Oceanography And Marine Biology

OCEANOGRAPHY AND MARINE BIOLOGY

AN ANNUAL REVIEW Volume 27

OCEANOGRAPHY AND MARINE BIOLOGY AN ANNUAL REVIEW Volume 27

HAROLD BARNES, Founder Editor MARGARET BARNES, Editor The Dunstaffnage Marine Research Laboratory Oban, Argyll, Scotland Assistant Editors A.D.Ansell R.N.Gibson T.H.Pearson

ABERDEEN UNIVERSITY PRESS

FIRST PUBLISHED IN 1989 This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” This book is copyright under the Berne Convention. All rights reserved. Apart from any fair dealing for the purpose of private study, research, criticism or review, as permitted under the Copyright Act, 1956, no part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, electrical, chemical, mechanical, optical, photocopying, recording or otherwise, without the prior permission of the copyright owner. Enquiries should be addressed to the Publishers. © Aberdeen University Press 1989 British Library Cataloguing in Publication Data Oceanography and marine biology: an annual review.—Vol. 27 1. Oceanography—Serials 2. Marine biology—Serials 551.46′005 ISBN 0-203-01629 -7 Master e-book ISBN

ISBN 0-203-19 130-7 (Adobe eReader Format) ISBN 0-08-036397-0 (Print Edition) ISSN 0075-3218

PREFACE

The demand for this series of Annual Reviews continues unabated. Many people are still willing to contribute and we are grateful to them. It is always a pleasure to acknowledge the willingness with which they accede to our editorial requests. The success of the series depends on their co-operation; without it our task would be impossible. We thank them all and also the publishers for maintaining the regular appearance of these Annual Reviews. In the first 27 volumes of the series titles of references have not been included. The reason was the space they would occupy which it was thought could be better spent on text. Present demand, however, indicates that a change is necessary. From Volume 28 onwards it is hoped to include full titles in reference lists.

CONTENTS

page PREFACE

iv

A Comparison of Marine Photosynthesis with Terrestrial Photosynthesis: a Biochemical Perspective GRAHAM J.KELLY

1

The Ecology and Behaviour of Ascidian Larvae IB SVANE AND CRAIG M.YOUNG

28

Egg Production in Cirripedes MARGARET BARNES

62

Biology of Marine Herbivorous Fishes MICHAEL H.HORN

134

The Benguela Ecosystem. Part VI. Seabirds A.BERRUTIN.J.ADAMS AND S.JACKSON

222

Review of Research Relevant to the Conservation of Shallow Tropical Marine Ecosystems B.G.HATCHER, R.E.JOHANNES AND A.I.ROBERTSON

284

AUTHOR INDEX

353

SYSTEMATIC INDEX

386

SUBJECT INDEX

398

Oceanogr. Mar. Biol. Annu. Rev., 1989, 27, Margaret Barnes, Ed. Aberdeen University Press

A COMPARISON OF MARINE PHOTOSYNTHESIS WITH TERRESTRIAL PHOTOSYNTHESIS: A BIOCHEMICAL PERSPECTIVE GRAHAME J.KELLY* CSIRO Marine Laboratories, G. P. O. Box 1538, Hobart, Tasmania 7001, Australia

ABSTRACT The maximum photosynthetic rate of marine microalgae—both cultures and natural populations—is often estimated as being two to four times higher than that of leaves of terrestrial plants, although the average estimate is one to two times. In an attempt to ascertain whether the higher estimates are biochemically feasible, the photosynthetic biochemistry of the two groups of plants is compared. A survey of the literature on the physiology and biochemistry of marine algae shows that: (1) their chlorophyll a content (dry weight basis) is similar to that of leaves of terrestrial plants, (2) the in vitro values obtained for photosynthetic electron transport and ribulosel,5-bisphosphate carboxylase activity have all been low, (3) carboxylation of phosphoenolpyruvate can account for no more than 10% of net carbon fixation, (4) photorespiration, which would have little influence on estimates of primary productivity, is possibly prevented by active uptake of inorganic carbon, (5) photoinhibition is either avoided or protections against it exist, and (6) the extent of dark respiration during photosynthesis is not known. Information in the literature indicates that the actual in situ (light-limited) photosynthesis of marine microalgae is 7% of its potential (at light saturation). Assuming that phytoplankton biomass (chlorophyll) has been correctly estimated, and that the biochemistry of marine algal photosynthesis is similar to that of terrestrial plants, then the current estimate of in situ marine photosynthesis is concluded to be biochemically reasonable.

INTRODUCTION This review is an attempt to put into some order the rambling thoughts of a peas-and-barley biochemist whose head was suddenly immersed in the ocean. Peas-and-barley biochemists are accustomed to expecting maximum rates of photosynthesis (Pmax: photosynthesis at saturating light and carbon dioxide and ambient temperature) of between 200 and 400 µmoles CO2 fixed·mg chlorophyll−1·h−1. These rates are equivalent to

2

GRAHAME J.KELLY

assimilation numbers (i.e. photosynthesis expressed as mg carbon fixed.mg chl. a−1·h−1) of between 2.4 and 4.8. I was therefore rather surprised to find that estimates of the Pmax of marine phytoplankton commonly yielded assimilation numbers between 5 and 10 (and sometimes approaching 20), and that investigators of marine photosynthesis often fear that assimilation numbers much below 5 may represent under-estimates of the true values. This difference between the marine and terrestrial sectors is illustrated by the contrast between the following two quotations: “Maximum rates of photosynthesis [by land plant leaves] depend on species and growth conditions; they can surpass 300 µmol CO2 reduced.mg chlorophyll−1·h−1” (i.e. assimilation number of 3.6) (Heber, Neimanis & Dietz, 1988); “Often cited as evidence of low growth rates has been the fact that assimilation ratios reported from open-ocean stations have almost always been 100

0.4a

8a

5

Seven cultures

–

–

34a

Amphidinium carterae

1.1

?

–

Pyrocystis noctiluca Bedford Basin Dunaliella tertiolecta Skeletonema costatum Synechococcus spp. in eddy surface waters

1.2 ? 5.4 0.5

4.0 1.6 6.8 3.3

30 14 79 15

1.0a

1.0a

100

Beardall & Morris, 1976 Beardall et al., 1976 Holdsworth & Colbeck, 1976 Glover & Morris, 1979 Glover & Morris, 1979 Appleby, Colbeck, Holdsworth & Wadman, 1980 Rivkin et al., 1982 Smith et al., 1983 Sukenik et al., 1987 Mortain-Bertrand et al., 1987 Prézelin et al., 1987

a

Approximate mean values.

turing CO2, but this idea lost popularity when it was found that algal photosynthetic rates correlated with RuBP carboxylase activity, but not with PEP carboxylase activity (Mukerji, Glover & Morris, 1978; Glover & Morris, 1979; Smith, Platt & Harrison, 1983). Terrestrial C4 plants and CAM plants utilize PEP carboxylase to capture CO2 in, respectively, the C4 pathway and crassulacean acid metabolism (CAM); hence the terms “C4 plants” and “CAM plants”. These mechanisms are not alternatives to the Calvin cycle because, by themselves, they would not permit plant growth. In any event, it is quite unlikely that any unicellular marine alga could be classed as a C4 plant or CAM plant, because the former requires co-operation between two types of cells in the one plant, while the latter is an adaptation for survival in arid environments. Consequently, marine algae are almost certainly “C3 plants”. The origin of the terms “C3” and “C4” have, however, caused some confusion. C3 refers to the fact that a three-carbon compound (glycerate-3-P, see Fig 3) is the first product of CO2 fixation, whereas C4 denotes that a four-carbon compound (oxaloacetate, see Fig 4) is initially observed. As phytoplankton synthesise more protein and less carbohydrates than do multicellular plants, much triose-P produced by the Calvin cycle must move via PEP to produce the carbon skeletons needed for amino-acid biosynthesis. A good portion of this PEP is carboxylated with CO2 to produce oxaloacetate (which is, itself, a carbon

PHOTOSYNTHESIS: A BIOCHEMICAL PERSPECTIVE

17

Fig 4.—The Calvin cycle as the producer of PEP for PEP carboxylation. The maximum long-term ratio of PEP carboxylation: RuBP carboxylation during growth is 1:3. If two thirds of the triose-P (representing the product of CO2 fixation) is used for the synthesis of carbohydrates, lipids, and amino acids not derived from oxaloacetate (as is usual), then the ratio reduces to 1:9. Consequently, attempts to credit PEP carboxylation with more than 10% of the total photosynthetic carbon fixation are fundamentally in error.

skeleton, and is a precursor of several others). Although this gives a “C4 flavour” to algal CO2 fixation, the principal purpose of oxaloacetate synthesis in marine algae (amino-acid synthesis) is entirely unrelated to that of C4 and CAM plants (a CO2-capturing mechanism). Nevertheless, the concept that PEP carboxylation may make up for a perceived insufficiency of RuBP carboxylation has persisted, and has supported a considerable amount of theoretical discussion (Morris, 1980; Priscu & Goldman, 1983; Smith et al., 1983). But, as emphasised by Figure 4, the concept itself is wrong. In growing photosynthetic cells, three CO2 must be fixed by RuBP carboxylase in order to generate a PEP before a CO2 can be fixed by PEP carboxylase or PEP carboxykinase. This is because there is no known biochemical pathway whereby the product of PEP carboxylation (oxaloacetate) can be metabolised so that three of its four carbons are used to regenerate CO2-acceptor (PEP) while the other is kept in the reduced form and used for plant growth (Kelly et al., 1976). Consequently, since one net PEP carboxylation requires three prior RuBP carboxylations, the maximum contribution of PEP carboxylation to total photosynthetic CO2 fixation is 25%. This theoretical limit reduces to about 10% in practice because not all amino-acid skeletons can be derived from oxaloacetate, and because protein is only one of the major products of cell growth; others, for example sugars, polysaccharides and lipids, cannot be synthesised from oxaloacetate without the loss of the CO2 that was fixed during PEP carboxylation. The practical value of about 10% for the contribution of PEP carboxylation is in good agreement with experimental values of 6.7 to 12.7% recently obtained for two diatoms (Mortain-Bertrand, Descolas-Gros & Jupin, 1988). Finally, there may be stages in the growth cycle of an alga where PEP carboxylation equals or exceeds RuBP carboxylation. Microalgal cultures in the stationary phase appear to have this attribute (Mukerji et al., 1978; Glover & Morris, 1979), but these cells are obtaining PEP from a carbohydrate reserve (e.g. mannitol in brown algae) previously synthesised entirely through RuBP carboxylation (Kremer, 1981b). PHOTORESPIRATION AND MARINE ALGAE DOES PHOTORESPIRATION OCCUR IN MARINE ALGAE? Photosynthetic cells of terrestrial C3 plants exhibit a major light-dependent set of reactions that results in O2 uptake, CO2 release, and the consumption of energy (Fig 5). Due to its light dependence and gas exchanges, it has been called photorespiration, but it is otherwise not at all like dark (mitochondrial) respiration. Photorespiration occurs when O2 replaces CO2 in the RuBP carboxylase (now called “carboxylase/ oxygenase”) reaction, leading to the formation of one glycerate-3-P, plus a smaller (2-carbon) compound called glycolate (not two glycerate-3-P, as in Fig 3). A series of six reactions then converts two glycolate molecules to one glycerate-3-P and one CO2 (Fig 5). If all molecules of RuBP were oxygenated, plant growth would cease. The combination of the enzyme’s relative affinities for CO2 and O2 and the relative concentrations of these two gases in air results, however, in a situation somewhat akin to Figure 5b where carboxylation exceeds oxygenation sufficiently for plant growth to proceed. The real function of photorespiration is not yet known. It may act as a “release valve” to dissipate photosynthetically generated energy, as in Figure 5a, when the CO2 supply is restricted (e.g. during periods

18

GRAHAME J.KELLY

Fig 5.—Photorespiration. a, at the CO2 compensation point, where the amount of CO2 lost by photorespiration is exactly balanced by the amount fixed in the Calvin cycle; at this state, where the ratio of RuBP carboxylated: oxygenated is 1:2, 9 ATP and 6 NAD(P)H are ‘wasted’, while plant growth is zero. b, the situation with terrestrial C3 plants in normal air, where carboxylation to oxygenation is approximately 1:0.7, and growth (represented by the synthesis of a molecule of sucrose from 12 CO2) in the presence of photorespiration is possible.

of water stress when stomates close), thereby preventing “over-reduction” of the photochemical apparatus and subsequent photoinhibition (see below; Osmond, 1981; but see Powles, Comic & Louason, 1984). Nevertheless, it still proceeds in C3 plants when stomates are open, reducing productivity by 30 to 50%. In C4 plants photorespiration is, however, inhibited because the C4 pathway concentrates CO2 in the vicinity of RuBP carboxylase, and because O2 and CO2 are straightforward competitors for entry onto the enzyme’s active site, the increased CO2 concentration overwhelms the oxygenase reaction so that it is effectively eliminated. A different mechanism appears to produce a similar consequence in marine algae. As mentioned on p. 26, the effects of pH and/or the active uptake of inorganic carbon lead to intracellular CO2 concentrations of 100 µM or more in marine algae. Consequently, the CO2:O2 ratio would be expected to increase such that, in these cells too, the oxygenase reaction would be overwhelmed. This does not mean that marine algae may not be capable of photo-respiration. They can synthesise glycolate (Lloyd, Canvin & Culver, 1977; Glover & Morris, 1981; Fogg, 1983) and they possess enzymes for converting it to glycerate (Paul & Volcani, 1976). But despite claims that photorespiration may occur (Burris, 1977; Morris & Glover, 1981), the likelihood of it occurring under natural conditions is small (Lloyd et al., 1977; Birmingham, Coleman & Colman, 1982; Burns & Beardall, 1987), and most recent work showing that O2 has little or no effect on the rates of photosynthesis by marine algae argues against it (Colman & Gehl, 1983; Colman, 1984; Bidwell & McLachlan, 1985; Beer & Israel, 1986; Cook & Colman 1987; Johnston & Raven, 1987). A study of coccoid marine cyanobacteria by Glover & Morris (1981) indicated that these marine photosynthetic cells have a greater potential for photorespiration than have eukaryotic marine algae, but even in this case the magnitude of photorespiration in the presence of 2 mM was near zero. One further consideration, however, counteracts these arguments. Build-up of CO2 in cells (due to effects of pH and/or active uptake) will lead to increased photosynthesis and, consequently, increased O2 evolution. This may produce locally high concentrations of O2 in chloroplasts, favouring RuBP oxygenase activity and photorespiration. Evidence for such an event in three freshwater green microalgae provided with inorganic carbon (CO2 plus HCO3) at concentrations up to 0.6 mM has recently been obtained (Yokota & Kitaoka, 1987). Similar experiments with marine microalgae growing in 2 mM inorganic carbon are required. PHOTORESPIRATION AND PRODUCTIVITY ESTIMATES If photorespiration in phytoplankton does occur, it would not affect productivity estimations based on measurements of O2 evolution (using the Winkler technique), as suspected by some workers (see Round, 1981, p. 259). Reference to Figure 5a shows that a zero change in CO2 (one CO2 taken up, and one released) correlates with a zero change in O2: three O2 are used— two during oxidation of two RuBP, and net of one during oxidation of two glycolate—while three O2 are liberated from water (see Fig. 1, p. 22) during the photosynthetic electron transport that generates the required reducing power (five NADPH and one NADH; Fig 5a). The amount of O2 used during glycolate oxidation is the same whether the glycolate is oxidised by an oxidase, as shown in Figure 5, or by a mitochondrial dehydrogenase linked to respiratory O2

PHOTOSYNTHESIS: A BIOCHEMICAL PERSPECTIVE

19

consumption; both types of oxidation have been detected in algae. With photorespiration proportionately less than photosynthesis, as when net CO2 fixation proceeds (Fig 5b), the net O2 evolution will equal the net CO2 assimilation, as if this CO2 fixation were occurring in the absence of photorespiration. Photorespiration would affect productivity estimations using the 14C-technique, but not dramatically. It is most probable that “total CO2 fixation minus photorespiratory CO2 evolution” would be estimated, because photorespiratory and photosynthetic metabolisms share common intermediates and are obligatorily interconnected (Fig 5), so that 14C entering the Calvin cycle will also enter photorespiration after a short delay. In a sense, the result would be the same as for O2 measurements, i.e. (net) CO2 fixation would be measured as if it were occurring in the absence of photorespiration.

PHOTOINHIBITION If photorespiration is absent, the question arises as to how phytoplankton dispose of excess photosynthetically generated energy. Failure to dispose of this excess energy is thought to be one cause of photoinhibition, whereby over-reduction of electron transport components causes a chain of events that results in damage to one or more of the components that make up the reaction centre of photosystem II (Allakhverdiev, Šetlíková, Klimov & Šetlik, 1987; Demeter, Neale & Melis, 1987; Lidholm, Gustafsson & Öquist, 1987; Kirilovsky, Vernotte, Astier & Étienne, 1988; Ohad, Koike, Shochat & Inoue, 1988). The damage can become so extensive as to be irreversible, and cell death follows. Other possible mechanisms for protecting against photoinhibition implicate (1) carotenoids; (2) use of energy to reduce O2 to the superoxide anion (Mehler reaction) which, however, must subsequently be dealt with; (3) spillover of energy from photosystem II to photosystem I; (4) cyclic electron flow around photosystem II; and (5) reduction in the level of light-harvesting pigments (Samuelsson et al., 1985; Ben-Amotz, Gressel & Avron, 1987; Demmig, Winter, Krüger & Czygan, 1987; Geider, 1987; Siefermann-Harms, 1987; Falkowski, Kolber & Fujita, 1988). In addition, as Harris (1978) has argued, phytoplankton populations may normally largely avoid photoinhibition, either passively (vertical mixing; sinking), or actively (motile microalgae may swim away form excessive light). Support for this viewpoint has come from Vincent, Neale & Richerson (1984), who describe an extensive (but still reversible) photoinhibition of phytoplankton each morning in the high-altitude Lake Titicaca (Peru-Bolivia) when the phytoplankton became temporarily trapped under extreme irradiance by a near-surface thermocline. On the other hand, algae that are continually exposed to high light intensities are apparently more resistant to photoinhibition. Seaweeds, Synechococcus, Thalassiosira pseudonana, and a population of benthic microalgae dominated by Euglena obtusa and Pleurosigma angulatum all showed no evidence of photoinhibition under almost full sea-surface irradiance (Penniman & Mathieson, 1985; Mills & Wilkinson, 1986; Prézelin, Putt & Glover, 1986; Kana & Glibert, 1987a; Sakshaug, Demers & Yentsch, 1987). Thus, there appears to be a broad spectrum of tolerance to photoinhibition in algae. At least some of this variation may be an adaptation, in that photosynthetic cells exposed daily to high irradiance contain less lightharvesting pigments per electron-transport chain (compared with “shade-adapted” cells), and are thus less likely to overload their light processing machinery (Osmond, 1981). To a certain degree, microalgae may avoid photoinhibition by using excess photochemical energy rather than allowing it to accumulate to the point where all reaction centres become occupied. Such utilisation could involve the assimilation of more CO2 than is required for growth. This may help to explain why stressed (e.g. nitrogen-deficient) cells of some species excrete large quantities of carbohydrates, as polysaccharides or glycolate, during photosynthesis (Jensen, 1984; Vieira & Myklestad, 1986; Al-Hasan &

20

GRAHAME J.KELLY

Fogg, 1987); these excretions would be the end-product of energy harvested in a process not linked to cell growth (see also Dring, 1982, p. 75). Exudation of low molecular weight organic compounds by nonstressed cells is, however, believed by Bjørnsen (1988) to be a consequence of passive diffusion through the cell membrane. Not only high intensities of photosynthetically active radiation, but also the much lower intensities of UVB light (290–320 nm) normally encountered in the solar spectrum cause photoinhibition of algae that occur high in water columns (Harris, 1978; Richardson, Beardall & Raven, 1983, Döhler, 1984; Jokiel & York, 1984; Vincent et al., 1984, Wood, 1987). The mechanism of this inhibition is not known. Likely explanations are that the nucleic acids or components of the photosynthetic electron transport chain are damaged. The photosynthetic cells of leaves of terrestrial plants are protected from UV-B-induced photoinhibition by flavonoid compounds in the non-photosynthetic cells of the leaf epidermis (Caldwell, 1981). DARK (MITOCHONDRIAL) RESPIRATION DOES MITOCHONDRIAL RESPIRATION PROCEED DURING PHOTOSYNTHESIS? In productivity experiments using the question arises whether gross photosynthesis (total CO2 fixation itself) or net photosynthesis (total CO2 fixation minus CO2 loss through respiration and/or photorespiration) is measured. Investigations of this question have reached differing conclusions: that it measures gross photosynthesis (Dring & Jewson, 1982), net photosynthesis (Smith & Platt, 1984), or some value in between, depending on the growth rate (Harris & Piccinin, 1983). This problem will not be solved until the extents of mitochondrial respiration (in the dark and in the light) and photorespiration are known. There are, however, some indications. First, as outlined above, photorespiration may not occur in phytoplankton living in natural conditions, and even if it did, it would not be likely to affect productivity measurements. Thus, only dark (mitochondrial) respiration need be considered. It is not known whether the dark (mitochondrial) respiration of photosynthetic cells is suppressed during photosynthesis, and if so, to what extent (Kelly, 1983). In one sense, respiration may be unnecessary in that its prime purpose (complete oxidation of carbohydrate to CO2 in order to generate energy) would seem inappropriate in a cell harvesting light energy and reducing CO2 to carbohydrate. A portion of the Krebs cycle (tricarboxylic acid cycle) would, however, need to proceed at a low rate in order to generate α-ketoglutarate, which is the precursor of four amino acids and of chlorophyll. This could proceed as shown in Figure 6, in which a net evolution of one CO2 per α-ketoglutarate synthesised would occur. But note that, for the growth of cells containing 50% protein and 1% chlorophyll (dry weight basis), the required rate of this partial Krebs cycle would lead to a CO2 evolution equivalent to only 2% of the photosynthetic CO2 fixation. Despite this requirement for some Krebs cycle activity for growth, a concomitant respiratory O2 uptake would not necessarily occur because the reducing power (NADH) generated during α-ketoglutarate synthesis could be delivered to the cytosol and used there, as shown in Figure 6. The respiratory electron transport chain would then remain inactive. Some recent experiments with the freshwater green microalgae Chlamydomonas reinhardtii and Selenastrum minutum suggest, however, that both the Krebs cycle and respiratory electron transport function during photosynthesis in these algae (Peltier & Thibault, 1985; Husic & Tolbert, 1987; Weger, Birch, Elrifi & Turpin, 1988), but the extent to which this is so in other algae is not known, and once again

PHOTOSYNTHESIS: A BIOCHEMICAL PERSPECTIVE

21

Fig 6.—Limited operation of the mitochondrial Krebs cycle during photosynthesis in order to generate α-ketoglutarate. Two molecules of glycerate-3-P (from photosynthetic CO2 fixation) are converted to one α-ketoglutarate and one CO2, and two NAD+ are reduced in the process. A metabolic shuttle may dispose of this reducing power in the cytosol (as shown) or respiratory electron transport may oxidise it, with consumption of one O2.

the question of function arises. With respect to the function, Husic & Tolbert (1987) have alluded to the alternative (cyanide-resistant) pathway of respiratory electron transport in plant mitochondria. This pathway is not coupled to ATP synthesis and is therefore a means of “burning off” excess carbohydrate (Lambers, 1982), an intriguing possibility in view of the suggestion that marine algae may not be able to dispose of excess energy by photorespiring as land plants do. The alternative pathway, however, generates superoxide anions and peroxide, thus mechanisms for the removal of these toxic substances would need to accompany any alternative pathway activity. All of these considerations have been subsequently disturbed by an experiment, elegant in its simplicity, in which an inhibitor of mitochondrial electron transport was shown to inhibit photosynthesis by barley leaf protoplasts, but to have no effect on the photosynthesis of chloroplasts isolated from these protoplasts (Krömer, Stitt & Heldt, 1988). It was concluded that mitochondrial electron transport is required during photosynthesis to generate ATP for cytosolic metabolism, and that the ultimate electron donor is photosynthetically-generated NADPH shuttled from chloroplasts to mitochondria, more or less by a reversal (and extension) of the malateoxaloacetate shuttle shown in Figure 6. The advantage of this system is that it is energetically efficient (3.8 ATP per 4 photons, compared with 1.3 ATP per 4 photons when the ATP is generated during either cyclic or pseudocyclic electron transport in chloroplasts). It could therefore be useful to phytoplankton cells growing in a low light environment, especially because these cells require cytosolic ATP for protein synthesis (an ATP supply rate of about 7% of the rate of CO2 fixation is needed only for the activation and attachment of amino acids to growing polypeptide chains), and to pump inorganic carbon into cells (possibly an even more ATP-demanding process; Yokota et al., 1987). Note that this system of producing ATP by oxidising NADPH of chloroplast origin involves mitochondrial electron transport only; Krebs cycle activity would not be required. Thus, respiratory oxygen uptake would occur, but respiratory CO2 evolution would not (other than that noted above for αketoglutarate synthesis). Phytoplankton appear to possess ample respiratory electron transport activity (Kenner & Ahmed, 1975). MITOCHONDRIAL RESPIRATION AND PRODUCTIVITY ESTIMATES Of course, the extent of mitochondrial respiration during the night also needs to be evaluated for productivity estimates. Most data indicate that the rate at which marine algae respire in darkness is equivalent to about 10% of the light-saturated rate of photosynthesis (Harris, 1978; Falkowski & Owens, 1980; Taguchi, 1981; Richardson et al., 1983; Geider, Osborne & Raven, 1985; Smith & Geider, 1985; Gerard, 1986). Assuming respiration occurs both day and night, then at maximum growth rates the net photosynthesis would equal 80% of gross photosynthesis. This percentage may, however, be higher if respiration is partially or completely inhibited during photosynthesis, as discussed above, and there is other evidence for this (Burris, 1980; Scherer, Stürzl & Böger, 1984). There is also evidence that any respiratory CO2 released in illuminated cells is quickly re-assimilated by photosynthesis before it can leave the cell (Scherer et al., 1984). This possibility, which was denied by Bidwell (1977), was proposed in the 1950s for Dunaliella euchlora (Ryther, 1956).

22

GRAHAME J.KELLY

Fig 7.—Efficiencies in marine production. Mean daytime surface irradiance of is equal to (Richardson et al., 1983), and thus to and to Average photosynthesis (Table I) of 69 g carbon·m−2·yr−1 is converted to its energy equivalent as described on p. 25. Average fish production is calculated by dividing total production at the end of food chains ( tonnes fresh weight [Ryther, 1969], equivalent to approximately tonnes carbon) by total ocean area and then converting to energy units as for the phytoplankton.

Although respiration appears to use only a small fraction of carbohydrate produced under light saturation, it may become more significant to the balance of carbon metabolism under light limitation. Phytoplankton commonly occur at depths where irradiance is low, so respiration could be important in assessing net photosynthesis in the ocean, but to what extent is not clear; it has been estimated that 40% of gross photosynthesis in the ocean is lost in dark respiration (see Burris, 1980) but some species, when growing slowly in dim light, appear able to adjust their respiration to a correspondingly low level (Falkowski & Owens, 1980; Geider, 1987; Langdon, 1987). Of course, if irradiance falls so low that photosynthesis and respiration are equal, then growth will cease. At this point, the irradiance is termed the “light compensation point”. Experiments indicate this point to range between low values near for diatoms grown in low light, to high values of over for dinoflagellates (Falkowski & Owens, 1980; Rivkin, Seliger, Swift & Biggley, 1982; Hobson & Guest, 1983; Geider & Osborne, 1986; Langdon, 1987). CONCLUDING COMMENT The efficiency with which solar energy is used to convert inorganic precursors into fish in the ocean is 0. 00014%, based on an average sea-surface irradiance of and an annual total fish production of tonnes carbon (Fig. 7). Taking the value for marine algal photosynthesis to be 69 g carbon fixed·m−2·yr−1 (Table I, see p. 13), then the global efficiencies for the two sections of this conversion (i.e. sunlight to algae, and algae to fish) are a little above 0.1 % each. For sunlight to algae, the value varies regionally from 0.02 to 5%; the overall mean is relatively low (0.12%) because much light energy is atenuated by the overlying water column before it reaches the algae (Kirk, 1983, p. 244), resulting in marine photosynthesis being, on average, markedly light-limited. For algae to fish, the low value is a consequence of having about three links, on average, in the marine food chain from the microscopic phytoplankton to the fish at the end of the food chains (Ryther, 1969). In time, the value for marine algal production, and thus the efficiency values shown in Figure 7, may be revised. From the biochemical viewpoint, revisions may be needed when the extents of respiration and photorespiration in marine phytoplankton are better understood. As emphasised in this review, upward revisions cannot, however, be confidently based on high estimates of Pmax values for marine photosynthetic cells unless it is demonstrated that these cells have a photosynthetic machinery with a significantly greater capacity for CO2 fixation than that of the photosynthetic cells of terrestrial plants. ACKNOWLEDGEMENTS I am grateful for many helpful discussions with S.W.Jeffrey who introduced me to the marvellous minute plants of the ocean, and with J.S.Parslow, who taught me much about their growth. REFERENCES Aizawa, K. & Miyachi, S., 1984. FEBS Lett., 173, 41–44.

PHOTOSYNTHESIS: A BIOCHEMICAL PERSPECTIVE

23

Aizawa, K., Tsuzuki, M. & Miyachi, S., 1986. Plant Cell Physiol, 27, 37–47. Akagawa, H., Ikawa, T. & Nisizawa, K., 1972. Plant Cell Physiol, 13, 999–1016. Al-Hasan, R.H. & Fogg, G.E., 1987. Mar. Ecol. Prog. Ser., 37, 305–307. Allakhverdiev, S.I., Šetlíková, E., Klimov, V.V. & Šetlík, I., 1987. FEES Lett., 226, 186–190. Appleby, G., Colbeck, J., Holdsworth, E.S. & Wadman, H., 1980. J. Phycol., 16, 290–295. Avron, M., 1981. In, The Biochemistry of Plants, Vol.8, Photosynthesis, edited by M.D.Hatch & N.K.Boardman, Academic Press, New York, pp. 163–191. Badger, M.R. & Andrews, T.J., 1982. Plant Physiol, 70, 517–523. Balch, W.M., Garside, C. & Renger, E.H., 1987. Deep-Sea Res., 34, 221–236. Barlow, R.G. & Alberte, R.S., 1985. Mar. Biol. 86, 63–74. Beardall, J. & Morris, I., 1976. Mar. Biol., 37, 377–387. Beardall, J., Mukerji, D., Glover, H.E. & Morris, I., 1976. J. Phycol, 12, 409–17. Beer, S. & Eshel, A., 1983. J. Exp. Mar. Biol. Ecol., 70, 99–106. Beer, S. & Israel, A., 1986. Plant Physiol., 81, 937–938. Ben-Amotz, A., Gressel, J. & Avron, M., 1987. J. Phycol., 23, 17601–181. Bidwell, R.G. S., 1977. Can. J.Bot., 55, 809–818. Bidwell, R.G.S. & McLachlan, J., 1985. J. Exp. Mar. Biol. Ecol., 86, 15–16. Bienfang, P.K. & Takahashi, M., 1983. Mar. Biol., 76, 213–218. Birmingham, B.C., Coleman, J.R. & Colman, B., 1982. Plant Physiol., 69, 259– 262. Björkman, O., 1981., Encycl. Plant Physiol., 12A, 57–107. Bjørnsen, P.K., 1988. Limnol. Oceanogr., 33, 151–154. Boardman, N.K., 1977. Annu. Rev. Plant Physiol., 28, 355–377. Boynton, W.R., Hall, C.A., Falkowski, P.G., Keefe, C.W. & Kemp, W.M., 1983. Encycl. Plant Physiol., 12D, 305–327. Brechignac, F. & Andre, M., 1984. Plant Physiol., 75, 919–923. Brown, J.S., 1988. J. Phycol., 24, 96–102. Bunt, J.S., 1975. In, Primary Productivity of the Biosphere, edited by H.Lieth & R.H.Whittaker, Springer-Verlag, Berlin, pp. 169–183. Burns, B.D. & Beardall, J., 1987. J. Exp. Mar. Biol. Ecol., 107, 75–86. Burris, J.E., 1977. Mar. Biol., 39, 371–379. Bums, J.E., 1980. In, Primary Productivity in the Sea, edited by P.G.Falkowski, Plenum Press, New York, pp. 411–432. Caldwell, M.M., 1981. Encycl. Plant Physiol., 12A, 169–197. Cassie, R.M., 1963. Limnol. Oceanogr., 8, 433–439. Chow, W.S. & Andersen, J.M., 1987a. Aust. J. Plant Physiol., 14, 1–8. Chow, W.S. & Anderson, J.M., 1987b. Aust. J. Plant Physiol., 14, 9–19. Coats, D.W. & Harding, L.W., 1988. J. Phycol., 24, 67–77. Colman, B., 1984. Plant Cell Environ., 7, 619–621. Colman, B. & Gehl, K.A., 1983. J.Phycol., 19, 216–219. Cook, C.M. & Colman, B., 1987. Plant Cell Environ., 10, 275–278. Cook, C.M., Lanaras, T. & Colman, B., 1986. J. Exp. Bot., 37, 977–984. Cullen, J.J., Zhu, M. & Pierson, D.C., 1986. Limnol. Oceanogr., 31, 1364–1373. Delaney, M.E. & Walker, D.A., 1978. Biochem. J., 171, 477–482. Demeter, S., Neale, P.J. & Melis, A., 1987. FEBS Lett., 214, 370–374. Demmig, B., Winter, K., Krüger, A. & Czygan, F.-C., 1987. Plant Physiol., 84, 218– 224 . Descolas-Gros, C. & Fontugne, M.R., 1985. Mar. Biol., 87, 1–6. Dietz, K.-J., 1986. Planta, 167, 260–263. Dietz, K.-J., Neimanis, S. & Heber, U., 1984. Biochim. Biophys. Acta., 767, 444–450. Dixon, G.K. & Merrett, M.J., 1988. New Phytol., 109, 47–51. Döhler, G., 1984. Z. Naturforsch., 39C, 634–638.

