Friedla¨nder et al., 2001). Despite these differences, there are indications that at least some of the controls that regulate sperm motility may be the same for both eupyrene and apyrene sperm. These controls, however, probably differ between lepidopteran species. Most is known about sperm activation in the commercial silkworm, B. mori. Oˆmura (1936a, 1938) was the first to note that the sperm of this insect depend on a secretion of the prostatic region of the male reproductive tract for their activity. This observation has been greatly extended in a series of papers by Osanai and his colleagues (Aigaki and Osanai, 1985; Aigaki et al., 1987, 1988, 1994; Kasuga et al., 1987; Osanai et al., 1987a,b, 1988, 1989a,b, 1991; Osanai and Chen, 1993; Osanai and Baccetti, 1993; Osanai and Kasuga, 1990; Osanai and Nagaoka, 1990). In summary, activation of Bombyx sperm appears to be a


275

two-stage process involving the presence of both proteases and cAMP in the reproductive tract. 8.1.2

Regulation by endogenous proteases

Oˆmura (1936a) observed that sperm taken from the seminal vesicle of Bombyx were incapable of fertilisation in an artificial insemination procedure. But if incubated with an extract of spermatophores taken from mated females, or of the prostatic region of the male tract (other regions were ineffective), then these sperm were activated to be fertile. Oˆmura (1938) noted that the epithelial cells of the prostatic region of the male tract have a secretory appearance. It was also shown that the prostatic-activating factor was water soluble, and although Oˆmura concluded that the factor was ‘‘stable in some degree under heat treatment’’, the data in his paper actually show that its activity was quickly lost when subjected to high temperatures. Two distinct steps towards activation of Bombyx sperm can be mimicked by treatment of stored semen from the seminal vesicle with bovine trypsin (Osanai et al., 1989a). Treatment with this enzyme causes the dissociation of eupyrene bundles releasing individual eupyrene sperm. At the same time, trypsin also causes both eupyrene and apyrene sperm to become motile (Fig. 13). It is not entirely clear whether the action of trypsin in dispersing the eupyrene bundles is a necessary step in causing motility, since the eupyrene sperm become active before the bundles are fully dissociated, but this does not necessarily rule out the possibility that dissolution and motility are in some way connected. Trypsin has similar actions on the stored sperm of M. sexta (Friedla¨nder et al., 2001). The actions of trypsin on Bombyx sperm mimic those of an endogenous proteinase, initiatorin, found in the prostatic region of this insect’s male reproductive tract, the most distal region of the ejaculatory duct. Initiatorin has been purified and characterised (Aigaki et al., 1994). It is a trypsin-like arginine C-endopeptidase. The purified protease is capable of sperm activation (Osanai et al., 1989a). However, it is nevertheless possible that other proteases are present in the male tract and may also contribute to sperm activation. In Manduca, sperm activation also involves an initiatorin-like protease, but its origin in the reproductive system appears to be completely different. Although endogenous proteinase activity has been found that can cause both dissolution of sperm bundles and acquisition of motility, the activating enzyme(s) appears to be produced not in the reproductive tract of the male insect but in that of the female, in the walls of the bursa copulatrix (A. Jeshtadi, R.K. Seth, and S.E. Reynolds, unpublished). 8.1.3

Role of cyclic AMP

In Bombyx, it appears that acquisition of motility, at least in apyrene sperm, requires an additional step, involving the actions of 30 ,50 -cyclic adenosine

276


FIG. 13 Trypsin treatment of spermatozoa taken from the male tract. (a) Spermatozoa from the seminal vesicle of 1-day-old unmated M. sexta, showing an intact eupyrene bundle and isolated apyrene sperm (arrow heads) after 30 min culture in clean HEPES buffer. Scale bar ¼ 0:1 mm. (b) Eupyrene bundle from the seminal vesicle of 1-day-old M. sexta, showing partial dissociation after 15 min culture in 1 mg ml1 trypsin solution in HEPES buffer. Scale bar ¼ 0:2 mm. (c) Eupyrene bundle from seminal vesicle of 1-day-old M. sexta, showing almost complete dissociation after 30 min culture in 1 mg ml1 trypsin solution in HEPES buffer. Scale bar ¼ 0:2 mm. From Friedla¨nder et al. (2001).


277

monophosphate (cAMP). Although protease alone was sufficient to cause acquisition of motility when added to semen from the seminal vesicle, this treatment was not sufficient to cause motility when washed apyrene spermatozoa, i.e. sperm free from seminal plasma, were used (Osanai et al., 1989b). In this case, both trypsin and cAMP were necessary for the acquisition of motility. To explain this result it was supposed that the seminal plasma contains sufficient cAMP to allow motility to be gained, without adding any more, but that in washed sperm the absence of cAMP in the medium prevents the acquisition of motility in response to protease action alone. A similar situation probably exists in at least some other Lepidoptera. We have found (A. Jeshtadi, R.K. Seth and S.E. Reynolds, unpublished) that in both the sphingid M. sexta and the noctuid S. litura, exogenous dibutyryl cAMP (but not dibutyryl cGMP) conferred motility on apyrene sperm in semen taken from the male storage organ. Although cAMP is also known to influence sperm motility in a number of orthopteran insects, cAMP may not play a universal role in Lepidoptera, since cAMP apparently has no effect on sperm motility in the saturniids H. cecropia and A. pernyi (Osanai and Baccetti, 1993). The cellular role of cAMP in the activation of lepidopteran sperm is not understood. It seems likely that exogenous cAMP enters the sperm where it activates a cAMP-dependent protein kinase. cAMP-dependent protein kinases are required for activation of sperm in a number of taxonomically diverse animals (Morisawa, 1994). Probably once inside the sperm, the cAMP acts by regulating protein phosphorylation, as is also the case in mammalian sperm (Tash and Bracho, 1998). How does exogenously applied cAMP (normally an intracellular second messenger) get into the sperm? Trypsin has been shown to cause the formation of micropores within the flagellar membrane of individual apyrene Bombyx spermatozoa (Osanai and Kasuga, 1990) and cAMP could enter the sperm through these openings. Removal of the lacinate appendages of eupyrene sperm could have a similar role in permitting the entry into the sperm cell of cAMP and other regulators. It is nevertheless still not clear whether cAMP acts endogenously to initiate sperm motility, and if so, whether it is synthesised within the sperm cell itself or whether it is made elsewhere. It is also unclear whether cAMP influences eupyrene sperm motility. These spermatozoa are of course confined within bundles when semen is taken from the seminal vesicle, and it is possible that they are unable to move without prior proteinase-mediated dissolution of the bundle. Treatment of eupyrene bundles with cAMP alone does not cause motility. 8.1.4

Dissolution of eupyrene bundles

As previously described (Section 4.3), following release from the testes into the upper vas deferens, a matrix of electron-opaque wavy fibres surrounds the

278


sperm cells within the eupyrene sperm bundles. No such matrix is present around the apyrene sperm, which dissociate soon after their release into the UVD. It appears that it is the wavy fibres that hold the eupyrene sperm together in their bundles, although proof of this is lacking. The wavy fibre meshwork is retained during storage of the spermatozoa in the seminal vesicle of Bombyx, Manduca and Lymantria, but by the time that the sperm reach the female’s spermatheca the wavy fibre matrix is no longer evident. These wavy fibres are lost during transfer of the sperm from the male to the female at mating, presumably due to the action of endogenous activators within the spermatophore. Some of this dissociating action is undoubtedly due to proteases, such as initiatorin in Bombyx. In agreement with this interpretation, when trypsin is used to experimentally activate sperm from the seminal vesicle of Manduca (Fig. 13), the wavy fibres disappear progressively with the increase of either incubation time or trypsin concentration, at the same time as the spermatozoa of the eupyrene bundle acquire motility (Friedla¨nder et al., 2001). On the other hand, Osanai and Isono (1997) have shown that in Bombyx, organic acids present in the spermatophore contents may also act to promote eupyrene bundle dissociation. Succinate appears to be the main contender for this role, being formed in the spermatophore by extracellular glycolysis and coupled arginine degradation under hypoxic conditions. Osanai and Isono (1997) propose that protease action may be most important in dissociating the head part of the eupyrene bundle, while succinate may be more important in dissociating the trunk and tail parts of the bundle. 8.1.5

Maturational changes in the sperm during activation

Again as noted above (Section 4.3), the passage of sperm down the male reproductive tract is accompanied by marked changes in their surface appearance. The detail of this surface decoration varies somewhat between species, but in general the eupyrene sperm in particular undergo particularly marked changes in the male tract, acquiring both inner and outer envelopes, but retaining the already existing long thin reticular appendage (Fig. 10). The apyrene sperm are less ornately attired: they gain an outer sheath but lack the reticulate appendage, the inner sheath is less prominent than in eupyrene sperm and is not present along the whole length of the sperm as in eupyrene sperm. By the time that the sperm reach the female’s spermatheca, further marked surface changes have occurred in the eupyrene but not the apyrene sperm (Fig. 14). It seems likely that at least some of these additional changes are connected with the acquisition of motility. Others may be more analogous with what is called in mammals ‘‘capacitation’’; these late changes are concerned with the ability of the sperm to fertilise the egg rather than with movement (Guraya, 2000). This is not, however, meant to imply that the mechanism of such capacitation is the same in insects and mammals.


279

In Manduca, the most notable change in the appearance of eupyrene sperm, as they are transferred to the female (Friedla¨nder et al., 2001), is that these cells lose their reticulate appendage, and the gap in the outer envelope through which it formerly bulged now appears as an open slit in the sperm coat (Fig. 14). All eupyrene sperm are already modified in this way by the time that they reach the spermatheca, and it seems likely that this structural change is related to the acquisition of motility. With their continued presence in the spermatheca, however, many eupyrene sperm undergo a further change in appearance that cannot be concerned with motility, since these sperm were motile before they left the spermatophore. This further change occurs as the sperm ‘‘hatch’’ from their envelopes, emerging through this slit as ‘‘naked’’ spermatozoa, which leave the spermatheca to fertilise eggs in the oviduct (Fig. 15). They leave behind in the spermatheca large numbers of envelope ‘‘ghosts’’, in which the envelope no longer contains a sperm. Other eupyrene sperm fail to hatch in this way, however, and are left behind in the spermatheca to degenerate within their envelope. It is not clear what distinguishes those sperm that fail to hatch and those that do. However, it is possible that an imperfect opening of the envelope slit may prevent some spermatozoa from hatching. Once in the spermatheca all the apyrene sperm behave in the same way. They show no obvious change in their surface structure while they are in the spermatheca, and they neither hatch nor leave behind empty envelopes. Eventually, they do show signs of degeneration, but this does not involve any obvious change in the structure of their envelope. Similar changes to those occurring to eupyrene sperm in vivo are seen when sperm from the seminal vesicle are treated with trypsin (Friedla¨nder et al., 2001). The proteinase causes the reticulate appendage of the eupyrene sperm to swell and to lift off from the envelope (Fig. 14). With longer incubations and/or higher concentrations of enzyme, there is a progressive loss of material from the eupyrene inner envelope. Eventually, the outer envelopes of some eupyrene sperm are seen to open widely, and the naked sperm within begin to hatch,

FIG. 14 Structure of eupyrene and apyrene spermatozoa in the female reproductive tract of M. sexta. (a) Transverse sections through eupyrene spermatozoa in spermatheca. Note absence of both extracellular matrix and reticular appendages. A longitudinal slit (arrows) in the outer envelope marks the location at which the reticular appendages bulged while spermatozoa were in seminal vesicle. D, mitochondrial derivatives; I, inner envelope; O, outer envelope. (b) Longitudinal sections through eupyrene spermatozoa in spermatheca. Note the parallel, equidistant striation in tangential section of spermatozoon ðÞ and connections between outer (O) and inner (I) envelopes. A, axoneme; D, mitochondrial derivative. (c) Transversal sections through apyrene spermatozoa in spermatheca. The sections are similar in appearance to those of apyrene spermatozoa in seminal vesicle. Note absence of slit in envelope (E). D, mitochondrial derivatives. Scale bar in each case ¼ 200 nm. From Friedla¨nder et al. (2001).

280



281

FIG. 15 Changing structure of sperm in the female reproductive tract of M. sexta. (a) Spermatheca showing sections of empty envelopes from which eupyrene spermatozoa ‘‘hatched’’ before leaving spermatheca to fertilise the eggs. I, inner envelope; O, outer envelope. Scale bar ¼ 200 nm. (b) Section of spermatheca showing various stages of decay of both eupyrene (E) and apyrene (A) spermatozoa within their envelopes. Scale bar ¼ 200 nm. (c) Section through spermatheca showing eupyrene spermatozoa at their late stages of decay. Note open envelopes still containing spermatozoa, empty envelopes and isolated mitochondrial derivatives (D). Scale bar ¼ 200 nm. From Friedla¨nder et al. (2001).


282

leaving ghosts behind. As with the situation in vivo, this experimental exposure to trypsin in vitro appears to have no obvious effect on the apyrene sperm. The observed in vitro changes in sperm surface structure are not totally identical to those seen in vivo, and it appears that trypsin treatment in vitro, especially with high concentrations of enzyme, may damage the inner envelope of eupyrene sperm more extensively than is seen during activation in vivo. This may perhaps reflect the fact that activation in vivo probably involves more than just one proteolytic enzyme (A. Jeshtadi, R.K. Seth and S.E. Reynolds, unpublished). However, it is plain that the consequence of both natural and artificial proteolytic activation of lepidoteran sperm is the removal of the double surface envelope of many eupyrene sperm, while the apyrene sperm remain within their single-layered envelope. Since both types of sperm become motile during activation, however, it is difficult to argue that acquisition of motility per se is due simply to the loss of the envelope. Moreover, it is also clear that hatching of the eupyrene sperm does not occur until the sperm have either been in the spermatheca some time, or have been extensively exposed to trypsin. Therefore, we suggest that proteolytic activation of sperm motility is not achieved through envelope removal, but by some more subtle change, perhaps producing a change in the outer envelope of both types of sperm that enables them to take up nutrients (or perhaps an activating signal such as cAMP) from the seminal plasma. It seems probable that the removal of the double envelope of eupyrene sperm once they reach the spermatheca has some other purpose not directly connected with motility. Since only ‘‘hatched’’ eupyrene sperm can leave the spermatheca and enter the oviduct, it seems likely that the surface decoration may have some role in sperm sorting (see above, Section 7.2). Perhaps the removal of the outer eupyrene sperm coat allows them to escape from an adhesive ‘‘trap’’ within the spermatheca. Accordingly, those eupyrene spermatozoa that do not ‘‘hatch’’ and the apyrene spermatozoa that never ‘‘hatch’’, remain in the spermatheca. Alternatively, the removal of the eupyrene sperm envelope may be required to allow the sperm to acquire the ability to fertilise the egg. Interestingly, mammalian spermatozoa also lose their outer coats during capacitation (Roselli et al., 1990). Further more, it is even possible that these two propositions are not mutually exclusive. 8.1.6

Other factors

In the saturniids, H. cecropia and A. pernyi, Shepherd (1974a,b, 1975) showed that sperm motility is acquired during mating, but that all the necessary controlling factors are derived from the male. Male moths of these species could be induced to form a spermatophore and fill it with sperm, even in the absence of a female (Shepherd, 1974a). The sperm so transferred acquired motility similar to those transferred during a normal mating, showing that the necessary substance(s) responsible for initiation and maintenance of motility must be derived


283

from the male tract. Shepherd (1975) identified this substance as a polypeptide on the basis of its susceptibility to digestion by proteinases, but nothing more was learned of its nature. It is not clear whether the regulation of sperm motility and fertility in saturniids differs from the protease/cAMP model presented above.