24

GRAHAME J.KELLY

Dring, M.J., 1982. The Biology of Marine Plants. Edward Arnold Ltd., London, 199 pp. Dring, M.J. & Jewson, D.H., 1982. Proc. R. Soc. London Ser. B., 214, 351–368. Dubinsky, Z., Falkowski, P.G. & Wyman, K., 1986. Plant Cell Physiol., 27, 1335– 1349. Eppley, R.W., 1972. Fish. Bull., 70, 1063–1085. Evans, E.L. & Alien, M.M., 1973. J. Bacteriol., 113, 403–08. Evans, J.R., 1986. Planta, 167, 351–358. Evans, J.R., 1987. Aust. J. Plant Physiol., 14, 157–170. Everest, S.A., Hipkin, C.R. & Syrett, P.J., 1986. Mar. Biol., 90, 165–172. Falkowski, P. & Kiefer, D.A., 1985. J. Plankt. Res., 7, 715–731. Falkowski, P.G., 1981. J. Plankt. Res., 3, 203–216. Falkowski, P.G., Kolber, Z. & Fujita, Y., 1988. Biochim. Biophys. Acta, 933, 432– 443. Falkowski, P.G. & Owens, T.G., 1980. Plant Physiol., 66, 592–595. Fleischhacker, P. & Senger, H., 1978. Physiol. Plant., 43, 43–51. Fogg, G.E., 1983. Bot. Mar., 26, 3–14. Forbes, J.R., Denman, K.L. & Mackas, D.L., 1986. Mar. Ecol. Prog. Ser., 32, 181– 191. Fu, C.F. & Gibbs, M., 1987. Plant Physiol., 83, 849–855. Furnas, M.J., 1983. J. Plankt. Res., 5, 637–655. Furnas, M.J. & Mitchell, A.W., 1987. Cont. Shelf Res., 7, 1049–1062. Furuya, K., Hasumoto, H., Nakai, T. & Nemoto, T., 1986. Deep-Sea Res., 33A, 621– 630. Geider, R.J., 1987. New Phytol., 106, 1–34. Geider, R.J. & Osborne, B.A., 1986. Mar. Biol., 93, 351–360. Geider, R.J., Osborne, B.A. & Raven, J.A., 1985. J. Phycol., 21, 609–619. Gerard, V.A., 1986. Mar. Biol., 90, 473–82. Gieskes, W.W. & Kraay, G.W., 1984. In, Flows of Energy and Materials in Marine Ecosystems, edited by J.R.Fasham, Plenum Press, New York, pp. 171–190. Glibert, P.M., Dennett, M.R. & Goldman, J.C., 1985. J. Exp. Mar. Biol. Ecol., 85, 21–36. Glover, H.E., 1980. J. Plankt. Res., 2, 69–79. Glover, H.E., Keller, M.D. & Gullard, R.R.L., 1986. Nature (London), 319, 142– 143. Glover, H.E. & Morris, I., 1979. Limnol. Oceanogr., 24, 510–519. Glover, H. E. & Morris, I., 1981. Arch. Microbiol., 129, 42–46. Goldman, J.C., 1986. Limnol. Oceanogr., 31, 1358–1363. Goldman, J.C., McCarthy, J.J. & Peavey, D.G., 1979. Nature (London), 279, 210– 215. Gounaris, K., Barber, J. & Harwood, J.L., 1986. Biochem. J., 237, 313–326. Harris, G.P., 1978. Ergeb. Limnol., 10, 1–171. Harris, G.P. & Piccinin, B.B., 1983. J. Phycol., 19, 185–192. Heber, U., Neimanis, S. & Dietz, K.-J., 1988. Planta, 173, 267–274. Hiller, R.G. & Goodchild, D.J., 1981. In, The Biochemistry of Plants, Vol. 8, Photosynthesis, edited by M.D.Hatch & N.K.Boardman, Academic Press, New York, pp.1–49. Hiller, R.G., Larkum, A.W.D. & Wrench, P.M., 1988. Biochim. Biophys. Acta, 932, 223–231. Hitchcock, G.L., 1982. J. Plankt. Res., 4, 363–377. Hobson, L.A. & Guest, K.P., 1983. Mar. Biol., 74, 1–7. Hobson, L.A., Morris, W.J. & Guest, K.P., 1985. Plant Physio!., 79, 833–837. Holdsworth, E.S., 1985. Mar. Biol., 86, 253–262. Holdsworth, E.S. & Bruck, K., 1977. Arch. Biochem. Biophys., 182, 87–94. Holdsworth, E S. & Colbeck, J., 1976. Mar. Biol., 38, 189–199. Holm-Hansen, O. & Riemann, B., 1978. Oikos, 30, 438–47. Husic, D.W. & Tolbert, N.E., 1987. Proc. Nat. Acad Sci: U.S.A., 84, 1555–1559. Iturriaga, R. & Mitchell, B.G., 1986. Mar. Ecol. Prog. Ser., 28, 291–297. Jeffrey, S.W. & Hallegraeff, G.M., 1987. Mar. Ecol. Prog. Ser., 35, 293–304.

PHOTOSYNTHESIS: A BIOCHEMICAL PERSPECTIVE

25

Jenkins, W.J., 1982. Nature (London), 300, 246–248. Jensen, A., 1984. In, Marine Phytoplankton and Productivity, edited by O.Holm-Hansen et al., Springer-Verlag, Berlin, pp. 61–72. Johnston, A.M. & Raven, J.A., 1987. Phycologia, 26, 159–166. Joint, I.R. & Pomroy, A.J., 1986. Mar. Biol., 92, 465–474. Joiris, C., 1985. Bull. Mar. Sci., 37, 767–768. Jokiel, P.L. & York, R.H., 1984. Limnol. Oceanogr., 29, 192–199. Jørgensen, E.G., 1970. Physiol. Plant., 23, 11–17. Kana, T. M. & Glibert, P.M., 1987a. Deep-Sea Res., 34A, 497–516. Kana, T. M. & Glibert, P.M., 1987b. Deep-Sea Res., 34A, 479–95. Karekar, M. D. & Joshi, G.V., 1973. Bot. Mar., 16, 216–220. Keller, A. A., 1987. Mar. Biol., 96, 107–114. Kelly, G. J., 1983. Trends Biochem. Sci., 8, 38 only. Kelly, G. J., Latzko, E. & Gibbs, M., 1976. Annu. Rev. Plant Physiol., 27, 181–205. Kenner, R. A. & Ahmed, S. L, 1975. Mar. Biol., 33, 129–133. Kerby, N. W. & Evans, L.V., 1983. J. Phycol., 19, 1–3. King, F. D., 1986. Deep-Sea Res., 33A, 733–754. Kirilovsky, D., Vernotte, C., Astier, C. & Étienne, A.-L., 1988. Biochim. Biophys. Acta, 933, 124–131. Kirk, J.T.O., 1983. Light and Photosynthesis in Aquatic Ecosystems. Cambridge University Press, Cambridge, 401 pp. Kishino, M., Okami, N., Takahashi, M. & Ichimura, S., 1986. Limnol. Oceanogr., 31, 557–566. Kobza, J. & Edwards, G.E., 1987. Plant Cell Physiol., 28, 1141–1152. Kremer, B.P., 1981a. Oceanogr. Mar. Biol. Annu. Rev., 19, 41–94. Kremer, B.P., 198Ib. Phycologia, 20, 242–250. Kremer, B.P. & Berks, R., 1978. Z. Pflanzenphysiol., 87, 149–165. Krömer, S., Stitt, M. & Heldt, H.W., 1988. FEES Lett., 226, 352–356. Lambers, H., 1982. Physiol. Plant., 55, 478–85. Langdon, C., 1987. J. Plankt. Res., 9, 459–482. Laws, E.A., DiTullio, G.R. & Redalje, D.G., 1987. Limnol. Oceanogr., 32, 905– 918. Laws, E.A., Redalje, D.G., Haas, L.W., Bienfang, P.K., Eppley, R.W., Harrison, W.G., Karl, D.M. & Marra, J., 1984. Limnol. Oceanogr., 29, 1161–1169. Leftley, J.W., Bonin, D.J. & Maestrini, S.Y., 1983. Oceanogr. Mar. Biol. Annu. Rev., 21, 23–66. Legendre, L., Demers, S., Garside, C., Haugen, E.M., Phinney, D.A., Shapiro,L. P., Therriault, J.-C. & Yentsch, C.M., 1988. J. Plankt. Res., 10, 1–6. Legendre, L., Demers, S., Yentsch, C.M. & Yentsch, C.S., 1983. Limnol. Oceanogr., 28, 996–1003. Li, W.K.W., 1985. Deep-Sea Res., 32, 1381–1391. Li, W.K.W. & Goldman, J.C., 1981. Microbiol. Ecol., 7, 113–121. Lidholm, J., Gustafsson, P. & Öquist, G., 1987. Plant Cell Physiol., 28, 1133–1140. Lieth, H., 1973. Hum. Ecol., 1, 303–332. Lilley, R.McC. & Walker, D.A., 1975. Plant Physiol., 55, 1087–1092. Lloyd, N.D.H ., Canvin, D. T. & Culver, D. A., 1977. Plant Physiol., 59, 936–940. Lohrenz, S.E., Wiesenburg, D. A., DePalma, I. P., Johnson, K. S. & Gustafson, D. E., 1988. Deep-Sea Res., 35A, 793–810. Lorenzen, C. J., 1976. In, The Ecology of the Seas, edited by D.H.Cushing & J.J. Walsh, Blackwell Scientific Publications, Oxford, pp. 173–185. Lucas, W. J. & Berry, J. A., 1985. Physiol. Plant., 65, 539–543. Makino, A., Mae, T. & Ohira, K., 1984. Plant Cell Physiol., 25, 511–521. Malone, T.C., 1980. In, Primary Productivity in the Sea, edited by P.G.Falkowski, Plenum Press, New York, pp. 301–319. Martin, J.H. & Fitzwater, S.E., 1988. Nature (London), 331, 341–343.

26

GRAHAME J.KELLY

McMeekin, T.A., Chandler, R.E., Doe, P.E., Garland, C.D., Olley, J., Putro, S. & Ratkowsky, D. A., 1987. J. Appl. Bacterial., 62, 543–550. Mills, D.K. & Wilkinson, M., 1986. Bot. Mar., 29, 125–129. Mimuro, M. & Fujita, Y., 1977. Biochim. Biophys. Acta., 459, 376–389. Moal, J., Martin-Jezequel, V., Harris, R.P., Samain, J.-F. & Poulet, S.A., 1987. Oceanol. Acta, 10, 339–346. Morris, I., 1980. In, Primary Productivity in the Sea, edited by P.G.Falkowski, Plenum Press, New York, pp. 139–159. Morris, I. & Glover, H., 1981. Limnol. Oceanogr., 26, 957–961. Mortain-Bertrand, A., Descolas-Gros, C. & Jupin, H., 1987. J. Exp. Mar. Biol. Ecol., 112, 11–26. Mortain-Bertrand, A., Descolas-Gros, C. & Jupin, H., 1988. J. Plankt. Res., 10, 199– 217. Mukerji, D., Glover, H.E. & Morris, I., 1978. J. Phycol., 14, 137–142. Murphy, D.J., 1986. Biochim. Biophys. Acta, 864, 33–94. Myers, J., 1980. In, Primary Productivity in the Sea, edited by P.G.Falkowski, Plenum Press, New York, pp. 1–16. Ohad, I., Koike, H., Shochat, S. & Inoue, Y., 1988. Biochim. Biophys. Acta, 933, 288–298. Osborne, B.A. & Geider, R.J., 1987. Plant Cell Environ., 10, 141–149. Osmond, C.B., 1981. Biochim. Biophys. Acta, 639, 77–98. Palmisano, A.C., SooHoo, J.B., SooHoo, S.L., Kottmeier, S.T., Craft, L.L. & Sullivan, C.W., 1986. J. Plankt. Res., 8, 891–906. Paul, J.S. & Volcani, B.E., 1976. Arch. Microbiol., 110, 247–252. Peltier, G. & Thibault, P., 1985. Plant Physiol., 79, 225–230. Penniman, C.A. & Mathieson, A.C., 1985. Bot. Mar., 28, 427–435. Pirt, S.J., 1986. New Phytol, 102, 3–37. Plante-Cuny, M.-R. & Bodoy, A., 1987. Oceanol. Acta, 10, 223–237. Platt, T., Subba Rao, D.V. & Irwin, B., 1983. Nature (London), 301, 702–704. Popovic, R., Colbow, K., Vidaver, W. & Bruce, D., 1983. Plant Physiol., 73, 889– 892. Powles, S.B., Cornic, G. & Louason, G., 1984. Physiol. Vég., 22, 437–446. Prézelin, B.B., 1981. In, Physiological Bases of Phytoplankton Ecology, edited by T. Platt, Bulletin 210, Department Fisheries & Oceans, Ottawa, pp. 1–3. Prézelin, B.B. & Alberte, R.S., 1978. Proc. Nat. Acad. Sci. U.S.A., 75, 1801–1804. Prézelin, B.B., Glover, H.E. & Campbell, L., 1987. Mar. Biol., 95, 469–480. Prézelin, B.B., Putt, M. & Glover, H.E., 1986. Mar. Biol., 91, 205–217. Priscu, J.C., 1984. J. Plankt. Res., 6, 531–542. Priscu, J.C. & Goldman, C.R., 1983. Limnol. Oceanogr., 28, 1168–1181. Ramachandra Reddy, A. & Das, V.S.R., 1987. Photosynthetica, 21, 341–348. Raven, J.A., 1970. Biol. Rev., 45, 167–221. Richardson, K., Beardall, J. & Raven, J.A., 1983. New Phytol., 93, 157–191. Richardson, K., Samuelsson, G. & Hällgren, J.-E., 1984. J. Exp. Mar. Biol. Ecol., 81, 241–250. Rivkin, R.B. & Putt, M., 1987. Science, 238, 1285–1288. Rivkin, R.B., Seliger, H.H., Swift, E. & Biggley, W.H., 1982. Mar. Biol., 68, 181– 191. Rochet, M., Legendre, L. & Demers, S., 1986. J. Exp. Mar. Biol. Ecol., 101, 211– 226. Round, F.E., 1981. The Ecology of Algae. Cambridge University Press, Cambridge, 653 pp. Ryther, J.H., 1956. Nature (London), 178, 861–863. Ryther, J.H., 1969. Science, 166, 72–76. Sakshaug, E., Demers, S. & Yentsch, C.M., 1987. Mar. Ecol. Prog. Ser., 41, 275– 282. Sakshaug, E. & Holm-Hansen, O., 1986. J. Plankt. Res., 8, 459–473. Samuelsson, G., Lönneborg, A., Rosenqvist, E., Gustafsson, P. & Öuist, G., 1985. Plant Physiol., 79, 992–995. Samuelsson, G. & Prézelin, B.B., 1985. J. Phycol., 21, 453–57. Sandmann, G., 1985. Physiol. Plant., 65, 481–486. Scherer, S., Stürzl, E. & Böger, P., 1984. Physiol. Plant., 60, 557–560. Sheldon, R.W. & Sutcliffe, W.H., 1978. Limnol. Oceanogr., 23, 1051–1055.

PHOTOSYNTHESIS: A BIOCHEMICAL PERSPECTIVE

27

Shulenberger, E. & Reid, J.L., 1981. Deep-Sea Res., 28A, 901–919. Siefermann-Harms, D., 1987. Physiol. Plant, 69, 561–568. Smith, J.C. & Platt, T., 1985. Mar. Ecol. Prog. Ser., 25, 31–37. Smith, J.C., Platt, T. & Harrison, W.G., 1983. Prog. Oceanogr., 12, 425–459. Smith, R.E.H. & Geider, R.J., 1985. J. Exp. Mar. Biol. Ecol., 93, 191–210. Smith, R.E.H. & Platt, T., 1984. Mar. Ecol. Prog. Ser., 16, 75–87. Steemann Nielsen, E., 1952. J. Con., Cons. Int. Explor. Mer, 18, 117–140. Steemann Nielsen, E. & Hansen, V.K., 1961. Physiol. Plant., 14, 595–613. Stitt, M., 1986. Plant Physiol, 81, 1115–1122. Stockner, J.G. & Antia, N.J., 1986. Can. J. Fish. Aquat. Sci., 43, 2472–2503. Strickland, J.D.H., Eppley, R.W. & Rojas de Mendiola, B., 1969. Bol. Inst. Mar Peru, 2, 1–45. Strickland, J.D.H. & Parsons, T.R., 1972. Bull. Fish. Res. Board Can., No. 167 (2nd edition), see pp. 267–278. Sukenik, A., Bennett, J. & Falkowski, P., 1987. Biochim. Biophys. Acta, 891, 205– 215. Suzuki, R. & Fujita, Y., 1986. Mar. Ecol. Prog. Ser., 28, 81–85. Taguchi, S., 1981. J. Exp. Mar. Biol. Ecol., 55, 115–131. Terashima, I. & Evans, J.R., 1988. Plant Cell Physiol., 29, 143–155. Thinh, L.-V. & Griffiths, D.J., 1985. Bot. Mar., 28, Thomas, W.H., Seibert, D.L.R., Alden, M., Neori, A. & Eldridge, P., 1984. Biomass, 5, 181–209. Usuda, H., Ku, M.S.B. & Edwards, G.E., 1984. Aust. J. Plant Physiol, 11, 509– 517. Van Baalen, C., 1985. Plant Physiol., 78, 799–802. Venrick, E.L., Beers, J.R. & Heinbokel, J.F., 1977. J. Exp. Mar. Biol. Ecol, 26, 55–76. Vesk, M. & Jeffrey, S.W., 1977. J. Phycol., 13, 280–288. Vesk, M. & Jeffrey, S.W., 1987. J. Phycol., 23, 322–336. Videau, C., 1987. Mar. Ecol. Prog. Ser., 35, 141–151. Vieira, A.A.H. & Myklestad, S., 1986. J. Plankt. Res., 8, 985–994. Vincent, W.F., Neale, P.J. & Richerson, P.J., 1984. J. Phycol., 20, 201–211. von Caemmerer, S. & Farquhar, G.D., 1981. Planta, 153, 376–387. Walker, D., 1979. Energy, Plants and Man. Packard Publishing Ltd, Funtington, Sussex, 32 pp. Walker, D.A., 1974. In, Plant Carbohydrate Biochemistry, Proc. 10th Symp. Phytochem. Soc., Edinburgh, 1973, edited by J.B.Pridham, Academic Press, London, pp. 7–26. Weger, H.G., Birch, D.G., Elrifi, I.R. & Turpin, D.H., 1988. Plant Physiol., 86, 688–692. Whittaker, R.H. & Likens, G.E., 1975. In, Primary Productivity of the Biosphere, edited by H.Lieth & R.H.Whittaker, Springer- Verlag, Berlin, pp. 305–328. Williams, P.J.le B., Heinemann, K.R., Marra, J. & Purdie, D.A., 1983. Nature (London), 305, 49–50. Wood, L.W., 1985. Can. J.Fish. Aquat. Sci., 42, 38–3. Wood, W.F., 1987. Mar. Biol., 96, 143–150. Yentsch, C.S., 1974. Oceanogr. Mar. Biol. Annu. Rev., 12, 41–75. Yokota, A. & Canvin, D.T., 1985. Plant Physiol., 77, 735–739. Yokota, A., Iwaki, T., Miura, K., Wadano, A. & Kitaoka, S., 1987. Plant Cell Physiol., 28, 1363–1376. Yokota, A. & Kitaoka, S., 1987. Planta, 170, 181–189. Zenvirth, D. & Kaplan, A., 1981. Planta, 152, 8–12.

Oceanogr. Mar. Biol. Annu. Rev., 1989, 27, 45–90 Margaret Barnes, Ed. Aberdeen University Press

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE* IB SVANE Kristineberg Marine Biological Station, Kristineberg 2130, S-450 34 Fiskebäckstil, Sweden and CRAIG M.YOUNG† Harbor Branch Oceanographie Institution, 5600 Old Dixie Highway, Fort Pierce, Florida 34946, USA

ABSTRACT Recent studies of ascidian larval biology reveal a diversity of structure and behaviour not previously recognised. The introduction of direct methods for observing ascidian larvae in situ has provided important insights on larval behaviour, mortality, and dispersal not possible with the microscopic larvae of most other marine invertebrates. In the context of these recent advances, this review considers the ecology of pelagic phases (egg, embryo, and larva) of the ascidian life cycle and relates aspects of reproduction and larval biology to the recruitment, abundance, and distribution of adult populations. INTRODUCTION The ascidian larva functions in site selection and dispersal while transporting adult rudiments between generations (Berrill, 1955, 1957; Millar, 1971; Cloney, 1987). Interest in the ecology of ascidian larvae has recently been stimulated by the introduction of new ideas and methods, most notably in situ observations of living larvae. The resulting advances justify a reconsideration of ascidian larval ecology. From an ecological standpoint, all pelagic phases of the ascidian life cycle (which may include eggs, embryos and larvae) influence adult populations. Thus, our definition of larval ecology is broad; it includes reproductive processes leading to the production of gametes and larvae, as well as post-settlement consequences of larval processes. The literature on larval evolution as it relates to ascidian and chordate phylogeny is large, and will not be considered in the present paper. Reviews on ascidian biology dealing with aspects of reproduction and larval ecology have been published by Millar (1971), Berrill (1975), and Cloney (1987), and detailed reviews on

* Harbor Branch Contribution No. 678.

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

29

Fig 1.—Generalised life cycles of typical solitary ascidians (top) and compound ascidians (bottom). The diagrams incorporate characteristics of many species; not all aspects (e.g., fusion of colonies) are present in the life cycle of any given species.

ascidian metamorphosis have been presented by Cloney (1978, 1982). More recently, ascidian metamorphosis and larval locomotion have been treated as parts of general comprehensive reviews by Burke (1983) and Chia, Buckland-Nicks & Young (1984), respectively. It is our goal to bring together a scattered literature that comprises ascidian larval ecology. While emphasising ecological aspects of larval release, development, dispersal, behaviour, and settlement in relation to the distribution and abundance of adult populations, we also deal with reproductive, morphological, and developmental problems that help to elucidate aspects of ascidian larval ecology. REPRODUCTION, FERTILISATION, AND LARVAL ANATOMY MODES OF REPRODUCTION Ascidians may be either oviparous (typical of most solitary ascidians), ovoviviparous (typical of most compound ascidians), or viviparous (Fig 1). True viviparity in which nutrients are exchanged between parent and embryo across a placenta-like structure is known to occur in only a single species, Hypsistozoa fasmeriana (Brewin, 1959). Much of the earlier literature, including the classic works of Berrill (e.g. 1929, 1935, 1950) uses the term in reference to ovoviviparous species that retain embryos to hatching, but without apparent transfer of nutrients after ovulation. In oviparous forms, fertilisation and all subsequent developmental stages occur externally. Ascidian embryology and developmental biology have been the subjects of numerous important studies which have been reviewed by Cloney (1987) and Berrill (1975), and provided the theme for a recent symposium (Lambert, 1982). Conklin’s (1905) study is still the classic description of cell lineage, although there has been a late resurgence of cell lineage studies using modern fluorescent markers and immunological techniques (Whittaker, 1987). The most common developmental pattern (urodele development) terminates in the production of a swimming tadpole larva. Many molgulids and a few species in the family Styelidae (Pelonia corrugata and Polycarpa tinctor) undergo direct development, bypassing the larval stage completely (reviewed by Berrill, 1931; Millar, 1971; Young et al., 1988). In ascidians, such direct development is known as anural development (Fig 2). As in many other kinds of animals, larval size is correlated with the degree of parental protection. Compound ascidians releasing fully developed larvae generally produce relatively few large larvae, whereas the eggs released by oviparous solitary ascidians are relatively more numerous and result in the production of much smaller, less differentiated tadpoles. In many invertebrates, there appears to be a correlation between adult size and brooding (Menge, 1975; Strathmann & Strathmann, 1982); within a given taxon, species with smaller individuals are more likely to brood than those with larger individuals. If we define an individual as a single zooid, not a colony, the relationship also seems to hold in ascidians. Accordingly, most colonial ascidians, having small zooids, brood relatively large complex larvae which develop from large eggs. Among solitary ascidians, brooders tend to have small bodies and produce relatively large eggs and larvae, whereas oviparous species often

† Order of authorship is alphabetical.

30

IB SVANE AND CRAIG M.YOUNG

Fig 2.—Generalised life cycle of a molgulid ascidian with anural development.

have relatively larger bodies and produce small eggs. Exceptions to this pattern are found in Boltenia echinata (Berrill, 1948b; Svane, 1983) and Corella inflata (Child, 1927; Lambert, Lambert & Abbott, 1981), both of which brood small eggs and produce small, simple tadpole larvae. At least one solitary ascidian, Polycarpa pomaria, is observed in the laboratory to brood facultatively. It is normally oviparous and produces small eggs, but developing larvae have not been found in the atrial cavity in nature (Berrill, 1950). One explanation for the relationship between brooding and zooid size is that cost of reproduction prohibits formation of numerous eggs in small individuals (Vance, 1973; Chia, 1974; Menge, 1976; Heath, 1977; Strathmann & Strathmann, 1982). This explanation, however, begs the question of why very large colonies of genetically identical zooids invest their resources for sexual reproduction in few large tadpoles rather than numerous small ones. Strathmann, Strathmann & Emson (1984) have suggested that selffertilisation can lead to brooding in echinoderms. The same argument could be applied to ascidians if we knew the extent of self-fertilisation. Botryllus schlosseri, which brood their embryos in the atrial cavity, can be either semelparous or iteroparous (R.Grosberg, pers. comm.). When fully developed larvae are removed surgically from the atrial cavities of a semelparous colony, the brood is replaced, suggesting that reproduction is not constrained by energy limitations. When a brood is replaced with inert glass beads, the colony dies shortly thereafter, just as in a typical semelparous colony (R.Grosberg, pers. comm.). Based on these findings, Grosberg has suggested that large broods prevent water flow through the atrium, killing the parent colony before a second brood is formed. FERTILISATION Sperm morphology and the physiology of ascidian fertilisation have been reviewed by Lambert (1982). Ascidian sperm are among the simplest in the animal kingdom in that they lack a midpiece. The excentric mitochondrion lies next to the nucleus in the head region. Paradoxically, this simple sperm must penetrate a complex egg covering consisting of two layers of somatic cells and a chorion. Ascidian eggs have an effective block to polyspermy that appears to involve the egg membrane rather than the accessory cells or chorion (Lambert & Lambert, 1981). Moreover, many species demonstrate complete or partial blocks to self-fertilisation (Morgan, 1938, 1942). Stolidobranchs tend to have a better block to self-fertility than phlebobranchs (Cloney, 1987). Many ascidiids and corellids are completely self-fertile. Several species in the genus Ascidia demonstrate a partial block to self-fertility when first spawned, but this block wears off during the first few hours, ultimately permitting virtually 100% self-fertilisation (Scofield, Schlumpberger, West & Weismann, 1982 for A. ceratodes; Young, unpubl. data on A. callosa and A. mentula). Although self-sterility clearly implies the absence of self-fertilisation in nature, the converse (absence of a block to self-fertility) does not necessarily indicate that eggs are routinely self-fertilised in the field. Indeed, the extent to which self-fertilisation occurs in nature is unknown. Ovulation and subsequent development have been observed in isolated zooids of Ecteinascidia turbinata, suggesting that selffertilisation is likely (Lambert, 1982). Some species release sperm seconds to minutes before eggs are spawned, providing an opportunity for sperm to disperse and mix with that of other individuals before fertilisation (Lambert & Brandt, 1967; Lambert et al., 1981). Recent electrophoretic evidence for two selffertile species, Chelyosoma production and Corella inflata, indicates that populations are not inbred; both populations have high levels of heterozygosity (S.Cohen, pers. comm.). Although the observed

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

31

Fig 3.—Representative tadpole larvae of compound ascidians, including Diplosoma macdonaldi (A), Botryllus planus (B), Distaplia occidentalis (C), Ecteinascidia turbinata (D) and Eudistoma capsulatum (E). Only the styelid Botryllus has a modified sensory vesicle. BB: branchial basket. BS: oral siphon. AS: atrial siphon. Other labels as in Fig 4. Fig 4.—Representative tadpole larvae of solitary ascidians. Boltenia villosa (A), Corella inflata (B), and Halocynthia igaboja (C) have both a statocyte and an ocellus in the sensory vesicle. Styela coriacea (D) has a compound sensory organ known as a photolith, whereas Molgula occidentalis (E) has a statocyte but no ocellus. N: notochord; O: ocellus; S: statocyte; P: adhesive papilla ; PL: photolith; A: ampulla; TF: tail fin; F: tail tunic.

heterozygosity is not given, all alleles of Ciona intestinalis reported by Schmidtke, Weiler, Kunz & Engel (1977) were highly polymorphic. Lyerla & Lyerla (1978), however, found that the colonial species Clavelina picta and C. oblonga were monomorphic and may experience high levels of inbreeding. Whether this is due to obligate self-fertilisation or almost total asexual reproduction, or a combination of both, is unknown (Lyerla & Lyerla, 1978). Ascidian sperm demonstrate chemotactic attraction to eggs under laboratory conditions (Miller, 1975, 1982). Although there is some species specificity in these interactions, many stolidobranchs demonstrate low levels of specificity (Miller, 1982). LARVAL ANATOMY Ever since the chordate nature of the ascidian tadpole larva was recognised by Kowalevsky (1866), phylogenetic questions have motivated detailed studies of tadpole structure and development. Because details of tadpole morphology have been reviewed in depth several times during the past two decades (Millar, 1971; Berrill, 1975; Katz, 1983; Cloney, 1978, 1987), we shall deal only with structures that control larval habitat selection and swimming behaviour that are directly relevant to the ecology of larvae themselves. Those portions of the larva that have been called the “adult action system” (developing and fully developed adult structures present in the larva) will not be reviewed here. The overall sizes of ascidian larvae range up to 4.5 mm, with the largest being found among colonial ascidians in the order Aplousobranchiata: Hypsistozoa fasmeriana (Brewin, 1959), Polycitor circes (Millar, 1963), and Eudistoma digitatum (Millar, 1964). The largest tadpoles are also among the most complex, as they contain well-developed rudiments of adult structures (Fig 3). By contrast, the simple larvae typically produced by oviparous species are generally about 1 mm long (see Fig 4). Cloney (1978) has calculated that the large complex larva of Hypsistozoa fasmeriana is approximately 3500 times greater in volume than Molgula occidentalis which has a simple tadpole larva. The typical tadpole of a solitary ascidian consists of a trunk (also called the “body” or “head” in the earlier literature), which is 150–250 µm long and 100–125 µm wide, and a tail about 750 µm in length, though this may vary. A detailed review of larval morphology in a simple tadpole is given for ciona intestinalis by Katz (1983). Katz, following earlier workers (Grave, 1944; Scott, 1946) classified the larval tissues into two categories. The six organ systems that comprise the larval action system are the tunic, epidermis, notochord, tail musculature, adhesive organs and nervous system. Four additional organ systems carried by the larva will ultimately become organs of the adult stage, but without immediate ecological relevance during the larval period (Grave, 1944; Scott, 1946). The simple type of larva is found in the phlebobranchiate familes Cionidae, Ascidiidae, Corellidae, and Diazonidae, but also has been retained in the phylogenetically more derived stolidobranchiate families Pyuridae, Styelidae, and Molgulidae. In colonial ascidians (Aplousobranchiata and the stolidobranchiate subfamily Botryllinae), a more elaborate larva is found wherein oral papillae occur in great variety, and numerous adult structures, non-functional in