9 9.1

The evolutionary rationale of dichotomous spermatogenesis EVOLUTION OF APYRENY

Ever since Meves’ (1903) discovery of apyrene sperm in Lepidoptera, their function has been a topic of keen interest. What possible use is a sperm without a nucleus? It was argued in Section 1.2 that the dichotomous spermatogenesis of lepidopterans must confer an adaptive benefit, and that the characteristics of these anucleate cells, which are maintained throughout the Order, must have evolved as a result of natural and/or sexual selection. Certainly it is now clear, at least in the case of B. mori, that apyrene sperm play an essential role in male fertility; Sahara and Kawamura (2002) have shown experimentally that eupyrene sperm alone are unable to fertilise eggs in vivo, and that apyrene sperm must also be transferred to the female. Since it is self-evident that anucleate sperm can play no direct role in enhancing the male moth’s fitness, apyrene sperm must therefore be ‘‘helper’’ sperm that enhance the fitness of the male moth that produces them by assisting eupyrene sperm of the same insect to fertilise an egg. In other words, apyrene sperm are gametic ‘‘eunuchs’’ that serve the interests of their nucleate counterparts (Silberglied et al., 1984); they are in effect altruists whose evolutionary interests can only be assured through inclusive fitness (Trivers, 1985). It is nevertheless far from self-evident that such altruism can evolve in haploid gametes, because sperm compete among one another within an ejaculate, and the haploid condition of the sperm means that it will be hard for altruists to benefit from inclusive fitness. This situation leads to a conflict of interest between the male parent and his sperm. This is not the place to give a detailed treatment of this theoretical evolutionary issue, and the reader is referred to the paper by Swallow and Wilkinson (2002), which summarises the arguments. Modelling by Kura and Nakashima (2000) has nevertheless shown that the conditions for the evolution of ‘‘soldier sperm’’ (i.e. those contributing only indirectly to fitness) are not particularly hard to achieve. Satisfyingly, they conclude that soldier sperm are likely to be smaller than fertile morphs, as is universally found. They further conclude that the evolution of soldier sperm is favoured if control of sperm phenotype resides with parental, diploid cells, rather than with the haploid sperm. As we have seen, this is in fact the case, it being clear that the decision to develop along the apyrene pathway is taken


284

before the first meiotic division has occurred, at a time when the spermatocyte’s phenotype is determined by genes expressed in the diploid condition. Further, the hormonal signal ASIF that heralds the switch from eupyreny to apyreny circulates throughout the body, and is likely to originate outside the testes, as it is present also in the female (see Section 5.2). Moreover, the possibility of an individual spermatocyte selfishly ‘‘defecting’’ from the apyrene to the eupyrene developmental pathway is limited by the fact that all of the spermatocytes are linked together by intercytoplasmic bridges (see Section 2.2), and therefore share the same intracellular controls on the progress of meiosis. 9.2

POSSIBLE FUNCTIONS OF APYRENE SPERM

A number of hypotheses as to the nature of the ‘‘help’’ provided by apyrene sperm to related eupyrene sperm have been advanced, and have been reviewed by Silberglied et al. (1984), and Swallow and Wilkinson (2002). Our list largely concurs with the latter paper and adds some variants. (a) Apyrene sperm may directly assist through their own active motility in the transfer of eupyrene sperm from the male to the female, or their transport within the female. It is notable that apyrene sperm are already dispersed from their bundles in the male tract, while eupyrene bundles are maintained intact until they are transferred to the female. Further, there is good evidence that apyrene sperm are activated to be motile more quickly than eupyrene sperm, and that apyrene sperm leave the bursa and reach the spermatheca sooner than eupyrene sperm (see Section 7.1). These observations are all consistent with the idea that the function of apyrene sperm is to facilitate the movement of eupyrene sperm, perhaps by providing motive power. Unfortunately for this idea, however, there are no reports showing that motility of lepidopteran apyrene sperm can cause the directed movement of the corresponding eupyrene sperm, whether the latter are still in bundles or after they have been dispersed. Cook and Wedell (1999) dismissed this idea on the grounds that if apyrene sperm provided motive power for eupyrene sperm then their relative proportions should vary little between species, whereas in fact they vary greatly. Nevertheless, co-operation among fertile sperm to enhance motility is known in other insects (Hayashi, 1998) and some types of non-fertilising parasperm of prosobranch snails are apparently highly modified for a transport function (Buckland-Nicks, 1998), so that it would be premature to rule out this possibility in Lepidoptera. A sophisticated variant of this hypothesis is that the apyrene sperm could induce reflex peristaltic movements of the muscular walls of either the male or the female tracts (or both), thereby faciliting eupyrene sperm movement. There is no evidence for or against this idea, as far as we are aware. Such stimulation of the musculature could be achieved by apyrene


285

sperm either through their own movements, or perhaps through their release of some contraction-initiating chemical messenger. (b) Apyrene sperm may indirectly assist in the transfer or transport of eupyrene sperm by activating or helping to activate motility in eupyrene sperm. Both apyrene and eupyrene sperm are known to be activated by factors in the male or female reproductive tract. In B. mori, in which the matter has been most thoroughly studied, activation is caused by a two-stage process involving both proteolytic action and cyclic nucleotide signalling (Osanai and Baccetti, 1993). Since individual eupyrene sperm are bound together within their bundles, however, proteolytic digestion of the restraining bundle fibres is a prerequisite for the acquisition of full motility. Moreover, the fact that the spermatophore is packed with glycogen granules and other materials, means that escape from the spermatophore is also dependent on the dissolution of the spermatophore matrix. It has therefore been suggested by Osanai et al. (1987b) that one function of apyrene sperm may be to facilitate the dissolution of eupyrene bundles and the escape of eupyrene sperm from the spermatophore matrix by agitating the contents of the spermatophore. In effect, the apyrene sperm act as miniature stir-bars, facilitating the actions of initiatorin and other reproductive tract proteases (see Section 8.1) in removing proteinaceous barriers to eupyrene sperm motility. This would explain why apyrene sperm escape from their bundles and also acquire motility before eupyrene sperm. On the other hand, it is hard to see why quite so many apyrene sperm are required. Even if it is true that they have this role, if this were their only function then it is not obvious how apyrene sperm could have evolved. (c) Apyrene sperm may directly help eupyrene sperm to fertilise eggs. There is no evidence for this. Apyrene sperm do not leave the spermatheca to enter the genital duct (Friedla¨nder and Gitay, 1972), so that direct assistance in fertilisation of the egg is improbable. (d) Apyrene sperm may indirectly help eupyrene sperm to fertilise eggs by promoting the capacitation of eupyrene sperm. Eupyrene sperm undergo important changes in their surface structure while they are in the spermatheca (see Section 8.1). These changes are probably important or even essential in the capacitation of these cells (i.e. in their acquisition of fertilising ability). Since these changes occur within the spermatheca, it is probable that the conditions within that organ are necessary for surface modification. Since similar changes in surface appearance are caused by incubation with trypsin (Friedla¨nder et al., 2001), it is probable that an endogenous proteolytic enzyme is responsible in vivo. It cannot at present be ruled out that apyrene sperm could secrete factors that are either directly proteolytic, or which facilitate the actions of proteinases from other sources. Although lepidopteran apyrene sperm do not have the appearance of being actively secretory, prosobranch mollusc parasperm

286


have been reported to release a variety of secretion products into the seminal fluid (Buckland-Nicks, 1998), so that this must be considered at least possible in Lepidoptera. (e) Apyrene sperm may indirectly help related eupyrene sperm to fertilise the eggs of a multiply-mated female through sperm competition. Many lepidopterans have polyandrous mating systems, and post-copulatory sexual selection must therefore occur (Birkhead and Pizzari, 2002), either through classical sperm competition (Parker, 1970) or cryptic female choice (Eberhard, 1996). One simple model of competition between the sperm of more than one male is through a ‘‘fair raffle’’ among sperm that are randomly mixed in the female storage organ (Parker and Simmons, 1991). Although under such circumstances there is selective advantage to the male in the transfer of larger numbers of sperm to the female, there would be no advantage in augmenting sperm numbers with inexpensive apyrene sperm, since only eupyrene sperm can compete for fertilisation. There is however evidence that this lottery mechanism of sperm competition frequently does not apply. Instead, the second mating results in the effective disablement of fertile sperm from a previous mating. This can occur in two ways. First, the second mating may kill or at least physiologically incapacitate (reduce the ability to fertilise) previous sperm. Second (see (f) below), the female may eject previous sperm, so that raffle is not ‘‘fair’’. Incapacitation has not been shown in Lepidoptera, but it remains possible that it could occur. In this case, apyrene sperm could be the agents of incapacitation of unrelated eupyrene sperm. This is effectively the ‘‘kamikaze’’ or ‘‘soldier sperm’’ hypothesis (Baker and Bellis, 1988). There is disagreement as to whether soldier sperm really exist, but the ‘‘lancet’’ type of non-fertilising parasperm in prosobranch molluscs appears to have many of the characteristics expected of soldier sperm (Buckland-Nicks, 1998), and it has been shown theoretically that conditions for the evolution of soldier sperm are not stringent (Kura and Nakashima, 2000). Direct prevention of fertilisation by the eupyrene sperm of another male would seem to require apyrene sperm to be able to distinguish such sperm from their own eupyrene kin. A clue to how this might happen comes from work by Civetta (1999) and Price et al. (1999) on D. melanogaster. By using sperm with green fluorescent protein (GFP)–labelled tails, these authors showed that incapacitation contributes to second mating sperm precedence in D. melanogaster, but was only important when a sufficient interval had elapsed between the first and second mating. Because this incapacitation occurs even in the absence of transfer of sperm at the second mating, the adverse effect on previous sperm must be due to some component of seminal fluid other than sperm. Evidently, ‘‘new’’ sperm are able to resist the incapacitating effect of seminal fluid, while ‘‘old’’ sperm


287

are not. Despite the fact that this cannot be a soldier sperm effect, the difference in susceptibility of ‘‘old’’ and ‘‘new’’ sperm suggests one mechanism in which putative soldier sperm might be able to distinguish between related and unrelated fertilising sperm. At present we may say, however, that there is no evidence that lepidopteran apyrene sperm are killer or soldier sperm. Moreover, since in many Lepidoptera, apyrene sperm quickly disappear from the spermatheca (they are either destroyed or are ejected – see Section 7.3), they do not in practice have the opportunity to interfere with unrelated eupyrene sperm from subsequent matings. Of course, this latter point might well be viewed as evidence that selection favours the active removal by the female of apyrene sperm from previous matings, for precisely this reason. (f) Apyrene sperm may indirectly help eupyrene sperm to fertilise the eggs of a multiply mated female by mechanically hindering the ability of unrelated eupyrene sperm to reach the egg. A variant of the sperm competition hypothesis that seems worthwhile to distinguish is the idea that apyrene sperm are simply a mechanical hindrance to unrelated eupyrene sperm. Movement of an individual through a crowd is slower than free movement in an unoccupied space. Apyrene sperm greatly outnumber eupyrene sperm in all Lepidoptera, so that it might be supposed that apyrene sperm could significantly hinder the ability of eupyrene sperm from another mating to reach the egg, simply by their abundance. Such hindrance would of course be a double-edged sword, since an excess of apyrene sperm might also hinder the movements of related eupyrene sperm from the same mating. However, this might not matter if the prime object of the male was to prevent the success of another male. Sperm sorting by the female (see Section 7.2) might be a device to prevent this kind of hindrance. As for the previous hypothesis (e), the removal of apyrene sperm from the spermatheca after mating in some species might be considered as evidence either in favour or against this idea. There is currently no compelling evidence in favour of this hypothesis. (g) Apyrene sperm may indirectly help related eupyrene sperm to fertilise the eggs of a multiply mated female by masquerading as eupyrene sperm. There is evidence that in some Lepidoptera, female fitness is dependent on receiving a sufficient quantity of eupyrene sperm (e.g. A. assectella, Thibout, 1979; B. mori, Karube and Kobayashi, 1999; S. litura, Seth et al., 2002a). Therefore, females should seek to maximise the amount of sperm that they receive by soliciting multiple matings. But since mating itself is likely to impose fitness costs (e.g. Kawagoe et al., 2001), a trade-off will exist (Bell and Koufopanou, 1986), and females should cease to forage for matings when an optimum number of fertile sperm has been acquired (e.g. Arnqvist and Nilsson, 2000). It has indeed been found that the sexual receptivity of some female moths is inhibited by the presence of eupyrene sperm in the spermatheca (see Section 6.4).

288


On the other hand, since it is in the interest of a male to prevent female remating, but because the manufacture of eupyrene sperm is costly, males may seek to deceive females about the number of eupyrene sperm that they transfer. In this light, apyrene sperm may be seen as inexpensive mimics of eupyrene sperm. He et al. (1995) noted that the numbers of apyrene spermatozoa in the spermatheca of P. separata decline after mating in a fashion that is the inverse of the female’s tendency to remate. They suggested that apyrene sperm might act to inhibit receptivity. Cook and Wedell (1999) tested this hypothesis in P. napi. They found that those females that remated had significantly more apyrene sperm in their spermathecae than those females that did not remate. There was however no difference in the number of eupyrene sperm in the spermatheca between remating and non-remating females. Importantly for the test of the hypothesis, there was no effect of the previous mating history (virgin or not) of the male, either on the number of apyrene sperm stored, or on the female’s tendency to remate, despite the fact that mating history strongly influenced both spermatophore size and its content of eupyrene sperm (counterintuitively, previously mated males transferred more eupyrene sperm in a smaller spermatophore). These findings suggested to the authors that the evolutionary ‘‘purpose’’ of apyrene sperm, at least in P. napi, is to exploit a female system that monitors sperm number to ensure maximum fertility. In this scenario, apyrene sperm, being less costly to produce than eupyrene sperm are a ‘‘cheap filler’’, acting as a dishonest signal of male fertility. Swallow and Wilkinson (2002) have expressed some reservations about Cook and Wedell’s (1999) paper, pointing out that variation in the percentage of apyrene sperm retained was not independent of spermatophore size. Another piece of evidence supporting the ‘‘cheap filler’’ hypothesis is that some P. napi females have a genetic tendency actively to eject apyrene but not eupyrene sperm, and that these females also remate more readily (Wedell, 2001). This was suggested to indicate the probable existence in this species of a sexual evolutionary arms race. In this race, females have selected those males that provide most sperm per mating, while males have increased their fitness by reducing the cost of providing more sperm. Presumably, an optimum ratio of high-cost fertile eupyrene sperm and low-cost infertile apyrene sperm would arise as a result of such sexual conflict. So long as the female cannot distinguish between fertile and infertile sperm, this optimum ratio would be likely to favour a high proportion of infertile sperm. The proportion of apyrene sperm transferred to the female is indeed very high (88%) in P. napi (Swallow and Wilkinson, 2002). In some other lepidopteran species, however (e.g. A. assectella, Thibout, 1979; B. mori, Karube and Kobayashi, 1999; S. litura, Seth et al., 2002a), female moths have evidently evolved the ability to distinguish between eupyrene and apyrene sperm. In these cases it would be expected


289

that the proportion of apyrene sperm would be lower, and this is indeed the case (e.g. S. litura, 47%; Swallow and Wilkinson, 2002). If the prime evolutionary ‘‘purpose’’ of apyrene sperm is to compete at minimum cost with the eupyrene sperm of other males, then it would be expected that males would transfer more of them when the risk of such competition was greatest. This would be true when comparing between species, or between different mating situations within a species. Accordingly, the ratio of fertile to non-fertilising morphs was indeed adjusted in favour of apyrene sperm when P. separata were reared under crowded conditions (He and Miyata, 1997), although no change in the eupyrene:apyrene ratio was found when P. interpunctella were reared in crowded conditions on a low-protein diet (Gage and Cook, 1994), or when this insect was allowed to mate in the presence of rival males (Cook and Gage, 1995). On the other hand, P. interpunctella males transferred a higher proportion of apyrene sperm to young than to old females (Cook and Gage, 1995), perhaps because the risk of the young females remating is greater. Thus, it appears that in at least some cases, the proportion of apyrene sperm transferred can be adjusted according to situation, but in other cases this does not occur. We note by way of comparison that some molluscs have been noted to increase the proportion of parasperm when the risk of competition is great (Oppliger et al., 1998), but that flies of the Drosophila obscura group do not alter the proportion of short, less preferred sperm in response to female mating status (Snook, 1998). In evaluating these results, it ought to be borne in mind that the proportions of sperm developing according to the eupyrene and apyrene pathways in Lepidoptera are determined by the time of appearance of ASIF (see Section 5.2), an event that appears to be obligately linked to the timing of metamorphosis, and occurs long before mating. Moreover, crowded rearing conditions cannot be assumed to be a sure indicator of mate rivalry in adulthood. In this case, therefore, there may be developmental constraints on what can be achieved in the way of adjustment of sperm development in response to mating conditions. (h) Apyrene sperm may indirectly help related eupyrene sperm to fertilise the eggs of a multiply mated female by subverting cryptic female choice of sperm. Females may exercise post-copulatory sexual selection by choosing to utilise sperm from high-quality partners and rejecting others. In this case, multiple mating may be seen as conferring fitness because it allows this choice to be exercised, while perhaps avoiding the potential cost of resisting unwanted mating attempts by males (Birkhead and Pizzari, 2002). There is no doubt that such cryptic post-copulatory female choice occurs in polyandrous Lepidoptera. Sperm preference is exerted by many species in favour of males that supply large spermatophores. An excellent example of this (LaMunyon and Eisner, 1993, 1994) is the arctiid moth