32

IB SVANE AND CRAIG M.YOUNG

the larvae, may be present (Barnes, 1971; Millar, 1971; Cloney, 1977). Larvae of some colonial ascidians may even have a functional heart (Cloney, 1982). Although the general complement of sensory and locomotory structures is similar in the larvae of solitary and compound ascidians, many of the subtle differences could have ecological or behavioural significance. The surface of an ascidian larva is covered with a secreted, non-cellular tunic layer. This transparent tunic, also called the test or cuticle, is elaborated into two fins which, in tadpoles of solitary ascidians, are situated dorsally and ventrally on the trunk and tail. The fins are formed by extra-embryonic test cells present between the egg chorion and the embryo (Cloney & Cavey, 1982; Robinson, Kusten & Cloney, 1986). At hatching, the test cells often persist on the outside of the tunic fin for a period of time (Grave, 1921; Cloney & Cavey, 1982; Katz, 1983), although it is not known whether their persistence into the posthatching stage has any ecological significance. In any case, the test cells generally fall off before metamorphosis. Tadpole larvae of most solitary ascidians are laterally flattened for their entire length. In addition to the large dorsal and ventral fins, other smaller fins may be arranged in various positions on the lateral sides of the tail (Katz, 1983). Some colonial ascidians, but not all (see Grave & Woodbridge, 1924), have tails which are rotated through 90 degrees by a quarter twist at the base of the tail (Damas, 1904; Grave, 1921; Grave & McCosh, 1924). Berrill (1950) hypothesised that the twist results from limited space for development in the chorion. Although this appears to be an adequate proximate explanation, there has been no speculation as to why chorion size should be constrained in larger embryos. Perhaps it is limited in small zooids by the space available for brooding. If this is the case, one might expect larger chorionic spaces (and perhaps untwisted tails) in the embryos of atrial brooders than oviducal brooders. One consequence of the twisted tail is a more erratic, spiral swimming pattern. All ascidian larvae move by flexing muscles in the tail. These muscles are arranged in three longitudinal bands situated laterally (or, in the case of compound ascidians with twisted tails, dorsally and ventrally) and bounding the notochord. The notochord consists of about 40 squamous, matrix-filled cells surrounded by a fibrous sheath (Cloney, 1964; Mancuso & Dolcemascolo, 1977). The tail is innervated not only by the neuromuscular junctions that stimulate the muscle blocks (reviewed by Katz, 1983), but also by axons connected to pairs of ciliary sensory cells lying along the midline of the tail epidermis (Torrence & Cloney, 1982). It is not known whether these are proprioceptors that signal tail posture or if the receptors provide the larva with sensory input concerning some aspect of the outside environment. Most tadpoles respond quickly to tactile stimulation of the tail tunic (R.A.Cloney, pers. comm.). The anterior end of the trunk is equipped with three adhesive papillae. Cloney (1978) has identified at least nine different types of papillae, which he has classified into three major groups: non-everting glandular papillae, everting glandular papillae, and non-glandular papillae. The papilla structure may play a role in determining how well larvae select substrata. For example, some tadpoles with simple coniform papillae secrete adhesive even before contacting the substratum (Cloney, 1978). Although one might expect that this would reduce their capability for habitat selection, several species with this sort of papillae are known to discriminate among substrata at settlement (Young & Braithwaite, 1980a; Young, 1982). Also counter-intuitive is the observation that some tadpoles which can expose their adhesive abruptly by everting the papillae (reviewed by Cloney, 1978) seem to be quite indiscriminate with respect to attachment surface. The larvae of Botryllus schlosseri and Symplegma viride have a sucker-like holdfast mechanism, under nervous control, which reportedly allows the larva to attach and release its hold several times before selecting a final site for metamorphosis (Grave, 1934). At the opposite extreme, larvae of some urodele molgulids (e.g., Molgula citrina, M. manhattensis, M. occidentalis) either lack papillae (Cloney, 1978) or have tiny oral papillae that do not function in attachment. In these species, either the entire larval body becomes adhesive (Kingsley, 1882; Grave, 1944) or attachment is by means of a large primary ampulla near

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

33

the anterior end. The larvae of M. occidentalis, which fall into this latter category, are capable of discriminating among substratum types (Young, in press), presumably by delaying the release of ampullar adhesive until after encountering an appropriate habitat. It is not surprising, therefore, that Torrence & Cloney (1983, 1988) have reported sensory neurons in the anterior epidermis of this species. Attachment in some anural species occurs shortly after the egg contacts sea water (Young et al., 1988). Follicle cells surrounding the egg chorion of M. pacifica undergo a holocrine secretion of adhesive shortly after spawning. The embryo develops to hatching on the first substratum contacted by the egg, after which the juvenile moves away a short distance (Young et al., 1988). The adhesive papillae of most tadpoles are probably sensory as well as adhesive in function. Torrence & Cloney (1983,1988) have demonstrated the presence of sensory neurons in the oral papillae of Diplosoma macdonaldi and many other ascidians in several families. Similar neurons, as well as an associated papillar ganglion, have been reported at the light microscope level for the larvae of Botryllus schlosseri (Grave, 1934; Grave & Riley, 1935). Although the ultrastructural studies provide few cues as to the functions of the papillar receptors, behavioural evidence suggests that they may function both as mechanoreceptors and chemoreceptors (Cloney, 1978; Young & Braithwaite, 1980a; Young, 1982; Davis, 1987; Torrence & Cloney, 1988). The cerebral vesicle of an ascidian larva generally contains at least two sensory structures that control swimming behaviour and orientation: a multicellular ocellus for the reception of light, and a statocyte containing a statolith for the reception of gravity (Fig 4). A third type of possible sensory structure has been found in both phlebobranchs and stolidobranchs (Dilly, 1969; Eakin & Kuda, 1971; Reverberi, 1979; Svane, 1982). This structure is located in the dorso-posterior wall of the cerebral vesicle, imbedded in the tissue lateral to the cerebral vesicle, or confined to an auxiliary vesicle which communicates with the cerebral vesicle at the level of the statocyte (Svane, 1982; Vorontosova & Malakhov, 1984; Svane & Young, unpubl.). The cells carry 2 µm globular ciliary structures and resemble coronet cells (although more simple) typical for the saccus vasculosus found in elasmobranchs and many ganoid and teleost fishes. The globular cilia possess a modified (9+0) arrangement with tubules that open into the lumen of the cerebral vesicle (Eakin & Kuda, 1971). These structures have been variously interpreted as a second ocellus (Dilly, 1969; Torrence & Cloney, 1988) or hydrostatic pressure receptors (Eakin & Kuda, 1971). The only species of tadpole larva that has, however, been tested for a pressure sensitivity demonstrated no such response (Crisp & Ghobashy, 1971). Alternatively, if the ascidian “coronet cells” are homologous to those of fishes, they could have a secretory function as suggested by Lanzing & Lennep (1971) and Emanuelsson & Mecklenburg (1974) or they could function as a receptor for low molecular weight materials (Jansen & Flight, 1969). The actual function or functions, however, remain unknown. The morphology and ultrastructure of the ocellus and statocyte have been studied in detail (e.g. Dilly, 1961, 1962, 1964; Barnes, 1971; Eakin & Kuda, 1971; Katz, 1983; Vorontosova & Malakhov, 1984). Structural details of these organs vary substantially among species, and these differences have often been assumed to produce behavioural differences (Berrill, 1975). A typical ocellus, as exemplified by Ciona intestinalis, consists of three lens cells, a number of neurosensory retinal cells and a unicellular pigmented cup (Dilly, 1964; Eakin & Kuda, 1971; Turon, 1988). Similar ocelli are found in colonial aplousobranchs (Grave, 1921; Barnes, 1971, 1974). The larvae of most styelids have either simplified ocelli or compound sensory structures that incorporate elements of both the ocellus and the statocyte into a single structure. For example, in larvae of Styela partita, the ocellus is reduced to a small melanic granule surrounded by a thin layer of optically clear material (Grave, 1944; Whittaker, 1966), whereas colonial styelids such as Botryllus have the statocyte and ocellus combined into a “photolith” (Garstang & Garstang, 1928; Grave, 1934; Berrill, 1949; Torrence & Cloney, 1988). Larvae of the solitary styelid Cnemidocarpa finmarkiensis are

34

IB SVANE AND CRAIG M.YOUNG

reported to possess both an ocellus and a photolith (Vorontosova & Malakhov, 1984). The styelid genera Dendrodoa and Polycarpa (Millar, 1971), as well as the pyurid Herdmania momus (Svane & Young, unpubl.), and most urodele molgulids (Grave, 1926; Berrill, 1931) have statocytes, but no pigmented ocellus. The behavioural and ecological implications of these structural variations remain largely undocumented. TIMING AND SYNCHRONY OF REPRODUCTION SEASONAL PATTERNS OF REPRODUCTION Successful reproduction in cross-fertilising organisms depends on synchrony among members of the population. Although a few species, including both tropical (Goodbody, 1961, 1963) and temperature (Lambert, 1968; Berrill, 1957; Svane & Lundälv, 1981) ones, reproduce throughout the year, most ascidians reproduce in one or two discrete peaks during the annual cycle (reviewed by Millar, 1971; Berrill, 1975). Sexual and asexual portions of the reproductive cycle occur simultaneously in some compound ascidians and alternate in regular cycles in others (Millar, 1971). Previous workers have assumed that because reproductive cycles are correlated with temperature, temperature is the causative factor that entrains the cycles. Other factors that covary with temperature on an annual basis, including day length, phytoplankton concentrations, lunar periods, and tide levels, have not, however, been investigated. The important work of Pearse and his colleagues (Pearse & Eernisse, 1982; Pearse & Beauchamp, 1986; Pearse & Walker, 1986; and others) on the role of light in entraining circannual reproductive cycles in echinoderms should inspire similar studies in ascidians, particularly because many solitary ascidians are easily reared in laboratory seawater systems. DIEL PATTERNS OF LARVAL RELEASE AND SPAWNING In colonial ascidians, fully developed larvae may be incubated in the oviduct, in the atrial chamber of a zooid, in a specialised brood pouch (e.g. Distaplia spp.), or in the common cloacal chambers of the colony. Larval release may involve expulsion of the larva from an existing opening or directly through the tunic of the colony. The timing of larval release is controlled by light in every species that has been experimentally investigated (Table I). In general, spawning occurs in response to light following a period of dark adaptation. The duration of light exposure required, which has been called the dormant period (Huus, 1939) or the latency period (Lambert, Lambert & Abbott, 1981), varies somewhat among species. Olson (1983) tabulated the latency period for nine species of compound ascidians. We have expanded his table to include many additional species, including solitary ascidians and several colonial species not reported in the primary literature (Table I). Of 27 species for which observations or data are available, 16 release larvae in the early morning or within a short time after exposure to light in the laboratory. Five species release larvae all day (one of which continues to produce larvae throughout the night; Grave, 1936), five release between the late morning and early afternoon hours, and one releases in the afternoon (Table I). With a single exception, Podoclavella moluccensis (A.Davis, pers. comm.), the species releasing larvae at midday are didemnids which contain symbiotic algae, and which transfer the symbionts between generations during the larval stage (Duyl, Bak & Sybesma, 1981; Olson, 1983). The actual mechanism of larval release remains poorly understood for most colonial ascidians. Metandrocarpa taylori has been observed to eject larvae forcefully by means of a general contraction of body wall musculature (Abbott, 1955). Euherdmania claviformis and Pycnoclavella stanleyi move larvae through the atrium by a series of contractions in the thoracic region according to Trason (1957, 1963).

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

35

Larvae of Eudistoma ritteri have been observed to swim actively from the atrial chamber of the adult zooid (Levine, 1962). We do not, however, know whether release is under control of the adult nervous system or if, alternatively, larvae leave spontaneously following stimulation of their photoreceptors. In Distaplia occidentalis, adult control TABLE I Spawning or larval release in relation to light-dark cycles Species Polyclinidae Aplidium (Amaroucium) constallatum A. stellatum A. antillense Polycitoridae Distaplia occidentalis D. corolla Polycitor mutabilis Cystodytes lobatus Eudistoma hepaticum E. olivaceum E. capsulatum Clavelina oblonga Didemnidae Diplosoma listerianum D. similis Didemnum molle Leptoclinum mitsukurii Trididemnum solidum Lissoclinum voeltzkowi L. patella Perophoridae Perophora viridis Ecteinascidia turbinata Cionidae Ciona intestinalis

Corellidae Corella paralellogramma C. willmeriana

Average spawning latency Spawning time Settlement time References 20 min

morning

morning

Scott, 1954; Mast, 1921

2–3 h

morning morning

morning morning

Gotelli, 1987 Bingham, pers. comm.

15 min

morning

morning

late morning continuous all day morning morning morning all day

midday continuous all day morning morning morning all day

Watanabe & Lambert, 1973 Bingham, pers. comm. Oka, 1943 Lambert, 1979 Bingham, pers. comm. Bingham, pers. comm. Bingham, pers. comm. Bingham, pers. comm.

morning midday 1100–1400 morning midday midday 1100–1800

morning midday midday morning midday midday afternoon

Crisp & Ghobashy, 1971 Olson, 1983 Olson, 1983 Yamaguchi, 1975 Duyl et al., 1981 Olson, 1983 Olson & McPherson, 1987

morning morning

morning morning

Grave & McCosh, 1924 Young, unpubl.

3–4 h

min

0–3 h

27 min

morning

morning

Lambert & Brandt, 1967; Conklin, 1905 ; Whittingham, 1967

11–17 min

morning

morning

Huus, 1939, 1941

15.1 min (sperm)

morning

Lambert et al., 1981

36

IB SVANE AND CRAIG M.YOUNG

C. inflata Chelyosoma productum Styelidae Styela partita S. plicata (Japan) S. plicata (Calif.) Botryllus schlosseri

12.6 min (sperm)

20.8 min (eggs) Lambert et al., 1981 18.9 min (eggs) Young & Braithwaite, 1980a

morning morning

11–12 h

afternoon

10–12 h

evening afternoon all day

all day morning morning morning morning

B. planus Botrylloides mutabilis B. nigrum Metandrocarpa taylori

15 min

morning morning morning morning

Polyandrocarpa tincta P. misakiensis

morning afternoon (16.30)

morning

11.5h

24 h

continuous

Symplegma viride Pyuridae Halocynthia roretzi

11–12pm twice a day (2 types)

Molgulidae Molgula manhattensis

M. citrina M. pacifica

few min

Conklin, 1905; Rose, 1939 Yamaguchi, 1975 West & Lambert, 1976 Grave, 1934; Grave & Woodbridge, 1924 Bingham, pers. comm. Yamaguchi, 1975 Morgan, 1977 Watanabe & Lambert, 1973 Grave, 1936 Hashimoto & Watanabe, 1982 Grave, 1936 Hirai & Tsubata, 1956

Watanabe & Tokioka, 1972 morning

all day night

Berrill, 1931; Whittingham, 1967; Conklin, 1905 Grave, 1926 Young et al., 1988

seems unlikely, since the brood pouches (which originate from the adult oviducts; Berrill, 1948a) often occur in the tunic as isolated entities following the regression of the zooids (Watanabe & Lambert, 1973). Reese (1967) and Woollacott (1979) have shown that isolated sperm ducts of Ciona intestinalis can be induced to spawn by exposure to light. Whether Distaplia brood pouches release larvae by a similar mechanism is not known. Because larvae of D. occidentalis are not active at the moment of release, larval activity seems, however, an unlikely explanation in this species (Watanabe & Lambert, 1973). The physiology of larval release has been investigated in greatest detail for two temperate species, D. occidentalis and Metandrocarpa taylori (Watanabe & Lambert, 1973). Under natural illumination, larvae were released about 1 h after sunrise, but with higher intensities of artificial illumination, the average latency period was reduced to 15 min. Because colonies could be stimulated to release larvae at any time of the day (following adequate dark adaptation), it was concluded that endogenous rhythms play no role in larval release. The rate of larval release was correlated positively with the length of the dark adaptation period. Colonies kept in continuous darkness released relatively fewer larvae than colonies maintained in

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

37

continuous light. Larval release rate has also been correlated with temperature in Podoclavella moluccensis (Davis, in press). Among solitary ascidians, many phlebobranchs have large gonoducts in which masses of eggs and sperm are stored before spawning, whereas most stolidobranchs have no such storage gonoducts. In the latter group, eggs and sperm are released directly from the gonads. Ciona intestinalis has been reported to spawn in approximately three-day cycles in which gametes are shed into the gonoduct for several successive days, then spawned when the oviduct is full (Yamaguchi, 1975). Although few case studies are available, the mechanisms of spawning in solitary ascidians are understood better than the mechanisms of larval release in colonial ascidians. C. intestinalis, which spawns shortly after exposure to light, has been investigated most intensively (Lambert & Brandt, 1967; Whittingham, 1967). Because the photoreceptor that mediates spawning is located at the distal tip of the spermoduct (Reese, 1967), and the spermoduct possesses an extensive network of microtubules (Woollacott, 1979) the spermoduct acts as an independent effector system. The isolated spermoduct can be stimulated to spawn in the same manner as the intact animal (Reese, 1967; Woollacott, 1979). The action spectrum for spawning peaks at wavelengths that implicate cytochrome c as the photoreceptor pigment (Lambert & Brandt, 1967). Other solitary ascidians that spawn in the early morning include several phlebobranchs in the family Corellidae (Huus, 1939; Lambert et al., 1981), as well as the molgulid Molgula manhattensis. It may be significant that every one of these, irrespective of taxonomic affinity, has transparent or transluscent tunic. The reverse side of this correlation also holds; every solitary stolidobranch species with opaque tunic that has been examined spawns many house after the onset of light stimulation (as in Styela plicata; Yamaguchi, 1975; West & Lambert, 1976 and Halocynthia roretzi; Hirai & Tsubata, 1956). Only one solitary ascidian, Molgula pacifica, has been reported to spawn in the dark (Young et al., 1988). As gametes of the latter species are entirely self-fertile, its reproductive pattern may not require diel synchrony among individuals in a population. Likewise, Grave (1926) observed no pattern of release in response to light in Molgula citrina. Yamaguchi (1975) reports that Styela plicata and Ciona intestinalis both spawn after dark in Japan. Both of these species spawn, however, earlier in the day in other parts of the world. The ecological significance of morning larval release by most ovoviviparous ascidians has not been investigated. Thorson (1964) speculated that early morning release allows the larvae maximal time to distinguish between light and dark substrata at settlement. Most compound ascidian larvae swim for no more than a few hours; presumably they select habitats long before the sun has set. With only a few exceptions, all reported data on ascidian spawning times come from laboratory settings in which the adults were exposed to relatively high light intensities. Watanabe & Lambert (1973) have shown with at least one species that the latency period is a function of the cumulative quanta of energy with which the ascidian is illuminated. Thus, spawning can be delayed by illuminating with lower light intensities. Although diel changes in underwater light regimes have not been related quantitatively to any of the species with “morning release”, it seems possible that release could actually occur long after sunrise. Water turbidity, depth, cloud cover, season, and other factors influencing underwater irradiance should all play a role in determining when compound ascidians release larvae. Larval release has been observed in situ by divers for several tropical didemnids, all of which support extracellular algal symbionts of the genus Prochloron or Synechocystis (Duyl et al., 1981; Olson, 1983; Olson & McPherson, 1987) and all of which release larvae at midday. The algae are transferred between generations by the larvae (Kott, 1981; Lafargue & Duclaux, 1979; Olson, 1980). When a larva is released into the common cloacal cavity, algae residing in the cavity become entangled in filaments of the rastrum, a specialised structure located at the postero-dorsal end of the larval trunk (Kott, 1981). Olson (1983) has provided evidence that juveniles, which lack adult pigmentation, require lower light levels than adult colonies can tolerate, but also must select habitats where there is

38

IB SVANE AND CRAIG M.YOUNG

adequate light for photosynthesis of the symbionts. He argues that selection of such habitats is best done when conditions for ultraviolet irradiation are harshest, in the middle of the day. Didemnum molle and other didemnids with short (about 15 min) larval lives accomplish this by releasing brooded larvae between late morning and early afternoon (Duyl et al., 1981; Olson, 1983). Because light is an important settlement cue for the tadpole larvae of many solitary ascidians (Young, 1982; Young & Chia, 1984; Svane, 1987), one might expect that spawning of these oviparous species would be timed so that larvae would attain metamorphic competency during daylight hours (Thorson, 1964). Data are insufficient to support strong conclusions in this regard, but it is interesting to note that most solitary ascidians that reportedly spawn in the morning hatch after about 24 h, but that Styela plicata which has a 12h latency period for spawning (West & Lambert, 1976), develops in half that time (Yamaguchi, 1975). Thus, many solitary ascidians attain metamorphic competency at about the same time of day, but accomplish it by different means. This observation suggests that spawning times and developmental periods could have evolved jointly to increase opportunities for habitat selection. Additional work is needed on the interactions linking these various factors.

THE PELAGIC PERIOD DISPERSAL POTENTIAL AND DURATION OF THE PLANKTONIC PERIOD The potential for ascidian dispersal in the larval stage is correlated with the kind of development they exhibit. Direct developers, viviparous, and ovoviviparous species generally spend no more than a few hours in the plankton (Berrill, 1935), whereas oviparous species may sometimes prolong larval life for as much as a week and still undergo successful metamorphosis (Svane, 1984; Young; unpubl. data on Ascidia callosa). In one respect this correlation between the length of larval life and developmental period is unexpected. The large larvae of compound ascidians, being orders of magnitude greater in volume than those of solitary ascidians, presumably have energy stores sufficient to support a longer larval life. Variability in the total length of pelagic life can originate from processes and adaptations occurring in any of the pre-settlement stages (egg, embryo, or larva). In oviparous ascidians, a portion of the dispersal period is passed in the egg and embryonic stages. Although these life-history stages cannot swim; they demonstrate several possible adaptations that may affect dispersal (Fig 5). At a given sea-water density and viscosity, the sinking rate of a sphere at low Reynolds number is determined by the balance between its density and its drag, the latter being a function of cross-sectional area. If an organism can increase its drag without a proportional increase in density, it will sink more slowly (Vogel, 1981; Chia, Buckland-Nicks & Young, 1984). Ascidians seem to have accomplished this in three different ways. First, many ascidian eggs undergo an osmotic expansion of the perivitelline space shortly after contacting sea water (Berrill, 1929, 1975). Secondly, additional diameter is conferred by the follicle cells, which are heavily vacuolated (hence, light) cells, extraembryonic in origin, that cover the outside of the chorion. Finally, the follicle cells of a few phlebobrachs (Ascidiella aspersa, Corella inflata, C. willmeriana, C. parallelogramma) are lighter than sea water and cause the eggs to float. Lambert & Lambert (1978) have demonstrated that the follicle cells of C. inflata float because their vacuoles contain ammonia, which is lighter than sea water. In the latter species, however, egg flotation facilitates brooding rather than dispersal (Child, 1927; Young, 1988). The duration of the embryonic (i.e. pre-hatching) period in ascidians is partly a function of egg-size and water temperature (Figs 6 and 7), so temperate and cold-water species with oviparous development

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

39

Fig 5.—Ascidian egg adaptations. Ova of Styela montereyensis (A and B), an ascidian with typical extraembryonic cell layers. Newly spawned eggs (C) and 16-celled embryos (D) of Halocynthia aurantium. Note that the chorion swells after spawning, increasing the effective surface area and drag of the egg without a commensurate increase in mass. Eggs of Corella inflata (E) showing floating, ammonia-filled follicle cells that facilitate brooding. Eggs of Molgula pacifica (F) attached to the substratum by a sticky adhesive coat secreted by follicle cells. FC: follicle cells; CH: chorion; TC: test cell; OV: ova; EM: embryo; AC: adhesive coat. Fig 6.—Relationship between temperature and duration of embryonic development (to completion of tadpole). Species: Ascidia atra, A.curvata, A.mentula, A.prunum, A.mammillata, Ascidiella aspersa. Redrawn from Berrill (1935). Fig 7.—Relationship between egg-size and duration of embryonic development (up to completion of tadpole) at 16°C. Species: Molgula, Styela, Ascidia, Ciona, Diazona, Corella, Boltenia, Polycarpa, Tethyum, Syplegma, Dendrodoa, Distomus, Stolonica. Redrawn from Berrill (1935).

probably spend more time in the plankton than similar species in the tropics. As an example, Styela gibbsii takes 30 h to hatch as a competent larva at 10°C in Washington (Cloney, 1987), whereas the congeneric S. plicata spends only 12 h in the embryonic stage at temperatures between 18 and 26°C (Yamaguchi, 1975). The same relationship applies within single species that occur over wide geographic ranges or in areas where temperatures fluctuate over a wide range. For example, the time from fertilisation to hatching in Ascidiella aspersa may be as short as 15 hours at 23°C or as long as 39 hours at 8°C (Knaben, 1952). Within a species, there may be substantial variability in the duration of the free-swimming period (Grave, 1920, 1928; Grave & McCosh, 1924; Grave & Woodbridge, 1924; Levine, 1962; Olson & McPherson, 1987, and many others). At least three major monographs (Berrill, 1935; Grave, 1935; Grave & Nicoll, 1939), attempted to explain this variability over a half century ago; nevertheless, many of the variables influencing length of larval life remain undetermined. Habitat selection may play a major role in determining the length of larval life. Some species with specific habitat requirements delay metamorphosis in the absence of suitable cues (Young, 1982); others settle faster in the presence of conspecifics (or extracts of conspecifics; Grave, 1935; Svane, Havenhand & Jørgensen, 1987) or under particular light conditions (Crisp & Ghobashy, 1971) than they otherwise might. Larval life may be shortened or prolonged artificially by a wide array of chemical reagents, hormones, and other artificial treatments (Grave, 1935; Grave & Nicoll, 1939; Lynch, 1961), but it is doubtful if these have any ecological relevance. In some oviparous species, only a fraction of larvae delay settlement for a prolonged period (e.g., up to 10 days in Ascidia mentula and 6 days in Ciona intestinalis; Svane, 1984, and unpubl.). The remainder settle much earlier. Indeed, settling curves for many species plotted from the literature demonstrate a marked negative skew (long tails on the upper end of the curve) (Millar, 1971). Grave & Nicoll (1939) observed that larval longevity of Ascidia nigra decreased as the breeding season advanced but did not discover the factor controlling this phenomenon. Although they hypothesised the existence of an hormonal “aging factor” that controlled the onset of metamorphosis, the fact that many larvae settled shortly after hatching indicated that such a factor was not necessary (Grave & Nicoll, 1939). Larvae of some colonial ascidians are incapable of settlement immediately upon release, whereas others (e.g., Didemnum molle; Olson, 1985) apparently have no pre-competent period. Within a given species, there is some variation in the time to settlement. For example, Olson (1985) reported that some larvae of D. molle settled immediately, whereas others swam for as long as 20 min. It is not known whether the latter individuals were competent to settle upon release. It seems likely that ascidians demonstrate the same inverse relationship between larval survival and time spent in the plankton as other invertebrates (Thorson, 1950; Menge, 1972; Vance, 1973; Chia, 1974; Emlet, McEdward & Strathmann, 1987), although this question has not been specifically addressed for ascidians. Ascidians in general may have short larval life because there is greater advantage to survival than dispersal.

40

IB SVANE AND CRAIG M.YOUNG

Certainly the advantages of dispersal to benthic invertebrates have been difficult to demonstrate (see Jackson & Strathmann, 1981; Palmer & Strathmann, 1981; Svane, 1984). Because the larvae of many compound ascidians are large enough to observe underwater, dispersal of several such species has been studied in situ. Olson (1985) followed numerous larvae of D. molle at Lizard Island on the Great Barrier Reef. Direction and speed of dispersal were related directly to current speed. Although the larvae swam during the dispersal phase, their swimming appeared virtually ineffectual because of the dominant influence of currents. Most recruitment occurred on the upstream edges of patch reefs downstream from source populations. Active swimming against the current was only observed close to the bottom or in the lee of coral heads, where the current moved very slowly (Olson, 1985). This observation suggests that swimming may function primarily in site selection rather than dispersal. Young (1985), noted that the larvae of Ecteinascidia turbinata spent a large portion of their pelagic period drifting passively without any spontaneous movements whatsoever, but Davis & Butler (in press) have found that larvae of Podoclavella moluccensis swim almost continuously, with only brief resting periods. In both species, dispersal distance seemed, however, to correlate with current speed. Observations made by Olson & McPherson (1987) confirm that dispersal depends mainly on currents, not swimming. There is evidence that larval swimming time is severely over-estimated by experiments conducted in closed containers of sea water. Olson (1985) found that swimming times of larvae released naturally on the reef ranged from 40 to 370 s, whereas larvae captured and retained underwater in closed Plexiglass boxes had settlement times distributed around a mean of 15 min, even when a dark surface was present in the box. Likewise, Duyl et al. (1981) noted that although some beaker-reared larvae of Trididemnum solidum settled in about 15 min, others delayed settlement as long as 3 h. More field observations are needed before we completely understand the nature and extent of the laboratory bias. In quiet harbours, dispersal of ovoviviparous ascidians may be very short indeed. Grosberg & Quinn (1986) investigated recruitment of Botryllus schlosseri on a 1-m square settling panel in the Eel Pond at Woods Hole, Massachusetts. A genetically characterised adult having a rare allele in the population was placed in the middle of the panel, and positions of recruits having the same allele were plotted. Settlement was found to be a function of distance from the adult colony, with most individuals settling within a few centimetres. Berrill (1955) has attributed aggregation in the brooding styelid Dendrodoa grossularia to similar short distance dispersal. Lambert (1968) also noted consistently high local recruitment near parents in Corella inflata, a solitary corellid that retains larvae until they are competent to settle. Although maximum dispersal potential can be estimated by knowing something about the combined lengths of the pre-competent and competent periods (Jackson & Strathmann, 1981) and the speeds of currents larvae are likely to encounter, such estimates may not be representative of the actual dispersal distances achieved by the larvae. Dispersal can be shortened by mortality, early settlement, or entrapment in areas with little or no current. Olson & McPherson (1987) have provided a convincing in situ demonstration of this point for the larvae of Lissoclinum patella on the Great Barrier Reef. Based on in vitro field measurements of the length of larval life, it was predicted that a larva should be able to disperse several hundred metres over a period of about two hours. In observing free-swimming larvae in the field, it was, however, discovered that the majority were consumed by predatory fishes and cnidarians, and the remainder settled less than 10 m from the parent colonies (Olson & McPherson, 1987). Thus, the potential dispersal was approximately an order of magnitude greater than the realised dispersal. Many colonial marine animals, including bryozoans, sponges, and some cnidarians demonstrate very short dispersal distances as a result of brooding (reviewed by Jackson, 1986). Philopatry (as short dispersal distance is sometimes called) may result in greater temporal stability of local populations than is found in

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

41

species with longer dispersal (Chernoff, 1985). Other ecological and evolutionary consequences of philopatry have been discussed by Shields (1982) and Jackson (1986). DYNAMICS OF LOCOMOTION From the standpoint of anatomy and ultrastructure, the tail is the best understood portion of a tadpole (reviewed by Katz, 1983). It is, therefore, surprising that there have been no careful studies of the mechanics of larval swimming in ascidians. In general terms, tadpoles move by alternating contractions and relaxations of the lateral muscle blocks that run alongside the tail. The elastic notochord antagonises the contractions of muscle blocks, translating the longitudinal forces into lateral (or sometimes dorso-ventral) flexions of the tail. Tadpole larvae, because they are small in size, function at low Reynolds numbers where inertial forces are overshadowed by the effects of viscosity. Consequently, tadpoles only move when their tails beat; there is virtually no glide stroke. The trunk of a tadpole does not remain stationary as the head of a fish does. Instead, it moves from side to side, with a displacement from the midline that may be either equal to or slightly less than the amplitude of the tail flexions. Cloney (1959) has noted that a swimming larva of Boltenia villosa flexes around two stationary nodes in the tail region, one just behind the trunk and the other about one-third of the way forward from the end of the tail fin. That portion of the tail between the nodes moves with about the same amplitude as the tip of the fin and the anteriormost portion of the trunk. In addition, larvae of many colonial ascidians have a 90-degree twist in the tail, causing the tail to beat dorsally and ventrally rather than laterally. Grave (1920) was apparently the first to note that larvae of ascidians spin on their longitudinal axis while swimming. He was not able to ascertain experimentally the cause of this spin, but proposed several possible mechanisms. In Aplidium constellation, Grave (1920) attributed the spin to an interaction among three mechanisms: a horizontal tail fin, lateral asymmetry in the trunk, and oblique alignment of the myofibrils in the muscle blocks of the tail (Grave, 1920). Both the horizontal tail fin and the lateral trunk asymmetry apparently resulted from development of the tail within the narrow confines of a small chorionic space. The manner in which the tail folds against the trunk during development molds the trunk tunic into a sort of short spiral (Grave, 1920, 1921). Because of this corkscrew shape, the tadpoles of Aplidium constellatum rotate even as they sink passively through the water column (Grave, 1920). In Botryllus schlosseri (which has a vertical tail fin; Grave & Woodbridge, 1924) and Perophora viridis (horizontal fin; Grave & McCosh, 1924), clockwise rotation during swimming was attributed in part to asymmetrical insertion of the tail at the posterior end of the trunk (Grave & McCosh, 1924; Grave & Woodbridge, 1924) and in part to the axial torsion imparted to the tail movements by a spiral arrangement of myofibrils. Now that methods are available for analysis of movements and flow effects at low Reynolds numbers (Vogel, 1981), the time is right for a re-analysis of locomotory methods in ascidian tadpoles. Ascidian tadpoles move faster than invertebrate larvae which employ ciliary locomotion, but slower than some crustacean larvae (reviewed by Chia et al., 1984). Berrill (1931) demonstrated a positive correlation between swimming speed and larval size. Swimming speeds of tadpoles, as measured in the laboratory, range from 0.2–2.5 cm/s (Berrill, 1931; Chia et al., 1984; Olson, 1985). LARVAL ORIENTATION WITH RESPECT TO PHYSICAL CUES The sensory structures in the tadpole cerebral vesicle (ocelli, statocysts or modifications thereof) mediate both kinetic and tactic (sensu Fraenkel & Gunn, 1940) responses. Although some workers (e.g., Millar, 1971) have referred to ascidian phototaxis as a phototropism, the term is incorrect. Tropisms (orientation of new