290


Utetheisa ornatrix, in which almost complete sperm precedence is obtained by the male that supplies the larger spermatophore, regardless of parental age, between mating interval, duration of copulation or alkaloid content (in this species, the male supplies the female with sequestered defensive pyrrolizidine alkaloids). Since the order of mating does not matter, in this species the female must retain sperm from both matings. However in some moths, active displacement of previous sperm from the spermatheca is known to occur. This has been shown in S. litura (Etman and Hooper, 1979) and B. mori (Suzuki et al., 1996). Although the ejection of previous sperm is evidently effected by the female, since stored sperm are ejected from the spermatheca before the arrival of new sperm, the decision to eject is likely to be made on the basis of an assessment of the quality of the second male. Although it does not inevitably follow that apyrene sperm are the agents of this decision, it is possible. (i) Apyrene sperm may represent a nutritional gift from the male to the female that increases the fitness of the female receiving them, thereby also enhancing the fitness of the male. There is a limited amount of evidence in favour of this hypothesis in the butterfly P. napi. In this polyandrous species, males transfer an appreciable quantity of nutritional resources to the female with the spermatophore (see Section 6.3). These nutrients significantly augment female fertility (Karlsson, 1998), and females forage for matings to increase the acquisition of resources (Kaitala and Wiklund, 1994). However, because in P. napi there is no correlation between spermatophore mass and sperm number transferred (Cook and Wedell, 1999), there is no direct evidence from these data that the apyrene sperm are themselves a nutritional gift. It has been shown, however, that receipt by female P. napi of a larger quantity of apyrene sperm leads to greater inhibition of remating (Cook and Wedell, 1999), and also that apyrene sperm disappear from the spermatheca (Wedell, 2001), two observations that are at least consistent with the idea that in this species apyrene sperm may function at least in part as a nutritional gift. Nevertheless, Cook and Wedell (1999) dismissed the idea that apyrene sperm might be a nutritional gift on the grounds that in those Lepidoptera that are known to provide resources to enhance female reproductive success, these resources are found in a part of the spermatophore that is separate from that in which sperm are contained. This does not seem to us to be a good argument. Nevertheless, there is not much enthusiasm for the idea that apyrene sperm are a nutritional gift. The idea would be easier to contemplate if the apyrene sperm had the appearance of specialisation for nutritional transfer, for example, the presence of stored protein or carbohydrate, as is the case for the parasperm of stinkbugs (Schrader, 1960) and cockroaches (Richards, 1963), as well as some kinds of parasperm in prosobranch molluscs (Buckland-Nicks, 1998; Buckland-Nicks et al., 1999). The apyrene sperm of those lepidopterans that have been studied in detail


291

do not have such an appearance. However, it would be interesting to examine the apyrene sperm of P. napi. Note, however, that a different interpretation of the evolutionary function of P. napi apyrene sperm is possible (see (g) above). (j) Cost of apyrene sperm. Most of the above hypotheses make the initial assumption that apyrene sperm are less costly than eupyrene sperm. This assumption follows from the universal observation that apyrene sperm are smaller than eupyrene sperm, from the findings that the nuclear materials lost from apyrene sperm are expelled by peristaltic squeezing from the developing apyrene bundles (Kawamura et al., 2000) and are therefore potentially recoverable by the male, and that apyrene sperm are transferred to the female accompanied by less associated materials (i.e. they lack bundle associated extracellular materials – see Section 6.1). Swallow and Wilkinson (2002) attempted to test the idea that apyrene sperm are less costly by regressing the (log) numbers of apyrene sperm against the (log) numbers of eupyrene sperm for a sample of 22 different species of Lepidoptera. They argued that if apyrene sperm were less expensive to produce than eupyrene sperm, then the best fit line should be significantly greater in slope than 1, ‘‘since for every incremental increase in [eupyrene] sperm a proportionally larger increase in [apyrene] sperm should be realized’’. Unfortunately, the slope of the line ð1:16 0:52Þ did not differ significantly from 1, and the authors concluded that factors other than energetic ones were likely to be involved in determining ejaculate composition. This conclusion seems safe, given that we know that other factors do indeed operate, but fails to answer the original question. One can only conclude that correlation tests of this type lack the statistical power to reject the null hypothesis of no association.

10

Conclusion

In this review, we have aimed to show that dichotomous spermatogenesis in Lepidoptera is a physiologically and developmentally interesting topic. Research into how apyrene sperm development occurs and is regulated will pay important dividends in understanding of normal spermatogenesis. The existence of apyrene spermatocytes that continue developing in a predictable way to produce spermatozoa after their nuclei become apoptotic presents a very convenient system to study differentiation of eukaryotic cell in absence of newly synthesised informative RNA. Furthermore, the dichotomous eupyrene– apyrene dichotomy shows cells having the same genetical background but differentiating separatedly through a series of well characterised morphological markers that are easy to correlate with stages on molecular experimentation. Genetic techniques are poorly developed in the Lepidoptera and are likely to remain so. But molecular genetic techniques are now very powerful, and the

292


recent availability of a draft B. mori genome sequence (Xia et al., 2004), and the application of RNAi techniques (Novina and Sharp, 2004) to Lepidoptera (e.g. Kramer and Bentley, 2003), promise to make up for a great deal in genetically intractable organisms. In this review, we have given some indications of where it might be profitable to look for genes controlling apyrene spermatogenesis and look forward to advances in this research area. Lepidopteran insects are important subjects of research in many areas of reproductive ecology, especially those of sperm allocation, sperm competition and post-copulatory sexual selection. The existence of two types of sperm in this group of insects is of fundamental importance to their reproductive biology, and there is no doubt that the evolutionary biology of sperm dichotomy will continue to be of great interest. Swallow and Wilkinson (2002) have made the point that ecological and comparative studies have yielded only relatively disappointing insights into the adaptive value of sperm dichotomy, and recommend that experimental approaches to manipulate eupyrene: apyrene ratios should be used. To do this, ecologists and physiologists will have to get together. Sperm dichotomy is important in applied fields too. Attempts to control lepidopteran pests by interfering with their fertility (e.g. using g-radiation to induce F1 sterility, Carpenter and Gross, 1993; Seth and Reynolds, 1993) require a proper understanding of the impact of the sterilising agent on dichotomous spermatogenesis in the target insect. Since sublethal doses of many insecticides, but especially insect growth regulators, interfere with male fertility, it is important to determine these effects in full knowledge that the two types of sperm may be differentially effected (e.g. Seth et al., 2004). Another area of interest that we have had neither the time nor space to review is the potential for interaction between lepidopteran dichotomous spermatogenesis and the existence of sex-ratio distorting selfish genetic elements like the intracellular bacterial parasite, Wolbachia. It is known that Wolbachia is present in the spermatogenetic cysts of Drosophila spp. (Clark et al., 2002) and this is associated with cytoplasmic incompatibility. Intensive sampling has revealed that an unexpectedly high proportion of lepidopteran species is infected with Wolbachia (Jiggins et al., 2001). Such associations hold rich possibilities for interactions with dichotomous spermatogenesis, and indeed the very well-studied case of the F1-hybrid sterility observed between H. virescens and H. subflexa, which is due to aberrant spermatogenesis, has since been discovered to be associated with the presence of a Wolbachia-like bacterium in the testes of these two species (Degrugillier, 1994). We predict that this will prove to be a fertile area of research.

Acknowledgements We thank Mr Neil Maskell (Department of Zoology, University of Cambridge) and Mr William Reynolds (Oriel College, Oxford) for help in acquiring and


293

processing images of Figs. 1 and 2, and Drs Tim Karr and Rob Butcher (Department of Biology & Biochemistry, University of Bath) for helpful discussions about spermatozoa. References Abe´, S.-I., Asakura, S. and Ukeshima, A. (1988). Formation of flagella during interphase in secondary spermatocytes from Xenopus laevis in vitro. J. Exp. Zool. 246, 65–70. Aigaki, T., Kasuga, H., Nagaoka, S. and Osanai, M. (1994). Purification and partial amino acid sequence of initiatorin, a prostatic endopeptidase of the silkworm, Bombyx mori. Insect Biochem. Mol. Biol. 24, 969–975. Aigaki, T., Kasuga, H. and Osanai, M. (1987). A specific endopeptidase, BAEE esterase, in the glandula prostatica of the male silkworm, Bombyx mori. Insect Biochem. 17, 323–328. Aigaki, T. and Osanai, M. (1985). Arginase activity in the silkworm, Bombyx mori – developmental profiles and physiological role. J. Comp. Physiol. B Biochem. System. Env. Physiol. 155, 653–657. Aigaki, T., Osanai, M. and Kasuga, H. (1988). Arginine carboxypeptidase activity in the male reproductive glands of the silkworm, Bombyx mori. Insect Biochem. 18, 295–298. Akawa, T. D., Nakagawa, H., Sugihara, K., Narisawa, S., Ohyama, C., Nishimura, S.-T., O’Brien, D. A., Moreman, K. W., Millań, J. L. and Fukuda, M. N. (2002). Germ cell survival through carbohydrate-mediated interaction with Sertoli cells. Science 295, 124–127. Alsheimer, M., von Glasenapp, E., Hock, R. and Benavente, R. (1999). Architecture of the nuclear periphery of rat pachytene spermatocytes: distribution of nuclear envelope proteins in relation to synapatonemal complex attachment sites. Mol. Biol. Cell 10, 1235–1245. Andre´, J. (1959). E´tude au microscope e´lectronique de l’e´volution de chondriome pendant la spermatogeńe`se du Papillon du chou, Pieris brassicae. Ann. Sci. Nat. Zool. 1, 283–305. Andre´, J. (1962). Contribution a` la connaissance du chondriome, e´tude de ses modifications ultrastructurales pendant la spermatogeńe`se. J. Ultrastruct. Res. 3 (Suppl), 1–185. Arnqvist, G. and Nilsson, T. (2000). The evolution of polyandry: multiple mating and female fitness in insects. Animal Behav. 60, 145–164. Au, D. W. T., Reunov, A. A. and Wu, R. R. S. (1998). Four lines of spermatid development and dimorphic spermatozoa in the sea urchin Anthocidaris crassipina (Echinodermata, Echinoidea). Zoomorphology 118, 159–168. Baccetti, B. (1979). The evolution of the acrosomal complex. In: The Spermatozoon, (eds Fawcett, D. W. and Bedford, J. M.), pp. 305–329. Baltimore-Munich: Urban and Schwarzenberg, Inc. Baccetti, B. (1991). Comparative Spermatology: 20 Years After. New York: Raven Press. Baker, R. R. and Bellis, M. A. (1988). ‘‘Kamikaze’’ sperm in mammals?. Anim. Behav. 36, 936–939. Bass, H. W. (2003). Telomere dynamics unique to meiotic prophase: formation and significance of the bouquet. Cell Mol. Life Sci. 60, 2319–2324. Bebas, P. and Cymborowski, B. (1999). Effect of constant light on male sterility in the cotton leafworm Spodoptera littoralis. Physiol. Entomol. 24, 165–170.

294


Bebas, P., Cymborowski, B. and Giebultowicz, J. M. (2001). Circadian rhythm of sperm release in males of the cotton leafworm Spodoptera littoralis: in vivo and in vitro studies. J. Insect Physiol. 47, 859–866. Bebas, P., Cymborowski, B. and Giebultowicz, J. M. (2002). Circadian rhythm of acidification in insect vas deferens regulated by rhythmic expression of vacuolar H+-ATPase. J. Exp. Biol. 205, 37–44. Bell, G. and Koufopanou, V. (1986). The cost of reproduction. Oxford Surveys in Evol. Biol. 3, 83–131. Birkhead, T. R. and Pizzari, T. (2002). Postcopulatory sexual selection. Nat. Rev. Genet. 3, 262–273. Bissoondath, C. J. and Wiklund, C. (1996). Male butterfly investment in successive ejaculates in relation to mating system. Behav. Ecol. Sociobiol. 39, 285–292. Boggs, C. L. and Gilbert, L. E. (1979). Male contribution to egg production in butterflies: evidence for transfer of nutrients at mating. Science 206, 83–84. Boi, S., Fascio, U. and Ferragutti, M. (2001). Nuclear fragmentation characterises paraspermatogenesis in Tubifex tubifex (Annelida, Oligochaeta). Mol. Reprod. Dev. 59, 442–450. Bollenbacher, W. E., Smith, S. L., Goodman, W. and Gilbert, L. I. (1981). Ecdysteroid titer during larval–pupal–adult development of the tobacco hornworm, Manduca sexta. Gen. Comp. Endocrinol. 44, 302–306. Bressac, C. and Hauschteck-Jungen, E. (1996). Drosophila subobscura females preferentially select long sperm for storage and use. J. Insect Physiol. 42, 323–328. Brick, D., Lifschitz, E. and Friedla¨nder, M. (1979). Mitochondria differentiation during spermatogenesis in Drosophila melanogaster: an ultrastructural analysis in the male sterile mutant ms(1)413. J. Ultrastruct. Res. 66, 151–163. Brown, J. J. and Friedla¨nder, M. (1995). Influence of parasitism on spermatogenesis in the codling moth, Cydia pomonella L. J. Insect Physiol. 41, 957–963. Buckland-Nicks, J. (1998). Prosobranch parasperm: sterile germ cells that promote paternity?. Micron 29, 267–280. Buckland-Nicks, J., Bryson, I., Hart, L. and Partridge, V. (1999). Sex and a snail’s sperm: on the transport, storage and fate of dimorphic sperm in Littorinidae. Inver. Reprod. Dev. 36, 145–152. Carlton, P. M. and Cande, W. Z. (2002). Telomeres act autonomously in maize to organize the meiotic bouquet from a semipolarized chromosome orientation. J. Cell Biol. 157, 231–242. Carpenter, J. E. and Gross, H. R. (1993). Suppression of feral Helicoverpa zea (Lepidoptera: Noctuidae) populations following the infusion of inherited sterility from released substerile males. Environ. Entomol. 22, 1084–1091. Carson, H. L. (1945). A comparative study of the apical cell of the insect testis. J. Morphol. 77, 141–155. Chang, C. S., Chen, Y. T., Yeh, S. D., Xu, Q. Q., Wang, R. S., Guillou, F., Lardy, H. and Yeh, S. Y. (2004). Infertility with defective spermatogenesis and hypotestosteronemia in male mice lacking the androgen receptor in Sertoli cells. Proc. Natl. Acad. Sci. USA 101 (18), 6876–6881. Chippendale, G. M. and Alexander, B. R. (1973). Spermatogenesis of the southwestern corn borer Diatraea grandiosella. 2. Resumption in diapause larvae. Ann. Ent. Soc. Am. 66, 747–752. Civetta, A. (1999). Direct visualization of sperm competition and sperm storage in Drosophila. Curr. Biol. 9, 841–844. Clark, M. E., Veneti, Z., Bourtzis, K. and Karr, T. L. (2002). The distribution and proliferation of the intracellular bacteria Wolbachia during spermatogenesis in Drosophila. Mech. Dev. 111, 3–15.


295

Cook, P. A. and Gage, M. J. G. (1995). Effects of risks of sperm competition on the numbers of eupyrene and apyrene sperm ejaculated by the moth Plodia interpunctella (Lepidoptera, Pyralidae). Behav. Ecol. Sociobiol. 36, 261–268. Cook, P. A. and Wedell, N. (1996). Ejaculate dynamics in butterflies: a strategy for maximizing fertilization success?. Proc. Roy. Soc. Lond. B 263, 1047–1051. Cook, P. A. and Wedell, N. (1999). Non-fertile sperm delay female remating. Nature 397, 486. Daub, A.-M. and Hauser, M. (1988). Primary cilia in locust spermatocytes: formation, fate, and possible function. Zeitschr. Naturforsch. 43c, 455–462. Degrugillier, M. E. (1994). Testicular microorganisms in Heliothis: ultrastructure and distribution in H. viresens, H. subflexa and backcross males. J. Invert. Pathol. 64, 77–88. Delisle, J. and Bouchard, A. (1995). Male larval nutrition in Choristoneura rosaceana (Lepidoptera, Tortricidae): an important factor in reproductive success. Oecologia 104, 508–517. Delisle, J. and Hardy, M. (1997). Male larval nutrition influences the reproductive success of both sexes of the spruce budworm, Choristoneura fumiferana (Lepidoptera: Tortricidae). Funct. Ecol. 11, 451–463. Denlinger, D. L. (1985). Hormonal control of diapause. In: Comprehensive Insect Physiology, Biochemistry and Pharmacology, Vol. 8 (eds Kerkut, G. A. and Gilbert, L. I.), pp. 353–412. Oxford: Pergamon Press. Dubrovsky, E. B., Dubrovskaya, V. A. and Berger, E. M. (2004). Hormonal regulation and functional role of Drosophila E75A orphan nuclear receptor in the juvenile hormone signaling pathway. Dev. Biol. 268, 258–270. Dumser, J. B. (1980). The regulation of spermatogenesis in insects. Ann. Rev. Entomol. 25, 341–369. Dumser, J. B. and Davey, K. G. (1975). The Rhodnius testis: hormonal effects on germ cell division. Can. J. Zool. 53, 1682–1689. Eberhard, W. G. (1996). Female control: sexual selection by cryptic female choice. Princeton, NJ: Princeton University Press. Engebretson, J. A. and Mason, W. H. (1980). Transfer of 65 Zn at mating in Heliothis virescens. Environ. Entomol. 9, 119–121. Etman, A. A. M. and Hooper, G. H. S. (1979). Sperm precedence of the last mating of Spodoptera litura. Ann. Entomol. Soc. Am. 72, 119–120. Fain-Maurel, M.-A. (1966). Aquisitions rećentes sur les spermatoge`neses atypiques. Anneè Biol. 5, 513–564. Fox, P. D., Hewitt, G. M. and Hall, D. J. (1974). DNA replication and RNA transcription of euchromatic and heterochromatic chromosome regions during grasshopper meiosis. Chromosoma 45, 43–62. Friedla¨nder, M. (1976). The role of the transient perinuclear microtubules during spermiogenesis of the warehouse moth Ephestia cautella. J. Submicrosc. Cytol. 8, 319–326. Friedla¨nder, M. (1982). Juvenile hormone regulation of dichotomous spermatogenesis during the larval diapause of the codling moth. J. Insect Physiol. 28, 1009–1012. Friedla¨nder, M. (1983). Phylogenetic branching of Trichoptera and Lepidoptera: an ultrastructural analysis on comparative spermatology. J. Ultrastruct. Res. 83, 141–147. Friedla¨nder, M. (1989). 20-Hydroxyecdysone induces glycogen accumulation within the testicular sheath during in vitro spermatogenesis renewal in diapausing codling moth (Cydia pomonella). J. Insect Physiol. 35, 29–39. Friedla¨nder, M. (1991). Spermatogenesis. In: Tortricid Pests, their Biology, Natural Enemies and Control, (eds van der Geest, L. P. S. and Evenhuis, H. H.), pp. 149–155. Amsterdam: Elsevier Science Publishers B.V.

296


Friedla¨nder, M. (1993). Phylogenetic position of rhyacophiloid caddisflies (Insecta, Trichoptera): a spermatological analysis of Rhyacophilidae and Glossosomatidae. Zool. Scripta 22, 299–304. Friedla¨nder, M. and Benz, G. (1981). The eupyrene–apyrene dichotomous spermatogenesis of Lepidoptera: organ culture study of the timing of apyrene commitment in the codling moth. Int. J. Invert. Reprod. 3, 113–120. Friedla¨nder, M. and Benz, G. (1982). Control of spermatogenesis resumption in the post-diapausing larvae of the codling moth. J. Insect Physiol. 28, 349–355. Friedla¨nder, M. and Brown, J. J. (1995). Tebufenozide (Mimic), a nonecdysteroidal ecdysone agonist, induces spermatogenesis reinitiation in isolated abdomens of diapausing codling moth larvae (Cydia pomonella). J. Insect Physiol. 41, 403–411. Friedla¨nder, M. and Gershon, J. (1978). Reaction of surface lamella of moth spermatozoa to vinblastine. J. Cell Sci. 30, 353–361. Friedla¨nder, M. and Gitay, H. (1972). The fate of the normal-anucleated spermatozoa in inseminated females of the silkworm Bombyx mori. J. Morphol. 138, 121–130. Friedla¨nder, M. and Hauschteck-Jungen, E. (1982a). The regular divisions of the spermatocytes as related to a meiotic lysine-rich protein fraction. Chromosoma 82, 109–118. Friedla¨nder, M. and Hauschteck-Jungen, E. (1982b). Differential basic nucleoprotein kinetics in the two kinds of Lepidoptera spermatids: nucleate (eupyrene) and anucleate (apyrene). Chromosoma 82, 387–398. Friedla¨nder, M. and Hauschteck-Jungen, E. (1986). Regular and irregular divisions of Lepidoptera spermatocytes as related to the speed of meiotic prophase. Chromosoma 93, 227–230. Friedla¨nder, M., Jans, P. and Benz, G. (1981). Precocious reprogramming of eupyrene–apyrene spermatogenesis commitment induced by allatectomy of the penultimate larval instar of the moth Actias selene. J. Insect Physiol. 27, 267–269. Friedla¨nder, M., Jeshtadi, A. and Reynolds, S. E. (2001). The structural mechanism of trypsin-induced motility in Manduca sexta spermatozoa in vitro. J. Insect Physiol. 47, 245–255. Friedla¨nder, M. and Miesel, S. (1977). Spermatid anucleation during normal-atypical spermatogenesis of the warehouse moth Ephestia cautella. J. Submicrosc. Cytol. 9, 173–185. Friedla¨nder, M. and Reynolds, S. E. (1988). Meiotic metaphases are induced by 20-hydroxyecdysone during spermatogenesis of the tobacco hornworm, Manduca sexta. J. Insect Physiol. 34, 1013–1019. Friedla¨nder, M. and Reynolds, S. E. (1992). Intratesticular ecdysteroid titres and the arrest of sperm production during pupal diapause in the tobacco hornworm, Manduca sexta. J. Insect Physiol. 38, 693–703. Friedla¨nder, M. and Wahrman, J. (1965). The independence of spermatid differentiation from the meiotic divisions. Exp. Cell Res. 38, 680–684. Friedla¨nder, M. and Wahrman, J. (1970). The spindle as a basal body distributor. A study in the meiosis of the male silkworm moth, Bombyx mori. J. Cell Sci. 7, 65–89. Friedla¨nder, M. and Wahrman, J. (1971). The number of centrioles in insect sperm: a study in two kinds of differentiating silkworm spermatids. J. Morphol. 134, 383–398. Fuller, M. T. (1993). Spermatogenesis. In: The Development of Drosophila melanogaster, Vol. 1 (eds Bate, M. and Martinez Arias, A.), pp. 71–147. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Gage, M. J. G. and Cook, P. A. (1994). Sperm size or numbers? Effects of nutritional stress upon eupyrene and apyrene sperm production strategies in the moth Plodia interpunctella (Lepidoptera, Pyralidae). Funct. Ecol. 8, 594–599.


297

Garvey, L. K., Gutierrez, M. and Krider, H. M. (2000). Ultrastructure and morphogenesis of the apyrene and eupyrene spermatozoa in the gypsy moth (Lepidoptera: Lymantriidae). Ann. Entomol. Soc. Am. 93, 1147–1155. Gelbic, I. and Metwally, M. M. (1981). Changes in the development of male germinal cells in Spodoptera littoralis caused by the effects of juvenoids (Lepidoptera, Noctuidae). Acta Entomol. Bohemos. 78, 10–17. Giebultowicz, J. M., Bell, R. A. and Imberski, R. B. (1988). Circadian rhythm of sperm movement in the male reproductive tract of the gypsy moth Lymantria dispar. J. Insect Physiol. 34, 527–532. Giebultowicz, J. M., Blackburn, M. B., Thomas-Laemont, P. A., Weyda, F. and Raina, A. K. (1996). Daily rhythm in myogenic contractions of vas deferens associated with sperm release cycle in a moth. J. Comp. Physiol. A 178, 629–632. Giebultowicz, J. M. and Brooks, N. L. (1998). The circadian rhythm of sperm release in the codling moth Cydia pomonella. Entomol. Exp. Appl. 88, 229–234. Giebultowicz, J. M., Joy, J. E., Riemann, J. G. and Raina, A. K. (1994). Changes in protein-patterns in sperm and vas-deferens during the daily rhythm of sperm release in the gypsy-moth. Arch. Insect Biochem. Physiol. 27, 65–75. Giebultowicz, J. M., Loeb, M. J. and Borkovec, A. B. (1987). In vitro spermatogenesis in lepidopteran larvae: role of the testis sheath. Int. J. Invert. Reprod. Dev. 11, 221–226. Giebultowicz, J. M., Raina, A. K., Uebel, E. C. and Ridgway, R. L. (1991). Two-step regulation of sex-pheromone decline in mated gypsy-moth females. Arch. Insect Biochem. Physiol. 16, 95–105. Giebultowicz, J. M., Ridgway, R. L. and Imberski, R. B. (1990). Physiological basis for sterilizing effects of constant light in Lymantria dispar. Physiol. Entomol. 15, 149–156. Giebultowicz, J. M., Riemann, J. G., Raina, A. K. and Ridgeway, R. L. (1989). Circadian system controlling release of sperm in the insect testes. Science 245, 1098–1100. Giebultowicz, J. M., Weyda, F., Erbe, E. F. and Wergin, W. P. (1997). Circadian rhythm of sperm release in the gypsy moth, Lymantria dispar: ultrastructural study of transepithelial penetration of sperm bundles. J. Insect Physiol. 43, 1133–1147. Golubovskaya, I. N., Harper, L. C., Pawlowski, W. P., Schichnes, D. and Cande, W. Z. (2002). The pam1 gene is required for meiotic bouquet formation and efficient homologous synapsis in maize (Zea mays L.). Genetics 162, 1979–1993. Gould-Somero, M. and Holland, L. (1974). The timing of RNA synthesis for spermiogenesis in organ culture of Drosophila melanogaster testes. Wilhelm Roux Arch. 174, 133–148. Guraya, S. S. (2000). Cellular and molecular biology of capacitation and acrosome reaction in spermatozoa. Int. Rev. Cytol. 199, 1–64. Gvakharia, B. O., Bebas, P., Cymborowski, B. and Giebultowicz, J. M. (2003). Disruption of sperm release from insect testes by cytochalasin and b-actin mRNA mediated interference. Cell. Mol. Life Sci. 60, 1744–1751. Hamon, C. and Chauvin, G. (1992). Ultrastructural analysis of spermatozoa of Korscheltellus lupinus L. (Lepidoptera, Hepialidae) and Micropterix calthella L. (Lepidoptera, Micropterigidae). Int. J. Insect Morphol. Embryol. 21, 149–160. Hannah-Alava, A. (1965). The premeiotic stages of spermatogenesis. Adv. Genet. 13, 157–226. Hawley, R. S. (2002). Meiosis: how male flies do meiosis. Curr. Biol. 12, R660–R662. Hayakawa, Y., Komaru, A. and Munehara, H. (2002). Ultrastructural observations of eu- and paraspermiogenesis in the cottid fish Hemilepidotus gilberti (Teleostei: Scorpaeniformes: Cottidae). J. Morphol. 253, 243–254.

298


Hayashi, F. (1998). Sperm co-operation in the fishfly, Parachauliodes japonicus. Funct. Ecol. 12, 347–350. He, Y. B. and Miyata, T. (1997). Variations in sperm number in relation to larval crowding and spermatophore size in the armyworm, Pseudaletia separate. Ecol. Entomol. 22, 41–46. He, Y. B., Miyata, T. and Tanaka, T. (1995). Eupyrene and apyrene sperm and their numerical fluctuations inside the female reproductive tract of the armyworm, Pseudaletia separata. J. Insect Physiol. 41, 689–694. He, Y. B. and Tsubaki, Y. (1991). Effects of spermatophore size on female remating in the armyworm, Pseudaletia separata, with reference to larval crowding. J. Ethol. 9, 47–50. Henneguy, L.-F. (1897). Sur les rapports cils vibratiles avec les centrosomes. Arch. d’Anat. Microsc. 1, 481–496. Hennig, W. (1981). Insect Phylogeny. New York: Wiley. Henning, W. and Kremer, H. (1990). Spermatogenesis of Drosophila hydei. Int. Rev. Cytol. 123, 129–175. Hiroyoshi, S. (1995). Regulation of sperm quantity transferring to females at mating in the adult male of Polygonia-c-aureum (Lepidoptera, Nymphalidae). Appl. Entomol. Zool. 30, 111–119. Hiroyoshi, S. and Mitsuhashi, J. (1999). Sperm reflux and its role in multiple mating in males of a butterfly Polygonia c-aureum Linnaeus (Lepidoptera: Nymphalidae). J. Insect Physiol. 45, 107–112. Hodgson, A. N. (1997). Paraspermatogenesis in gastropod molluscs. Invert. Reprod. Dev. 31, 31–38. Holmes, S. (1979). Henderson’s Dictionary of Biological Terms. 9th ed. London and New York: Longman. Holt, G. G. and North, D. T. (1970). Spermatogenesis in the cabbage looper, Trichoplusia ni (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. 63, 501–507. Howell, B. J., Moree, B., Farrar, E. M., Stewart, S., Fang, G. W. and Salmon, E. D. (2004). Spindle checkpoint protein dynamics at kinetochores in living cells. Curr. Biol. 14, 953–964. Huang, F. and Subramanyam, B. (2003). Effects of delayed mating on reproductive performance of Plodia interpunctella (Hubner) (Lepidoptera: Pyralidae). J. Stored Prod. Res. 39, 53–63. Hunt, P. A. and Hassold, T. J. (2002). Sex matters in meiosis. Science 296, 2181–2183. Jamieson, B. G. M. (1987). The Ultrastructure and Phylogenesis of Insect Spermatozoa. Cambridge, UK: Cambridge University Press. Jans, P., Benz, G. and Friedla¨nder, M. (1984). Apyrene-spermatogenesis- inducing factor is present in the haemolymph of both male and female pupae of the codling moth. J. Insect Physiol. 30, 495–497. Jiggins, F. M., Bentley, J. K., Majerus, M. E. N. and Hurst, G. D. D. (2001). How many species are infected with Wolbachia? Cryptic sex ratio distorters revealed to be common by intensive sampling. Proc. Roy. Soc. Lond. B 268, 1123–1126. Jin, Z-Y. and Gong, H. (2001). Male accessory gland derived factors can stimulate oogenesis and enhance oviposition in Helicoverpa armigera (Lepidoptera: Noctuidae). Arch. Insect Biochem. Physiol. 46, 175–185. Joly, D. and Lachaise, D. (1994). Polymorphism in the sperm heteromorphic species of the Drosophila obscura group. J. Insect Physiol. 40, 933–938. Jones, R. T. (1978). The blood/germ cell barrier in male Schistocerca gregaria: the time of its establishment and factors affecting its formation. J. Cell Sci. 31, 145–164. Juberthie, C., Manier, J.-F. and Bossin, L. (1976). E´tude ultrastructurale de la double spermiogeńe`se chez l’opilion cyphophthalme Siro rubens Latreill. J. Microsc. Biol. Cell 25, 137–148.