42

IB SVANE AND CRAIG M.YOUNG

growth with respect to a physical scalar or vector) may occur in adult ascidians, but not in ascidian larvae. Photokinesis, which involves a change in activity level resulting from a light stimulus, is common in ascidian tadpoles, and there is also evidence for directional responses to the physical vectors of light and gravity. Such directional responses are correctly termed phototaxis and geotaxis, respectively. Following Thorson’s (1964) review in which generalised behaviours of invertebrate photoresponse were suggested, reviews on ascidian biology have adopted the generalisation that larvae are photopositive at release or hatching, then become progressively more photonegative before settlement (Millar, 1971; Berrill, 1975). While this paradigm seems to fit a few species, careful examination of the available data on ascidian phototaxis reveals so much variability within and among species that broad generalisations seem unwarranted. Even different broods of larvae from the same individual or population often show statistical differences in the proportions of larvae demonstrating particular behaviours. In this section, we will review the nature of this behavioural variability and discuss the ecological ramifications thereof. The light wavelengths and intensities that ascidian larvae can detect have been considered by Mast (1921), Grave (1935) and Young (1982). Mast (1921), found that the larvae of Aplidium constellation were unable to detect very low (30 m.c.) absolute intensities of white light, but demonstrated a shadow response when higher intensities were reduced abruptly by even 10 m.c. of energy. Tadpoles also responded to red light, though the spectral distribution of this light was not reported (Mast, 1921). The wavelength and intensity thresholds for the shadow response were studied by Young (1982) for 12 species of solitary ascidians. All species were able to sense light intensities as low as and tadpoles of one species, Ascidia callosa, responded at the lowest intensity tested, (Young, 1982). Using a monochromator to produce 10 nm band-width beams of light, it was shown that all 12 species were most sensitive to wavelengths in the blue and green regions of the spectrum. The upper visual threshold was between 575 and 650 nm (red-orange light), and none of the tadpoles tested responded to light below 425 nm (Young, 1982). Grave (1935) reported that the larvae of several colonial ascidians responded equally to all wavelengths in the human visible spectrum, although he did not report the bandpass data on the filters used for making these observations. Photokinesis In ascidian tadpoles, two kinds of photokinetic responses have been documented. In the first, tadpoles accumulate in regions of low light intensity just before settlement. Crisp & Ghobashy (1971) have demonstrated the kinetic nature of this phenomenon by studying the settlement distribution of Diplosoma listerianum in a chamber illuminated by horizontal light passing through a graded neutral density filter. Because the light source was diffuse and at right angles to the chamber, the authors interpreted the responses as a photokinesis. Larvae did not show any preference for a particular region of the light gradient while swimming, but accumulated at an intensity of 300– 500 lux at settlement (Crisp & Ghobashy, 1971). Dim light was preferred over complete darkness. The second manifestation of photokinesis, commonly called the shadow response, seems to be present in virtually all tadpoles with functional ocelli (Young & Chia, 1985) and at least one species, Molgula occidentalis, which lacks an ocellus (Torrence & Cloney, 1988). The only exception reported in the literature is Metandrocarpa taylori (Abbott, 1955). In the laboratory, ascidian larvae resting on the bottom of the culture vessel or drifting passively in the water column are stimulated to swim by an abrupt reduction in light intensity. This response was discovered independently by Mast (1921) and Grave (1920) in Aplidium constellatum. Mast (1921) also reported that swimming tadpoles respond to shadows by changing the position of the tail, a process which he used to explain phototactic orientation. Sudden increases in light

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

43

intensity do not elicit kinetic responses in resting ascidian tadpoles, though they may elicit changes in tail position in active ones (Mast, 1921). Only a few hypotheses have been advanced to explain the function of the shadow response. Crisp & Ghobashy (1971) and Grave (1935) noted that ascidian larvae settle sooner when stimulated to swim by frequent shadows than when held under constant illumination, but Young & Chia (1985) found this same relationship with only one of eight solitary ascidian species tested. For a species that needs shaded areas for survival (Olson, 1983; Young & Chia, 1984), it would seem reasonable that a tadpole should quickly undergo metamorphosis in a region with many shadows. Woodbridge (1924) extended this idea to explain how Botryllus schlosseri tadpoles locate seagrass blades, which is their normal habitat. She noted that larvae drifting passively in the water column would be stimulated to swim when passing through the shadow cast by a blade of seagrass. Larvae so stimulated swam upward 64% of the time and often contacted a seagrass blade. Because overhangs, rock crevices, and other shaded habitats tend to be excellent habitats for many ascidians (Millar, 1971; Berrill, 1975; Olson, 1983; Young & Chia, 1984; and others), one might suppose that the almost universally occurring shadow response could function in habitat location for many species of ascidians. This was tested experimentally in the laboratory for eight species by Young & Chia (1985). Tadpoles were offered choices between shaded overhangs and unshaded habitats under conditions of constant and fluctuating illumination. It was expected that larvae in the latter treatments would be stimulated to swim more often than larvae in the former, and would thereby be more likely to locate the optimal habitats. Only one species, Styela gibbsii, located the downward-facing surface significantly more often in fluctuating light than in constant light. None of the species located shaded habitats more often as a result of the shadow response. It was concluded that the shadow response does not facilitate selection of an optimal light regime at settlement (Young & Chia, 1985). Geotaxis and phototaxis A generalised description of ontogenetic changes in ascidian orientation behaviour has persisted in the literature for nearly 60 years. Millar (1971) has summarised the stereotyped behaviour as follows: “The ascidian larva has a characteristic pattern of behaviour consisting of an initial period when it swims upward (positive phototropism and negative geotropism) followed by a period when it swims or sinks downwards. The initial phase serves to distribute larvae, and it is in the second phase that the critical reaction is elicited in response to a decrease in light. The larva then swims towards dark areas, which in nature tend to be the vertical or lower surfaces of rocks etc.” This idea of ontogenetic change in orientation behaviour holds best for the compound ascidians that have been investigated; solitary ascidians seem to exhibit much more variability in their behaviour (Table II, Young & Braithwaite, 1980a; Young, 1982; Svane, 1987). Most data reporting the phototactic behaviour in ascidian larvae are casual observations made during studies of settlement and metamorphosis (Thorson, 1964). Many authors have, however, noted that all tadpoles do not behave the same way. There is often marked variability in behaviour among tadpoles of the same age and parentage, and also among individuals of different parentage within a given species. Thus, phototaxis in ascidians is best considered not as a stereotyped behaviour, but as a phenomenon with a statistical distribution. Ontogenetic changes in phototaxis and geotaxis have been documented convincingly in several ovoviviparous colonial ascidians which have a comparatively short planktonic life (Table II). The most detailed study, that of Crisp & Ghobashy (1971), indicates that Diplosoma listerianum conforms nicely to the classical generalisation. Larvae of this species remain strongly geonegative and slightly photopositive

44

IB SVANE AND CRAIG M.YOUNG

throughout the swimming period then switch to the opposite responses just before settlement (Crisp & Ghobashy, 1971). Swimming behaviour is modified by suboptimal temperatures and strong light, but is not influenced by hydrostatic pressures as high as two atmospheres (Crisp & Ghobashy, 1971). Botryllus schlosseri, like Diplosoma listerianum, remains weakly photopositive throughout most of its larval life (Grave & Woodbridge, 1924). Approximately two thirds of the larvae of Perophora viridis retain their positive phototaxis for their entire larval period; the remaining one third reportedly become photonegative towards the end of larval life (Grave & McCosh, 1924). In Aplidium constellatum, the photopositive period lasts for only a few seconds and the larvae spend most of their free-swimming period alternating periods of rest and activity in less illuminated regions (Grave, 1920). Several workers have made observations on behavioural changes of colonial ascidians in the field. Duyl, Bak & Sybesma (1981), Olson (1983, 1985) and Olson & McPherson (1987) all noted initial upwardswimming periods following release of didemnids on coral reefs. In Lissoclinum patella, the larvae moved upward for only about one minute (over a distance of about a metre) before they began swimming downward. In situ settlement experiments in a sealed chamber with a window at one end, however, indicated that larvae settle near the light (Olson & McPherson, 1987). Young (1986) made continuous in situ observations of behavioural changes in Ecteinascidia turbinata larvae of different ages. Contrary to the work with other compound ascidians, he found no evidence for an initial upward swimming period. Newly released larvae drifted passively more often than they swam. Throughout larval life, there were no significant differences in the amounts of time spent swimming upward and swimming downward. Swimming behaviour was extremely vari TABLE II Summary of ascidian tadpole photoresponses investigated to date. C: compound, S: solitary, X: response present, 0: response absent, +: responcses positive, −: respo onse negative, V: response variable, ?: not reported Shadow Phototaxis Family

Species

C/S response early late

Polyclinidae pellucidum

Aplidium (Amaroucium) C

X

+

–

A. constellatum

C

X

+

–

Didemnidae

C

X

+

Trididemnum solidum Polycitoridae Cionidae

Diplosoma listerianum C Distaplia sp. Ciona intestinalis

Grave, 1920; Mast, 1921 Grave, 1936; Mast, 1921 V

? C S

+ ? ?

– 0 +

Duyl et al., 1981 0 –

Perophoridae

Perophora viridis

C

?

+

V

Corellidae C. willmeriana Chelyosoma productum

Corella inflata S S

S X X

X 0 0

0 – V

– Young, 1982 Young & Braithwaite, 1980a

Reference

Crisp & Ghobashy, 1971 Berrill, 1948a Berrill, 1947; Castle, 1896; Dilly, 1964; Dybern, 1963 Grave & McCosh, 1924 Young, 1982

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

45

Shadow Phototaxis Family

Species

C/S response early late

Reference

Ascidiidae

Ascidia nigra

S

X

+

_

Goodbody, 1963; Grave, 1936

A. callosa A. paratropa A. mentula Styelidae

S S S Dendrodoa grossularia C

X X X S

0 0 – ?

– – – 0

Young, 1982 Young, 1982 Svane, 1987 0

X

+

–

Symplegma viride Polyandrocarpa tincta P. gravel Styela montereyensis

C C

X X

+ +

_ _

Grave & Woodbridge, 1924; Woodbridge, 1924 Grave, 1936 Grave, 1936

C S

X X

+ 0

_ V

S. coriacea S. gibbsii S. partita Cnemidocarpa finmarkiensis Metandrocarpa taylori Pyuridae Boltenia villosa Halocynthia igaboja Molgulidae

S S S S

X X X X

0 0 0 0

_ _ _ –

Grave, 1936 Young & Braithwaite, 1980b Young, 1982 Young, 1982 Grave, 1941, 1944 Young, 1982

C

0

–

_

Abbott, 1955

Pyura haustor S S Molgula citrina

S X X S

X 0 0 0

0 _ V 0

– Young, 1982 Young, 1982 0

Botryllus schlosseri

Berrill, 1950

Young, 1982

Grave, 1926

able among individual tadpoles, and the ontogenetic changes in swimming behaviours predicted by the standard paradigm (Thorson, 1964; Millar, 1971; Berrill, 1975) did not occur (Young, 1986). In solitary ascidians, behavioural responses to light and gravity are extremely variable. Two species, Ascidia nigra (Grave, 1935; Goodbody, 1963) and Ciona intestinalis (Berrill, 1947; Millar, 1953; Dybern, 1963), have been reported to pass through photopositive and photonegative phases, but this conclusion is based on casual observations and field distributions (Dybern, 1963) rather than careful behavioural experiments. Svane (1987, and unpubl.) has demonstrated experimentally that both C. intestinalis and Ascidia mentula remain photonegative during their entire free-swimming period. Styela partita is negatively geotactic under all light conditions, and demonstrates negative phototaxis early in larval life (Grave, 1941, 1944). Late-stage larvae cease swimming and sink passively to the bottom (Grave, 1941, 1944). Although Grave (1944) attributed this behaviour to a reduced ocellus in styelid larvae, similar behaviour is demonstrated by Ascidia mentula, a species with a complete 3-lens ocellus (Svane, 1987). The larva of Molgula citrina, which lacks an ocellus, displays short bouts of negative geotaxis followed by random swimming and periods of rest (Grave, 1926). No response to light was observed (Grave, 1926).

46

IB SVANE AND CRAIG M.YOUNG

Fig 8.—Phototactic orientation mechanism of Aplidium constellation, as explained by Mast (1921). The mechanism relies on the orientation of the pigment cup, which allows light to strike the retinal cells only from the anterior or right sides.

The larval ocellus of Ciona savignyi differentiates after hatching, and the larvae are reported to pass through at least four different behavioural phases that correspond to the ocellar changes (Kajiwara & Yoshida, 1985). Newly hatched larvae are strongly geonegative but do not respond to light. Within 30 min of hatching, larvae in culture vessels swarm at the surface in dense aggregations (Kajiwara & Yoshida, 1985). Larvae begin to exhibit the shadow response 1.5 h after hatching, and the response becomes stronger as larval life proceeds. Finally, larvae developed a strong negative phototaxis after 3.5 h according to Kajiwara & Yoshida (1985). In the San Juan Islands of Washington, photoresponses of 12 species of solitary ascidians have been studied (Young & Braithwaite, 1980a,b; Young, 1982). None of the species showed predictable positive phototaxis at hatching, although most demonstrated an intermittent negative geotaxis early in larval life. Photoresponses at settlement were extremely variable in some species (e.g., Chelyosoma productum, Styela montereyensis), while other species (e.g., Pyura haustor, Corella willmeriana, Cnemidocarpa finmarkiensis) chose shaded substrata more often than expected by chance. The settlement distributions indicative of geotactic responses (choices of upward-facing and downward-facing surfaces) were similarly variable within and among species (Young, 1982). The species differences in phototaxis and geotaxis correlated well with the distributions of adults in the field. Taken together, the data on taxes in solitary ascidians suggest that photoresponse is variable among species and that broad generalisations cannot be made. There is more consistency among the short-lived larvae of colonial ascidians, but even the latter differ in many important details. Mast (1921) is the only worker to have considered in detail the mechanism of phototactic orientation in ascidian tadpoles. Many invertebrate larvae that respond to directional light stimuli (e.g., crustacean zoeae, polychaete setigers) have two distantly positioned photoreceptors. The mechanism of orientation seems to involve moving back and forth until an equal amount of energy enters each eye. Using simple but elegant observational techniques with larvae of Aplidium constellatum, Mast postulated a credible mechanism by which tadpole larvae orientate with a single photoreceptor (Fig. 8). The pigment cup in the ocellus of A. constellatum is directed anterolaterally on the right side (which Mast termed the ocular side) of the trunk. Thus, the retinal cells are stimulated only by light directed from the front or from the ocular side; light directed from the rear of the animal or from the left side does not stimulate the ocellus. In a tadpole larva moving perpendicular to a beam of light, the ocellus is alternately illuminated and shaded by the trunk as the animal spins on its own longitudinal axis. Each time the ocellus is shaded, it stimulates the tail to bend slightly and briefly (while still vibrating) toward the ocular side. This causes the animal to turn slightly away from the light. With another half rotation, the retinal cells receive full illumination, causing the tail to bend toward the abocular side and producing the same effect as before. After this process has been repeated during several revolutions, the animal becomes situated such that its pigment cup shades the retinal cells continuously. In this position, no tail bending occurs, and the direction of movement is away from the light in a straight line. Photopositive animals orientate exactly the same way, except that they bend their tails toward the abocular side when the ocellus is shaded and toward the ocular side when it is abruptly illuminated (Mast, 1921). Both the shadow response and the phototactic response described by Mast (1921) depend on reflexes stimulated by abrupt changes in light intensity. There are, however, several subtle differences in addition to the net behavioural effect. First, the shadow response occurs only when an animal is at rest, whereas the phototactic stimulation occurs only in actively swimming animals. Secondly, the

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

47

shadow response occurs only when light intensity is decreased. An abrupt increase in intensity has no effect on resting tadpole larvae. Swimming animals, on the other hand, respond both to increases and decreases in illumination. Finally, the shadow response stimulates vibration of the tail, whereas phototaxis stimulates a bending of the tail while the latter continues to vibrate (Mast, 1921). Mast’s (1921) hypothesis on the mechanism of phototactic orientation is supported by the work of Kajiwara & Yoshida (1985) on Ciona savignyi. In this species, the pigment of the ocellus is diffuse and irregular in shape at hatching. It passes through a series of changes and increases in pigment density for the first 3.5 h of larval life until it finally takes on the characteristic cup shape. Behavioural changes correspond remarkably well to the changes in ocellus morphology. The shadow response begins when the ocellus becomes packed with dense pigmented tubules, 1.5 h after hatching, but directed swimming (negative phototaxis) does not occur until the pigment cup is completely formed (Kajiwara & Yoshida, 1985). LARVAL MORTALITY As with most invertebrate larvae, relatively little is known about the sources of mortality occurring during the pelagic phase. Most work to date has been on predators. Encounters between predatory fishes and three different species of larvae have been observed in the field (Olson, 1983; Olson & McPherson, 1987; Davis & Butler, in press). 87% of 133 larvae of Lissoclinum patella followed in the field were consumed by predatory fishes, zoanthids, or corals (Olson & McPherson, 1987). The major predators were territorial pomacentrids. Although the fish were equally abundant at all depths between 5 and 25 m, a higher percentage of tadpoles were consumed at shallow depths than at greater depths. Olson & McPherson (1987) attributed this difference to the fact that tadpoles passed pomacentrid territories repeatedly in shallow water as they were buffeted by the surge. One species of fish with a small mouth, Pomacentrus lepidogenys rejected larvae that they ingested. However, the egested larvae were always damaged too badly to continue swimming (Olson & McPherson, 1987). Predation rates on Podoclavella moluccensis in Southern Australia were much lower; of 270 larvae followed in the field, only two were mouthed by fishes, and both lived to settle successfully (Davis, unpubl.). Similarly, larvae of the didemnid, Didemnum molle were always rejected by fishes and always survived the attacks (Olson, 1983). Unidentified ascidian tadpoles have been taken in the guts of the pinfish Lagodon rhomboides in the Gulf of Mexico (J.Luzcovich, pers. comm.). This same fish species and several others accept tadpoles of many species of compound ascidians in laboratory aquaria (B.Bingham, unpubl.). Larval defences against predators (reviewed by Young & Chia, 1987) may be structural, behavioural, or chemical. The prevalence of both structural (Young, 1985) and chemical (Stoecker, 1978, 1980) defences in adult ascidians suggests that such mechanisms might also be present in larvae. However, besides anecdotal evidence (e.g., rejection of larvae by fishes in the field; Olson, 1983; Davis & Butler, in press), there is only one example of an ascidian larval defence in the literature: distastefulness in larvae of Ecteinascidia turbinata. The bright orange tadpole larvae of this perophorid ascidian are mouthed then rejected by several species of fish, including Lagodon rhomboides (Young & Bingham, 1987). By homogenising the larvae and embedding the homogenate in agar pellets, it was shown that the tadpoles are rendered unpalatable not by a structural defence but by a potent chemical substance (Young & Bingham, 1987). Fish that had experienced the taste of Ecteinascidia turbinata larvae learned quickly to avoid them as food, suggesting the possibility of aposematic (or “warning”) coloration. This hypotheses was tested by creating artifical mimics of palatable ascidian tadpoles (Clavelina oblonga) using orange stain. Fish that had not tasted unpalatable orange tadpoles of Ecteinascidia turbinata consumed the mimics readily, but experienced fish avoided all orange tadpoles regardless of species (Young & Bingham, 1987). The chemistry of the defence mechanism

48

IB SVANE AND CRAIG M.YOUNG

remains unknown, but Bingham (pers. comm.) has now found several other ascidians in which orange coloration is correlated with unpalatability to fishes. Several sessile benthic predators are known to prey on ascidian larvae. The temperate octocoral Alcyonium siderium and the sea anemone Metridium senile, both of which are common on subtidal rocks in New England, often contain large numbers of ascidian tadpoles in their guts (Sebens & Koehl, 1984). It has been proposed that these cnidarians reduce competitive interactions with adult compound ascidians (Aplidium stellatum) by consuming the competitors’ larvae (Sebens & Koehl, 1984). Davis & Butler (in press) observed ingestion of larval Podoclavella moluccensis by the hard coral Culicia tenella in the field. Corals also captured a small percentage of Lissoclinum patella larvae on the Great Barrier Reef (Olson & McPherson, 1987). Adult ascidians sometimes consume their own eggs and larvae or the offspring of other species of ascidians (Young, 1988). It has been shown that the tendency toward larval cannibalism is greater in species that occur as solitary individuals in the field than among species that live in aggregations and settle gregariously as larvae (Young, 1988). Rejection of conspecific eggs relative to those of other species has also been observed for Ciona intestinalis (Havenhand, pers. comm.). Species that reject their own tadpoles do so by means of the “crossed reflex” (Hecht, 1918), in which objects stimulating the oral tentacles or the inside of the incurrent siphon epithelium elicit closure of the excurrent siphon followed by a strong contraction of the body wall musculature (Young, 1988). Species with larger siphon diameters tend to reject eggs and larvae less often than species with small siphon diameters. Young (1988) has proposed that an efficient rejection mechanism might be an important prerequisite for gregarious settlement responses in ascidians. There are undoubtedly numerous sources of mortality other than predators that eliminate larvae or embryos during their planktonic period (Thorson, 1950, 1966). However, very few such factors have been investigated for ascidians. Goodbody & Fisher (1974) reared the embryos of Ascidia nigra in sea water collected from three different habitats and demonstrated higher hatching success in water from the open ocean than in water from two inshore habitats. Survival differences were attributed to pH and salinity (Goodbody & Fisher, 1974). Many larvae are probably lost by drifting away from appropriate settlement sites. However, the extent of this loss has not been estimated. SETTLEMENT The transition between pelagic and benthic existence involves two processes, settlement and metamorphosis. In ascidians, settlement may be defined as the process of locating and affixing to the juvenile habitat. It generally, though not always, precedes metamorphosis. Metamorphosis is defined as the sequence of morphological events that transform the larva into a sessile, feeding juvenile (Cloney, 1982). These processes may include (but are not limited to) attachment, resorption of the tail, rotation of the trunk, emigration of blood cells from the hemocoel to the tunic, extension of epidermal ampullae, retraction of the sensory vesicle, and destruction of larval structures (Cloney, 1982). A larva capable of undergoing these metamorphic changes successfully is termed “competent” (Jackson & Strathmann, 1981). The processes by which ascidian tadpoles select habitats may involve numerous characteristic behaviours including phototaxis, thigmotaxis, chemotaxis, kin recognition, and gregariousness. Behaviour only becomes important, however, given the appropriate ecological opportunities (Moore, 1975). Thus, inherent behavioural patterns must interact with the distribution of available habitats, with pre-settlement mortality, and with factors modifying larval distribution in the plankton to produce the pattern of settlement.

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

49

Settlement distributions are modified in turn by post-settlement mortality and migration (the latter in didemnids and some molgulids) to produce the distributional patterns of the juveniles and adults. Because microscopic juveniles are generally difficult to monitor in the field, there are few data on the settlement distributions of invertebrate larvae in general. Just as the number of recruits (where “recruit” is defined as an individual large enough to observe in the field) is not necessarily representative of the number of settlers because of mortality occurring between the two stages (Keough & Downes, 1982; Connell, 1985; Butler, 1986; Davis, 1987), so the distribution of recruits does not always reflect the settlement distribution. Nevertheless, the positive information contained in recruitment or distributional data should not be ignored; often, the only sites where we know with certainty that settlement occurred are those occupied by juveniles or adults. Colonial ascidians are often large enough to observe from the very moment of settlement, so the problem of distinguishing recruits from settlers is less acute in some members of the Ascidiacea than in many other invertebrate groups. Most behavioural studies have considered responses to single factors in isolation, whereas tadpoles are confronted in the field with interactive, complex suites of potential settlement cues. Whether settlement cues are ranked consistently or change in relative importance under different environmental conditions is not known. Some larvae, including those of soft-sediment molgulids, have been described as non-discriminating at settlement. Such are presumed to settle randomly within whatever regions competent larvae arrive. In the literature, three major spatial patterns have been attributed to larval behaviour: (1) small-scale singlespecies aggregations that result from gregarious settlement behaviour, (2) epibiosis and multiple-species aggregations, and (3) occurrence in shaded or cryptic habitats, generally attributed to negative phototaxis. LARVAL RESPONSES TO CONSPECIFICS Many solitary ascidians form clumps (Fig 9). Dense aggregations may occur on virtually any scale, only the smallest of which are likely to be influenced by settlement choices of the larvae (Butman, 1987). Aggregation is common on soft bottoms where other ascidians may be among the only available hard substrata for settlement (Young, 1985), but also occurs on rock surfaces where ascidian tunic is much less common than other available substrata (Young, 1982). Members of the genus Pyura, particularly in the southern hemisphere, commonly form dense beds in the lower intertidal zone. These include P. praeputialis in New South Wales, Australia (Dakin, Bennett & Pope, 1948; Underwood & Fairweather, 1986), P. chilensis and P. stolonifera in Chile (Gutierrez & Lay, 1965; Paine & Suchanek, 1983), P. pachydermatina in New Zealand (Batham, 1956), P. stolonifera in South Africa (Stephenson, 1942; Day, 1974) and P. haustor in Washington, USA (Young, 1982). Subtidal styelids and pyurids that aggregate include Bolteniopsis prenenti (Monniot, 1965), Microcosmus vulgaris (Monniot, 1965), Styela gibbsii (Young, 1985), Dendrodoagrossularia (Berrill, 1955) and Pyura haustor (Young, 1982, 1985). Molgula occidentalis form aggregations on virtually all scales from fourths of metres up to tens of metres on the relatively flat terrain of intertidal sandbars (Young, in press). Molgulids occurring in small-scale aggregations on hard substratum include M. complanata (Schmidt, 1982a, b), M.manhattensis (Monniot, 1965), M. occulata (Monniot, 1965) and the anural M. pacifica (Young et al., 1988). Subtidal aggregations have been reported in numerous phlebobranchs including Chelyosoma productum (Young & Braithwaite, 1980a), Corella inflata (Lambert, 1968), and Ascidia mentula (Havenhand & Svane, in press). Aggregations of some species tend to have unimodal size distributions, suggesting that they are formed by the nearly simultaneous settlement of larvae on a common substratum, whereas aggregations of other species are polymodal, consisting of numerous small individuals attached to the tunic of larger ones. These patterns suggest two different kinds

50

IB SVANE AND CRAIG M.YOUNG

Fig 9.—Small-scale aggregations of ascidians. A: subtidal clump of Ciona intestinalis in Gullmarsfjorden, Sweden, B: intertidal aggregation of Dendrodoa grossularia, probably formed by philopatric dispersal. Note two large adults (A) surrounded by numerous younger recruits (R). C: aggregation of Pyura haustor, a species with strongly gregarious larvae at 20 m depth in the San Juan Islands, Washington, USA. D: adult and juvenile Ascidia mentula in Gullmarsfjorden, Sweden. This species settles gregariously. E: “social” ascidian Clavelina picta at 15 m depth in the Bahamas. Apparent aggregations are formed by asexual budding, not larval processes. F: high-density clump of Styela montereyensis on an intertidal piling in Neah Bay, Washington. Larvae of this species are not gregarious (Young, 1988).

of behaviour, both of which have been documented in ascidians: aggregation among siblings or other individuals of the same age, and selection of adult conspecifics as a settlement site. Chemical cues associated with adults and juveniles stimulate metamorphosis and settlement in many ascidian species. Grave (1935) and Grave & Nicoll (1939) discovered that larvae of Ascidia nigra and Polyandrocarpa sp. were induced to settle sooner in adult or larval tissue extracts than in plain sea water. The same phenomenon has now been shown to occur in Ascidia mentula, Ascidiella scabra, (Svane, Havenhand & Jørgensen, 1987), and Pyura haustor (Young, unpubl. data). In all of these species, the time to settlement is related to the concentration of the adult extracts to which larvae are exposed. Significantly, extracts of tunic, the tissue larvae are most likely to contact, are more potent inducers than visceral extracts (Svane et al., 1987). Ascidiella scabra can be stimulated to undergo metamorphosis while still enclosed in their egg envelope (Svane et al., 1987), suggesting that larvae are competent to settle immediately after hatching. Although Grave & Nicoll (1939) reported a species specific effect (tissue extracts of Polyandrocarpa do not stimulate metamorphosis in Ascidia nigra), Svane et al. (1987) found little evidence for specificity between the two closely related ascidiids, Ascidiella scabra and Ascidia mentula. Larvae of Pyura haustor can be stimulated to undergo metamorphosis by exposure to sea water that has passed through the branchial sac and atrium of conspecific adults (Young, 1978). All of these phenomena are known only from the laboratory. Although their ecological significance remains unproven, the observations suggest not only that adult-associated cues could be important inducers of metamorphosis, but also that larvae may be able to detect water-borne (as opposed to surface-bound) cues. Larvae of several ascidians respond to newly settled juveniles. In dense cultures of Chelyosoma productum, larvae settle in a distinctly non-random manner (Young & Braithwaite, 1980a) and select areas with established juveniles much more often than regions colonised only by algae and bacteria (Young & Braithwaite, 1980a). Duyl, Bak & Sybesma (1981) showed that larvae of Trididemnum solidum settle sooner when even a single juvenile is present in a beaker than when no established juvenile is present. In related studies with Ascidia nigra, it was shown that curves of settlement versus time have steeper slopes when large numbers of larvae are present in the vials (Grave & Nicoll, 1939). Grosberg (1981) compared settlement density of Botryllus schlosseri among plates with varying densities of established juvenile colonies present. Although juveniles were avoided by many species of invertebrates, the presence of juveniles had no apparent effect on the settlement rates of conspecifics or any other species of ascidian. These data are supported by Schmidt (1982b) showing random spatial distributions among B. schlosseri settlers in Britain. Schmidt (1982b) also recorded the distribution of two other species of ascidians, Diplosoma listerianum, and Molgula complanata, on Perspex panels after submergence for four weeks; M. complanata was found to be aggregatively distributed while Diplosoma listerianum were distributed randomly. Schmidt (1982a,b) concluded that the observed distribution pattern resulted from tadpole choices or the absence thereof. In D. listerianum this conclusion was supported by nearest neighbour distances in laboratory settlement experiments, but the evidence for larval behaviour as a determinant of juvenile spatial pattern in Molgula complanata is equivocal, since post-settlement processes operating during the four week

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

51

submergence period were not addressed. Havenhand & Svane (in press) demonstrated that distributions of newly settled Ascidia mentula in Petri dishes were significantly different from random at densities higher than 5 per cm2, but not at lower densities. The second pattern of single-species aggregation, characterised by a polymodal age distribution within a given clump, implies that larvae attach directly to the surfaces of established adults. Because the same pattern can, however, result from differential mortality in which survival is greater in that portion of the population settled on adults, careful settlement experiments are required before behavioural preferences are implicated as the cause. Such choice experiments have been run with several species. Larvae of Pyura haustor, a species that forms aggregations in the rocky intertidal zone, and in both soft and hard subtidal habitats, demonstrate a strong preference for the tunic of adult conspecifics as a settlement site (Young, 1982). Other substrata, including rocks, mollusc shell, and the tunic from various other ascidians are selected much less often than P. haustor tunic. Experiments in which the grooves and crevices of adult tunic are removed by a clean cut indicate that the attraction is not to structural or textural features, but rather to some chemical aspect of the adult. The possible selective advantage of this behaviour has been investigated by allowing larvae to settle on living adults and on rocks in the laboratory, then outplanting the juveniles to intertidal habitats where the adults naturally occur (Young, 1983). Over a two-week period, survival on adults was significantly higher than survival on adjacent rocks. Young (1984) attributed this difference to higher moisture and lower temperatures in the ascidian clumps during periods of tidal exposure. Havenhand & Svane (in press) have investigated gregarious settlement of Ascidia mentula, a species that consistently aggregates in the fjords of western Sweden. Over an 11-year period, recruitment of juveniles at several subtidal sites was correlated with adult density. In the laboratory, three pieces of evidence were indicative of gregarious settlement responses: aggregation of juveniles, acceleration of metamorphosis in response to adult tunic extracts, and apparent attraction to adult tunic. In the latter experiments, larval distribution was examined after 10 min of incubation in a submerged tube with an adult affixed to one end. There were consistently more larvae at the adult end than the opposite end. Although this experiment is indicative of gregarious larval behaviour, it does not demonstrate whether distant chemotaxis is involved or if larvae simply remain in the vicinity of an adult following random encounter. Adults and juveniles of Molgula occidentalis cover themselves completely with sand by means of fine tunic filaments all over the body. Larvae offered a choice of sand collected from the environment, sand removed from adults, or adult tunic selected the adult sand significantly more often than the other substrata (Young, in press). This ability to select a habitat may seem surprising in that these molgulid larvae have very reduced oral papillae (Cloney, 1978), but sensory neurons have been located in the anterior epidermis of the apapillate larva (Torrence & Cloney, 1988) (Fig 10). Young (1988) tested eight ascidian species for gregarious settlement by offering natural rocks and adult tunic to larvae as alternative substratum choices. Besides those species already discussed (Pyura haustor, Chelyosoma productum), only one, Styela gibbsii, was gregarious in laboratory experiments. Four other species that sometimes attach to adults under field conditions (Styela montereyensis, Boltenia villosa, Ascidia callosa, and Halocynthia igabojd) did not demonstrate strong preferences for adult tunic in the laboratory. The remaining five species tested did not select adult tunic as a preferred substratum. This pattern was expected, as none of them are found attached to conspecifics in the field (Young, 1988). Young

52

IB SVANE AND CRAIG M.YOUNG

Fig 10.—Attachment process of the apapillate larvae of Molgula occidentalis (A). Newly metamorphosed juvenile (B) showing sticky primary ampulla which provides initial attachment to the substratum. Juvenile several days old (C) attached to adjacent sand grains with secondary ampullae. Sediment-covered adults (D) on an intertidal sandflat.