299

Kaitala, A. and Wiklund, C. (1994). Polyandrous female butterflies forage for matings. Behav. Ecol. Sociobiol. 35, 385–388. Kajiura, Z., Nakagaki, M. and Takei, R. (1993). Spermiogenesis of the testes of juvenile hormone-treated silkworm larvae, Bombyx mori. Comp. Biochem. Physiol. A 106, 495–499. Kambysellis, M. P. and Williams, C. M. (1971a). In vitro development of insect tissues. I. A macromolecular factor prerequisite for silkworm spermatogenesis. Biol. Bull. Wood’s Hole Mar. Biol. Lab. 141, 527–540. Kambysellis, M. P. and Williams, C. M. (1971b). In vitro development of insect tissues. II. The role of ecdysone in the spermatogenesis of silkworms. Biol. Bull. Wood’s Hole Mar. Biol. Lab. 141, 541–552. Karlsson, B. (1998). Nuptial gifts, resource budgets, and reproductive output in a polyandrous butterfly. Ecology 79, 2931–2940. Karube, F. and Kobayashi, M. (1999). Presence of eupyrene spermatozoa in vestibulum accelerates oviposition in the silkworm moth, Bombyx mori. J. Insect Physiol. 45, 947–957. Kasuga, H., Aigaki, T. and Osanai, M. (1987). System for supply of free arginine in the spermatophore of Bombyx mori: arginine liberating activities of contents of male reproductive glands. Insect Biochem. 17, 317–322. Katsuno, S. (1977a). Studies in eupyrene and apyrene spermatozoa in the silkworm Bombyx mori L. (Lepidoptera: Bombycidae). IV. The behaviour of the spermatozoa in the internal reproductive organs of female adults. Appl. Entomol. Zool. 12, 352–359. Katsuno, S. (1977b). Studies in eupyrene and apyrene spermatozoa in the silkworm Bombyx mori L. (Lepidoptera: Bombycidae). V. The factor related to the separation of eupyrene sperm bundles. Appl. Entomol. Zool. 12, 370–371. Katsuno, S. (1987). Time of differentiation and determination of eupyrene and apyrene spermatozoa and resulting spermatogenesis in the silkworm, Bombyx mori (Lepidoptera, Bombycidae). Appl. Entomol. Zool. 22, 13–18. Kawagoe, T., Suzuki, N. and Matsumoto, K. (2001). Multiple mating reduces longevity of females of the windmill butterfly Atrophaneura alcinous. Ecol. Entomol. 26, 258–262. Kawaguchi, E. (1928). Zytologische Untersuchungen an Seidenspinner und seinen Verwandten. I. Gametogenese von Bombyx mori L. und Bombyx mandarina M. und iheren Bastarde. Zeitschr. Zellforsch. Mikrosk. Anat. 7, 519–522. Kawamura, N. and Sahara, K. (2002). In vitro cultivation of spermatocysts to matured sperm in the silkworm Bombyx mori. Dev. Growth Differ. 44, 273–280. Kawamura, N., Sahara, K. and Fugo, H. (2002). Glucose and ecdysteroid increase apyrene sperm production in in vitro cultivation of spermatocytes of Bombyx mori. J. Insect Physiol. 49, 25–30. Kawamura, N., Yamashiki, N. and Bando, H. (1998). Behavior of mitochondria during eupyrene and apyrene spermatogenesis in the silkworm, Bombyx mori, investigated by fluorescence in situ hybridization and electron microscopy. Protoplasma 202, 223–231. Kawamura, N. and Yamashiki, N. (1999). Giant nebenkern produced by colcemid treatment of the spermatocytes of the silkworm, Bombyx mori. Protoplasma 207, 67–74. Kawamura, N., Yamashiki, N., Saitoh, H. and Sahara, K. (2000). Peristaltic squeezing of sperm bundles at the late stage of spermatogenesis in the silkworm, Bombyx mori. J. Morphol. 246, 53–58. Kawamura, N., Yamashiki, N., Saitoh, H. and Sahara, K. (2001). Significance of peristaltic squeezing of sperm bundles in the silkworm, Bombyx mori: elimination of irregular eupyrene sperm nuclei of the triploid. Zygote 9, 159–166.

300


Kerr, J. B. and de Kretser, D. M. (1981). The cytology of the human testis. In: The Testis, (eds Burger, M. D. and de Kretser, D.), pp. 141–169. NY: Raven Press. Kiger, A. A., Jones, D. L., Schulz, C., Rogers, M. B. and Fuller, M. T. (2001). Stem cell self-renewal specified by JAK-STAT activation in response to a support cell cue. Science 294, 2542–2545. Kingan, T. G., Bodnar, W. M., Raina, A. K., Shabanowitz, J. and Hunt, D. F. (1995). The loss of female sex pheromone after mating in the corn earworm Helicoverpa zea: identification of a pheromonostatic peptide. Proc. Natl. Acad. Sci. USA 92, 5082–5086. Klein, C. and Wolf, K. W. (1997). Spermatocytes of the caddisfly Potamophylax rotundipennis (Trichoptera, Insecta): a fine structure study with emphasis on synaptonemal complex plates associated with chromatin. Tissue & Cell 29, 283–291. Koehler, J. K. and Birky, C. W. (1966). An electron microscope study of the dimorphic spermatozoa of Asplanchna (Rotifera). Zeitschr. Zellforsch. 70, 303–321. Koval, T. M. (1983). Intrinsic resistance to the lethal effects of x-irradiation in insect and arachnid cells. Proc. Natl. Acad. Sci. USA 80, 4752–4755. Kramer, S. F. and Bentley, W. E. (2003). RNA interference as a metabolic engineering tool: potential for in vivo control of protein expression in an insect larval model. Metab. Eng. 5, 183–190. Kristensen, N. P. (1984). The male genitalia of Agathiphaga (Lepidoptera: Agathiphagidae) and the lepidopteran ground plan. Entomol. Scand. 15, 151–178. Kubo-Irie, M., Irie, M., Nakazawa, T. and Mohri, H. (1999b). Morphological changes in eupyrene and apyrene spermatozoa in the reproductive tract of the male butterfly Atrophaneura alcinous Klug. Invert. Reprod. Dev. 34, 259–268. Kura, T. and Nakashima, Y. (2000). Conditions for the evolution of soldier sperm classes. Evolution 54, 72–80. Kubo-Irie, M., Ohata, K., Meguro, T., Kanazawa, T. and Mohri, H. (1999a). Morphological evidence that spermatogenetic cells undergo apoptosis within testis of diapausing swallowtail pupae (Atrophaneura alcinous). J. Exp. Zool. 283, 194–201. LaChance, L. E. and Olstad, G. (1988). Spermiogenesis of eupyrene sperm in prepupae, pupae, and adults of Heliothis virescens (Lepidoptera: Nocyuidae): an ultrastructural study. Ann. Entomol. Soc. Am. 81, 292–300. La Chance, L. E., Richard, R. D. and Ruud, R. L. (1977). Movement of eupyrene sperm bundles from the testis and storage in the ductus ejaculatorius duplex of the male pink bollworm: effects of age, strain, irradiation and light. Ann. Entomol. Soc. Am. 70, 647–651. Lai-Fook, J. (1982a). Testicular development and spermatogenesis in Calpodes ethlius Stoll (Hesperiidae, Lepidoptera). Can. J. Zool. 60, 1161–1171. Lai-Fook, J. (1982b). The vasa deferentia of the male reproductive system of Calpodes ethlius Stoll (Hesperiidae, Lepidoptera). Can. J. Zool. 60, 1172–1183. Lai-Fook, J. (1991). Absorption of phosphorus from the spermatophore through the cuticle of the bursa copulatrix of the butterfly, Calpodes ethlius. Tissue & Cell 23, 247–259. LaMunyon, C. W. (2000). Sperm storage by females of the polyandrous noctuid moth Heliothis virescens. Anim. Behav. 59, 395–402. LaMunyon, C. W. and Eisner, T. (1993). Postcopulatory sexual selection in an arctiid moth (Utetheisa ornatrix). Proc. Natl. Acad. Sci. USA 90, 4689–4692. LaMunyon, C. W. and Eisner, T. (1994). Spermatophore size as a determinant of paternity in an arctiid moth (Utetheisa ornatrix). Proc. Natl. Acad. Sci. USA 91, 7081–7084. LaMunyon, C. W. and Huffman, T. S. (2001). Determinants of sperm transfer by males of the noctuid moth Heliothis virescens. J. Insect Behav. 14, 187–199.


301

LaMunyon, C. W. and Ward, S. (1998). Larger sperm outcompete smaller sperm in the nematode Caenorhabditis elegans. Proc. Roy. Soc. Lond. B 265, 1997–2002. Lanot, R., Sellam, M. O., Bucher, B. and Thiebold, J. J. (1990). Ecdysone-controlled meiotic reinitiation in oocytes of Locusta migratoria involves a decrease of cAMP levels. Insect Biochem. 20, 639–644. Lanot, R., Thiebold, J., Costet-Corio, M. F., Benbeniste, P. and Hoffmann, J. H. (1988). Further experimental evidence for the involvement of ecdysone in the control of meiotic reinitiation in oocytes of Locusta migratoria (Insecta, Orthoptera). Dev. Biol. 126, 212–214. Leclercq-Smekens, M. (1978). Reproduction d’Euproctis chrysorrhea L. (Le´pidopte`re Lymantridae). II. Aspects cytologiques et role de la cellule apicalle (cellule de Verson) des testicules. Bull. Sci. Acad. Roy. Belg. 64, 318–332. Lender, T. and Duveau-Hagege, J. (1963). La survie et la diffe´renciation en culture in vitro des gonads de larves de dernier aˆge de Galleria mellonella (Lepidopte`re, Pyralidae). Dev. Biol. 6, 1–22. Leviatan, R. and Friedla¨nder, M. (1979). The eupyrene–apyrene dichotomous spermatogenesis of Lepidoptera. The relationship with postembryonic development and the role of the decline in juvenile hormone titer towards pupation. Dev. Biol. 68, 515–524. Loeb, M. J., Brandt, E. P. and Masler, E. P. (1985). Modulation of the rate of spermatogenesis by the central nervous system of the tobacco budworm, Heliothis virescens. Int. J. Invert. Reprod. Dev. 8, 39–49. Loeb, M. J., Brandt, E. P., Woods, C. W. and Bell, R. A. (1988). Secretion of ecdysteroid by sheaths of testes of the gypsy-moth, Lymantria dispar, and its regulation by testis ecdysiotropin. J. Exp. Zool. 248, 94–100. Loeb, M. J., De Loof, A., Gelman, D. B., Hakim, R. S., Jaffe, H., Kochansky, J. P., Meola, S. M., Schoofs, L., Steel, C., Vafapoulou, X., Wagner, R. M. and Woods, C. W. (2001). Testis ecdysiotropin an insect gonadotropin that induces synthesis of ecdysteroid. Arch. Insect Biochem. Physiol. 47, 181–188. Logarinho, E., Bousbaa, H., Dias, J. M., Lopes, C., Amorim, I., Antunes-Martins, A. and Sunkel, C. E. (2004). Different spindle checkpoint proteins monitor microtubule attachment and tension at kinetochores in Drosophila cells. J. Cell Sci. 117, 1757–1771. Lui, W. Y., Mruk, D., Lee, W. M. and Cheng, C. Y. (2003). Sertoli cell tight junction dynamics: their regulation during spermatogenesis. Biol. Reprod. 68, 1087–1097. Lukhtanov, V. A. and Dantchenko, A. V. (2002). Principles of the highly ordered arrangement of metaphase I bivalents in spermatocytes of Agrodiaetus (Insecta, Lepidoptera). Chromosome Res. 10, 5–20. Machida, J. (1929). Eine experimentelle untersuchung u¨ber die apyrenen spermatozoen des seidenspinners Bombyx mori L. Z. Zellforsch. Mikr. Anat. 61, 466–510. Mancini, K. and Dolder, H. (2003). Sperm morphology and arrangement along the male reproductive tract of the butterfly Euptoieta hegesia (Insecta: Lepidoptera). Invert. Reprod. Dev. 44, 107–117. Marcotte, M., Delisle, J. and McNeil, J. N. (2003). Pheromonostasis is not directly associated with post-mating sperm dynamics in Choristoneura fumiferana and C. rosaceana females. J. Insect Physiol. 49, 81–90. Marec, F., Leutelt, J., Traut, W. and Wolf, K. W. (1993). Visualization of polyfusosomes in the gonads of a moth, Ephestia ku¨hniella Z. J. Insect Embryol. Morphol. 22, 487–496. Marec, F. and Traut, W. (1993). Synaptonemal complexes in female and male meiotic prophase of Ephestia ku¨hniella (Lepidoptera). Heredity 71, 394–404.

302


Marshall, L. D. (1982). Male nutrient investment in the Lepidoptera: what nutrients should males invest?. Am. Nat. 120, 273–279. Matsuda, R. (1976). Morphology and Evolution of the Insect Abdomen. Oxford: Pergamon Press. McKee, B. D. and Karpen, G. H. (1990). Drosophila ribosomal RNA genes function as an X–Y pairing site during male meiosis. Cell 61, 61–72. McLaren, A. (1995). Germ cells and germ cell sex. Philos. Trans. Roy. Soc. B 350, 229–233. Meves, F. (1903). U¨ber oligopyrene und apyrene Spermien und ihre Entstehung nach Beobachtungen an Paludina und Pygaera. Arch. Mikrosk. Anat. Entwicklungsmech. 61, 1–84(with 8 plates). Miskimen, G. W., Rodriguez, N. L. and Nazario, M. L. (1983). Reproductive morphology and sperm transport facilitation and regulation in the female sugarcane borer, Diatraea saccharalis (F.) (Lepidoptera, Crambidae). Ann. Entomol. Soc. Am. 76, 248–252. Morisawa, M. (1994). Cell signalling mechanisms for sperm motility. Zool. Sci. 11, 647–662. Munson, J. P. (1906). Spermatogenesis of the butterfly, Papilio rutulus. Proc. Boston Soc. Nat. Hist. 33, 43–124. Murugavel, K., Veerapandian, C. and Subramanian, A. (2002). Sperm abnormalities following black quarter and foot and mouth disease vaccination in buffalo bulls. Indian Vet. Med. J. 24, 289–291. Niwa, O., Shimanuki, M. and Miki, F. (2000). Telomere-led bouquet formation facilitates homologous chromosome pairing and restricts ectopic interaction in fission yeast meiosis. EMBO J. 19, 3831–3840. North, D. T. and Holt, G. (1968). Inherited sterility in progeny of irradiated male cabbage looper. J. Econ. Entomol. 61, 928–931. Novina, C. D. and Sharp, P. A. (2004). The RNAi revolution. Nature 430, 161–164. Oberhauser, K. S. (1989). Effects of spermatophores on male and female monarch butterfly reproductive success. Behav. Ecol. Sociobiol. 25, 237–246. Oberhauser, K. S. (1992). Rate of ejaculate breakdown and intermating intervals in monarch butterflies. Behav. Ecol. Sociobiol. 31, 367–373. Ogawa, M. (1993). Differentiation and proliferation of haemopoietic stem cells. Blood 81, 2844–2854. Oˆmura, S. (1936a). Artificial insemination of Bombyx mori. J. Fac. Agr. Hokkaido Imp. Univ. 38, 135–150. Oˆmura, S. (1936b). Studies on the reproductive system of the male of Bombyx mori I Structure of the testis and the intratesticular behaviour of the spermatozoa. J. Fac. Agr. Hokkaido Imp. Univ. 38, 151–181(with 3 plates). Oˆmura, S. (1938). Studies on the reproductive system of the male of Bombyx mori. II. Post-testicular organs and post-testicular behaviour of the spermatozoa. J. Fac. Agr. Hokkaido Imp. Univ. 40, 129–170(with 3 plates). Oppliger, A., Hosken, D. J. and Ribi, G. (1998). Snail sperm production characteristics vary with sperm competition risk. Proc. Roy. Soc. Lond. B 265, 1527–1534. Orr, A. G. (2002). The sphragis of Heteronympha penelope Waterhouse (Lepidoptera: Satyridae): its structure, formation and role in sperm guarding. J. Nat. Hist. 36, 185–196. Osanai, M., Aigaki, T., Kasuga, H. and Yonezawa, Y. (1986). Role of arginase transferred from the vesicula seminalis during mating and changes in amino acid pools of the spermatophore after ejaculation in the silkworm, Bombyx mori. Insect Biochem. 16, 879–885.