(1985) provided field evidence that aggregation in Styela gibbsii reduces predation by the subtidal asteroid Evasterias troschelii by limiting the predator’s access to a region of soft tunic on the posterior end. SETTLEMENT RESPONSES TO OTHER ORGANISMS On both hard and soft bottoms, many species of solitary ascidians occur in multiple species aggregations where individuals are attached directly to individuals of other species. Monniot (1965) described the ecological relationships in such aggregations, termed “blocs de Microcosmus” in which large individuals of Microcosmus vulgaris provide attachment surfaces for many species of ascidians and other invertebrates. In Washington, aggregations dredged from deep, muddy bottoms often include as many as eight different species of solitary ascidians. Boltenia villosa and Styela gibbsii live attached to the tunic of Halocynthia igaboja and Pyura haustor, where they obtain protection from the predatory gastropod Fusitriton oregonensis (Young, 1985, 1986). In laboratory experiments, it has been demonstrated that larvae of both these epizooitic species prefer the tunic of the host species over the tunics of other species, as well as common inorganic substrata from their environment including rock and shell (Young, 1982). Moreover, the larvae of both Boltenia villosa and Styela gibbsii delay metamorphosis in the absence of suitable substratum (Young, 1982). Davis (1987) documented the settlement choices of larvae of Podoclavella moluccensis by following, larvae underwater and noting the outcome (settlement or rejection) of their first encounter with a substratum. Besides wooden pier pilings, most of the available substrata were sponges of several species. When a substratum was acceptable, a larva beat its tail vigorously for 3–5 min after attachment. Attached larvae were capable of rejecting a substratum by detaching with a short flick of the tail. Larvae generally rejected sponges, particularly those of the genera Dendrocia and Mycale. During the first month after settlement, survival was significantly higher on bare space and on the preferred sponges than on those sponges most often rejected by the larvae, suggesting that the settlement choices had an adaptive component (Davis, 1987). Moreover, larval choices were good predictors of the relative proportions of recruits appearing on the various available substrata in the system. SETTLEMENT RESPONSES TO PHYSICAL CUES Ascidians often occur in cryptic habitats such as cracks, crevices, and the dark undersides of rocks or overhangs. It has often been assumed that these habitats are located by negtive photo taxis (Dybern, 1963; Thorson, 1964; Crisp & Ghobashy, 1971; Millar, 1971; Berrill, 1975). The same pattern could, however, result from other behaviours including the shadow response (Woodbridge, 1924; Young & Chia, 1984), rugophilia, and negative geotaxis. Many selective pressures operate more intensely in open than cryptic sites, particularly during the juvenile stage. These include predation (Keough & Downes, 1986), silt (Young & Chia, 1984; Svane, 1987), competition with diatoms or filamentous algae (Goodbody, 1963; Young & Chia, 1984), grazing by herbivorous snails (Young & Chia, 1984) and ultraviolet light (Olson, 1983). Thus, both behaviour and selective mortality must be considered as important determinants of a cryptophilic distribution. Indeed, photonegative behaviour probably evolved in response to the predictable differences in mortality between open and shaded habitats.

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

53

By integrating field recruitment studies with laboratory experiments on larval behaviour, Dybern (1963) demonstrated that negative phototaxis in the larval stage can explain the crytophilic distribution of Ciona intestinalis in Gullmarsfjorden, Sweden. Although he did not consider the possibility that post-settlement mortality on upward-facing surfaces could enhance the distributional pattern, Dybern’s (1963) observation that deep-water (low light) populations often occur on such surfaces supports his hypothesis that larval behaviour sets the pattern of distribution. On subtidal limestone outcroppings in the Gulf of Mexico, the compound ascidian Aplidium stellatum recruits primarily on vertical surfaces, which are much less common than horizontal ones (Gotelli, 1987). Laboratory experiments with the larvae demonstrated that vertical surfaces were chosen over horizontal ones in a consistent 2:1 ratio regardless of the relative proportions of the two kinds of surfaces. The nature of the behaviour producing this pattern has not been investigated. While following larvae of Didemnum molle underwater on the Great Barrier Reef, Olson (1983) observed that nearly all individuals settled on the undersurfaces of coral rubble and that recruits on horizontal settlement panels occur mostly on the undersides, within a few centimetres from the edge. In vitro experiments demonstrated, in agreement with the field observations, that larvae select shaded over unshaded surfaces. The advantage of photonegative settlement behaviour becomes apparent in the juvenile stage. Individuals reared in exposed reef habitats died within four days of settlement, possibly of exposure to ultraviolet light (Olson, 1983). The behaviour of D. molle larvae contrasts markedly with that of Lissoclinum patella, another didemnid containing symbiotic Prochloron algae (Olson & McPherson, 1987). Larvae of the latter species settled exclusively near the light end of a 1-m long light gradient. Olson & McPherson (1987) used this photopositive settlement behaviour to explain the low density of adult colonies at depths greater than 25 m. Alternative hypotheses (e.g., low survival of symbionts due to lower light levels) were, however, not tested. A comparative study of larval settlement behaviour in 12 species of ascidians in the Puget Sound region of Washington, USA, showed that some species were strongly photonegative at settlement, whereas others did not discriminate between illuminated and shaded substrata (Young, 1982). In most cases, larval behaviour was a good predictor of adult distribution in rocky subtidal habitats. Photonegative larval behaviour has also been implicated as a determinant of distribution in Diplosoma listerianum (Crisp & Ghobashy, 1971) and Botryllus schlosseri (Woodbridge, 1924; Dybern, 1963). Young & Svane (unpubl.) compared larval behaviour and field distribution of two solitary ascidians in Florida, one (Microcosmus exasperatus) of which lacks a pigmented ocellus, and one (Ascidia nigra) of which discriminates between light and dark regions at settlement. Recruitment of the two species on half-shaded Plexiglas plates in the field followed the laboratory predictions; Microcosmus exasperatus settled randomly, whereas Ascidia nigra demonstrated a preference for the dark portions of the plates (Young & Svane, unpubl.). ROLE OF REPRODUCTIVE AND LARVAL PROCESSES IN RECRUITMENT In recent years, population biologists working in the marine environment have placed an increasing emphasis on processes that control the abundance (or “supply”) of new recruits (Underwood & Denley, 1984; Connell, 1985; Roughgarden, Iwasa & Baxter, 1985). In ascidians, as in other benthic marine invertebrates with open populations, both temporal patterns of abundance (i.e. population dynamics) and spatial patterns of distribution are influenced by developmental mode, mortality occurring during the larval stage, advection and diffusion of larvae, and larval behaviour, as well as the distribution, fecundity and mortality of adult populations. We have already discussed many of these factors as isolated topics. In this

54

IB SVANE AND CRAIG M.YOUNG

section, we will consider the relative contributions of these factors in producing spatial and temporal patterns of ascidian recruitment. For our purposes, recruitment may be defined as appearance of a new generation of ascidians in a benthic population. The definition is operational rather than absolute, since different sampling protocols cause recruits to be first documented at different ages after settlement (Keough & Downes, 1982; Connell, 1985). Thus, a settler is the product of larval processes alone, but a recruit is the composite product of larval processes and events occurring during that portion of the post-settlement stage before the animal is first noticed by the ecologists. It is axiomatic that recruitment varies from place to place and from time to time. One might expect that recruitment events would be tied closely to reproductive seasonality (reviewed by Millar, 1971), but local patterns of mortality can result in no recruitment even when many larvae are produced. In a discrete population, it should be possible to estimate pre-recruitment mortality by comparing fecundity and recruitment. If recruitment were determined entirely by larval production, then one would expect a perfect correlation between patterns of fecundity and patterns of recruitment (Gotelli, 1987); lower levels of correlation should reflect variability in pre-recruitment losses. Most benthic invertebrates have, however, open populations where immigration and emigration are difficult or impossible to assess (Roughgarden et al., 1985). Quantitative comparisons of fecundity and recruitment become meaningful only where populations are closed or discrete, where dispersal time is very short, or where larval losses can be measured by direct obervations. Fortunately, some ascidian populations meet these criteria. Davis & Butler (in press) have presented evidence that the ascidian Podoclavella moluccensis has closed populations in the gulfs of South Australia. Using underwater observation of larvae, Davis (1988) estimated the number of individuals surviving to each major life history stage. Of 6430 larvae produced per m2, only 6.1 % settled. This estimate of larval mortality is about the same order of magnitude as estimates for other benthic invertebrates, including bivalves with much longer-lived larvae (reviewed by Strathmann, 1985; Young & Chia, 1987). About 63% of the settlers survived the first month to become “recruits” and 14% of recruits survived to become reproductive adults. Thus, only 0.56% of all larvae survived to sexual maturity (Davis, 1987). Gotelli (1987) studied a shallow subtidal population of Aplidium stellatum on a discrete limestone outcropping. Just over 50% of the temporal variation in recruitment could be explained by variation in the number of zooids brooding larvae in the previous month (Gotelli, 1987). The rest of the variability in recruitment was attributed to unstudied and variable losses in the larval and juvenile stages. Life table and survivorship curves have been calculated for many ascidians (Goodbody, 1962, 1963; Lambert, 1968; Goodbody & Gibson, 1974; Nomaguchi, 1974; Svane & Lundälv, 1981; Svane, 1983, 1987; Young & Chia, 1984; Young, 1985; Keough & Downes, 1986; Davis, 1987), but almost none of these consider mortality (“wastage”, in the terminology of Thorson, 1950, 1966) in the plankton, and only a few have characterised mortality immediately after settlement. The compound ascidian Podoclavella moluccensis is the only species for which mortality in all life history stages has been estimated (Davis, 1988). Olson & McPherson (1987) have estimated larval mortality of one additional species by direct observation, and several workers have studied the contribution of early post-settlement mortality. One method of documenting the mortality occurring during these early (and often microscopic) benthic stages is to transplant laboratory-settled juveniles on artificial substrata to the field habitats of interest. This method has been used for eight species of solitary ascidians in Washington (Young & Chia, 1984), for Ascidia mentula in Sweden (Svane, 1987), for Ascidia nigra in the Caribbean (Goodbody, 1963), and for Ciona intestinalis in Japan (Nomaguchi, 1974; Yamaguchi, 1975). High resolution monitoring of larger juveniles in the field has yielded similar data for several compound ascidians (Olson, 1983; Keough & Downes, 1986; Davis, 1987). Most of these studies indicated high habitat specific early mortality. All of those studies in which

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

55

Fig 11.—Cluster dendrogram of correlation coefficients of recruitment over a 12-year period for Ascidia mentula at three stations each at two depth levels in Gullmarsfjorden on the Swedish west coast. Redrawn from Svane (1984).

mortality was measured in different seasons or in different years showed high temporal variation. The causes of this temporal variation in survival have not been explored adequately in any system. Because temperature is correlated with the timing of reproduction in many ascidian species (Millar, 1971), populations of the same species living under different temperature regimes might be expected to either recruit at different times or demonstrate differences in the duration of the reproductive period. This is reflected in the size-frequency distributions of Scandinavian populations of Ciona intestinalis. Shallowwater populations which experience relatively high summer temperatures consist of only a single generation at a time, whereas deeper populations living where temperatures are relatively low and stable may have up to three generations represented in the age structure, because of multiple recruitment events (Dybern, 1963; Millar, 1971; Svane, 1983). Synchronous recruitment of Podoclavella moluccensis has been noted by Davis & Butler (in press) at several sites in South Australia. However, Keough (1983) compared recruitment of Ciona intestinalis, Botrylloides leachii, Didemnum sp. and other sessile animals at two sites in Southern Australia separated by a distance of approximately 100 km and found no synchrony between sites. Svane (1983, 1984, 1988) found good synchrony in recruitment of Ascidia mentula, Ciona intestinalis, and Ascidiella sp. between sites separated by distances of 110 km and 9 km in Sweden. When comparing 12 years of recruitment patterns from stations in Gullmarsfjorden and in the archipelago off the Swedish west coast, Svane (1984) showed that recruitment correlated significantly between stations in the inner fjord and that portion of the central fjord greater than 20 m deep, while the station in the exposed archipelago correlated significantly with the shallow (15 m) station in the central portion of the fjord (Fig 11). This pattern was explained by the hydrographic properties of the area which may cause entrapment of eggs and larvae for long periods of time. Although temperature or other factors may help to explain the absence of recruitment synchrony between widely separated sites, many other factors (e.g., larval mortality; aggregation, local physical factors) are likely to be important; as always, correlation is fraught with the danger of misinterpretation! The solitary ascidians on subtidal rock walls of the Swedish west coast (Svane & Lundälv, 1981, 1982; Svane, 1983) have been monitored more intensively than those at any other site. Over a 12-year period, repeated sampling revealed major differences in the longevity and recruitment dynamics of four major species, Ciona intestinalis, Boltenia echinata, Ascidia mentula, and Pyura tessellata. The most common annual species in the system, Ciona intestinalis, fluctuated widely from year to year (Svane, 1983), whereas the longest-lived species, Pyura tessellata, had a constant population size over the entire study (Svane & Lundälv, 1982; Svane, 1983). Those species with the most stable populations also had low fecundity and recruitment. Solitary ascidians with relatively short (less than two years) generation time such as Corella inflata (Lambert, 1968), Ascidia nigra (Goodbody, 1962), and Ciona intestinalis (Dybern, 1963; Gulliksen, 1972; Svane, 1983) often demonstrate large local variations in population size from year to year. Current paradigms would predict that such species should be excellent colonisers, having high fecundities and fast growth to reproductive maturity (Jackson, 1977). These species are “weedy”; by producing many gametes, they are likely to have larvae present whenever an ecological opportunity (e.g. in the form of a new substratum) becomes available for settlement. Many short-lived species are poor competitors for space (Goodbody, 1965; Lambert, 1968) or are readily preyed upon due to their thin, gelatinous tunic (Lambert, 1968; Gulliksen & Skjaeveland, 1973; Young, 1986). Lambert (1968) has suggested that the brooding mechanism of Corella inflata is a method for producing multiple generations at a good site before individuals

56

IB SVANE AND CRAIG M.YOUNG

begin to lose ground to superior competitors such as compound ascidians (but see Young, 1988, for an alternative hypothesis). Annual, or short lived, ascidians are commonly found in fouling communities occurring on submerged hard substrata such as experimental panels (e.g. Sutherland, 1974; Field, 1982; Kay & Butler, 1983; Todd & Turner, 1986, and many others). Recruitment onto these hard substrata is highly unpredictable and the communities themselves are highly unstable both temporally and spatially (Dayton, 1984). Longer lived ascidians may, however, stabilise these communities for a period of time (Sutherland, 1981). Recruitment onto fouling panels and subsequent community development is influenced strongly by substratum size, quality, and the orientation of the panels (e.g., Keough, 1983; Todd & Turner, 1986). The developing communities are rarely comparable to those found on natural hard substrata since different organising forces operate (Svane, 1988). Kay & Keough (1981) found that relatively small isolated patches (the pen shell Pinna bicolor) were colonised mainly by larval recruitment while cleared patches on pilings nearby were colonised by the vegetative extension of adjacent sponges and colonial ascidians. Submerged artificial hard substrata may therefore be regarded as larval filters in which the effective pore size is determined by the physical properties of the material and by the surrounding environment, on both small and large scales. Fecundity is related inversely to tunic dry weight (Svane, 1983). Thus, species with low fecundity (e.g. Pyura tessellata) compensate by increasing longevity (more than 11 years) with a tough, protective adult tunic. Although these species seem to recruit rarely because of their low fecundity, their populations are often more stable than populations of species with higher recruitment potential that risk more of their available energy by sending larvae into the plankton. EPILOGUE Over the past half century, studies of ascidian tadpole larvae have progressed along many fronts. In recent years, ultrastructural studies have shed light on attachment, locomotory, and behavioural mechanisms, and behavioural studies indicate functions for known structures. Studies of long-term recruitment dynamics integrated with laboratory studies of larvae show how various life-history stages integrate to produce temporal and spatial patterns of variation which are of interest to the ecologist. Ascidians have proved ideal for general studies of larvae for many reasons, among which are short larval life, lecithotrophy (larvae need not be fed in culture), diversity of behaviours and structures, and ease of obtaining material for experiments. The large size of some colonial ascidian larvae allows us to follow, observe, and manipulate larvae in their natural environment. Logistic problems prohibit the use of such direct methods in most other invertebrate groups. Although in the past five years a rapid development in the use of in situ techniques has been seen, there remains a variety of important and unresolved questions in ascidian larval ecology specifically and invertebrate larval ecology in general that could be addressed using ascidian tadpole larvae. ACKNOWLEDGEMENTS We wish to thank our colleagues for valuable comments and criticism. We are particularly indebted to Brian Bingham, Lane Cameron, Andy Davis, Jon Havenhand, and Randy Olson for providing assistance and valuable information. One of us (I.Svane) is grateful to Dr T.Brattegård and the Marine Biological Station, University of Bergen, for providing excellent working facilities during the preparation of this work. The study was supported by the Swedish Natural Science Research Council, contract no. B-BU 8526–300 to I.Svane and in part by National Science Foundation grant no. OCE-8400406 to C.M.Young. The project was conceived and completed during reciprocal transatlantic visits supported by our respective institutions.

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

57

REFERENCES Abbott, D.P., 1955. J. Morphol., 97, 569–593. Barnes, S.N., 1971. Z. Zellforsch. Mikrosk. Anat., 11, 71–16. Barnes, S.N., 1974. Cell Tissue Res., 155, 27–45. Batham, E.J., 1956. Trans. R. Soc. N.Z., 84, 447–465. Berrill, N.J., 1929. Phil. Trans. R. Soc. Ser. B., 218, 37–78. Berrill, N.J., 1931. Phil. Trans. R. Soc. Ser. B., 219, 281–346. Berrill, N.J., 1935. Phil. Trans. R. Soc. Ser. B., 225, 255–326. Berrill, N.J., 1947. J. Mar. Biol. Assoc. U.K., 26, 616–625. Berrill, N.J., 1948a. Q. J. Microsc. Sci., 89, 253–289. Berrill, N.J., 1948b. J. Morphol, 82, 269–285. Berrill, N.J., 1949. J. Mar. Biol. Assoc. U.K., 28, 633–650. Berrill, N.J., 1950. The Tunicata. The Ray Society, London, 354 pp. Berrill, N.J., 1955. The Origin of Vertebrates. Clarendon Press, Oxford, 257 pp. Berrill, N.J., 1957. In, Plymouth Marine Fauna, Marine Biological Association of the United Kingdom, Plymouth, pp. 371–382. Berrill, N.J., 1975. In, Reproduction of Marine Invertebrates, Vol. II, edited by A.C. Giese & J.S.Pearse, Academic Press, New York, pp. 241–282. Brewin, B.I., 1959. Q. J. Microsc. Sci., 100, 575–589. Burke, R.D., 1983. Can. J. Zool., 61, 1701–1719. Butler, A.J., 1986. J. Exp. Mar. Biol. Ecol., 97, 13–36. Butman, C.A., 1987. Oceanogr. Mar. Biol. Annu. Rev., 25, 113–165. Castle, W.E., 1896. Bull. Mus. Comp. Zool., 27, 203–280. Chernoff, H., 1985. MS Thesis, Florida State University, Tallahassee, 136 pp. Chia, F.-S., Buckland-Nicks, J. & Young, C.M., 1984. Can. J. Zool., 62, 1205–1222. Chia, F.S., 1974. Thal. Jugosl., 10, 121–130. Child, C.M., 1927. J. Morphol. Physiol., 44, 467–517. Cloney, R.A., 1959.PhD Dissertation, University of Washington, Seattle, 111 pp. Cloney, R.A., 1964. Acta Embryol. Morphol. Exp., 7, 111–130. Cloney, R.A., 1977. Cell Tissue Res., 183, 423–444. Cloney, R.A., 1978. In, Settlement and Metamorphosis of Marine Invertebrate Larvae, edited by F.-S.Chia & M.Rice, Elsevier North-Holland Biomedical Press, Amsterdam, pp. 255–282. Cloney, R.A., 1982. Am. Zool., 22, 817–826. Cloney, R.A., 1987. In, Reproduction and Development of Marine Invertebrates of the Northern Pacific Coast, edited by M.F.Strathmann, University of Washington Press, Seattle, pp. 607–639. Cloney, R.A. & Cavey, M.J., 1982. Cell Tissue Res., 222, 547–562. Conklin, E.G., 1905, J. Acad. Nat. Sci. Philadelphia, 8, 1–119. Connell, J.H., 1985. J. Exp. Mar. Biol. Ecol., 93, 11–45. Crisp, D.J. & Ghobashy, A.F.A.A., 1971. In, Proceedings of the Fourth European Marine Biology Symposium , edited by D.J.Crisp, Cambridge University Press, pp. 443–465. Dakin, W.J., Bennett, I., & Pope, E., 1948. Aust. J. Sci. Res. Ser. B, 1, 173–230. Damas, D., 1904. Arch. Biol., 20, 745–833. Davis, A.R., 1987. J. Exp. Mar. Biol. Ecol., 106, 57–71. Davis, A.R., 1988. J. Exp. Mar. Biol. Ecol., 117, 157–167. Davis, A.R., in press. In, 23rd European Marine Biology Symposium, edited by J.S. Ryland & P.A.Tyler, Olsen & Olsen, Helsingør, Denmark. Davis, A.R. & Butler, A.J., in press. J. Exp. Mar. Biol. Ecol. Day, R.W., 1974. Zool. Afr., 9, 35–58.

58

IB SVANE AND CRAIG M.YOUNG

Dayton, P.K., 1984. In, Ecological Communities. Conceptual Issues and the Evidence, edited by D.Strong et al., Princeton University Press, Princeton, pp. 181–197. Dilly, P.N., 1961. Nature (London), 191, 786–787. Dilly, P.N., 1962. Q. J. Micr. Sci., 103, 393–398. Dilly, P.N., 1964. Q. J. Micr. Sci., 105, 13–20. Dilly, P.N., 1969. Z. Zellforsch Mikrosk. Anat., 95, 331–346. Duyl, F.C.van, Bak, R.P.M. & Sybesma, J., 1981. Mar. Ecol. Prog. Ser., 6, 35– 42. Dybern, B.I., 1963. Zool. Bidr. Uppsala, 35, 589–602. Eakin, R.M. & Kuda, A., 1971. Z. Zellforsch. Mikrosk. Anat., 112, 287–312. Emanuelsson, H. & Mecklenburg, C.von, 1974. Cell Tissue Res., 148, 27–44. Emlet, R.B., McEdward, L.R. & Strathmann, R.R., 1987. In Echinoderm Studies, Vol. 2, edited by M.Jangoux & J.M.Lawrence, Balkema, Rotterdam, pp. 55– 136. Field, B., 1982. J. Exp. Mar. Biol. Ecol., 57, 25–33. Fraenkel, G.S. & Gunn, D.L., 1940. The Orientation of Animals. Oxford Press, Oxford, 352 pp. Garstang, W. & Garstang, S.L., 1928. Q. J. Miscrosc. Sci., 72, 1–50. Goodbody, L, 1961. Proc. Zool. Soc. London, 136, 403–109. Goodbody, L, 1962. Biol. Bull. (Woods Hole, Mass.), 122, 40–51. Goodbody, I., 1963. Biol. Bull. (Woods Hole, Mass.), 124, 31–4. Goodbody, I., 1965. Biol. Bull. (Woods Hole, Mass.), 129, 128–133. Goodbody, I. & Fisher, E., 1974. Biol. Bull. (Woods Hole, Mass.), 146, 206–216. Goodbody, I. & Gibson, J., 1974. Biol. Bull. (Woods Hole, Mass.), 146, 217–237. Gotelli, N.J., 1987. Mar. Biol., 94, 45–51. Grave, C. & McCosh, G.K., 1924. Wash. Univ. Stud. Scient. Ser., 11, 89–117. Grave, C.A., 1920. J. Exp. Zool., 30, 239–257. Grave, C.A., 1921. J. Morphol., 36, 71–101. Grave, C.A., 1926. J. Morphol., 42, 453–71. Grave, C.A., 1934. Pap. Tortugas Lab., 28 143–156. Grave, C.A., 1935. Pap. Tortugas Lab., 29, 211–291. Grave, C.A., 1936. Pap. Tortugas Lab., 29, 209–291. Grave, C.A., 1941. Biol. Bull. (Woods Hole, Mass.), 81, 287. Grave, C.A., 1944. J. Morphol., 75, 173–191. Grave, C.A. & Nicoll, P.A., 1939. Pap. Tortugas Lab., 32, 1–16. Grave, C.A. & Riley, G., 1935. J. Morphol., 57, 185–211. Grave, C.A. & Woodbridge, H., 1924. J. Morphol., 39, 207–247. Grosberg, R.K., 1981. Nature (London), 290, 700–702. Grosberg, R.K. & Quinn, J.F., 1986. Nature (London), 322, 456–459. Gulliksen, B., 1972. Sarsia, 51, 83–96. Gulliksen, B. & Skjaeveland, S.H., 1973. Sarsia, 52, 15–20. Gutierrez, M. & Lay, S., 1965. Estud. Oceanol, 1, 1–31. Hashimoto, K. & Watanabe, H., 1982. Proc. Jpn. Acad., 58, 9–12. Havenhand, J.N. & Svane, I., in press. In, 23rd European Marine Biology Symposium, edited by J.S.Ryland & P.A.Tyler, Olsen & Olsen, Helsingør, Denmark. Heath, D.J., 1977. J. Theor. Biol., 64, 363–373. Hecht, S., 1918. J. Gen. Physiol., 1, 147–166. Hirai, E. & Tsubata, B., 1956. Bull. Biol. Stn. Asamushi, 8, 1–4. Huus, J., 1939. Avh. Nor. Vidensk. Akad. Oslo, Mat. Naturvidensk. Kl., 4, 5–49. Huus, J., 1941. Avh. Nor. Vidensk. Akad. Oslo, Mat. Naturvidensk. Kl., 8, 1–13. Jackson, G.A. & Strathmann, R.R., 1981. Am. Nat., 118, 16–25. Jackson, J.B.C., 1977. Am. Nat., 111, 743–767.

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

Jackson, J.B.C., 1986. Bull. Mar. Sci., 39, 588–606. Jansen, W.F. & Flight, W.F.G., 1969. Z. Zellforsch. Mikrosk. Anat., 100, 439–465. Kajiwara, S. & Yoshida, M., 1985. Biol. Bull. (Woods Hole, Mass.), 169, 565–577. Katz, M. J., 1983. Biol. Bull. (Woods Hole, Mass.), 164, 11–27. Kay, A.M. & Butler, A.J., 1983. Oecologia (Berlin), 56, 70–78. Kay, A.M. & Keough, M.J., 1981. Oecologia (Berlin), 48, 123–130. Keough, M.J., 1983. J. Exp. Mar. Biol. Ecol, 66, 213–245. Keough, M.J. & Downes, B.J., 1982. Oecologia (Berlin), 54, 348–352. Keough, M.J. & Downes, B.J., 1986. Mar. Ecol. Prog. Ser., 33, 279–285. Kingsley, J.S., 1882. Proc. Boston Soc. Nat. Hist., 21, 440–451. Knaben, N., 1952. Avh. Nor. Vidensk-Akad. Oslo, Mat.-Naturwidensk. Kl., 3, 1–24 Kott, P., 1981. Proc. 4th Int. Symp. Coral Reefs, 2, 721–723. Kowalevsky, A.O., 1866. Mem. Acad. Sci. St. Petersbg., 10, 1–19. Lafargue, F. & Duclaux, G., 1979. Ann. Inst. Océanogr. Paris, 55, 163–184. Lambert, C.C., 1982. Am. Zool, 22, 841–849. Lambert, C.C. & Brandt, C.L., 1967. Biol. Bull. (Woods Hole, Mass.), 132, 222–228. Lambert, C.C. & Lambert, G., 1978. Science, 200, 64–65. Lambert, C.C. & Lambert, G., 1981. J. Exp. Zool., 217, 291–295. Lambert, G., 1968. Biol. Bull. (Woods Hole, Mass.), 135, 296–307. Lambert, G., 1979. Biol. Bull. (Woods Hole, Mass.), 157, 464–477. Lambert, G., Lambert, C. & Abbott, D.P., 1981. Can. J. Zool., 59, 1493–1504. Lanzing, W.J.R. & Lennep, E.W.van, 1971. Aust. J. Zool., 18, 353–371. Levine, E.P., 1962. J. Morphol, 111, 105–138. Lyerla, T.A. & Lyerla, J.H., 1978. Comp. Biochem. Physiol., 59B, 111–116. Lynch, W.F., 1961. Am. Zool., 1, 59–66. Mancuso, V. & Dolcemascolo, G., 1977. Acta Embryol Exp., 2, 207–220. Mast, S.O., 1921. J. Exp. Zool., 34, 149–187. Menge, B.A., 1972. Ecol. Monogr., 42, 25–50. Menge, B.A., 1975. Mar. Biol., 31, 87–100 Menge, B.A., 1976. Ecol. Monogr., 46, 355–393. Millar, R.H., 1953. Ciona. L.M.B.C. Mem. XXXV. University Press, Liverpool, 123 pp. Millar, R.H., 1963. Proc. Zool. Soc. London, 141, 689–746. Millar, R.H., 1964. Vidensk. Medd. Dan. Naturhist. Foren., 127, 159–180. Millar, R.H., 1971. The biology of ascidians. Adv. Mar. Biol., 9, 1–100. Miller, R.L., 1975. Nature (London), 254, 244–245. Miller, R.L., 1982. Am.Zool., 22, 827–840. Monniot, C., 1965. Vie Milieu, 16, 819–849. Moore, P.G., 1975. Mar. Behav. Physiol., 3, 97–100. Morgan, T.H., 1938. J. Exp. Zool., 78, 271–334. Morgan, T.H., 1942. Biol. Bull (Woods Hole, Mass.), 82, 161–171. Morgan, T.O., 1977. MSc thesis, University of Puerto Rico, Mayagüez Puerto Rico, 96pp. Nomaguchi, T.A., 1974. Exp. Gerontol., 9, 231–234. Oka, H., 1943. Annot. Zool Jpn., 22, 54–58 Olson, R.R., 1980. Am. Zool., 20, 778. Olson, R.R., 1983. Biol. Bull (Woods Hole, Mass.), 165, 221–240. Olson, R.R., 1985. Ecology, 66, 30–39. Olson, R.R. & McPherson, R., 1987. J. Exp. Mar. Biol. Ecol., 110, 245–256. Paine, R.T. & Suchanek, T.H., 1983. Evolution, 37, 821–831. Palmer, A.R. & Strathmann, R.R., 1981. Oecologia (Berlin), 48, 308–318.

59

60

IB SVANE AND CRAIG M.YOUNG

Pearse, J.S. & Beauchamp, K.A., 1986. Int. J. Invert. Reprod. Dev., 9, 289–297. Pearse, J.S. & Eernisse, D.J., 1982. Mar. Biol., 67, 121–125. Pearse, J.S. & Walker, C.W., 1986. Int. J. Invert. Reprod. Dev., 9, 71–77. Reese, J.P., 1967. MS thesis, San Diego State College, San Diego, California, 81 pp. Reverberi, G., 1979. Acta Embryol. Exp., 1, 91–99. Robinson, W.E., Kusten, K. & Cloney, R.A., 1986. J. Exp. Zool., 237, 63–72. Rose, S.M., 1939. Biol. Bull. (Woods Hole, Mass.), 77, 216–232. Roughgarden, J., Iwasa, J., & Baxter, C., 1985. Ecology, 66, 54–67. Schmidt.G.H., 1982a. Mar. Ecol. Prog. Ser., 9, 97–100. Schmidt, G.H., 1982b. Biol. Bull. (Woods Hole, Mass.), 162, 195–201. Schmidtke, J., Weiler, C., Kunz, B. & Engel, W., 1977. Nature (London), 266, 532– 533. Schofield, V.L., Schlumpberger, J.L., West, L.A., & Weismann, I.L., 1982. Nature (London), 295, 499–502. Scott, F.M., 1946. Biol. Bull. (Woods Hole, Mass.), 91, 66–80. Scott, F.M., 1954. J. Exp. Zool., 127, 331–366. Sebens, K.P. & Koehl, M.A.R., 1984. Mar. Biol., 81, 255–271. Shields, W.M., 1982. Philopatry, Inbreeding, and the Evolution of Sex. State University of New York Press, Albany, 245 pp. Stephenson, T.A., 1942. Proc. Linn. Soc. London, 54, 219–232. Stoecker, D., 1978. Biol Bull. (Woods Hole, Mass.), 155, 615–626. Stoecker, D., 1980. Ecology, 61, 1327–1334. Strathmann, R.R., 1985. Annu. Rev. Ecol. Syst., 16, 339–361. Strathmann, R.R. & Strathmann, M.F., 1982. Am. Nat., 119, 91–101. Strathmann, R.R., Strathmann, M.F. & Emson, R.H., 1984. Am. Nat., 123, 796– 818. Sutherland, J.P., 1974. Am. Nat., 108, 859–873. Sutherland, J.P., 1981. Am. Nat., 118, 499–519. Svane, L, 1982. Acta Zool. (Stockholm), 63, 85–89. Svane, L, 1983. Sarsia, 68, 249–255. Svane, L, 1984. Biol. Bull. (Woods Hole, Mass.), 167, 630–646. Svane, I., 1987. Ophelia, 27, 87–100. Svane, L, 1988. Ophelia, 29, 25–41. Svane, I., Havenhand, J.N. & Jrgensen, A.J., 1987. J. Exp. Mar. Biol. Ecol., 110, 171–181. Svane, I. & Lundälv, T., 1981. J. Exp. Mar. Biol. Ecol., 50, 163–182. Svane, I. & Lundälv, T., 1982. Sarsia, 67, 249–257. Thorson, G., 1950. Biol. Rev., 25, 1–45. Thorson, G., 1964. Ophelia, 1, 167–208. Thorson, G., 1966. Neth. J. Sea Res., 3, 267–293. Todd, C.D. & Turner, S.J., 1986. J. Exp. Mar. Biol. Ecol., 99, 199–231. Torrence, S.A. & Cloney, R.A., 1982. Zoomorphology, 99, 103–115. Torrence, S.A. & Cloney, R.A., 1983. Zoomorphology, 102, 111–123. Torrence, S.A. & Cloney, R.A., 1988. In, Marine Biodeterioration, edited by M.F.Thompson, R.Sarojini & R.Nagabhushanam, Oxford & IBH Publishing Company, New Delhi, pp. 150–163. Trason, W.B., 1957. J. Morphol., 100, 5–45. Trason, W.B., 1963. Univ. Calif. Berkeley Publ. Zool., 65, 81–113. Turon, X., 1988. Bull. Inst. Cat. Hist. Nat., 55, 109–118. Underwood, A.J. & Denley, E.J., 1984. In, Ecological Communities: Conceptual Issues and the Evidence, edited by D.Strong et al., Princeton University Press, Princeton, pp. 151–180. Underwood, A.J. & Fairweather, P.G., 1986. Proc. Ecol. Soc. Aust., 14, 7–13. Vance, R.R., 1973. Am. Nat., 107, 339–352. Vogel, S., 1981. Life in Moving Fluids. Princeton University Press, Princeton, 352 pp.