303

Osanai, M. and Baccetti, B. (1993). 2-step acquisition of motility by insect spermatozoa. Experientia 49, 593–595. Osanai, M. and Chen, P. S. (1993). A comparative study on the arginine degradation cascade for sperm maturation of Bombyx mori and Drosophila melanogaster. Amino Acids 5, 341–350. Osanai, M. and Isono, M. (1997). Dissociation of eusperm bundles by acids, especially by succinate accumulated in the spermatophore of the silkmoth, Bombyx mori. Invert. Reprod. Dev. 31, 99–108. Osanai, M. and Kasuga, H. (1990). Role of endopeptidase in motility induction in apyrene silkworm spermatozoa: micropore formation in the flagellar membrane. Experientia 46, 261–264. Osanai, M., Kasuga, H. and Aigaki, T. (1987a). The spermatophore and its structural changes with time in the bursa copulatrix of the silkworm, Bombyx mori. J. Morphol 193, 1–11. Osanai, M., Kasuga, H. and Aigaki, T. (1987b). Physiological role of apyrene spermatozoa of Bombyx mori. Experientia 43, 593–596. Osanai, M., Kasuga, H. and Aigaki, T. (1988). Functional morphology of the glandula prostatica, ejaculatory valve, and ductus ejaculatorius of the silkworm, Bombyx mori. J. Morphol. 198, 231–241. Osanai, M., Kasuga, H. and Aigaki, T. (1989a). Induction of motility of apyrene sperm bundles of the silkmoth, Bombyx mori, by initiatorin and trypsin. Invert. Reprod. Dev. 15, 97–103. Osanai, M., Kasuga, H. and Aigaki, T. (1989b). Isolations of eupyrene sperm bundles and apyrene spermatozoa from seminal fluid of the silkmoth, Bombyx mori. J. Insect Physiol. 35, 401–408. Osanai, M., Kasuga, H. and Aigaki, T. (1991). Motility-related ultrastructural changes in the flagellar membrane of apyarene spermatozoa of the silkworm, Bombyx mori, induced by Arg-C endopeptidases. Invert. Reprod. Dev. 19, 193–201. Osanai, M. and Nagaoka, S. (1990). Adenine compounds in the male reproductive tract and the spermatophore of the silkmoth, Bombyx mori. Comp. Biochem. Physiol. B. Biochem. Mol. Biol. 102, 49–55. Osanai, M., Kunimoto, Y. and Nagaoka, S. (1995). Glycogen phosphorylase in the male reproductive system and spermatophore of the silkmoth, Bombyx mori. J. Sericul. Sci. Jpn. 64, 523–533. Page, S. L. and Hawley, R. S. (2003). Chromosome choreography: the meiotic ballet. Science 301, 785–789. Park, Y. I., Ramaswamy, S. B. and Srinivasan, A. (1998). Spermatophore formation and regulation of egg maturation and oviposition in female Heliothis virescens by the male. J. Insect Physiol. 44, 903–908. Parker, G. A. (1970). Sperm competition and its evolutionary consequences in the insects. Biol. Rev. 45, 525–567. Parker, G. A. and Simmons, L. W. (1991). A model of constant random sperm displacement during mating: evidence from Scatophaga. Proc. Roy. Soc. Lond. B 246, 107–115. Pasini, M. E., Redi, C. A., Caviglia, O. and Perotti, M. E. (1996). Ultrastructural and cytochemical analysis of sperm dimorphism in Drosophila subobscura. Tiss. Cell, 28, 165–175. Paulus, W. (1989). Ultrastructural investigation of spermatogenesis in Spongilla lacustris and Epydatia fluvialis (Porifera, Spongillidae). Zoomorphology 109, 123–130. Phillips, D. M. (1970). Insect sperm: their structure and morphogenesis. J. Cell Biol. 44, 243–277.

304


Phillips, D. M. (1971). Morphogenesis of the lacinate appendages of lepidopteran spermatozoa. J. Ultrastruct. Res. 34, 567–585. Pivnick, K. A. and McNeil, J. N. (1988). Puddling in butterflies: sodium effects on reproductive success in Thymelicus lineola. Physiol. Ecol. 12, 461–472. Pizzari, T. and Birkhead, T. R. (2002). The sexually selected sperm hypothesis: sexbiased inheritance and sexual antagonism. Biol. Rev. 77, 183–209. Price, C. S., Dyer, K. A. and Coyne, J. A. (1999). Sperm competition between Drosophila males involves both displacement and incapacitation. Nature 400, 449–452. Print, CG. and Loveland, L. L. (2000). Germ cell suicide: new insights into apoptosis during spermatogenesis. Bioessays 22, 423–430. Proshold, F. I. (1991). Number of sperm bundles in the duplex of tobacco budworms (Lepidoptera, Noctuidae) as a function of age. J. Econ. Entomol. 84, 1485–1491. Proshold, F. I. (1995). Remating by gypsy moths (Lepidoptera: Lymantriidae) mated with F1 sterile males as a function of sperm within the spermatheca. J. Econ. Entomol. 88, 644–648. Proshold, F. I. (1997). Sperm transfer in gypsy moths: effects of constant or cyclic temperature and constant light or darkness during the pupal stage. J. Entomol. Sci. 32, 321–331. Proshold, F. I., LaChance, L. E. and Richard, R. D. (1975). Sperm production and transfer by Heliothis virescens, H. subflexa, and the sterile hybrid males. Ann. Entomol. Soc. Am. 68, 31–34. Raina, A. K. (1989). Male-induced termination of sex pheromone production and receptivity in mated females of Heliothis zea. J. Insect Physiol. 35, 821–826. Raina, A. K., Kingan, T. G. and Giebultowicz, J. M. (1994). Mating-induced loss of sex-pheromone and sexual receptivity in insects with emphasis on Helicoverpa zea and Lymantria dispar. Archiv. Insect Biochem. Physiol. 25, 317–327. Rebollo, E., Llamazares, S., Reina, J. and Gonzalez, C. (2004). Contribution of noncentrosomal microtubules to spindle assembly in Drosophila spermatocytes. PLOS Biol. 2 (1), 54–64. Reinecke, L. H., Reinecke, J. P. and Adams, T. S. (1983). Morphology of the male reproductive tract of mature larval, pupal, and adult tobacco hornworms (Lepidoptera: Sphingidae), Manduca sexta. Ann. Entomol. Soc. Am. 76, 365–375. Reinholdt, L. G., Gutierrez, G. M. and Krider, H. M. (2002). Meiotic chromosome missegregation during apyrene meiosis in the gypsy moth, Lymantria dispar, is preceded by an aberrant prophase I. Chromosoma 111, 139–146. Richards, A. G. (1963). Giant sperm in the cockroach Periplaneta americana. Entomol. News 74, 57–60. Riddiford, L. M. and Ashenhurst, J. B. (1973). The switchover from virgin to mated behaviour in female cecropia moths: the role of the bursa copulatrix. Biol. Bull. Mar. Biol. Lab. Woods Hole 144, 162–171. Riddiford, L. M., Cherbas, P. and Truman, J. W. (2001). Ecdysone receptors and their biological actions. Vitam. Horm. 60, 1–73. Riddiford, L. M., Hiruma, K., Zhou, X. F. and Nelson, C. A. (2003). Insights into the molecular basis of the hormonal control of molting and metamorphosis from Manduca sexta and Drosophila melanogaster. Insect Biochem. Mol. Biol. 33, 1327–1338. Riemann, J. G. and Giebultowicz, J. M. (1991). Secretion in the upper vas deferens of the gypsy moth correlated with the circadian rhythm of sperm release from the testes. J. Insect Physiol. 37, 53–62. Riemann, J. G. and Giebultowicz, J. M. (1992). Sperm maturation in the vasa deferentia of the gypsy-moth, Lymantria dispar (Lepidoptera, Lymantriidae). Int. J. Insect Morphol. Embryol. 21, 271–284.


305

Riemann, J. G. and Thorson, B. J. (1971). Sperm maturation in the male and female genetial tract of Anagasta ku¨hniella (Lepidoptera: Pyralidae). Int. J. Insect Morph. Embryol. 1, 11–19. Riemann, J. G., Thorson, B. J. and Ruud, R. L. (1974). Daily cycle of release of sperm from the testes of the Mediterranean flour moth. J. Insect Physiol. 20, 195–207. Roeder, G. S. and Bailis, J. M. (2000). The pachytene checkpoint. Trends Genet. 16, 395–403. Roosen-Runge, E. C. (1977). The Process of Spermatogenesis in Animals. Cambridge: Cambridge University Press. Roselli, M., Marchini, M., Soldati, G., Campana, A. and Balerna, M. (1990). Removal of sperm coat from human spermatozoa by interaction with cervical mucus or a capacitating medium. Andrologia 22, 543–547. Royer, L. and McNeil, J. N. (1993). Male investment in the European corn-borer, Ostrinia nubilalis (Lepidoptera Pyralidae) – impact on female longevity and reproductive performance. Funct. Ecol. 7, 209–215. Rutowski, R. L. and Gilchrist, G. W. (1986). Copulation in Colias eurytheme (Lepidoptera, Pieridae) – patterns and frequency. J. Zool. 209, 115–124. Sado, T. (1961). Spermatogenesis of the silkworm and its bearing on the radiation induced sterility. Jap. J. Genet. 36 (Suppl), 136–151. Sado, T. (1963). Studies on the eupyrene and apyrene spermatozoa in the silkworm Bombyx mori L. (Lepidoptera: Bombycidae) I. The intratesticular behaviour of the spermatozoa at various stages from the 5th instar to the adult. Appl. Entomol. Zool. 12, 142–153. Sahara, K. and Kawamura, N. (2002). Double copulation of a female with sterile diploid and polyploid males recovers fertility in Bombyx mori. Zygote 10, 23–29. Sahara, K. and Kawamura, N. (2004). Roles of actin networks in peristaltic squeezing of sperm bundles in Bombyx mori. J. Morphol. 259, 1–6. Sasaki, M. and Riddiford, L. M. (1984). Regulation of reproductive behaviour and egg maturation in the tobacco hawk moth, Manduca sexta. Physiol. Entomol. 9, 315–327. Schrader, F. (1960). Evolutionary aspects of aberrant meiosis in some Penatominae (Heteroptera). Evolution 14, 498–508. Seth, R. K., Kaur, J. J., Rao, D. K. and Reynolds, S. E. (2002a). Sperm transfer during mating, movement of the sperm in the female reproductive tract, and sperm precedence in the common cutworm Spodoptera litura. Physiol. Entomol. 27, 1–14. Seth, R. K., Kaur, J. J., Rao, D. K. and Reynolds, S. E. (2004). Effects of larval exposure to sublethal concentrations of the non-steroidal ecdysone agonists RH-5849 and tebufenozide (RH-5992) on male reproductive physiology in Spodoptera litura. J. Insect Physiol. 50, 505–517. Seth, R. K., Lovell, K. V. and Reynolds, S. E. (2003). Effect of gamma irradiation on proliferation and survival of Sf9 cells: radioresistance in an insect cell line. J. Nucl. Agr. Biol. 32, 179–191. Seth, R. K., Rao, D. K. and Reynolds, S. E. (2002b). Movement of spermatozoa in the reproductive tract of adult male Spodoptera litura: daily rhythm of sperm descent and the effect of light regime on male reproduction. J. Insect Physiol. 48, 119–131. Seth, R. K. and Reynolds, S. E. (1993). Induction of inherited sterility in the tobacco hornworm Manduca sexta (Lepidoptera, Sphingidae) by substerilizing doses of ionizing radiation. Bull. Entomol. Res. 83, 227–235. Shepherd, J. G. (1974a). Activation of saturniid moth sperm by a secretion of the male reproductive tract. J. Insect Physiol. 20, 2107–2122. Shepherd, J. G. (1974b). Sperm activation in saturniid moths: some aspects of the mechanism of activation. J. Insect Physiol. 20, 2321–2328.

306


Shepherd, J. G. (1975). A polypeptide sperm activator from male saturniid moths. J. Insect Physiol. 21, 9–22. Silberglied, R. E., Shepherd, J. G. and Dickinson, J. L. (1984). Eunuchs – the role of apyrene sperm in Lepidoptera. Am. Nat. 123, 255–265. Smith, D. S. (1968). Insect Cells. Edinburgh, UK: Oliver & Boyd. Snook, R. R. (1997). Is the production of multiple sperm types adaptive? Evolution 51, 797–808. Snook, R. R. (1998). The risk of sperm competition and the evolution of sperm heteromorphism. Anim. Behav. 56, 1497–1507. Snook, R. R. and Karr, T. L. (1998). Only long sperm are fertilisation-competent in six sperm-heteromorphic Drosophila species. Curr. Biol. 8, 291–294. Snook, R. R., Markow, T. A. and Karr, T. L. (1994). Functional nonequivalence of sperm in Drosophila pseuoobscura. Proc. Natl. Acad. Sci. USA 91, 11222–11226. Sonnenschein, M. and Haüser, C. L. (1990). Presence of only eupyrene spermatozoa in adult males of the genus Micropterix Hu¨bner and its phylogenetic significance (Lepidoptera: Zeugloptera: Micropterigidae). Int. J. Insect Morphol. Embryol. 19, 269–276. Sonoli, S. B. and Hooli, M. A. (1992). Histological and histochemical studies of the apical cells of the cotton bollworm, Heliothis armigera (Hueber) (Lepidoptera: Noctuidae). Int. J. Insect Morphol. Embryol. 21, 263–270. Sorour, J. and Larink, O. (2001). Toxic effects of benomyl on the ultrastructure during spermatogenesis of the earthworm Eisenia fetida. Ecotoxicol. Environ. Safety 50, 180–188. Sullivan, A. A. and Thummel, C. S. (2003). Temporal profiles of nuclear receptor gene expression reveal coordinate transcriptional responses during Drosophila development. Mol. Endocrinol. 17, 2125–2137. Suzuki, N., Okuda, T. and Shinbo, H. (1996). Sperm precedence and sperm movement under different copulation intervals in the silkworm, Bombyx mori. J. Insect Physiol. 42, 199–204. Svard, L. and Wiklund, C. (1991). The effect of ejaculate mass on female reproductive output in the European swallowtail butterfly, Papilio machaon (L.) (Lepidoptera, Papilionidae). J. Insect Behav. 4, 33–41. Swallow, J. G. and Wilkinson, G. S. (2002). The long and short of sperm polymorphism in insects. Biol. Rev. 77, 153–182. Syrova, Z., Sauman, I. and Giebultowicz, J. M. (2003). Effects of light and temperature on the circadian system controlling sperm release in moth Spodoptera littoralis. Chronobiol. Int. 20, 809–821. Szo¨llo¨si, A. (1982). Relationship between germ and somatic cells in the testes of locusts and moths. In: Insect Ultrastructure, Vol. 1 (eds King, R. C. and Akai, H.), pp. 32–60. NY: Plenum Press. Szo¨llo¨si, A., Riemann, J. and Marcaillou, C. (1980). Localization of the blood–testis of the moth, Anagasta kuehniella. J. Ultrastruct. Res. 72, 189–199. Takamori, H. and Kurokawa, H. (1986). Ultrastructure of the long and the short sperm of Drosophila bifasciata (Diptera: Drosophilidae). Zool. Sci. 3, 847–858. Takeuchi, S. (1969). Endocrine studies on spermatogenesis in the silkworm, Bombyx mori L. Dev. Growth Diff. 11, 8–28. Tash, J. S. and Bracho, G. E. (1998). Identification of phosphoproteins coupled to initiation of motility in live epididymal mouse sperm. Biochem. Biophys. Res. Commun. 251, 557–563. Taylor, S. S., Scott, M. I. F. and Holland, A. J. (2004). The spindle checkpoint: a quality control mechanism which ensures accurate chromosome segregation. Chromosome Res. 12, 599–616.


307

Tazima, Y. (1967). The Genetics of the Silkworm. London: Logos Press. Thibout, E. (1979). Stimulation of reproductive activity of females of Acrolepiopsis assectella (Lepidoptera: Hyponomeutoidea) by the presence of eupyrene spermatozoa in the spermatheca. Entomol. Exp. Appl. 26, 279–290. Topfer-Petersen, E., Wagner, A., Friedrich, J., Petrunkina, A., Ekhlasi-Hundrieser, M., Waberski, D. and Drommer, W. (2002). Function of the mammalian oviductal sperm reservoir. J. Exp. Zool. 292, 210–215. Toyama, K. (1894). On the spermatogenesis of the silkworm. Bull. Coll. Agr. Tokyo Imp. Univ. 2, 125–157(with 2 plates). Trelles-Sticken, E., Dresser, M. E. and Scherthan, H. (2000). Meiotic telomere protein Ndj1p is required for meiosis-specific telomere distribution, bouquet formation and efficient homologue pairing. J. Cell Biol. 151, 95–106. Trivers, R. (1985). Social Evolution. Menlo Park, CA: Benjamin/Cummings. Truman, J. W. and Riddiford, L. M. (1974). Physiology of insect rhythms. III. The temporal organization of the endocrine events underlying pupation of the tobacco hornworm. J. Exp. Biol. 60, 371–382. Tschudi-Rein, K. and Benz, G. (1990). Mechanisms of sperm transfer in female Pieris brassicae (Lepidoptera, Pieridae). Ann. Ent. Soc. Am. 83, 1158–1164. Tulina, N. and Matunis, E. (2001). Control of stem cell self-renewal in Drosophila spermatogenesis by JAK-STAT signalling. Science 294, 2546–2549. Vazquez, J., Belmont, A. and Sedat, J. (2002). The dynamics of homologous chromosome pairing during male meiosis. Curr. Biol. 12, 1473–1483. Verson, E. (1889). Zu¨r spermatogenesis. Zool. Anz. 12, 100–103. Voinov, D.-N. (1903). Sur l’existence d’une double spermatogeńe`se chez les papillons. Arch. Zool. Exp. Gen. ð4th seriesÞ 1903, 49–52. Watanabe, M. (1988). Multiple matings increase the fecundity of the yellow swallowtail butterfly, Papilio xuthus L., in summer generations. J. Insect Behav. 1, 17–29. Watanabe, M., Bon’no, M. and Hachisuka, A. (2000). Eupyrene sperm migrates to spermatheca after apyrene sperm in the swallowtail butterfly, Papilio xuthus L. (Lepidoptera : Papilionidae). J. Ethol. 18, 91–99. Watanabe, M. and Hirota, M. (1999). Effects of sucrose intake on spermatophore mass produced by male swallowtail butterfly Papilio xuthus L. Zool. Sci. 16, 55–61. Watanabe, M. and Sato, K. (1993). A spermatophore structured in the bursa copulatrix of the small white Pieris rapae (Lepidoptera, Pieridae) during copulation and its sugar content. J. Res. Lepidoptera 32, 26–36. Watanabe, M., Wiklund, C. and Bon’no, M. (1998). The effect of repeated matings on sperm numbers in successive ejaculates of the cabbage white butterfly Pieris rapae (Lepidoptera: Pieridae). J. Insect Behav. 11, 559–570. Wedell, N. (2001). Female remating in butterflies: interaction between female genotype and nonfertile sperm. J. Evol. Biol. 14, 746–754. Wedell, N. and Cook, P. A. (1999a). Strategic sperm allocation in the small white butterfly Pieris rapae (Lepidoptera : Pieridae). Funct. Ecol. 13, 85–93. Wedell, N. and Cook, P. A. (1999b). Butterflies tailor their ejaculate in response to sperm competition risk and intensity. Proc. Roy. Soc. Lond. B 266, 1033–1039. Wedell, N., Gage, M. J. G. and Parler, G. A. (2002). Sperm competition, male prudence and sperm-limited females. Trends Ecol. Evol. 17, 313–320. White, M. J. D. (1973). Animal Cytology and Evolution. 3rd ed. Cambridge, UK: Cambridge University Press. White, L. D., Proshold, F., Holt, G. G., Mantey, K. D. and Hutt, R. B. (1975). Codling moth: mating and sperm transfer in females paired with irradiated and unirradiated males. Ann. Entomol. Soc. Am. 68, 859–862. Wilson, E. B. (1925). The Cell in Development and Heredity. New York: Macmillan.

308


Witalis, J. and Godula, J. (1993). Postembryonal development of the testes in the cotton leaf worm, Spodoptera littoralis (Boisd.) (Noctuidae, Lepidoptera). Acta Biol. Hung. 44, 281–295. Wolf, K. W. (1990a). Sheathed nuclear division in primary spermatocytes of Orgyia antiqua (Lepidoptera, Insecta). Biosystems 24, 5–15. Wolf, K. W. (1990b). Mitotic and meiotic spindles from two insect Orders, Lepidoptera and Diptera, differ in terms of microtubule and membrane content. J. Cell Sci. 97, 91–100. Wolf, K. W. (1991). The structure of Verson’s cell in Ephestia kuehniella Z. (Pyralidae, Lepidoptera). Cell Tiss. Res. 266, 525–534. Wolf, K. W. (1992). Spindle membranes and microtubules are coordinately reduced in apyrene relative to eupyrene spermatocytes of Inachis io (Lepidoptera, Nymphalidae). J. Submicr. Cytol. Pathol. 24, 381–394. Wolf, K. W. (1994). The unique structure of lepidopteran spindles. Int. Rev. Cytol. 152, 1–48. Wolf, K. W. (1995). The intraspindle membrane system of metaphase II spermatocytes in Phragmatobia fuliginosa (Lepidoptera, Arctiidae) appears to form a continuum. J. Struct. Biol. 115, 41–49. Wolf, K. W. (1996). Cytology of Lepidoptera. 8. Acetylation of alpha-tubulin in mitotic and meiotic spindles of two Lepidoptera species, Ephestia kuehniella (Pyralidae) and Pieris brassicae (Pieridae). Protoplasma 190, 88–98. Wolf, K. W. and Bastmeyer, M. (1991). Cytology of Lepidoptera. V. The microtubule cytoskeleton in eupyrene spermatocytes of Ephestia kuhniella (Pyralidae), Inachis io (Nymphalidae), and Orgyia antiqua (Lymantriidae). Eur. J. Cell Biol. 55, 225–237. Wolf, K. W., Baumgart, K. and Traut, W. (1987). Cytology of Lepidoptera. II. Fine structure of eupyrene and apyrene primary spermatocytes in Orgyia thyellina. Eur. J. Cell Biol. 44, 57–67. Wolf, K. W. and Joshi, H. C. (1996). Microtubule organization and distribution of g-tubulin in male meiosis of Lepidoptera. Mol. Reprod. Dev. 45, 547–559. Woods, W. A. and Stevenson, R. D. (1996). Time and energy costs of copulation for the sphinx moth. Manduca sexta. Physiol. Zool. 69, 682–700. Xia, Q. et al. (93 authors) (2004). A draft sequence for the genome of the domesticated silkworm (Bombyx mori). Science 306, 1937–1940. Yamashiki, N. and Kawamura, N. (1997). Behaviors of nucleus, basal bodies and microtubules during eupyrene and apyrene spermiogenesis in the silkworm, Bombyx mori (Lepidoptera). Dev. Growth Diff. 39, 715–722. Yamashiki, N. and Kawamura, N. (1998). Behavior of centrioles during meiosis in the male silkworm, Bombyx mori (Lepidoptera). Dev. Growth Diff. 40, 619–630. Yasuzumi, G. and Oura, C. (1965). Spermatogenesis in animals as revealed by electron microscopy - XV. The fine structure of the middle piece in developing spermatid of the silkworm, Bombyx mori Linne. Z. Zellforsch. Mikr. Anat. 67, 502–520. Zirkin, B. R. (1998). Spermatogenesis: its regulation by testosterone and FSH. Semin. Cell Dev. Biol. 9, 417–421. Zylberberg, L. (1969). Contribution a l’e´tude de la double spermatogeńe`se chez Lepidoptera (Pieris brassicae L. Pieridae). Ann. Sci. Nat. Zool. Ser. 12 11, 569–626.

Index acetylcholine esterase (ACHE) 113, 114 Acheta 86, 88, 90, 94, 124, 126, 128 Acheta domesticus 51, 54, 87, 89, 93, 95, 96, 143 Acrolepiopsis assectella 268 Acromyrmex octospinosus 28 acrosome 232 Actias selene 249 active mechanical sensing 174–6 active tactile sensors, neurobiology 49 antennal mechanoreceptors central neurophysiology of 122 distribution 84 information processing 122–6 interneurons 126–9 physiology of neurons 100–3 sensory physiology of 74 sensory structures and transduction 74 antennal mechanosensory and motor pathways connections to CNS 109–13 immunocytochemistry 113–18 mechanosensory neuropils 103–9 motoneurons in insects 118–22 neuroanatomy of 103 antennal motor system antennal innervation 65–70 biomechanics and kinematics 59–65 morphological types 51–8 neuron and muscular physiology 70–4 behavioural physiology 129 active sensing 143, 174–6 passive sensing 130, 174–6 biomimetics and antennal engineering 168 steering insects and robots 169–70 tactile sensors, engineering 170–2 adaptation 21 Aedes 2–3 Aedes aegypti 23 Agrodiaetus 222 Anochetus 71 Anopheles gambiae 28, 29, 32 ADVANCES IN INSECT PHYSIOLOGY VOL. 32 ISBN 0-12-024232-X

antagonistic pleiotropy 23 antenna motor system antennal motoneurons and muscles, in insects 70–4 biomechanical and kinematic considerations 59–65 efferent innervation, antenna 65–70 model systems 51 crustacean 55–8 insect 52–5 antennal lobe (AL) 110 antennal mechanoreceptors, sensory physiology antennal mechanosensory neurons physiology 100–3 mechanoreceptors distribution campaniform sensilla 87–8 chordotonal organs 89–92 hair plates 84–7 sensory hairs 92–100 sensory structures and transduction 74 campaniform sensilla 78–9 chordotonal organs 79–82 crustacean antenna 82–4 mechanosensory hairs 75–7 strand receptors 82 antennal mechanosensory and motor center (AMMC) 103, 113, 118, 120, 121, 123, 126 antennal mechanosensory and motor pathways antennal motoneurons, in insects dendritic arborisation pattern 120–2 general organisation 118 identified neurons 118–19 soma positions 119–20 connections to CNS 109 local brain neurons 110–11 motoneurons 110 neurons 111–13 immunocytochemistry 113–18 mechanosensory neuropils Crustacea 107–9 insects 103–7 Copyright r 2005 by Elsevier Ltd All rights of reproduction in any form reserved

310

antennal motoneurons and muscles, in insects excitatory innervation 70–2 inhibitory and modulatory innervation 72–4 antennal movements 49 antennal scanning 157 antennal sensory afferents 104 antennal tactile sense behavioural physiology 129 active sensing 143 passive sensing 130 antennation 165–6, 166–7, 168 Antheraea pernyi 270, 277, 282 antimicrobial peptides 9, 10, 13, 15, 17 Apis 57, 67, 87, 88, 110 Apis dorsata 165 Apis florea 165 Apis mellifera 24, 28, 29, 51, 163, 165 apyrene 208 apyrene sperm functions 284–91 in Lepidoptera 209–14 behaviour in female 269 apyrene spermatogenesis inducing factor (ASIF) 248–9, 249, 250, 251, 284, 289 apyreny, evolution 283–4 Aretaon asperrimus 150 Armigeres subalbatus 26 Ascogaster quadridentata 249 Asobara tabida 23, 26 Aspergillus nidulans 228 Astacus 55 Atemeles pubicollis 166 Atrophaneura alcinous 254, 260, 274 ball-joint-like structure 55 basipodite–exopodite (BE) joint 58 basipodite–ischiopodite–meropodite (BIM) joint 57, 58, 60, 61 behavioural complexity, levels 176–8 biomimetics and antennal engineering 168 steering insects and robots 169–70 tactile sensors, engineering 170–2 Blaberus 138, 168 Blaberus craniifer 137, 168 Blaberus discoidalis 149, 168 Blaberus giganteus 168 Blatella 79, 133, 168 Blatella germanica 137

INDEX

blood–testis barrier 218 Bo¨hm bristles see hair plates Bombus terrestris 27, 29, 31 Bombyx culture system 244 Bombyx mori 6, 169–70, 213, 235, 244, 250, 263, 271, 272, 274 Bombyx sperm 264, 271, 275 Bothroponera tesserinoda 139, 166 boundary defence 5 cuticle 6–8 digestive system 8–9 integumental epithelium 8 reproductive tract 9–10 sensilla 8 spiracles and respiratory system 9 bouquet 228, 229 branchiostegite 135 bundle-tip actin microfilaments 256 Calpodes ethlius 219, 260 Cambarus 55 cAMP 274–5, 277, 282 campaniform sensilla 78–9, 81, 83, 84, 91, 100, 103 honeybee flagellum 88 insect model organisms 87 Campodea 52 Camponotus sericeus 141, 166 Carausius 66, 67, 87, 88, 98, 121 Carausius morosus 51, 59, 86, 131, 144, 149 carpopodite–flagellum (CF) joint 57, 58, 60, 61, 69, 92–3, 134, 145 Cherax 57, 91 Cherax destructor 62, 91, 134–5, 145, 147, 148, 150, 152, 154 chordotonal organs 50, 52, 74, 79–82, 89 in Crustacea 91–2 in insect model organisms 89 Johnston’s organ 90–1 of pedicel 90 of scape 90 Choristoneura 263, 265, 270 Choristoneura fumiferana 263, 265, 266, 270 Choristoneura rosaceana 263, 265, 266, 270 chromosome pairing failure 226–7 genetic control, meiosis 229–30 clear band 236 clotting and wound closure 11–12

INDEX

coevolution 22 Colias eurytheme 264 Colias philodice 264 communication, antennal tactile sense 159 agonistic behaviour 160–3 courtship and mating 166–8 resources, transferring information 163–6 contact-chemosensory hairs 76–7, 82, 94, 98 contact context 61, 62, 63 copepods 31 corpora allata (CA) 246 coxopodite–basipodite (CB) joint 57, 58 Crithidia bombi 31 Crustacea 11, 50, 150 antenna, mechanoreceptors 82–4 antennal tactile sense 173–4 chordotonal organs 91–2 mechanosensory neuropils 107–9 model systems 55–8 muscle innervation 69–70 reflex modulation 135 reflex types 134–5 crustacean antenna 55, 107, 109, 135, 174 chordotonal organs 91 mechanoreceptors 82–4 crustaceans reflex modulation 135 reflex types 134–5 cuticle 6–8 30 ,50 -cyclic adenosine monophosphate see cAMP cyclic AMP role in sperm activation 275–7 trypsin 277 see also cAMP Cydia pomonella 218, 240, 243, 245, 246, 249, 252, 253 cysts 216–17 eupyrene and apyrene, co-existence 250–1 Danaus plexippus 265 Daphnia 31, 33 defence components model 3 degenerative evolution, at cellular level 212 dendritic arborisation pattern 120–1 dendritic sheath 75 diapausing testes 218 Diatraea saccharalis 270

311

dichotomous spermatogenesis 206, 207–9 evolution 212 evolutionary rationale apyreny 283–4 apyrene sperm, functions 284–91 in Lepidoptera 220 dichotomous spermatogenesis, regulation eupyrene to apyrene, switch from 248–51 apyrene spermatogenesis, inducing factor 248–50 apyreny determining factors 250 eupyrene and apyrene cysts, co-existence 250–1 mitosis and meiosis, control ecdysteroids, indirect role 245–6 juvenile hormone, role 246–7 meiosis control, by ecdysteroids 242–5 meiosis initiation 241 oogenesis and male reproductive systems 247–8 spermatogenesis, timetable 236, 238, 239 spermatogonial proliferation 239, 241 spermatogenesis control during diapause 253–4 spermiogenesis control 251–3 dorsal unpaired median (DUM) neurons 65 do-si-do 161 double-stranded b-actin RNA 256 Drosophila 2, 9, 13, 15, 29, 65, 107, 208, 224, 227, 228 Drosophila melanogaster 23, 24, 27, 34, 64, 216, 228, 286 Drosophila subobscura 208 DUM neuron activity 73 ecdysteroids control of meiosis 242–5 indirect role of 245–6 Ectomyelois ceratoniae 221, 241, 252 effector systems antimicrobial peptides 17 enzyme cascades and cytotoxins nitric oxide 16–17 phenoloxidase 16 reactive oxygen species 17