THE ECOLOGY AND BEHAVIOUR OF ASCIDIAN LARVAE

61

Vorontsova, M.N. & Malakhov, V.V., 1984. Zool. Zh., 63, 1036–1045. Watanabe, H. & Lambert, C.C., 1973. Biol. Bull. (Woods Hole, Mass.), 144, 556– 566. Watanabe, H. & Tokioka, T., 1972. Publ. Seto Mar. Biol. Lab., 19, 327–345. West, A.B. & Lambert, C.C., 1976. J. Exp. Zool., 195, 263–270. Whittaker, J.R., 1966. Dev. Biol., 14, 1–39. Whittaker, J.R., 1987. Am. Zool., 27, 607–622. Whittingham, D.G., 1967. Biol. Bull. (Woods Hole, Mass.), 132, 292–298. Woodbridge, H., 1924. Biol. Bull. (Woods Hole, Mass.), 47, 223–230. Woollacott, R.M., 1979. In, Reproductive Ecology of Marine Invertebrates, edited by S.E.Stancyk, University of South Carolina Press, Columbia, pp. 157–172. Yamaguchi, M., 1975. Mar. Biol., 29, 253–259. Young, C.M., 1978. MS Thesis, Brigham Young University, Utah, 98 pp. Young, C.M., 1982. PhD dissertation, University of Alberta, 260 pp. Young, C.M., 1983. Am. Zool., 23, 1021. Young, C.M., 1984. In, Proceedings of the Fifth International Echinoderm Conference, Galway, edited by B.F.Keegan & B.D.S.O’Connor, Balkema, Rotterdam, pp. 577–583. Young, C.M., 1985. Mar. Biol., 84, 309–321. Young, C.M., 1986. Bull. Mar. Sci., 39, 279–289. Young, C.M., 1988. J. Exp. Mar. Biol. Ecol., 117, 9–26. Young, C.M., in press. Bull. Mar. Sci. Young, C.M. & Bingham, B.L., 1987. Mar. Biol., 96, 539–544. Young, C.M. & Braithwaite, L.F., 1980a. J. Exp. Mar. Biol. Ecol., 42, 157–169. Young, C.M. & Braithwaite, L.F., 1980b. Biol Bull. (Woods Hole, Mass.), 159, 428–440. Young, C.M. & Chia, F.-S., 1984. Mar. Biol., 81, 61–68. Young, C.M. & Chia, F.-S., 1985. J. Exp. Mar. Biol. Ecol., 85, 165–175. Young, C.M. & Chia, F.-S., 1987. In, Reproduction of Marine Invertebrates, Vol. 9: General Aspects: Seeking Unity in Diversity, edited by A.C.Giese et al. Blackwell/Boxwood, Palo Alto and Pacific Grove, California, pp. 384–463. Young, C.M., Gowan, R.F., Dalby Jr, J., Pennachetti, C.A. & Gagliardi, D., 1988. Biol. Bull (Woods Hole, Mass.), 174, 39–6.

Oceanogr. Mar. Biol. Annu. Rev., 1989, 27, 91–166 Margaret Barnes, Ed. Aberdeen University Press

EGG PRODUCTION IN CIRRIPEDES MARGARET BARNES The Scottish Marine Biological Association, The Dunstaffnage Marine Research Laboratory, Oban, Argyll, PA34 4AD, Scotland

ABSTRACT gE g production in three orders, Acrothoracica, Rhizoce phala, and Thoracica has been dealt with in this review. The Acrothoracica are burrowing forms, the Rhizocephala are parasitic, and the Thoracica the true barnacles. uM ch of the information on egg production concerns the Thoracica. There are, however, scattered references to egg production in the other two orders. Cirripedes may be hermaphroditic or have separate sexes in which case dwarf males are needed for fertilisation. In some cases hermaphrodites have complemental or apertural males. Transfer of spermatozoa, ac tivated by secretions of the oviducal gland, penetrate the oviducal sac and fertilisation results. gE gs a re developed outside the body but within the mantle cavity of the adult. The relative virtues of self- and cross-fertilisation have been discussed. Animals producing lecithotrophic nauplii usually have broods of fewer than usual eggs of a larger siz e than normal. The rhiz ocephalans are an exception; their eggs are small and numerous. Acrothoracicans also have lecithotrophic nauplii but the eggs are bigger and fewer than in the rhizoce phalans. Some lepadomorphs have lecithotrophic larvae and within the balanomorphs the teT racl ita species produce some anomalies; three species are lecithotrophic and the rest, as far as is known, are planktotrophic. At high latitudes and low temperature the fertilisation of eggs and release of nauplii has to be carefully controlled with probably only one brood a year. At lower latitudes, where temperature is higher and environmental conditions remain favourable longer, many broods can be produced in quick succe ssion. INTRODUCTION As members of the Crustacea, cirripedes have an exclusively sessile habit either as free-living filter-feeders or as highly specialised parasites. They may be estuarine or marine and are found at all depths of the ocean in a wide variety of habitats. The sub-class Cirripedia (a member of the Maxillopoda, see e.g. Newman,

EGG PRODUCTION IN CIRRIPEDES

63

Zullo & Withers, 1969) was traditionally regarded to comprise the orders Apoda, Ascothoracica, Thoracica, Acrothoracica, and Rhizocephala. The Apoda, represented by the parasite Protolepas bivincta, was removed from the Cirripedia after Bocquet-Védrine (1972a, 1979) showed that the parasite was an epicaridean. The Ascothoracica encompasses primitive crustaceans living as parasites of echinoderms and coelenterates and are generally found in deep water. They are suctorial rather than cirral feeding, the nauplii lack frontal horns (Zullo, 1979) and all stages are capable of feeding; this is not true of the other orders. The sperm morphology differs from that of the remainder of the cirripedes (Grygier, 1982, 1983a). There has also been some discussion about the position of the gonopores and whether carapace pores could be homologous to frontal horn glands (Grygier, 1983b). Hallberg, Elofsson & Grygier (1985) have now found that the ascothoracid larva described by Grygier (1983c) does indeed have functional compound eyes suggesting a close relationship to the Cirripedia. There is still, however, only a limited amount of material available for such comparisons. In view of the controversy that has been apparent for some time about the position of the Ascothoracica—the most recent account being that of Grygier (1987)—it seems reasonable at present to keep it separate from the Cirripedia but as a sub-class of the Maxillopoda (Newman, 1982). This account of egg production in cirripedes will, therefore, only be concerned with the orders Thoracica, Acrothoracica, and Rhizocephala. The Acrothoracica represent burrowing forms, the Rhizocephala parasitic forms without appendages or digestive tract, and the Thoracica the true barnacles. Most of the information regarding egg production concerns the last of these three orders. There are, however, numerous references, many of them anecdotal, to egg production in the other orders and an attempt will be made to summarise these. Systematics and nomenclature are not the concern of this article and it is, therefore, the intention to use the names given by the various workers in their published papers. GENERAL Cirripedes may be either hermaphroditic or have separate sexes (Darwin, 1851, 1873; Batham, 1945a; Henry & McLaughlin, 1965, 1967; McLaughlin & Henry, 1972; Gomez, 1974; Foster, 1978, 1983; Newman, 1980; Dayton, Newman & Oliver, 1982; Crisp, 1983; Hui & Moyse, 1984). When the sexes are separate the males are referred to as dwarf males. In some cases hermaphrodites may also have one or more males attached in which case they are called complemental males. Crisp (1983) introduced the term apertural males for individuals he found settled on the operculum of Chelonobia patula. Similar males have been found in Chirona tennis (Zevina & Poliakova, 1986). These apertural males are capable of feeding and appear to be potential hermaphrodites in which development has been arrested. The determination of sex in cyprids of barnacles carrying dwarf or complemental males has received some attention over the years as can be seen from the work of the following authors: Kühnert (1934), Callan (1941), Veillet (1956, 1961), Yanagimachi (1961a), Gomez, Faulkner, Newman & Ireland (1973), Gomez (1974, 1975), Høeg (1984), and Walker (1985, 1987). Complemental and dwarf males are greatly reduced and consist of little more than the reproductive structures; their comparative anatomy has been reviewed recently by Klepal (1987). They may be attached in or about the external aperture of a normal hermaphrodite or female. At maturity spermatozoa are introduced into the visceral tissue or mantle cavity (body chamber) of the mature hermaphrodite or female. The transfer of spermatozoa may be by an injection of cells from the male in the Rhizocerphala (see e.g. Ichikawa & Yanagimachi, 1958; Reischman, 1959) or via a penis in the Acrothoracica and Thoracica (first seen by R. Bishop and recorded by Bate, 1869). In some cases the penis is capable of great extension in length for this purpose. Because semen is introduced into a functional female near the oviducal opening but

64

MARGARET BARNES

not into the female genital tract exception may be taken to the use of the word copulation; it has, however, been widely used although some people prefer the term pseudo-copulation. In the Acrothoracica and Thoracica oviducts running from the ovaries enter the mantle cavity at the base of the first cirri. Movement of ova from the ovaries via the oviducts is stimulated by semen covering the exit of the oviducts in the mantle cavity. Spermatozoa, if not already motile, are activated by secretions from the oviducal gland and the oviducal sac containing the ova is penetrated (Walley, 1965; Barnes, Barnes & Klepal, 1977; Klepal & Barnes, 1977; Klepal, Barnes & Barnes, 1977; Walker, 1977a, 1980). Fertilisation takes place at this time and in all cirripedes the eggs develop external to the body but within the mantle cavity of the adult. In what follows the term egg will be used for the fertilised ovum plus its nutritive and protective tissues and from which, in the case of cirripedes, one of the planktonic stages emerges. The embryo will be regarded as the young organism in its stages of development within the egg. Within the eggs of cirripedes the embryos develop, eventually hatch and the young leave the mantle cavity. At this time it may be as a Stage I nauplius or, in some cases, as a cyprid. There are several nauplius stages (up to six) followed by a cyprid stage which eventually settles and becomes the adult animal. In some Rhizocephala the cyprid stage is followed by a kentrogon stage (which produces the reproductive adult) or a trichogon stage (Høeg, 1987). FERTILISATION Cross-fertilisation is obligatory in most cirripedes although several instances of self-fertilisation are known. Self-fertilisation may occur in some species when individuals are isolated by distances greater than the length of the extended penis (see e.g. Barnes & Crisp, 1956; Barnes & Barnes, 1958). In such cases parthenogenesis or activation by water-borne spermatozoa must be shown to be absent. The former could be eliminated by detailed cytological investigations. This has never been done but because male and female gonads are ripe simultaneously and cross-fertilisation is normal in the same species, parthenogenesis is unlikely. Fertilisation by water-borne spermatozoa also appears improbable (Barnes & Crisp, 1956). First, there is no decrease in the incidence of egg masses in isolated animals with increasing distance of isolation. Secondly, seminal fluid is rarely seen to be emitted except as a result of the stimulus of copulation and then takes place within the mantle cavity when the semen at first coagulates on contact with sea water. Thirdly, oviposition itself normally only occurs in association with the copulatory stimulus. Self-fertilisation, therefore, remains the only possible mechanism. The percentage of a population whose egg masses have reached a given stage of development is always greater in contiguous compared with isolated individuals, indicating that in the latter oviposition may be delayed. This delay is almost certainly dependent on the absence of the appropriate stimulus to oviposition associated with copulation. Eventually the threshold value of this stimulus must be reached and perhaps under these conditions movements of the penis within the mantle cavity are sufficient to initiate discharge of semen and to induce oviposition. Although many apparently viable nauplii may be obtained from the ripe egg masses contained in isolated, and presumably self-fertilised, individuals there are often many unsegmented and abnormal eggs. This suggests that self-fertilised eggs are less viable. The fact that the number of normal, viable nauplii from an animal of a given size is less when it is growing isolated than when it is in a position to be cross-fertilised also suggests that the latter is a more effective form of reproduction (Barnes & Crisp, 1956). Instances of self-fertilisation have been found in Verruca stroemia (Barnes & Crisp, 1956), Balanus amphitrite (Patel & Crisp, 1961), B. amphitrite communis (Pillay & Nair, 1972), B. balanus (Barnes & Barnes, 1954; Crisp, 1954), B. eburneus (Cheung & Nigrelli, 1969; Landau, 1976), B. improvisus

EGG PRODUCTION IN CIRRIPEDES

65

(E.Furman, pers. comm.), B. perforatus (Barnes & Crisp, 1956), B. trigonus (Werner, 1967), Octolasmis warwickii (Harker, 1975), and possibly in Platylepas ophiophilus (Zann, 1975) although this is not yet confirmed. There is evidence that several species of Chthamalus can self-fertilise (Barnes & Crisp, 1956; Barnes & Barnes, 1958; Tenerelli, 1958; Klepal & Barnes, 1975). In this genus many of the species extend to high intertidal levels and the capacity to resort ultimately to self-fertilisation may have a survival value. The population density at such levels may not always be sufficient to ensure proximity of individuals for mutual copulation. The advantages of hermaphrodites which normally cross-fertilise being able, in some circumstances, to self-fertilise have been discussed by Tomlinson (1966), Ghiselin (1969), and Sastry (1983). Ghiselin (1984) has discussed the use of dwarf or complemental males which can be regarded as a form of self-fertilisation when they originate from the adult to which they eventually become attached. In many instances it is assumed that this is not the case. REPRODUCTIVE GONADS The processes of spermatogenesis and oogenesis are outwith the scope of this review albeit spermatozoa and ova are necessary for egg production and deserve a mention. The most detailed experimental work on these processes has been done on members of the Thoracica. Early workers were aware that the spermatozoa of some cirripedes were finely filiform and capable of movement at some time in their life. In more recent times modern methods of study including scanning and electron microscopy have been used to extend our knowledge of cirripede spermatozoa and spermatogenesis. Rigo (1941) worked on Balanus amphitrite communis; Bocquet-Védrine & Pochon-Masson (1969) on B. perforatus; Munn & Barnes (1970a,b) and Barnes, Klepal & Munn (1971) on B. amphitrite amphitrite, B. balanoides, B. balanus, B. crenatus, B. eburneus, and B. perforatus, also Chthamalus stellatus, Verruca stroemia, and Elminius modestus; and Honma & Nakajima (1973) on Balanus eburneus. The relative inactivity of the spermatozoa of B. balanoides was commented on by several of these authors and also by Walley, White & Brander (1971) and Walker (1977b) who found that fluid from the oviducal gland stimulated activation at the time of oviposition. Turquier & Pochon-Masson (1969) worked on the spermatozoa of the acrothoracid Trypetesa (=Alcippe) nassarioides. References to the early work can be found in many of these recent papers. Gonad production involves biochemical synthesis with the formation of nucleic acids for spermatozoa and the mobilization of lipid and protein for ova. Food reserves may or may not be stored by the adult animal prior to gonad development. In boreo-arctic barnacles such as Balanus balanoides and B. balanus assimilation and storage of reserves begins in March to April at the time of the diatom increase (Barnes, Barnes & Finlayson, 1963). Gonad production in these species is relatively slow, beginning in early May and lasting until October in B. balanoides and early February in B. balanus. Final maturation takes place immediately before fertilisation (Barnes, Barnes & Klepal, 1977) in late October to November in the former species and February in the latter. The precise time of fertilisation varies by a few weeks depending on latitude. After oviposition and fertilisation these animals are devoid of spermatozoa and ova until feeding begins again. After fertilisation the penis of B. balanoides is shed in one or two moults (Klepal, Barnes & Barnes, 1975) and is gradually replaced, the length rapidly increasing from August to reach a maximum in October immediately prior to copulation (Barnes & Stone, 1972). In warm-water genera such as Chthamalus there are several cycles of gonad production in a year allowing for several broods of eggs (Crisp & Patel, 1969) in contrast to the one annual brood in boreo-arctic species. In general the behaviour of C. stellatus seems to be markedly independent of environmental conditions as far as food is concerned (Barnes, 1972). In some species the male gonads may remain active throughout the

66

MARGARET BARNES

year and ovaries may or may not carry mature ova or ova available for final maturation as soon as environmental conditions are favourable. In Balanus improvisus Blom (1965) reports large amounts of spermatozoa in winter but practically no oocytes— they begin to increase when sea-water temperature rises in April. In B. rostratus, however, oocytes are produced throughout the year according to Korn (1985) while male gonads only regenerate from June onwards and the animals fertilise in late September to October. A large number of oocytes degenerate before fertilisation; resorption of oocytes can also be seen in figures given for B. amphitrite and B. eburneus by Fyhn & Costlow (1975). Other species, although not breeding continuously, retain mature germinal cells throughout the year, reproduction being initiated as soon as conditions allow it. This appears to be so in B. perforatus and B. eburneus near Tarante, Italy (Lepore, Sciscioli & Gherardi, 1979) and in B. algicola in South Africa (Sandison, 1954). RHIZOCEPHALA The Rhizocephala are parasites of crustaceans, particularly decapods. The adult has a modified form and bears no resemblance to other cirripedes; the affinities are only shown by the larval stages and the spermatozoa. An up-to-date account of rhizocephalans has been given by Høeg & Lützen (1985). There are usually four nauplius stages, with the characteristic frontal horns but no alimentary canal, followed by a cyprid (Codreanu, 1959; Høeg, 1982). In some cases the hatching embryo is a nauplius and sometimes a cyprid. The more primitive rhizocephalans are ectoparasitic and the adult (reproductive) body develops in situ at the site of attachment of the cyprid to the host; there is no kentrogon stage. Such is the case in Duplorbis (Smith, 1906), Chthamalophilus delagei (Bocquet-Védrine, 1961), and Boschmaella balani (BocquetVédrine, 1968). The majority of rhizocephalans, however, develop from a kentrogon formed from the settled cyprid. The kentrogon penetrates the host and develops endoparasitically for several months, depending on the species. An adult reproductive body eventually emerges and becomes the externa while the internal nutritive structure (the interna) remains within the host’s body. This process has been known since the time of Delage (1884) but has recently been studied in detail by Høeg (1985a) for Lernaeodiscus porcellanae. The complete life history of this species has been determined by Ritchie & Høeg (1981) and Høeg & Ritchie (1985). The externa has a visceral mass, containing the ovaries, and is surrounded by a mantle. This is separated ventrolaterally from the visceral mass by a mantle cavity which develops as the externa matures. There is usually one mantle aperture at an extremity of the externa. Sylon, the sole genus of the Sylonidae, differs from most other rhizocephalans in having two mantle apertures (Lützen, 1981). In most rhizocephalans the aperture is open soon after the emergence of the externa but in some it may only open very late in the development of the externa or even when the embryos are ready for release. Such is the case in Clistosaccus paguri (Høeg, 1982, 1985b). In Thompsonia there is no mantle aperture (Reinhard & Stewart, 1956). Development of the externae depends on exposure to male cyprids. If virgin externae are kept in isolation and out of contact of male cyprids, they neither moult nor begin production of ova. In Lernaeodiscus porcellanae they can remain in this state almost indefinitely (Ritchie & Høeg, 1981). In Sacculina carcini virgin externae quickly die and drop off the host if no contact with male cyprids is made. Lützen (1981) has, however, reported ovulation in two full-grown Sylon hippolytes in which implantation of male cells had failed; the ova, however, did not develop. Rhizocephalans are now known to have separate sexes (see e.g. Reinhard, 1942a; Ichikawa &. Yanagimachi, 1958, 1960; Yanagimachi, 1961a,b; Yanagimachi & Fujimaki, 1967) although originally (Delage, 1884; Smith, 1906) they were thought to be hermaphroditic. Bocquet-Védrine (1961) described the

EGG PRODUCTION IN CIRRIPEDES

67

Chthamalophilidae as self-fertilising hermaphrodites and she later (Bocquet-Védrine, 1972b) expressed doubts about the function of male cyprids in Sacculina carcini. This has, however, now been settled by the work of Høeg (1984, 1987). He found that there are small (female) and large (male) cyprids in S. carcini. The female cyprids settle on the host and the males on the externae; this is in agreement with other rhizocephalans. Two sizes of cyprids were recognised by Veillet (1943a, 1945) in Triangulus galatheae and their separate roles were clarified by Yanagimachi (1961a,b) during experiments with Peltogasterella gracilis. A size difference (greater in some cases than others) between male and female cyprids is now established in the Peltogastridae and the Lernaeodiscidae (Ritchie & Høeg, 1981), the Sylonidae and the Clistosaccidae (Lùtzen, 1981; Høeg, 1982), as well as in the Sacculinidae (Høeg, 1984). Invasion by cells from male cyprids is necessary for further development of externae (see above). This invasion is generally thought to be through .the mantle aperture into one or two male cell receptacles, the testes (Ichikawa & Yanagimachi, 1960; Høeg, 1985b; Høeg & Ritchie, 1985). Høeg (1987) found a trichogon stage produced by male cyprids and has described what happens between settlement of the cyprid on an externa and the arrival of male cells in the receptacle of Sacculina carcini. A trichogon stage is present in the Sacculinidae, the Lernaeodiscidae and possibly in the Peltogastridae. A few genera lack receptacles, e.g. Mycetomorpha, Thompsonia, and Sylon. In those species that lack a mantle opening, or it only opens after spermatogenesis, the male cells must be injected through the integument and mantle of the virgin externa. Ichikawa & Yanagimachi (1958, 1960) do not regard the male “testis” in Sylonidae as a testis in the true sense of the word but rather that there is a female organ serving to hold a mass of male cells from the transformed male cyprid. There is no special receptacle for these cells and where they are found varies according to the genus. Spermatogenesis may, therefore, take place in the mantle cavity as in Duplorbis (Smith, 1906), in the colleteric gland in Sylon (Veillet, 1962; Lützen, 1981) or in the mantle of Mycetomorpha (Reinhard & Evans, 1951) and Thompsonia (Yanagimachi & Fujimaki, 1967). In the Peltogasteridae (Reinhard, 1942a; Ichikawa & Yanagimachi, 1958), Sacculinidae (Ichikawa & Yanagimachi, 1960), and Lernaeodiscidae (Ritchie & Høeg, 1981) spermatogenesis begins in the male cell receptacles and they become the only source of spermatozoa. This process begins in Clistosaccus paguri, which only has one receptacle, when ova within the ovary reach about 60 µm diameter. When the ova are about 100 µm spermatogonia differentiate and when 165 µm the receptacle is filled with spermatozoa. Ovulation and release of spermatozoa into the mantle cavity are simultaneous and in this species the mantle aperture opens at the same time (Høeg, 1982). A similar synchrony is recorded for other species although the mantle aperture may be already open (Lützen, 1981; Ritchie & Høeg, 1981). Ova pass through the atrium of the colleteric gland and are fertilised in the mantle cavity. After ovulation the externa exhibits peristaltic movements which serve to mix spermatozoa and ova together and to ventilate the mantle cavity. Although the mantle aperture is technically open it remains compressed during embryogenesis. During ovulation in Peltogaster paguri the aperture is tightly closed (Reinhard, 1942b). In Lernaeodiscus porcellanae ventilation may also be assisted by the grooming action of the castrated host which treats the externae as its own brood. When this is prevented the externae soon become fouled. In particular, the grooming assists during the moult which occurs after the release of each batch of nauplii (or cyprids). The nauplii are also released into the current generated by the host during grooming; this aids dispersion of the cirripede young (Ritchie & Høeg, 1981). When embryogenesis is complete the embryos hatch and are expelled from the mantle cavity through the mantle aperture; Reischman (1959) has described this for Peltogasterella. Information on the sizes of ova and eggs produced by rhizocephalans is scattered throughout the literature from as early as Delage (1884) to the present day. Even Leuckart (1859) stated of Sacculina inflata “… ovarian eggs were usually smaller than those of the egg-tubes…” but gave no sizes. Descriptions of the rhizocephalans collected during the Siboga Expedition (Van Kampen & Boschma, 1925; Boschma, 1931)

68

MARGARET BARNES

mention the presence of eggs in the mantle cavity of many species but only give sizes in a few cases. These are given in Table I; sometimes sizes have been estimated from scaled drawings. Other records from the literature are also shown in Table I. Sometimes the size of nauplius stage I is given if no information about the egg is TABLE I Rhizocephala: summary of relevant literature on breeding seasons, sizes and numbers of young, cyp=cyprid ; St I=stage I nauplius ; +=or more ; *=embryo released as cyprid from mantle cavity of adult, if known Species

Place

Breeding season

Incubation time, days

Sizes (L×B), Number of µm eggs per brood

Number of broods per year

References

Chthamalo philus delagei* Clistosaccu s paguri*

–

–

–

cyp 60–70

–

–

Sweden

About 30

ova 165 cyp 184×90

–

1 or 1+ per life

Drepanorch is neglecta

–

All year, summer peak –

BocquetVédrine, 1961 Høeg, 1982, 1985b

–

cyp 140

–

–

Drepanorch is villosa Heterosacc us ruginosus Lernaeodis cus cornutus Lernaeodis cus galatheae

Siboga, Exped. –

–

–

egg 100

–

–

–

–

–

–

–

–

–

egg 135×108 St I 200 egg 94

–

–

Krüger, 1940

–

–

–

–

–

Veillet, 1943a,b

Lernaeodis cus porcellanae –

–

–

–

small cyp 130 large cyp 180 egg 160×126

–

–

Krüger, 1940

–

–

–

–

–

All year

10 to 14

100+ to 20 000

About every 15 days

Müller, 1862 Ritchie & Høeg, 1981

Peltogaster paguri

–

–

egg 200×120 small cyp 227×202 large cyp 255×232 30 to 40

St I 210 to 280

9800 to 28000

6

BocquetVédrine, 1961 Boschma, 1931 George, 1959

McMumch, 1917; NilssonCantell, 1921; Reinhard, 1942b, 1946

EGG PRODUCTION IN CIRRIPEDES

Species

Place

Breeding season

Incubation time, days

Sizes (L×B), Number of µm eggs per brood

Number of broods per year

References

Peltogaster sulcatus

–

–

–

St I 250

–

–

Peltogaster ella gracilis Hokkaido, Japan

Bering Sea

–

–

St I L=B

–

–

NilssonCantell, 1921 Reischman, 1959

–

–

Small ova 140 to 150 Large ova 158 to 160 Small egg 140 to 150 Large egg 160 to 170 Small St I 220 to 245 Large St I 255 to 275

–

–

Yanagimach i, 1961a

Peltogastere lla socialis

–

–

St I 207 to 247

–

–

Japan

N.W. Pacific, Alaska, Bering Sea –

About 14

–

–

–

Peltogastere lla subterminali s Sacculina carcini

N.W. Pacific, Alaska California Millport, Scotland

–

–

St I (190 to 216) ×(126 to 148)

–

Shirase & Yanagimac hi, 1957 –

–

–

–

–

R.Mersey, England Plymouth, England Roscoff, France

Mar–May

Max Aug– Dec Min Jan– Mar –

–

–

–

Day, 1935

Mar–Oct

–

–

–

–

Orton, 1936

June–Aug

28 to 35

–

2+

Delage, 1884

France

–

–

ova 80×60 egg 180×150 egg 120×100

–

–

Isefjord, Denmark

July–Oct

12 to 18

100 000 to 300 000

6

BocquetVédrine, 1964 Lützen, 1984

–

Reinhard, 1944

Reinhard, 1944

Foxon, 1940; Heath, 1971

69

70

MARGARET BARNES

–

–

–

egg 70×54

–

–

–

–

–

–

–

Sacculina micracantha Sacculina papposa Sacculina rotundata Madras, India Sacculina setosa Sacculina sulcata Sacculina sp.

Siboga Exped. Siboga Exped. Siboga Exped. –

–

St I 165×120 –

Krüger, 1940 Hoek, 1909

egg 134

–

–

–

–

egg 100

–

–

–

–

egg 100

–

–

–

–

–

–

St I 215×135 –

egg 100

–

George, 1949 –

–

–

egg 110

–

–

July–Jan

–

–

(75 ova in 2 mm adult)

–

–

–

–

–

–

Black Sea

–

9

Sylon challengir* Sylon hippolytes* Sylon schneideri Thompsonia cubensis

–

–

–

egg 182×129 egg 170×125 St I 190×135 ova 60

–

–

Hoek, 1888

–

–

–

–

1 per life

–

–

–

ova 80×60

20 000 to 200 000 –

Lützen, 1981 Hoek, 1888

Cuba

–

–

ova 95×80 egg 85+

–

–

Thompsonia sp.* Thompsonia sp.*

Great Barrier Reef –

–

–

egg 34

–

–

–

–

cyp 200

–

–

Triangulus galatheae Bocquetia rosea

–

–

–

–

–

S California

–

–

St I 257×100 cyp 88.9±1. 89

–

–

Septodiscus flabellum Septosaccus cuenoti

Siboga Exped. Siboga Exped. Mutsu Bay, Japan

–

Boschma, 1931 Boschma, 1931 Boschma, 1931

Boschma, 1931 Boschma, 1931 Elston, Wilkinson & Burge, 1985; Krüger, 1940 Codreanu, 1959

Reinhard & Stewart, 1956 Potts, 1915 BocquetVédrine, 1961 Veillet, 1945 Pawlik, 1987

EGG PRODUCTION IN CIRRIPEDES

71

available. The difference in size between the mature egg and nauplius stage I is very little. Egg size may also increase during embryogenesis and in many cases the stage of development is not recorded. Veillet’s (1943a, 1945) suggestion that the sex of a cyprid is determined at the egg stage was confirmed by Yanagimachi (1961a,b) who went further and said it was determined in the ovary. Yanagimachi (1960) also considered the chromosome numbers of large and small eggs. Ritchie & Høeg (1981) confirmed the earlier work on Peltogaster paguri and Sacculina senta (Ichikawa & Yanagimachi, 1960) and Peltogasterella gracilis (Yanagimachi, 1961b) that two sizes of eggs may be produced but added that Lernaeodiscus porcellanae does not always produce broods of one sex. They found that the sex of the broods varied with season and that during the transition broods of mixed sex could be produced. Broods were predominantly female in summer and male in winter—these males being necessary for virgin externae that would emerge in the following spring. In Sacculina carcini (Høeg, 1984) some broods contain only one sex while others may be mixed. Høeg suggests that this may reflect the nutritional state of the externa which may itself be determined by the condition of the host. It may also be that in summer broods are laid in quick succession thus tending to deplete the nutritional reserves of the externa. Walker (1985, 1987) has examined nauplii from broods of S. carcini taken at different times of the year; he also found a progressive change from male to female cyprids from spring to summer reverting to males again from autumn to winter. The sizes of externae are difficult to measure because of their shape and because they may be distended when carrying eggs in the mantle cavity—the amount of distension depending on the stage of development of the embryos (Reinhard, 1942b). The size of the externa at maturity may vary according to species or the number of externae per host. According to Boschma (1931) Peltogasterella socialis externae 5–7 mm long have eggs in the mantle cavity whereas Sacculina carcini externae do not carry eggs until they are between 10 and 12 mm long (Foxon, 1940; Heath, 1971). In Sylon hippolytes size at maturity may be 10.4, 8.4 or 7.2 mm long depending on whether the host is carrying 1, 2 or 4 externae (Lützen, 1981). In Peltogaster paguri (Reinhard, 1942b) externae of 6 mm length may be mature on small hosts; on larger hosts they are not mature until 9 mm long. Season of maturity can also vary depending on temperature or even the life cycle of the host (Foxon, 1940). Some species have peak periods of reproduction during the summer and a rest period in the winter, such as Sacculina carcini in Denmark (Lützen, 1984), although further south the reproductive period of this species may be longer (Delage, 1884; Day, 1935; Orton, 1936; Foxon, 1940; Heath, 1971; Walker, 1987). Some species may be continuously reproductive, such as Lernaeodiscus porcellanae (Ritchie & Høeg, 1981), with incubation time depending on temperature and the time of year. In some species the externa may produce only one brood and then die and fall off the host, such as in Sylon hippolytes (Lützen, 1981). Results are not expressed in the same way in the literature (see Table I) which often makes comparisons difficult. Day (1935) suggested that reproduction in Sacculina carcini is synchronised with that of its host so that food reserves normally used by the host can be appropriated by the parasite. While this is possible it is known (Heath & Barnes, 1970) that the ovaries of the host, Carcinus maenas, show their greatest increase in size in the summer and yet nauplii and cyprids of Sacculina carcini are most abundant in winter (Pyefinch, 1948). It is suggested, therefore, that it is the developing interna, rather than the externa, which requires optimal conditions for growth. Furthermore, as rhizocephalan nauplii are lecithotrophic their release at a time when there is little plankton in the winter, far from being disadvantageous, may be an advantage in reduced losses due to predation (Heath, 1971). The number of eggs found in the mantle cavity after each ovulation depends on the size of the externa and some authors are careful to give this. The most comprehensive account is that of Lützen (1981; see Table II) for Sylon hippolytes. Numbers of eggs found in other species are given in Table I. It is of interest

72

MARGARET BARNES

that Peltogaster paguri of externa size 12 mm produces 28 000 eggs per brood which, allowing say 6 broods per summer (Reinhard, 1942b) would be 168 000 eggs per year (life?). From Table II it can be seen that Sylon hippolytes of the same size externa produces 120 000–152 700 eggs in its one brood per life. The former species releases the embryo as a nauplius and the latter as a cyprid. It has been said that species releasing as cyprids may not produce as many offspring as those releasing as nauplii (Høeg, 1982). This is hardly borne out by this result. A better comparison may be with Lernaeodiscus porcellanae which releases as a nauplius and may produce several hundred to 20 000 eggs per brood (Ritchie & Høeg, 1981). In summer this can be every 10–14 days but will obviously be less frequent at lower winter temperatures. If one assumes a brood say every month on the average (it breeds continuously according to Ritchie & Høeg, 1981) then it will produce up to 240 000 eggs per year which is more than in Sylon hippolytes and may be an under-estimate. At two more broods per year the number would increase to 280 000; in this case there are many more nauplii produced than cyprids. The merits of releasing young as nauplii or cyprids is discussed by Høeg (1982). Dispersion of the species is ensured by nauplii which remain swimming in the water for several days. Reduction of the planktonic stages, and consequently of the length of planktonic life, reduces the risk of predation and increases the chance of survival. In species such as Clistosaccus paguri, TABLE II Number of eggs in six externae of Sylon hippolytes calculated from serial sections (Lützen, 1981) Size of externae (mm)

Approximate

Length

Width

Height

number eggs

5.9 7.6 8.1 10.0 14.0 15.2

3.4 5.2 6.1 8.3 10.0 9.6

4.1 4.8 4.9 7.0 5.0 9.0

18 900 75 500 105 900 120 000 152 700 203 300

producing only one or very few broods during the lifetime of an externa, the lack of dispersion and greater chance of survival ensures that the cyprids remain in the vicinity of the hermit crab (host) population. ACROTHORACICA Cirripedes of the order Acrothoracica are burrowing and non-parasitic. They have a soft mantle without calcareous plates and are dioecious with dwarf males. The order is predominantly found in warm-temperate regions and generally in shallow (30 m) water (Ross & Newman, 1969; Zullo, 1979). Newman & Ross (1971) have, however, reported a species, Australophialus tomlinsoni, from depths of 300–600 m south of the Antarctic Convergence and two truly deep-water species have been found: Weltneria hessleri at 1000 m depth off Bermuda (Newman, 1971) and W. exargilla in the Bay of Biscay at a depth of 1500 m (Newman, 1974). The females live in self-excavated burrows usually in limestone, coral skeletons or mollusc shells. W. exargilla is different in that it burrows into soft clayey siliceous substrata which are fairly low in calcium carbonate. The tiny males are generally found near the aperture of the females in a pocket of the mantle tissue near the area of the ovary. In some species the males are, however, found attached to the wall of the

EGG PRODUCTION IN CIRRIPEDES

73

burrow. The acrothoracids so far described are small usually being only a few millimetres long (see Table III). The females are about ten times longer than the males. The burrows of these soft bodied barnacles provide them with the protection necessary because of the absence of a calcareous shell. The burrow is largely formed by abrasion caused by chitinous teeth on the mantle surface of the female although a cyprid larva settling on a suitable substratum has no mantle teeth suggesting that the initial penetration may be by chemical dissolution (Tomlinson, 1969). Indeed Turquier (1968) mentioned the use of carbonic anhydrase in burrowing by Trypetesa nassarioides and Tomlinson (1973) found papillae on the external surface of the mantle. In view of Turquier’s work Tomlinson suggested that these papillae might be secreting carbonic anhydrase or other alkaline phosphatase and that teeth, developed later, act only after enzymatic softening of the substratum. A chemical action combined with abrasion was also mentioned by Kamens (1981). The attachment area of the female is a major structural element within the burrow and allows movement of the mantle. As the mantle is cemented to the burrow wall at this point it cannot moult in this area and remains of exuxiae build up forming a horny disk (see Grygier & Newman, 1985, for a discussion about this disk). Dwarf males may be attached to the margins of this disk embedded in the mantle of the female, e.g. in Alcippe lampas (=Trypetesa lampas) (Berndt, 1903), Lithoglyptes indiens and Trypetesa lateralis (Tomlinson, 1969), and Cryptophialus melampygos (Batham & Tomlinson, 1965). In some cases the males may be attached to the burrow walls, e.g. in Berndtia purpurea (Utinomi, 1950). There are also cases where males have been found attached to either the mantle of the female or the burrow wall, such as in Kochlorine floridana (Wells & Tomlinson, 1966), Berndtia nodosa (Tomlinson, 1967), and Cryptophialus coronophorus (Smyth, 1986). Utinomi (1964) failed to find any males in his study of Tryptesa habei; they TABLE III Acrothoracica: comparison of size of adult female and size and number of young, cyp=cyprid;? stage doubtful Species Trypetesidae Trypetesa lateralis Cryptophialidae Australophialus pecorus Cryptophialus melampygos 2.5 2.6 2.7 Cryptophialus minutus Lithoglyptidae Berndtia purpurea 2.2 2.4 2.5

Size of young (L×B), Size of female (L×B), Number of young per µm mm brood

References

1.2–1.3 also 2.0

egg 250×180 cyp 500×200

180

Tomlinson, 1953, 1955

largest 1.28×0.82

egg 220 to 270 cyp 350 to 370 egg 260×210

1 to 8

Turquier, 1985a

14

Batham & Tomlinson, 1965

19 to 60

Darwin, 1854

14

Utinomi, 1961

1.9 to 2.1 cyp 530×230

largest 2.54

1.7 cyp 470×260

33 44 31 egg 254 cyp 229 to 406 egg 320 65 21 85

74

MARGARET BARNES

Species 2.6 2.6 2.7 Lithoglyptes tectoscrobis Weltneria exargilla

Size of young (L×B), Size of female (L×B), Number of young per µm mm brood

5.7×(1.7 to 1.0) (6.2 to 11.4)× (1.5 to 3.9)

64 57 32 egg 180×130 St? 510×370

1200 estimated by volume About 150

References

Grygier & Newman, 1985 Newman, 1974

may have been lost during the preparation of the material for examination. The size of the adult male is less than that of the cyprid from which it develops. The adult male lacks feeding appendages and the maturation process is at the expense of larval food reserves. Mature males consist mainly of reproductive material and its maturation must be synchronised in some way with that of the female if fertilisation is to be effective. Kühnert (1934) raised larvae of Alcippe lampas (=Trypetesa lampas) in separate vessels and found that males would differentiate without the presence of females but would not become fully mature. Tomlinson (1969) suggests that part of the maturation of the male may be under the chemical influence of the female. The partial burial of the male in the female’s mantle tissue near the ovary may be significant in this respect. The dwarf males in acrothoracids usually have a penis although it has not been seen in some species and is even reported as absent in others. In some species of Lithoglyptes the presence or absence has not been resolved but according to Tomlinson (1969) it is usually present. He was, however, not sure of it in L. hirsutus (Turquier, 1963). Newman (1971) did not find a penis in Weltneria hessleri; this may have been due to the immaturity of the males examined as Tomlinson (1969) suspected in Cryptophialus heterodontus. In Kochlorine ulula and Cryptophialus rossi a penis could not be positively identified by Tomlinson (1973), but he did find a penis sheath in the latter species. In C. epacrus a penis was found. In Kochlorine floridana the penis matures late in life (Wells & Tomlinson, 1966) whereas in K. bocqueti Turquier (1977) found none. In Trypetesa lateralis (Turquier, 1967c) the male is greatly reduced and there is no penis sheath. There is only a short penis in T. nassarioides (Turquier, 1967c, 1971) and Utinomi (1964) records its presence as doubtful in T. habei but a penis sheath is present. The sheath is formed from the mantle of the male (Smyth, 1986). The spermatozoa of T. lateralis can be activated by sea water, the length of time depending on the salinity of the water (Tomlinson, 1969). Turquier & Pochon-Masson (1969) made an ultrastructural study of spermiogenesis in T. nassarioides and showed it to be normal for cirripedes; dwarf males have a true fertilising role. The method of insemination needs further study as details of copulation and fertilisation have not been described in detail for any acrothoracid (Tomlinson, 1960, 1969; but see also Turquier, 1977). By whatever method involved the mature ova and sperm come into contact and fertilised egg masses are found in the mantle cavity of the females. During its life a female may be served by a few or many males. In Kochlorine floridana Wells & Tomlinson (1966) found up to five relatively short-lived dwarf males per female. In Australophialus turbonis there may be about 12 per female; up to 17 have been seen (Tomlinson, 1969). In Cryptophialus zulloi Tomlinson (1973) found remains of at least 27 males in the area of the attachment disk of the female. According to Turquier (1972a) the dwarf males of Trypetesa, in spite of their short life, are able to fertilise several egg masses. When their testes are empty they may be replaced by other males. In Cryptophialus melampygos old spent males are replaced during the life of the female (Batham & Tomlinson, 1965).

EGG PRODUCTION IN CIRRIPEDES

75

The ovaries lie on the dorsal side of the female under the thickened layer of the disk and the oviducts each end in an atrium which opens into the mantle cavity. Ovaries and oviducts have been described in Alcippe lampas by Berndt (1903) and in Berndtia purpurea by Utinomi (1950, 1960). The fertilised ova (eggs) remain in the mantle cavity of the female as two separate egg lamellae until hatching takes place (Tomlinson, 1969). Embryonic development takes place within the egg case. As early as 1849 Hancock reported seeing eggs in the mantle cavity of Alcippe lampas and watched the hatching of swimming nauplii. Whether the embryos eventually hatch as nauplii or cyprid larvae varies in different genera. In the Crytophialidae where hatching has been described it is usually as a cyprid. Although in Australophialus turbonis (Tomlinson, 1969) whether the nauplius is free-swimming or retained to the cyprid stage is not known; in A. pecorus the larvae is liberated as a cyprid (Turquier, 1985a). In the Lithoglyptidae species of the genus Weltneria are released as cyprids (Tomlinson, 1969: W. spinosa, W. reticulata; Turquier, 1985a: W. zibrowii; Newman, 1974: W. exargilla). Balanodytes taiwanus (Tomlinson, 1969) also hatches as a cyprid. In species of the genera Berndtia and Lithoglyptes so far described the embryos, however, hatch as nauplii (Tomlinson, 1969: L. indicuns, L. mitis, L. scamborachis; Utinomi, 1950, 1960: Berndtia purpurea). In the Trypetesidae some members of the same genus may hatch as nauplii (Berndt, 1903; Genthe, 1905; Kühnert, 1934: Alcippe (=Trypetesa) lampas; Turquier, 1976b: Trypetesa nassarioides) whereas T. lateralis hatches as a cyprid (Tomlinson, 1955). Whether hatched as nauplii, which develop through four stages (Visscher, 1938, says six in Alcippe lampas but this has never been confirmed by other workers) or cyprids the larvae are expelled from the female by contractions of the mantle cavity as seen by Genthe (1905) in Alcippe (=Trypetesa) lampas. The cyprids may be free swimming or have reduced swimming appendages as, for example in Cryptophialus melampygos (Batham & Tomlinson, 1965). In this species the cyprids can only crawl and so are splashed about in the sea water and as a result dispersion is reduced and dense infestations can occur (Tomlinson, 1969). Newman & Tomlinson (1974) have discussed the dispersion of species which lack pelagic larvae and have non-swimming or only weakly swimming cyprids. According to Turquier (1972a) the sex in Trypetesa species is determined before the cyprid settles although there is no sexual dimorphism in the cyprids. A biometrical study of the larvae of T. nassarioides (Turquier, 1972b) indicated a seasonal variation in size but these variations have no relation to sex. White (1970) found “No obvious chromosomal mechanism of sex determination” in T. lampas. In T. lateralis a female cyprid settles on a suitable substratum and attaches itself by its antennules. It proceeds to burrow into the substratum. The cyprid carapace is shed and the young barnacle quickly passes beneath the surface of the substratum. Developing male cyprids lose their carapace and become progressively smaller decreasing from to as low as (Tomlinson, 1955). There are only a few records of the number of ova produced or eggs incubated in the mantle cavities of female acrothoracids. Those available are summarised in Table IV. As can been seen the number of eggs varies from 1–8 in Australophialus pecorus (Turquier, 1985a) to 1200 estimated volumetrically by Grygier & Newman (1985) in Lithoglyptes tectoscrobis. It is not known how accurate this number is but it is far above other maximum counts of 150 or 180 in Weltneria exargilla (Newman, 1974) or Trypetesa lateralis (Tomlinson, 1955). TABLE IV

76

MARGARET BARNES

Acrothoracica: summary of relevant literature on breeding seasons, sizes and numbers of young, cyp=cyprid; St I=stage I nauplius; +=or more; *=embryo released as cyprid from mantle cavity of adult, if known Species

Place

Breeding season

Incubation time, days

Sizes (L×B), µm

Number of References eggs per brood

Japan

–

–

egg 240

numerous

Utinomi, 1964

–

–

–

egg 340×250 St I 450

–

Trypetesa (=Alcippe) lampas Trypetesa lampas

Heligoland, North Sea

May–Sept

–

St I 550×450

–

Roscoff, France

–

9 to 10

–

Florida, USA Trypetesa lateralis* Trypetesa nassarioides

June+? California

– Nov–Jan and June Feb–Oct Max Apr– Aug

– –

Australophial us pecorus*

Mediterranean Sea, N. Africa

–

–

Cryptophialus coronophorus Cryptophialus lanceolatus*

Guam

–

–

–

–

–

egg 380×280 St I 400 cyp 600 – egg 250×180 cyp 500×200 egg 280×210 St I 340 cyp 450×170 egg 220 to 270 cyp 350 to 370 cyp 300 to 310 cyp 500×200

Berndt, 1903; Kühnert, 1934; Krüger, 1940 NilssonCantell, 1921, 1978 Turquier, 1967a

Trypetesidae Trypetesa habei Alcippe lampas

Roscoff, France

6 to 12

Spivey, 1979 180 Turquier, 1967a,b,c

Tomlinson, 1953, 1955 Cryptophialid ae

1 to 8

Turquier, 1985a

–

Smyth, 1986

–

Tomlinson, 1969

14 to 44

Batham & Tomlinson, 1965

19 to 60

Darwin, 1854

–

Newman & Ross, 1971

14–85

–

ova 220×160 cyp 470×260 egg 320 – cyp 450×170

Utinomi, 1950, 1960, 1961 Turquier, 1977

–

–

–

Cryptophialus melampygos

Otago, New Zealand

–

–

Cryptophialus minutus* Cryptophialus tomlinsoni* Lithoglyptidae Berndtia purpurea Kochlorine bocqueti Kochlorine floridana

Chile

Jan+?

–

Antarctica

–

–

Japan

Aug–Sept

Madagascar –

egg 260×210 St I? 320×200 cyp 500×200 to 560×260 egg 254 cyp 230 to 406 cyp 500×200

cyp 550×190

–

Wells & Tomlinson, 1966

EGG PRODUCTION IN CIRRIPEDES

Kochlorine hamata Lithoglyptes mitis Lithoglyptes scamborachis

–

–

–

–

–

–

–

cyp 420×225

–

–

–

Lithoglyptes tectoscrobis Weltneria exargilla Weltneria reticulata* Weltneria spinosa* Weltneria zibrowii

Tongo

–

–

cyp (338 to 538) ×(153 to 184) egg 180×130

Bay of Biscay

July+

–

–

–

– Mediterranean Sea, Algeria, North Africa

“einige Dutzend” – –

77

Noll, 1875 Tomlinson, 1969 Tomlinson, 1969

St? 510×370

1200 estimated by volume about 150

Grygier & Newman, 1985 Newman, 1974

–

cyp 545×245

–

–

–

cyp 870×365

–

–

–

cyp 800×360

–

Tomlinson, 1969 Tomlinson, 1969 Turquier, 1985b

Comparison between size of female and number of eggs in brood incubated are given in Table III together with the size of the egg, nauplius and/or cyprid, whichever measurement is given in the original reference. Lithoglyptes tectoscrobis appears to have one of the smallest eggs which may be one of the reasons for their greater number compared with Weltneria exargilla. Both these genera belong to the same family Lithoglyptidae, and the females of the species are within the same size range (Table III). Berndtia pur purea also belongs to the Lithoglyptidae but the female is somewhat smaller (Table III) and contains 14–85 eggs depending on size of parent (Utinomi, 1961). The size of the hatched nauplius is similar to that of Alcippe lampas (Kühnert, 1934) and, as a matter of interest, is about twice as big as the corresponding stage of e.g. Balanus crenatus ( Herz, 1933) and B. amphitrite albicostatus Ishida ( & Yasugi 1937). The Cryptophialidae have much smaller females and fewer eggs (1–60) which are larger than those of Lithoglyptes tectoscrobis. The only species of the Trypetesidae, in which the number of eggs is given, is very small, similar to Australophialus pecorus but has a slightly smaller egg and a much larger brood. Because of the general uncertainty of the data it is difficult to come to any definite conclusions about size of female and the size and number of eggs produced. Development of any batch of eggs is synchronous (Turquier, 1985a) and the time from fertilisation to hatching ranges from 6–12 days depending on temperature in Trypetesa nassarioides and T. lampas (Turquier, 1967a,b,c). Cyprids leave the mantle cavity before a new batch of ova are fertilised in Cryptophialus melampygos (Batham & Tomlinson, 1965). In this species a female of 1.2 mm length may get an immature male but will not be ovigerous until 1.6–1.8 mm long. Three out of four females more than 1.7 mm long had eggs or larvae in the mantle cavity at all times of the year; the majority of breeding females were 1.9–2.1 mm long. Information on the breeding seasons of the various species is also fragmentary (Table IV). C. melampygos breeds throughout the year at Otago, New Zealand, according to Batham & Tomlinson (1965) so that, although the broods are relatively small, the total number of nauplii produced per year could be several hundred. The C. minutus found to contain 19–60 eggs by Darwin (1854) had been collected in early January in Chile. Trypetesa nassarioides and T. lampas from the Roscoff region of France breed from

78

MARGARET BARNES

February to October with a maximum from April to August (Turquier, 1967a,b). T. lampas at Heligoland was sexually mature and fertilised from May to August (Kühnert, 1934). Cyprids may be found as early as June, in great numbers in August, and none in February (Nilsson-Cantell, 1978). At Woods Hole, Visscher (1938) found larvae of T. lampas in June but they were rare in July and August. In Florida, T. lampas contains late embryos and first stage nauplii in June (Spivey, 1979) suggesting perhaps a similar breeding season to that of the same species in Europe. The greatest larval settlement of T. lateralis in California was in November to January with a minor peak in June (Tomlinson, 1953). Allowing for the time of development of the larvae this would mean the peak of the breeding season was in August to September. Turquier (1972b) gives the most detailed account of the reproductive cycle of T. nassarioides in the Baie de Morlaix, France. He sampled the population for 16 months from May to August of the following year. Young females settling early in the year reached a length of 3 mm by August and 5 mm by the end of the summer. Maturity was reached three to four months after settling so their reproductive activity was short in the first year. Growth was reduced or even suspended during the winter but survival was good. In animals settling late in the first year, and perhaps becoming only 1 mm long by December, there was considerable mortality. Reproductive activity started again in February and throughout the season the females continued to produce broods of 200–350 larvae each time until the end of summer when about 7– 8 mm long. Their lifespan appears to be about 18 months. The start of reproductive activity in February coincides, off the Atlantic coast of France, with minimum sea-water temperature and continues as the sea-water temperature rises to a maximum in August and begins to decrease again. The effect of temperature on reproduction seems less dramatic than in the Thoracica. The effect of light may also be less important than in the Thoracica as the cirripedes are buried in the interior of the host shells or substratum where the level of light will be low. The larvae are lecithotrophic and while food may be necessary for the adults for vitellogenesis, it is not essential for the larvae to be liberated when food is available in the sea water. The release of a brood seems to depend on an internal mechanism. It is possible to find females with mature males in which vitellogenesis has apparently been achieved and which are breeding but remain incapable of releasing the brood (Turquier, 1972b). THORACICA The largest and most diverse order of the Cirripedia is the Thoracica which contains three living suborders, the Lepadomorpha, Verrucomorpha, and Balanomorpha. The first of these has the greatest bathymetrie range from intertidal to at least 5000 m (Nilsson-Cantell, 1950). It contains the stalked barnacles belonging to the living superfamilies Ibloidea, Heteralepadoidea, Lepadoidea, and Scalpelloidea (Newman, 1987). The Verrucomorpha are also widely distributed both geographically and bathymetrically—up to 4630 m according to Nilsson-Cantell (1950). The living genus Verruca is asymmetrical and sessile. The classification of the Balanomorpha has been repeatedly revised over the years since Darwin’s (1854) monograph, a recent attempt being that of Newman & Ross (1976, 1977), who give three superfamilies: Chthamaloidea, Coronuloidea, and Balanoidea. Following Newman (1987), however, there are five living superfamilies, Chionelasmatoidea, Pachylasmatoidea, Chthamaloidea, Coronuloidea, and Balanoidea. The sessile acorn barnacles belong to the Balanomorpha. LEPADOMORPHA The Lepadomorpha contains barnacles having a capitulum and a peduncle. The capitulum is often protected by calcareous plates but in some cases these may be vestigial or even absent. The prosoma is contained within

EGG PRODUCTION IN CIRRIPEDES

79

the capitulum; the peduncle may or may not have calcareous scales. It contains the testes in males and the ovaries in females and hermaphrodites. Ibloidea Species of Ibla are tiny pedunculate barnacles with weakly calcified plates often seeking protection in crevices or shells of other animals including cirripedes. They may be intertidal or in shallow water sometimes attached to Pollicipes species or associated with Tetraclita species (Achituv & Klepal, 1981). Five species are now recognized and of these two are hermaphrodites with complemental males—Ibla quadrivalvis and I. pygmaea (=I. segmentata)—and the other three are females with dwarf males—I. cumingi (=I. sibogae), I. idiotica, and I. atlantica (Klepal, 1985). Egg sizes (Table V ) in all species are relatively large compared with those of cirripedes in general and because the animals are small the number of eggs per brood is also small. Only a limited number of large eggs can be accommodated in the mantle cavity of the parent. According to Broch (1922) ovigerous I. pygmaea may have about 16 large eggs at a time in the mantle cavity whereas I. quadrivalvis may have 100 to 300 (Anderson, 1965). Batham (1945a) found 20 ova in the ovary of a female I. idiotica that was 3 mm long. The number of eggs in I.cumingi depends on the size of parent, about 400 being the average (Klepal, 1985) which agrees with Hoek’s (1907) number of 480 in I. sibogae (=I. cumingi). The highest number recorded for this species is 1225 by Gaonkar & Karande (1980). In I. quadrivalvis the released nauplius stages are free-swimming but in I. idiotica the swimming habit is eliminated and development to the cyprid stage takes place within the mantle cavity of the parent. In both cases the larvae are lecithotrophic. In Bombay waters I. cumingi 5 mm long or over contain ovarian tissue at all times of the year but the main breeding season is in mid-February to April and from August into November (Gaonkar & Karande, 1980). During the monsoon months when water quality, especially salinity, is changing there is apparently no fertilisation. In the Philippines, Resell (1967) found that females 11.5 to 15 mm long were ovigerous. I. quadrivalvis breeds throughout the year except for May and June (late autumn and early winter) in Australian waters with peaks of release of young in September and February to March (Wisely & Blick, 1964). This is confirmed by Anderson (1965) who gives 16– 17 days at 23°C from fertilisation to release. He says that although only 100– 300 larvae are produced by brood, throughout the season an adult may release about 10 000 young. Heteralepadoidea This is a superfamily containing small pedunculates without calcareous plates or scales and about which there is not much information on reproduction. Heterolepas cornuta from the eastern Pacific was found to contain eggs but because of lack of data in earlier reports on this species Ross (1975) was not able to detect any seasonal variation in the reproductive cycle. He did not give any numbers or size of eggs except to say that the size was comparable with that given by Stubbings (1965). The sizes that could be found in the literature for Heterolepas, Paralepas and Anelasma species are included in Table VI. TABLE V

80

MARGARET BARNES

Ibloidea: summary of available literature on breeding seasons and size and number of young per brood. St I=stage I nauplius Species

Place

Breeding season Size of young (L×B), µm

Number of young References per brood

Ibla sibogae

Indonesia

–

480

Hoek, 1907

Ibla cumingi India

Molo Strait Mar–May not during monsoons; ova all year

– St I 300

Hoek, 1907

Gulf of Elat

–

ripe ova 110×85 up to 500×300

Ibla idiotica Otago, New Zealand Ibla pygmaea

Australia –

– egg 340 St I 360 –

400 depending on parent size – 20

– Karande, 1974a; Karande & Thomas, 1976; Gaonkar & Karande, 1980 Klepal, 1985

20 Batham, 1945a

Anderson, 1965

–

about 16

Broch, 1922

egg 320 egg 400×275

– 100–300

Krüger, 1940 Anderson, 1965; Wisely & Blick, 1964

Ibla quadrivalvis

38°12′S: 149°40′ E – Australia

– Aug–May

large female egg 210×160 small female egg 140×160 egg 180×130 max 1225

TABLE VI Heteralepadoidea: size of young and places found, cyp=cyprid; St I=stage I nauplius Species

Place

Size of young (L×B), µm

References

Anelasma squalicola – From sharks in North Sea Heteralepas cornuta Heteralepas minuta Heteralepas smilius Paralepas minuta Paralepas scyllarusi

– egg 400×280 egg 558 Senegal? Gulf of Guinea Chinese waters Senegal Kyusyu

St I 660×220 Krüger, 1940 Darwin, 1851 egg 190×111 cyp 1120×560 egg 270×120 egg 199×107 egg 233

Hoek, 1909

Stubbings, 1965 Stubbings, 1961 Ren, 1983 Stubbings, 1965 Utinomi, 1967

Lepadoidea The Lepadoidea are all hermaphrodites, without complemental males and are oceanic in occurrence. They are often attached to floating objects and may be widely dispersed from the initial populations. Lepas species are mainly known from their appearance on material found by chance either drifting in the sea or washed up on the shore.

EGG PRODUCTION IN CIRRIPEDES

81

A full account of the development of Lepas fascicularis from egg to adult has been given by WillemöesSuhm (1876) who said that the cyprid was probably short-lived as only a few were found even where there were many hundreds or thousands of nauplii. It may be, however, that there is great mortality or predation of late stage nauplii. The size of eggs and larvae given by Willemöes-Suhm compare favourably with those of later workers but the number of eggs is lower; this does, however, depend on the size of the parent animal. Bigelow (1896) gave an account of the early development of L. fascicularis and of Lepas sp. (Bigelow, 1902) but did not give any sizes. Information on the number and sizes of eggs and/or nauplius stages so far available on Lepas sp. is summarised in Table VII. There is virtually nothing known about L. anaserifera and L. hilli except that immature ova have been seen in the former (Darwin, 1851) and in the latter 29 out of 57 adults of 16.5–19 mm capitulum length after 60 days attached to a ship plying between Dakar and Barbados contained egg lamellae (Evans, 1958). L. anatifera var. testudinata on buoys in New Zealand waters may contain hatching eggs 4–5 weeks after settlement at capitulum lengths ranging from 23–27 mm (Skerman, 1958a). For L.australis it is only known that in New Zealand coastal waters nauplii are found 6–8 weeks after settlement of the parents, these having reached a capitulum length of 21 mm (Skerman, 1958a). In this species the cyprid is 2000–2250 long (Darwin, 1851; Nilsson-Cantell, 1930a) compared with 1250 µm for L. pectinata (Stebbing & Fowler, 1904) and 1300 µm for L. fascicularis (Willemöes-Suhm, 1876). The ova, eggs, and stage I nauplius all seem to be smaller in L. anatifera than in L. fascicularis (see Table VII). These sizes may depend somewhat on the temperature at which development takes place (Patel, 1959). The number of developing eggs in these two species is given as 1400 to many thousands or even tens of thousands in L. anatifera (Hoek, 1884; Witschi, 1935; Zann & Harker, 1978) and in L. fascicularis the figure of 4000 is given by Burmeister (1843) as quoted by Willemöes-Suhm (1876) or up to 300 000 found by Thorner & Ankel (1966) in a single animal of 27 mm capitulum length found in the tropics. This figure is much higher than any others quoted. Krüger (1927) quotes Burmeister (spelling it Burmester) (1843) as finding 40 000 eggs in L. anatifera whereas Willemöes-Suhm (1876) said the species was L. fascicularis and the number 4000. It has not been possible to trace the original 1843 reference in order to settle this discrepancy. Willemöes-Suhm (1876) also suggests that the Pacific and Atlantic forms of L. fascicularis may be different. In the laboratory Patel (1959) found that L. anatifera produced “ripe” eggs (presumably eggs containing larvae ready to hatch) of 290 µm long×136 µm wide at 19–20°C and 266×122 µm at 24–25°C. At this temperature breeding was continuous with embryonic development taking about two weeks. He found, however, that penis activity was inhibited below 19°C and because such a temperature is rarely reached in the North Sea the failure to reproduce in northern European waters is probably based on a failure to copulate. Patel (1959) did suggest that copulation in L. fascicularis might be possible at a lower temperature than in L. anatifera. Broch (1924, 1927), however, reports that L. fascicularis, like all lepadids cannot reproduce effectively in northern waters. Thorner (1967) says that the adults can exist in the North Sea but that reproduction is not possible except under the most optimal conditions. In the North Sea mature animals with ova in the peduncle and eggs in the mantle cavity are found on floating objects but developing stages are never found (Broch, 1924; Thorner & Ankel, 1966; Thorner, 1967) and there is a lack of larval stages in North Sea plankton. The species has a “high rate of reproduction” in warmer water. Larvae are found in the Mediterranean (Groom, 1894, quoted by Nilsson-Cantell, 1978, for L. anatifera) and off the southwest of Great Britain (Bainbridge & Roskell, 1966, for L. fascicularis). Support for this is also given by Plankton Recorder samples taken in the North Atlantic. The 16°C isotherm reaches 49°N in July and 51°N in August but falls back below 50°N in September. Larval stages of Lepas are only found during July and are limited to regions where the mean monthly sea-surface temperature is above 16°C.