312

haemocytes drawback 18 encapsulation 20, 21 phagocytosis 19 efferent innervation, antenna muscle innervation in Crustacea 69–70 in insects 65–9 electromyogram (EMG) 133 encapsulation 20, 21, 23 endoplasmic reticulum (ER) 224, 225 Entognatha, segmented antennae 52 envelope ghosts 279 enzyme cascades and cytotoxins 16–17 Ephestia ku¨hniella 222, 225, 227, 259, 260 Escherichia coli 29 Euastacus 57, 70, 135, 145 Euastacus armatus 58, 70, 135, 145 Euptoieta hegesia 260 eupyrene and apyrene sperm behaviour in female 269–74 descent along male tract morphological changes 260–1 movement and storage 259–60 release from testis into male tract circadian rhythm 259 differential release 257 disruption and maintenance 257–9 spermiation 255–7 sperm migration sperm movement 269–70 spermatophore and source of energy 271 spermatheca sperm loss 273–4 sperm sorting 271–3 transfer to female differential transfer 267–8 mating 262–3 number of sperm transferred 266 sperm allocation 267 sperm transfer 263–4 spermatophore 264–6 eupyrene sperm, in Lepidoptera 209–14 behaviour in female 269 evolutionary terminology 22 evolutionary trade-off 22

INDEX

excitatory junction potentials (EJPs) 68, 70, 71 exploratory behaviour and tracking of objects 144–9 fitness 22 flagellum biomechanics 61–5 distribution 93–100 flow cytometry 18, 247 follicle 214, 216, 217 follicle-stimulating hormone (FSH) 248 frequency-dependent selection 21, 30 Galleria mellonella 245 gametic eunuchs 283 gamma-aminobutyric acid (GABA) 116 germarium 214 germline stem cells (GSC) 215–16 giant interneuron (GI) 141–2 glutamate 115, 116 Gram-negative binding peptides 13 graviception 130–1 Gromphadorhina portentosa 105, 107, 169 Gryllus 65, 67, 86, 88, 90, 104 Gryllus bimaculatus 94, 98, 105, 106, 107, 131, 161 Gryllus campestris 127, 167 haemocoelic defence clotting and wound closure 11–12 effector systems antimicrobial peptides 17 enzyme cascades and cytotoxins 16–17 haemocytes 17–21 self/non-self recognition 12–14 signal transduction 14–15 haemocytes 17 drawback 18 encapsulation 20, 21 phagocytosis 19 hair plates 84–7, 132, 150 head–coxopodite (HC) joint 53, 57, 58 head–scape (HS) joint 52, 53, 54–5, 59, 60, 65, 69, 143, 157 head-stand 143 Helianthus annuus 154 Heliconius 265 Helicoverpa armigera 268 Helicoverpa zea 268–9 Heliothis 245

INDEX

Heliothis subflexa 270, 292 Heliothis virescens 214, 236, 266, 268, 270, 292 helper sperm 283 Hick’s organ 78 higher insects annulated antennae 52–5 hinge-joint construction 53, 55, 60 Homarus 55 Homarus americanus 70, 91, 134, 163 honeybee flagellum 88, 99–100 hosts and parasites, relationships 30 hub cells 216 Hyalophora cecropia 260 hyperpyrene 209 Hypoponera 139 Imd pathways 13, 15 immune defence evolutionary cost 23–6 hosts and parasites relationships 30 physiological cost 26–30 immunocytochemistry antennal mechanoreceptors acetylcholine (ACH) 113–14 nitric oxide 114 serotonin 114 taurine 114–15 antennal motoneurons and muscles gamma-aminobutyric acid 116 glutamate 115 proctolin 116 immunohistochemical results 116–17 inhibitory junction potentials (IJPs) 68, 70, 71 initiatorin 275, 278, 285 innervation excitatory 70–2 inhibitory and modulatory 72–4 insect antenna 50, 51, 74, 75 insect immune system boundary defence 6–10 haemocoelic defence 10 insect immunity, evolutionary ecology 1 boundary defence 5–10 defence via behaviour 4–5 ecological immunology and variation hosts and parasites relationships 30 life history and costs 22–30 haemocoelic defence 10 immune function, plasticity 33

313

memory 31 multiple infections 32–3 specificity and receptors 31–2 insects 103–7 antennal reflexes 132–4 antennal tactile sense 173–4 mechanosensory neuropils 103–7 model systems Entognatha, segmented antennae 52 higher insects, annulated antennae 52–5 integumental epithelium 8 intratesticular sperm maturation 217 see also peristaltic squeezing ischiopodite–meropodite (IM) joint 58, 91 Janet’s organ 89 Japyx 52 Johnston’s organ 64, 65, 80, 89, 90–1, 161 juvenile hormone, role 246–7 kinematics 59–61 antenna 56 lacinate appendages 236, 239, 261, 277 lateral antennular neuropile (LAN) 108, 109 Lepidoptera dichotomous spermatogenesis 220 regulation 236 eupyrene and apyrene sperm 209–14 sperm maturation 274 sperm movement and transfer 255 spermatogenesis spermatocytes 219 spermatogonia 218 testes 214–18 spermiogenesis peristaltic squeezing 234–5 sperm surface structures 235–6 spermatozoa formation 231–4 Lepidoptera, dichotomous spermatogenesis 206, 221 apyrene, cellular mechanism chromosome pairing failure 226–7 chromosome pairing in meiosis 229–30

314

synaptonemal complex 227–8 telomeres interaction 228–9 eupyrene and apyrene difference 220–6 divisions, duration 230–1 Lepidoptera, sperm movement and transfer eupyrene and apyrene sperm descent along male tract 259–61 release from testis into male tract 255–9 transfer to female 262–8 female moth, sexual behaviour sperm retention 268–9 lepidopteran apyrene differentiation 253 Leptinotarsa decemlineata 149 Leptopilina boulardi 24 life history trait 21, 22 lipopolysaccharides 29 locomotion steering course control flight 135–7 walking and running 137–41 tactile elicited emergency behaviour 141–3 locomotion and tactile localisation guidance 149–51 locust flagellum 101 Locusta 54, 65, 66 Locusta migratoria 51, 54, 61, 136 locusts 54, 55, 62, 65, 68, 72, 127, 132, 136–7 luteinising hormone (LH) 248 Lymantria 278 Lymantria dispar 226, 227, 259, 260, 268, 270 macromolecular factor (MF) 218 Man2a2 217 Manduca 55, 65, 110, 114, 116, 118, 119, 120, 242, 243, 244, 246, 254, 278 Manduca sexta 14, 81, 214, 240, 260, 270, 274, 279 marginal sensilla 88 mATPase 71 Matrona basilaris 27 MCF chordotonal organ 91, 134 mechanoreception 49 mechanoreceptors distribution campaniform sensilla 87–8 chordotonal organs 89–92

INDEX

hair plates 84–7 sensory hairs 92–100 mechanosensory hairs contact-chemosensory hairs 76–7 insect model organisms 93 tactile hairs 75–6 medial antennular neuropile (MAN) 109 meiosis control, by ecdysteroids 242–5 initiation 241 membrana basilaris 215, 255 memory, insect immunity 31 meropodite–carpopodite (MC) joint 53, 57, 91, 92, 134, 145 Metarhizium anisopliae 7 microelectric mechanical systems (MEMS) 171 Micropterix, Zeugloptera 212 microtubule organising center (MTOC) 224 Mnais costalis 28 muscle innervation in Crustacea 69–70 in insects 65–9 Myrmecocystus mimicus 160 naked spermatozoa 279 Nauphotea cinerea 161 Nebenkern 232, 234, 252 negative geotaxis 131, 132 Nervus antennalis 67, 104 Nervus musculus tentorio scapalis 65 nitric oxide (NO) 16–17, 114 nitric oxide synthase (NOS) 16–17, 114 nodule formation 19, 20, 21 non-contact context 61, 63 non-mitotic somatic support cells see hub cells nuclear elongation 252, 253 Odontomachus 71 oligopyrene 208 operant conditioning paradigm 154, 155, 164 Orchesella 52 Orconectes 57, 163 Orconectes limosus 148 Orconectes virilis 163 oscilloscope (OSC) 159 Ostrinia nubilalis 265

INDEX

Pachycondyla villosa 165 Palinurus 57, 70, 91, 134, 145 Palinurus vulgaris 58, 82, 134 PAMP 11, 12–13, 16 see also pathogen-associated molecular patterns Panulirus 57, 148 Panulirus argus 141 Papilio machaon 265 Papilio xuthus 263, 265, 268, 270, 273 parasperm 207, 208, 289, 290 passive mechanical sensing 174–6 pathogen-associated molecular patterns 11 pathogen-associated motifs see PAMP pattern recognition and learning antennal tactile sense 151 Pectinophora gossypiella 259 pedicel–flagellum (PF) junction 52–3, 61, 62, 64, 91, 133, 136, 164 peptidoglycan-recognition proteins (PGRPs) 13 periodic acid–Schiff (PAS) staining 261, 271 Periplaneta 54–5, 86, 87, 88, 90, 92, 93, 107, 112, 114 Periplaneta americana 79, 118, 125, 133, 137, 140, 149, 169 peristaltic squeezing 234–5, 238, 244 peu probable 209 PGRP–LC 13 PGRP–SA 13 phagocytosis 13, 18, 19 phantom contact point 151 phenoloxidase (PO) 13–16 phenotypic plasticity 33 Phragmatobia fuliginosa 224 physiological trade-off 22 Pieris brassicae 225, 270 Pieris rapae 263, 264, 265 Plasmodium 23, 32 Plasmodium gallinaceum 23 plasticity, immune function 33, 34 Plodia interpunctella 24, 266 polyfusosomes 219 Polygonia c-aureum 260 prespermatogonia 215–16 Procambarus 57 Procambarus alleni 148, 161 Procambarus clarkii 108, 148 Procambarus spiculifer 62, 63, 64 proctolin 68, 72, 73, 116

315

prohaemocytes 18 prophase checkpoint, lifting 243, 244, 245 Pseudaletia separata 265, 270, 273, 288, 289 Pygaera bucephala 209, 211 pyriform lobuli see follicle QTL markers 23, 26 radial mantle 236 reactive oxygen species 17, 29 reflex chain 132, 133 reflexes avoidance, assistance and resistance antennal reflexes, in insects 132–4 reflex modulation, in crustaceans 135 reflex types, in crustaceans 134–5 respiratory burst 17 reticular appendages 236, 239, 258, 261, 278, 279 RNA interference (RNAi) 256, 292 satellite body 236 scapal chordotonal organ (sCO) 101, 123 scape–pedicel (SP) joint 53, 55 Scatophaga stercoraria 25 Schistocerca 133 Schistocerca gregaria 30, 136 Schizosaccharomyces pombe 228 seminal vesicle (SV) 260, 276, 279 sensilla 8, 59 sensilla chaetica 75, 98, 114 sensilla trichoidea see sensilla chaetica sensillum 74, 100 sensory hairs 92 flagellum distribution 93–100 functional significance 100 sensory structures, transduction 74 campaniform sensilla 78–9 chordotonal organs 79–82 crustacean antenna, mechanoreceptors 82–4 mechanosensory hairs 75–7 strand receptors 82 serotonin 114, 117 Serratia 33 Sertoli cell 217

sheathed nuclear division 224 single fast flexor motoneuron 157 slit sensilla 83, 84 soldier sperm 283, 286, 287

316

soma positions 119–20 Spa¨tzle 14, 15 sperm hatch 279, 282 sperm maturation, Lepidoptera 217 sperm activation cyclic AMP role 275–7 endogenous proteases regulation 275 eupyrene bundles dissolution 277–8 maturational changes during activation 278–82 sperm motility acquisition 274–5 spermatocytes, Lepidoptera 220, 224 spermatogonia, Lepidoptera 218–19, 241 spermatophragma 262, 264 spermiation 255–7, 259 Sphodromantis lineola 127 sphragis see spermatophragma spiracles 9 Spodoptera exempta 26 Spodoptera littoralis 255, 256 Spodoptera litura 259, 270 steering insects and robots 169–70 stilt-stand 143 strand receptors 82, 101, 102 stress receptors 79 suboesophageal ganglion (SOG) 68, 106, 111, 113, 118, 127 synaptonemal complex (SC) 227–8 tactile acuity 59–61 tactile hairs 75–6, 77, 84, 134 tactile localisation 59, 149–51, 152 taurine 114–15 Teleogryllus 90, 137 Teleogryllus commodus 94, 166, 167

INDEX

Teleogryllus oceanicus 137, 167 telomere clustering 229, 230 telomeres nuclear envelope, interaction 228–9 Tenebrio molitor 20, 27, 28, 29, 34 tentorium 52, 65, 82, 89 testes, Lepidoptera anatomy 214 blood–germ cell barrier 217–18 cysts 216–17 Verson’s cell 214–16 tetanic contractions 71, 73 thioester-containing peptides (TEPs) 14, 19 Toll pathways 13, 15 tonic motoneuron 69, 134, 135 Trichoplusia ni 241, 270 tritocerebrum (TC) 108, 113, 118, 123 tritocerebral commissure giant (TCG) 112, 127 trypsin 275, 277, 278, 282, 285 tubular body 75, 78, 79, 84 TUNEL technique 254 typical haploid sperm see eupyrene sperm upper vas deferens (UVD) 257, 260, 261 ventral area of flagellar afferents (VFA) 104, 107, 112, 124, 173 ventral unpaired median (VUM) neurons 69 Verson’s cell 214–16 waggle-dancing honeybee 64, 163, 178 window discriminator (WD) 159 workspace 59–61, 133, 175

Advances in Insect Physiology, Volume 32 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 33 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 31 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 28 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 35 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 25

Advances in Insect Physiology, Volume 14

Advances in Insect Physiology, Volume 05

Advances in Insect Physiology, Volume 25

Advances in Insect Physiology, Volume 15

Advances in Insect Physiology, Volume 08

Advances in Insect Physiology, Volume 1

Advances in Insect Physiology, Volume 18

Advances in Insect Physiology, Volume 23

Advances in Insect Physiology, Volume 30

Advances in Insect Physiology, Volume 18

Advances in Insect Physiology, Volume 7

Advances in Insect Physiology, Volume 2

Advances in Insect Physiology, Volume 21

Advances in Insect Physiology, Volume 06

Advances in Insect Physiology, Volume 9

Advances in Insect Physiology, Volume 7

Advances in Insect Physiology, Volume 29

Advances in Insect Physiology, Volume 10

Advances in Insect Physiology, Volume 24

Advances in Insect Physiology, Volume 35

Advances in Insect Physiology, Volume 20

Advances in Insect Physiology, Volume 26

Advances in Insect Physiology, Volume 18

Advances in Insect Physiology, Volume 39

Advances in Insect Physiology, Volume 17

Advances in Insect Physiology, Volume 32 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 33 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 31 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 28 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 35 (Advances in Insect Physiology)

Advances in Insect Physiology, Volume 25

Advances in Insect Physiology, Volume 14

Advances in Insect Physiology, Volume 05

Advances in Insect Physiology, Volume 25

Advances in Insect Physiology, Volume 15

Advances in Insect Physiology, Volume 08

Advances in Insect Physiology, Volume 1

Advances in Insect Physiology, Volume 18

Advances in Insect Physiology, Volume 23

Advances in Insect Physiology, Volume 30

Advances in Insect Physiology, Volume 18

Advances in Insect Physiology, Volume 7

Advances in Insect Physiology, Volume 2

Advances in Insect Physiology, Volume 21

Advances in Insect Physiology, Volume 06

Advances in Insect Physiology, Volume 9

Advances in Insect Physiology, Volume 7

Advances in Insect Physiology, Volume 29

Advances in Insect Physiology, Volume 10

Advances in Insect Physiology, Volume 24

Advances in Insect Physiology, Volume 35

Advances in Insect Physiology, Volume 20

Advances in Insect Physiology, Volume 26

Advances in Insect Physiology, Volume 18

Advances in Insect Physiology, Volume 39

Advances in Insect Physiology, Volume 17

Recommend Documents