82

MARGARET BARNES

In New Zealand waters L. anatifera var. testudinata breeds at 17–20°C (Skerman, 1958a). Zann & Harker (1978) found L. anatifera containing egg masses in February off north Queensland, Australia. Breeding L. hilli were in temperatures of 24–26°C en route from Dakar to Barbados (Evans, 1958). In the Bass Strait L. anserifera produced ova at temperatures normally about 17–20°C (Darwin, 1851). Other genera of the Lepadoidea about which there is some information on egg production are included in Table VIII. In all the species listed the eggs are relatively small and given as numerous. This is in agreement with the Lepas species (Table VII) and is in contrast to the Scalpellum species (Table IX) which produce larger and fewer eggs. Lepas anatifera (Patel, 1959) and Octolasmis aymonini geryonophila (Colón-Urban et al., 1979) have been shown not to be capable of self-fertilisation under laboratory conditions. Rasmussen (1980) examined several hundred Conchoderma auritum taken from fishing vessels off Namibia between May and October. 72% of animals with a scutum length greater than 7 mm contained eggs. It is not clear in how TABLE VII Lepas species: notes on breeding seasons, size and number of young, cyp=cyprid ; St I, St II=stage I and II nauplius, respectively Species

Notes

Sizes of young (L×B), Number of young per µm brood

References

L. anatifera

–

Hoek, 1884, 1909

–

egg (166 to 189)× (113 to 120) St I 250

ova 140×100 egg 240 St I 250 40 000

Nauplii found in Mediterranean Sea in Feb, Mar, May, Oct. In British waters in Sept In culture

– In Mediterranean reproduces in May From beacons in New Zealand waters mature after about 40 days –

L. anserifera

“may be 1000s or 10, 000s”

–

Krüger, 1922, 1927, 1940 Nilsson-Cantell, 1978

egg (266 to 290)× (122 to 136) depending on temperature – –

–

Patel, 1959

“many 1000s” –

Witschi, 1935 Le Reste, 1965

–

–

Skerman, 1958a

–

Adult 12.5 mm capitulum, 783 20.5 mm, 22 840

Zann & Harker, 1978

Reproduces mid Apr to June in Mediterranean; 6 days between broods

St I 308×158

–

Le Reste, 1965

EGG PRODUCTION IN CIRRIPEDES

L. australis

South Shetland Is., cyp 2550 Antarctica From beacons in New – – Zealand waters mature after about 50 days – cyp 1500–2500 – L. fascicularis – egg 260×117 Pacific between St I 350 – Japan and Hawaii in June-July; off S.W. Great Britain in Sept St II 220 – N.Pacific between 180° and 156°W— many with lamellae ova 260 300 000 in adult with St I 1300×700 capitulum 27 mm North Sea plankton St I 350 – – ova 260 – St I 350 cyp max 1300×700 L. hilli 50% animals had egg – lamellae after 60 days L. pectinata North Sea plankton St I 260 – egg 170×113 – St I 260 – Reproduces in Mediterranean in Jan, Mar, Apr, May, Oct. Off S.W. Great Britain in Sept – cyp 1250 –

–

83

Nilsson-Cantell, 1930a

Skerman, 1958a

Darwin, 1851 – Nilsson-Cantell, 1978

Krüger, 1940

Bainbridge & Roskell, 1966 Thorner & Ankel, 1966 Hoek, 1909 Willemöes-Suhm, 1876 –

Evans, 1958

– Krüger, 1940 Nilsson-Cantell, 1978

Hoek, 1909

Stebbing & Fowler, 1904

TABLE VIII Lepadoidea other than Lepas species: summary of breeding time, size and number of young per brood, cyp=cyprid; St I=stage I nauplius Species

Notes

Alepas intermedia

Collected Dec 5°36′ egg 220 S: 132°52′E Collected Feb 7°19– egg 300×130 5′S: 1 16°49′E Tasmania in Jan egg 190×110 Arabian Gulf en route St I 278×173 to N.Europe – egg 186×132

Alepas morula Alepas pacifica Conchoderma attrition Conchoderma virgatum

Sizes of young (L×B), Number of young per References µm brood numerous

Hoek, 1907

numerous

Hoek, 1907

“large egg masses” –

Tubb, 1946 Dalley, 1984

numerous

Hoek, 1884, 1909

84

MARGARET BARNES

Species

Notes

Sizes of young (L×B), Number of young per References µm brood

– Dichelaspis darwini

St I 290 Mediterranean, 3 larvae Jan –

– –

Krüger, 1940 –

Le Reste, 1965

ova 100 St I 213 egg 230 egg 227×115

–

Coker, 1901

– –

Pilsbry, 1907 Krüger, 1940

Dichelaspis mülleri Megalasma minus Megalasma (Megalasma) minus Microlepas diademae Octolasmis aymonini geryonophila Octolasmis grayii Octolasmis lowei Octolasmis mülleri Beaufort, North Carolina, USA gravid late May to mid July Octolasmis warwickii North Queensland, Australia gravid between April and Oct Poecilasma carinatum Poecilasma obliquum Trilasmis (Poecilasma) obliqua

– – Sumba, April Gulf of Mexico, cultured, reproduces July to Sept Sumatra

egg (260 to 280)×100 numerous St I 230 –

Tubb, 1946 Colón-Urban et al., 1979

–

some

egg 160×80 St I 840

– –

Nilsson-Cantell, 1930b Krüger, 1940 Lang, 1976

21–459 depending on size of adult

Jeffries & Voris, 1983

South Africa egg 180

cyp 750 smallest adult had 325 largest 12 777

– Barnard, 1924 Zann & Harker, 1978

Atlantic, West Indies and Ascension Is. S. Hemisphere, Dec –

egg 250

Hoek, 1883

gravid Mar-Apr

egg (200–220)×80 egg 200×80

numerous –

Hoek, 1907 Krüger, 1940

– South Carolina, USA gravid in June to Oct –

many of these animals measurements or counts of embryos were made but a figure of 161 µm for egg length and 164 000 for number of eggs is given. Egg number depends on the size of the parent and Rasmussen (1980) gives a graph showing values for C. auritum ranging from about 5000 eggs at a scutum length of 6 mm up to about 300 000 at 12 mm. This compares with that given by Thorner (1967) for Lepas fascicularis. Lang (1976) found gravid Octolasmis mülleri (probably=O. lowei) in South Carolina, USA, every month of the year ranging from less than 10% of the population in February to March to more than 40% in August to November. Breeding and settlement was restricted to summer and autumn as release of eggs rarely took place at temperatures below 15°C. At Beaufort in North Carolina Jeffries & Voris (1983) found gravid animals from late May to mid-July and suggested that breeding was continuous during the summer; their sampling was restricted to the summer. They found brood size to depend on the capitulum length of the adult, the number of eggs ranging from about 200 at 2 mm capitulum length to 3000 at 4–5 mm. These same authors (Jeffries & Voris, 1979) examined O. grayii from the Malay peninsula and found that about 20% of mature

EGG PRODUCTION IN CIRRIPEDES

85

adults of 3.3–17.3 mm capitulum length . contained ova or eggs on the dates they sampled in December to April and in August. Scalpelloidea Members of the Scalpelloidea may be hermaphrodites with complemental males or females with dwarf males. In the genera Calantica and Smilium the males have a well separated capitulum and peduncle and resemble the free-living juvenile adults. In Euscalpellum the males are more sac-like and the distinction between capitulum and peduncle is not obvious; they still have cirri and a mouth. Males of the genus Scalpellum are even more degenerate being sac-like with neither peduncle nor mouth and only vestigial cirri and digestive system. Nilsson-Cantell (1932) gives an account of the males of Scalpellum species and the literature on those already described. The most recent and comprehensive review of the comparative anatomy of all cirripede males has been given by Klepal (1987). Embryonic development proceeds in egg lamellae held in the mantle cavity of the hermaphroditic or female adult. When hermaphroditic the embryos are released as stage I nauplii and lead a planktonic life up to the cyprid stage. Embryonic development seems to be abbreviated in adults that are female with embryos developing to the cyprid stage before hatching or leaving the mantle cavity (see Table IX). Barnard (1974) thought that there was no difference in life history between shallow-and deep-water Scalpellum species. Only “ova” were found in two of the deepest forms he studied and only “early metanauplius” in a third. All the species were, however, female with dwarf males so that it is probable that all had an abbreviated embryonic development. There has been some discussion on the merits or otherwise of abbreviated larval development (see Nilsson-Cantell, 1921; Kaufmann, 1965). Advantages include the reduced chance of predation which is present during a prolonged planktonic life and as a consequence fewer larvae need to be produced. A disadvantage may be lack of dispersal of the species by the larvae but because TABLE IX Scalpelloidea: summary of literature on Scalpellum and related species, cyp=cyprid; St I—stage I nauplius. St ?=stage not known; F=female; H=hermaphrodite Species

Place

Date collected or breeding

Size of adult (L×B), mm

Size of young (L×B), µm

Number of young per brood

References

Calantica pollicipedoide s Acroscalpellu m sp.

–

–

–

egg 340×280

–

Krüger, 1940

58°15′N: 48° 36′W

–

14.8×7.8

–

Newman & Ross, 1971

65°53′S: 70° 56′W 61°18′S: 56°9′ W 63°51′S: 62° 38′W 65°51'S: 82° 40’W

–

–

St ? (850×650) to (900×750) cyp 1300×650

–

–

F, 7.7×0.9

cyp 750×250

50 to 60

Newman & Ross, 1971 Newman & Ross, 1971

–

F, 85×34.5

egg 1500×1000

about 300

Acroscalpellu m africanum A. angulare

A. darwinii

Newman & Ross, 1971

86

MARGARET BARNES

Species

Date collected or breeding

Size of adult (L×B), mm

Size of young (L×B), µm

Number of young per brood

References

67°55’S: 90° 43’W –

–

–

egg 420×350

–

Krüger, 1940

–

–

–

egg 400×250

–

Krüger, 1940

–

–

–

egg 800×550

–

Krüger, 1940

Kerguelen Is. – – –

– – – –

6.0 – – –

– egg 900×600 egg 500×330 egg 430×300

39 – – –

Zevina, 1974 Krüger, 1940 Krüger, 1940 Krüger, 1940

– – Kerguelen Is.

– – –

– – 6.0

– S. Africa

– –

– F, 5.5

egg 500×300 – egg 1030×870 – egg 580 59 to 82 cyp 1100×520 egg 470×270 – – 14

Krüger, 1940 Barnard, 1924

5°42′S: 132° 25′E S.Africa

Sept

–

egg 420–350

about 20

Hoek, 1883

–

F, 4.0

cyp 1000×500

9

S.Africa

–

F, 10.0

cyp 750×500

24

Barnard, 1924; Thorner, 1967 Barnard, 1924

S.Africa

–

F, 4.0

cyp 400×250

8

S. cancellatum S.Africa

–

F, 6.5

cyp 750×500

S. capense S. chiliense

– Oct

F, 5.0 7.0

cyp 700×400 egg 500×400

S. compactum

S.Africa 32°18′S: 71° 50′W –

“small number” 15 to 17 21

–

–

cyp 1030

–

S. compressum

2°55′N: 124° 53′E

Oct

–

egg 800

“not numerous”

Scalpellum (Acroscalpell um) balanoides S.(A.) brevicaulis S.(A.) compression S.(A.) eugenie S.(A.) eumitos S.(A.) gracile S.(A.) hexagonum S.(A) micrum S. (A.) regium S.(A.) sergi S.(A.) sessile Scalpellum agulhense S. balanoides S. bolellinae

S. brachiutmcancri S. brevicaulis

Place

Krüger, 1940 Krüger, 1940 Zevina, 1974

Barnard, 1924; Thorner, 1967 Barnard, 1924 Barnard, 1924 Zevina, 1972 NilssonCantell, 1921 Hoek, 1883

EGG PRODUCTION IN CIRRIPEDES

S. convexum

–

–

–

S. Georgia

–

–

cyp 940

egg 780 cyp 1380 numerous

S. cornutum

–

–

–

cyp 750

S. eumitos S. faurei S. gibberum

S. Africa S. Africa E. Falkland Is.

– – May

F, 10.0 F, 6.0 –

cyp 900×600 cyp 750×400 egg 1090 to 1500

S. gracile

5°40′S: 120° 45′E S. Africa

Sept

F

egg 500×330

–

F, 9.0

egg 500×300

S. Africa S. Africa 37°17′S: 53° 52′W

– – Feb

F, 3.5 F, 6.5 –

cyp 600×300 cyp 800×400 egg 750

24°04′S: 70° 45′W 5°28′S: 134° 55′E

Sept

6.0

St ? 460

20

Zevina, 1972

Dec

H, 6.3×4.0

egg 340×280

6

Hoek, 1907

–

–

H

cyp 1250×510

–

June

H

ova 600 egg 1030×870

about 400

S. scalpellum

34°54′N: 56° 38′W 35°29′N: 50° 53′W British waters

Kaufmann, 1971 Hoek, 1883, 1884

Apr to Nov. Max inAug

H

St I 475×431

–

–

–

H, 16 to 26

110 to 582

–

Breed 3 times a year

H, 6.5 to 18.5

egg 500×(340 to 380) egg 433×352 St I 470

S. sessile

–

F

S. sinuatum

4°24′S: 129° 49′E –

28 to 3569 av. of 229 animals=402 egg 470×270

Krüger, 1922, 1927, 1940 Kaufmann, 1962, 1965

–

F, 6.0

–

S. slearnsi

–

–

H

–

S. micrum S. natalense S. ornatum S. parallelogram ma S. perlongum S. pollicipedoide s S. regina S. regium

– NilssonCantell, 1930a –

– less than 20 NilssonCantell, 1921, 1930a 53 “small number” 12 30 or less “not numerous”

NilssonCantell, 1921

Auivillius, 1894; NilssonCantell, 1978 Barnard, 1924 Barnard, 1924 cyp 1230 to 1920 Hoek, 1907 Barnard, 1924 Barnard, 1924 Barnard, 1924 Hoek, 1883

Bassindale, 1936; NilssonCantell, 1978

17

Hoek, 1907

“small number” eggs seen

Barnard, 1924 Hoek, 1906

87

88

MARGARET BARNES

S. stroemi

–

–

F

cyp 700

S. subalatum

S. Africa

–

F, 5.0

cyp 800×500

S. triangulare

Feb

F

–

S. uncinatum S. valvulifer

37°17′S: 53° 52′W S. Africa S. Africa

– –

F, 6.0 F, 6.0

S. vetutinum

–

–

–

cyp 1000×600 ova 500×300 cyp 750×400 –

–

–

June – – – H, 4.0 –

S. – ventricosum S. vulgare Mediterranean – – Scalpellum sp. – Smilium – hypocrites S. Africa – S. Africa Smilium pollicipedoide s

“large number but not numerous” “small number” “full”

Hoek, 1883, 1909; NilssonCantell, 1978 Barnard, 1924 Hoek, 1883

15 up to 30

Barnard, 1924 Barnard, 1924 Barnard, 1925

cyp 1330

Several hundred ova –

– ova 300 – –

St I 625×400 – – ova 200

– Hoek, 1884 45 –

ova 200 H, 1.5 to 12

20–30 ova 200

Barnard, 1924 at least 150

NilssonCantell, 1921 Le Reste, 1965 Thorner, 1967 Krüger, 1940

Barnard, 1924

the adults are female and dwarf males are essential for reproduction this lack of dispersal ensures that an adequate supply of cyprids capable of becoming dwarf males is always available. There has been very little work done to determine whether there are two types of cyprids, male and female. Stewart (1911) said there were two sizes of cyprids in S. squamuliferum, the smaller becoming complemental males and the larger the hermaphrodites. This was confirmed for the same species by Stubbings (1936). Svane (1986) working with S. scalpellum, an hermaphrodite with complemental males, has made some experimental observations. He found that isolated adult animals containing no complemental males did not produce egg lamellae and, therefore, self-fertilisation did not take place. He also found that all cyprids were potential hermaphrodites but that only about 50% also possessed the ability to metamorphose into males if they encountered an hermaphrodite with available receptacles. He concluded that genetic and environmental conditions influenced the determination of sex in this species. Similar work on female adults and dwarf males would be of value. As with many of the Rhizocephala and Acrothoracica what information there is on egg production is scattered throughout the literature describing material collected during expeditions. Practically no experimental work has been done on the Scalpelloidea. Kaufmann (1965) observed copulation in S. scalpellum in the laboratory and compared the size of the adult hermaphrodite and number of eggs in 229 cases. He found an average of 402 eggs per animal, with a range of 28 to 3569 (see Table IX), depending on the size of adult and time of the breeding season. Krüger (1927) counted eggs in the same species ranging from 16 to 26 mm in size and found 122 to 582 eggs depending on parental size. The number of eggs in other species can range from eight up to “several hundred” (Table IX). When a very small number is recorded it is not possible to know whether any eggs were lost during collection or preservation. Indeed this

EGG PRODUCTION IN CIRRIPEDES

89

can apply in many cases. When a reference states 150 eggs in each lamella then one can be sure there were 300 eggs present, as in Acroscalpellum dawrinii (Newman & Ross, 1971), and in Scalpellum regium, with approximately 200 eggs in each lamella (Hoeck, 1883) there would be a total of about 400. The times of breeding have only been followed closely by Kaufmann (1965) for S. scalpellum found in the Gullmarfjord, Sweden. He took regular samples and also examined animals of known age in the laboratory. He found that in Sweden an animal can probably carry three broods a year. There were relatively more animals with fewer eggs in the autumn than in early summer which suggests that the numbers of eggs may depend on the food available for ovarian development. Nutrition will decline as the adults use up their reserves and are not able to replenish them as food in this northern environment declines in late summer and autumn. In British waters, Bassindale (1936) gives the breeding peaks for S. scalpellum as mid-April, September, and November; Nilsson-Cantell (1978) gives the maximum in mid-August. For several of the other genera the information on the time of breeding is very sparse (see Table IX). Dates of collection of material in a reproductive state vary from 17 June to 20 October in the Northern Hemisphere and 14 February to 26 December in the Southern. Much of the material is from deep water where environmental conditions are relatively stable. Barnard (1924) was of the opinion that there was no particular breeding season. In shallower water the effects of temperature, light, and availability of nutrients can cause seasonal variation as shown in S. scalpellum. In general, the size of eggs produced by these genera is larger and the number less than in the Lepadoidea with a large number of very small eggs. Species of the genus-Pollicipes are hermaphrodites without complemental males. They are usually found in dense groups in the intertidal zone orientated so that they can feed as the backwash of the tide rushes over them. The length of peduncle can be adjusted to bring the capitulum into a suitable position for feeding. The egg lamellae form two thin sheets one on either side of the prosoma in the mantle cavity. The three main species about which there is information on the reproduction are P. polymerus from the Pacific coast of North America, P. cornucopia from southern European shores, and P. spinosus from New Zealand. The available data on egg sizes and numbers are given in Table X. Smith & Weldon (1920) thought that self-fertilisation might take place in P. polymerus but this has now been dis-proved by Hilgard (1960) and Lewis & Chia (1981). The last authors never found egg lamellae in animals separated by 11 cm or more from the nearest neighbour. On the Pacific coast of North America the time and length of breeding season of this species varies throughout its range of distribution (Cimberg, 1981; Lewis & Chia, 1981). Page (1986) found differences in size and age at maturity depending on the habitat. Reproductive patterns, percentage of adults with egg lamellae and the size of the lamellae have been studied by Page (1984). When standardised for size of adult, he found seasonal changes in the percentage of adults carrying egg lamellae; the weight of the lamellae ranged from 4 to 21% of the body weight. He has also investigated the effect of sea-water temperature and food on energy allocation in P. polymerus (Page, 1983). According to Cimberg (1981) there are two physiological races of P. polym erus on the Pacific coast of North America. The northern race shows maximum breeding activity at sea-water temperatures of 14°C or less and the southern race at 20°C. He found three types of adult: Type 1, north of Point Conception, which breeds during the summer as sea-water temperature approaches about 14°C; Type 2, at Latigo Point, south of Point Conception, which breeds during the summer as sea-water temperature increases up to 20°C; and Type 3, also found south of Point Conception, at Goleta Point and on Santa Catalina Island, which breeds during the winter as sea-water temperature falls to about 14°C. Types 1 and 3 are, therefore, breeding at the same temperature but not in the same season and constitute the northern race; Type 2 represents the southern race. He suggests that the similarity between Types 1 and 3 is because larvae north of Point

90

MARGARET BARNES

Conception are transported in the south-flowing California Current during the oceanic and upwelling seasons. Lewis & Chia (1981) have summarised data on the number of broods produced per year by P. polymerus. It can vary between about three to four (presumably the northern race) and eight at Santa Barbara (southern race). This figure of eight broods at Santa Barbara is given by Lewis & Chia (1981) as quoting Straughan (1971) although the figure does not seem to be in the latter reference. The number of eggs per brood also varies; compared with P. spinosus (New Zealand) and P. cornucopia (Europe), P. polymerus produces TABLE X Pollicipes and Lithotrya species: summary of breeding seasons, sizes and number of young, and number of broods per year. St I=stage I nauplius; o.d.=oven dry Species

Place

Breeding season

Size of young (L×B), µm

Number of Number of References eggs per brood broods per year

Pollicipes cornucopia

S. Spain and Gibraltar

63% with eggs in July

egg 245

15 400 per 25 mg o.d. body wt

–

Policipes polymerus Monterey Bay, Calif.

Santa Monica, Calif. April onwards, peaks in June, Sept, Dec –

–

St I 173×82

–

–

ova 100–127

104 000–240 000

4 to 5

Hilgard, 1960

–

144 000–288 000

2–4 depending

San Juan Is.

Friday Harbor, Wash.

–

ova 105×80 egg 250×130

–

–

Santa Barbara, Calif. Pollicipes spinosus

–

St I 207×114

–

8

New Zealand

5 years to reach maturity

1500–7000

Lithotrya dorsalis Barbados, West Indies

–

–

egg early 600×445 egg late 690×565 egg 570

Lewis & Chia, 1981 on situation Lewis, 1975 and Bodega Bay, Calif. Lewis & Chia, 1981 2

–

–

short breeding season, June– July Collected Sept–Oct –

St I 320

few hundred

–

Lewis, 1960

St I 362±25

–

–

Dineen, 1987

egg (235 to 276)×98 –

–

–

Sewell, 1926

egg 570

–

–

Indian Key, Florida Lithotrya nicobarica Lithotrya truncata

Philippines

Darwin, 1851; Barnes & Barnes, 1966, 1968 Barnes & Barnes, 1965

Batham, 1945b, 1946

Krüger, 1940

Darwin, 1851

EGG PRODUCTION IN CIRRIPEDES

91

many more eggs per brood and has more broods per year (Table X). It is not clear whether the low number of 100 young produced per brood at Santa Barbara (Straughan, 1971) is due to the effect of oil seepage or a gradual reduction towards the southern end of the range of distribution but it is 1000 times (not 100 times as quoted by Straughan, 1971) less than the figure given by Hilgard (1960) for Monterey Bay. Lewis & Chia (1981) say that “approximately seven times as many eggs are produced per brood by P. cornucopia at Cabo Silleiro and Gibraltar (Barnes & Barnes, 1968) compared with P. polymerus at San Juan Island”. This is incorrect: Barnes & Barnes (1968) give a figure of 15 400 eggs from a moderate size P. cornucopia in Europe which is seven times less (not more) than found at San Juan Island. P. cornucopia at Cabo Silleiro, Spain, contain egg lamellae in July and at Biarritz, France, 22% of the population contain egg lamellae in April, 85– 90% in July, and 16% in October (Barnes & Barnes, pers. obs.) but it is not known how many broods are produced per year. P. spinosus produces a very much larger egg than the European or American species (see Table X) and the larvae are lecithotrophic. It is not surprising, therefore, that it produces fewer eggs per brood (Batham, 1945b, 1946) in an animal of about the same size as P. polymerus. In addition, P. spinosus only produces two broods a year and so far fewer eggs per year than does P. polymerus with say three to four broods per year. Reproduction in P. spinosus at Otago, New Zealand, begins in December to January, as sea-surface temperature rises to about 14°C, reaches a peak in February to April and by August has finished by which time the sea-surface temperature has dropped to about 8°C. This temperature compares well with that of the northern race of P. polymerus. Lithotyra species are generally found in tropical seas in association with soft coral reef limestone into which they burrow. They are hermaphrodites, although Sewell (1926) says that L. nicobarica may be protandrous and only hermaphroditic when fully adult. The body is lodged partly in the peduncle with the ovary filling the peduncle. Darwin (1851) found eggs in L. truncata from the Philippines and remarked on their large size. Lewis (1960) records L. dorsalis in Barbados as having a short breeding season in June and July and producing a few hundred eggs. He says that “larvae are typical cyprids” and that a newly hatched cyprid is 320 µm long and 640 µm after 12 hours. He presumably means that the larvae are typical cirripede nauplii and that stage I is 320 µm lone. Dineen (1987) gives the length of L. dorsalis stage I nauplius as 362 ±25 µm The few sizes that are given in the literature are summarised in Table X. VERRUCOMORPHA The Verrucomorpha contains asymmetrical sessile barnacles which may have a membranous or a calcareous base. They are hermaphrodites. Darwin (1854) recorded four species “from Iceland to Cape Horn”. According to Newman (1987) the genus Verruca now contains about 60 species. They usually occur in deep water up to near inshore and shallow area (Zevina, 1987, 1988). V. stroemia is found in depths of 10–100 m in British waters and is common in northern Europe and the northern Mediterranean but has not been recorded from the Baltic Sea (Stone & Barnes, 1973). The reproductive TABLE XI Verruca species: summary of breeding and sizes of young. St I=stage I nauplius Species

Place

Notes

Sizes of young (L×B), µ m References

Verruca striata Verruca stroemia Northern waters

– Adriatic Sea Larval stages in plankton in summer

– One brood per year egg 180

egg 322×252 – Nilsson-Cantell, 1921, 1978

Krüger, 1940 Kolosváry, 1947

92

MARGARET BARNES

Species Millport, Scotland

Place

Larval stages in plankton most of year except Nov. Most in spring and early summer Oban, Scotland Several broods a year, main one in spring, ceases breeding in October British waters – British waters – Marseilles, France Reproduces from mid May to August

Notes St I 260×120 –

–

St I 270×120 egg 203 –

Sizes of young (L×B), µ m References Pyefinch, 1948; Barnes & Crisp, 1956; Barnes, 1958

Stone & Barnes, 1973; Barnes & Stone, 1973; Barnes & Barnes, 1975 Bassindale, 1936 Darwin, 1854 Le Reste, 1965

cycle of V. stroemia in Scotland has been considered in detail by Stone & Barnes (1973, 1974), Barnes & Stone (1973), and Barnes & Barnes (1975). Because of the asymmetry of the animal neither the ovaries nor the egg lamellae are equal in size, nor do they lie symmetrically on either side of the mantle cavity although both are appressed to the prosoma. The larger lamella lies above the prosoma (as this lies parallel to the substratum) and the smaller one lies below the prosoma. The larger lamella contains about 1.5 times as many eggs as the smaller one. Breeding ceases during the autumn and early winter in natural populations because of lack of food but if fed in laboratory cultures breeding will continue. Breeding is not inhibited in either constant light or dark but a temperature of 20°C is lethal even over moderate periods. Because of the lack of breeding in the natural population in the winter there is virtually complete synchronisation of the first brood of the year in early spring. Thereafter, synchrony is gradually lost as the animals produce further broods throughout the summer. Most authors report several broods per year in V. stroemia except Kolosváry (1947) who records only one in the Adriatic Sea. Verruca, being essentially a deep-water animal is subjected to a cold and relatively stable environment and information on the breeding behaviour of other species would be of interest. Ova and egg or stage I nauplius sizes are given in Table XI for V. stroemia from several habitats but information on only one other species, V. striata, could be found and this has a slightly larger egg than V. stroemia. BALANOMORPHA Members of the Balanomorpha are sessile barnacles with bilaterally sym-metrical shells and may have a membranous or a calcareous base. They are hermaphrodites; a very few species have been found with complemental males. Of the five superfamilies suggested by Newman (1987) nothing could be found in the literature regarding egg production in the Chionelasmatoidea and the Pachylasmatoidea. In contrast a considerable literature has developed on the other three. This involves general ecological studies and also a great amount of experimental work on the factors and conditions regulating egg production and on the biochemistry of the eggs themselves. The general ecological aspects will be considered first. Chthamaloidea Within this superfamily most of the work has been done on Chthamalus a warm-water genus some of which, such as C. fragilis, C. stellatus and C. dalli, extend into the north-temperate area. Southward & Southward

EGG PRODUCTION IN CIRRIPEDES

93

(1967) found C. dalli, at what is the northernmost recorded place for a chthamalid, in the Chukchi Sea. Here the shore is only ice-free from May to October and breeding takes place at a sea-water temperature of 6°C. Some animals even attempted to copulate at 4°C suggesting that low temperature is little handicap to fertilisation. These authors found 54% of the population with egg lamellae in late July and they suggested that there might be two broods a year. Korn & Kolotukhina (1983) found that C. dalli had two broods per year between March (when the sea-water temperature is 2°C) and June in the Sea of Japan. TABLE XII Chthamalus species summary of breeding seasons, sizes and number of young. St I=stage I Nauplius; o.d. =oven dry ; +=or more Species

Place

Breeding

Size of young (L×B), µm

Number of eggs per brood

References

C. anisopoma

Gulf of California

egg 163×82

–

Barnes & Barnes, 1965; Malusa, 1986

C. antennatus

New South Wales, Australia Hakodate, Japan

Breeds all year, at 13–33°C; no seasonal pattern. 44% fertilised in June, 52% July, 76–94% Dec– Feb 80% fertilised in Jan (summer) Fertilised Apr– July to June and late Aug to early Oct. Low levels continuous Mar to Oct. At least 2 broods perhaps 3 at low levels –

–

–

ova 110×80

–

Wisely & Blick, 1964; Pope, 1965 Iwaki, 1975

–

54% with eggs in July 1+broods a year –

–

Luckens, 1968, 1969 –

C. challengeri

Asamushi, Japan C. dalli

Upper levels breeds April Chukchi Sea

Sea of Japan

Two broods a year

–

Vladivostok

–

St I 218×112

–

Pacific coast, USA C. dentatus

–

St I 183×102

–

Cape Town, South Africa

Ripe eggs Oct to May, peak in Mar. Ripe again in Aug with peak in Dec

St I 140×100

Korn & Kolotukhina, 1983 Korn & Ovsyannikova, 1979 Barnes & Barnes, 1965 –

Southward & Southward, 1967

Bokenham, 1938; Sandison, 1954

94

MARGARET BARNES

Species

Place

Breeding

Size of young (L×B), µm

Number of eggs per brood

False Bay, South Africa

Breeds Sept to April but not in May to Aug

egg 191×111 St I 201×108

–

Gulf of Guinea

–

–

C. depressus

Mediterranean

egg (166 to 196) ×(90 to 102) Breeds May to Sept

Achituv & Wortzlavski, 1983; Achituv, 1986 Stubbings, 1961

Euraphia depressa (=C. depressus) C. fissus

Ligurian Sea

1298 eggs per 0.5 mg o.d. body wt –

Le Reste, 1965; Barnes & Barnes, 1968 Relini, 1983

egg 130×95

802–2640 depending on age. 990 eggs per 0.5 mg o.d. body wt

Hines, 1974, 1978, 1979

3 broods a year

–

–

Villalobos, 1979a

10–15% with eggs for 9 months. In culture submerged 80– 90% had eggs all year –

–

–

Page, 1984

egg 153×82

–

Eyed embryos in eggs in Jan Breeds all year except during monsoons July– Sept. In Jan temp, may be too low

–

–

C. malayensis

Papua New Guinea Bombay, India

Barnes & Barnes, 1965 Pope, 1965

St I 205 to 216

980 in adult of 4 mm diameter 1890 in one of 9 mm, also give max values of 3500 and 4300

Queensland, Australia

Eyed embryos in eggs Nov to

–

–

Western Australia

Eyed embryos at end Dec

–

–

Pope, 1965; Hines, 1966 June, peak Nov to Feb Pope, 1965

Ripe eggs all year, late stages in June to Sept Morro Bay, Calif. Peaking breeding June to Sept estimated 16 broods between Mar and Oct Santa Barbara, Calif. Goleta Pt., Calif.

California C. intertextus

St I 200×149

References

–

Karande & Palekar, 1963; Wagh, 1965; Karande, 1967, 1974a; Karande & Thomas, 1976; Gaonkar & Karande, 1980

EGG PRODUCTION IN CIRRIPEDES

C. montagui

Ligurian Sea

C. stellatus

Adriatic Sea

Marseilles, France

3–5 broods from June to Oct. A few fertilised in Mar to May Arcachon, France Breeds from Apr to May Ligurian Sea All year except Oct and Nov Mediterranean Breeds Mar to Oct Northern waters – Great Britain More than one brood from June onwards to Oct

Shetland and Fair Breeds early Isle, GB June to Sept Europe in general –

C. withersi

Bombay, India

Eggs all year – except Oct and Nov Eggs July–Dec, 2 – broods – –

Le Reste, 1965

–

–

Dessenoix, 1962

–

–

Relini, 1983

–

1528

St I 220 ova 130 egg 191×94 St I 190×90

– –

egg 230×130

–

Tenerelli, 1958, 1959 Hoek, 1909 Bassindale, 1936; Crisp, 1954; Barnes & Barnes, 1965; Nilsson-Cantell, 1978 Powell, 1954

–

500 to 2500 per 0.5 mg o.d. body wt depending on place and shore level St I 205 to 215

Breeds mainly Mar–June also in Sept–Nov and Feb–Mar

–

Relini, 1983

–

Kolosváry, 1947

95

Barnes & Barnes, 1968

7000 max 7100

Karande & Palekar, 1963; Karande & Thomas, 1976

Similarly, C. challenged has a restricted breeding season in April to July near the northern limits of its distribution at Hakodate (Iwaki, 1975), At the northernmost limits of its distribution in Scotland C. stellatus produces two broods annually between May and September (Crisp, 1950; Powell, 1954; Barnes, 1972; Achituv & Barnes, 1976) but only the second reaches maturity. It should be noted that because Southward (1976) found that some European C. stellatus should have been called C. montagui some papers referring to C. stellatus may have been concerned with a mixture of the two species. Chthamalids in the more southern parts of their distribution generally produce more than two broods per year depending on temperature and availability of food but rarely is reproduction continuous. There is spermatogenetic activity in C. malayensis throughout the year according to Karande & Palekar (1963) but it is greatest during periods of fertilisation. The available data from various parts of the world are summarised in Table XII. The time of breeding and number of broods may depend on shore level as in C. challenged (Luckens, 1968, 1969). In similar situations C. malayensis has more broods than C. withersi but produces fewer eggs per brood (Karande & Palekar, 1963; Karande, 1974a; Karande & Thomas, 1976; Gaonkar &

96

MARGARET BARNES

Karande, 1980). The 16 broods said to be produced by C. fissus in central California (Hines, 1978) is grossly in excess of the number found in any other Chthamalus species including C. fissus at Santa Barbara in California (Villalobos, 1979a) and may be an over-estimate. It is based on the assumption that for eight consecutive months an animal produces a new brood immediately the previous brood has been released; it must also be assumed that adequate food reserves and ideal environmental conditions are available. Barnes & Barnes (1965, 1968) have recorded the size of egg and/or stage I nauplius of C. stellatus and C. depressus on European shores as well as the number produced by both species. Egg sizes are about the same (190–200 µm) in each case. C. depressus produces about 1300 eggs per 0.5 mg oven dry body weight throughout its range of distribution in Europe whereas the number in C. stellatus varies with latitude and intertidal position from about 500 to 2500 per 0.5 mg oven dry body weight. In general the fecundity in C. stellatus is relatively lower in protected areas such as harbours than in more exposed situations. This may be due to turbidity in harbours and quiet bays. There is, however, the overriding effect of nutrient supply, the highest number of eggs being found at Arcachon, France, where there is shelter and the water is turbid but nutrient level is very high. These numbers are comparable with those given by Hines (1974, 1978, 1979) for C. fissus in California. The only other numbers given in the literature are for C. withersi, 7000 per brood and C. malayensis, 4300 per brood (Gaonkar & Karande, 1980). Karande & Palekar (1963) give 980 in an adult C. malayensis of 4 mm basal diameter and 1890 eggs in one of 9 mm basal diameter. These two species are from Bombay waters. Available information on other species of the Chthamaloidea is summarised in Table XIII. Chamaesipho columna near Sydney, Australia, is said to breed mainly from June to October with no egg lamellae present for about six months of the year (Wisely & Blick, 1964). At Leigh in New Zealand, Moore (1944), Foster (1967), and Luckens (1970, 1976) report the same species as breeding throughout the year whereas C. brunnea, also at Leigh, breeds only in spring and summer (November to April). The cultured stage TABLE XIII Chthamaloidea other than Chthamalus species: breeding seasons and sizes of young. St I=stage I nauplii Species

Place

Breeding season

Size of young (L×B), µm

References

Catophragmus polymerus

Sydney, Australia

–

Wisely & Blick, 1964

Chamaesipho brunnea Leigh, N.Z.

New Zealand

Peak breeding seasons Jan to Apr, June to Sept Breeding Sept to Feb

–

Moore, 1944

New Zealand

Chamaesipho columna Leigh, N.Z.

New Zealand

Breeds mainly in Nov to Apr –

New Zealand Breeds all year, settles all year at low levels but only in Apr to Jan at high levels –

St I 350×190 St I cultured (190 to 210)× (90 to 100) Breeds all year

Foster, 1967; Luckens, 1970 Barker, 1976

–

St I 300×170

Foster, 1967; Luckens, 1970, 1976

St I cultured

Barker, 1976

Moore, 1944

EGG PRODUCTION IN CIRRIPEDES

Species

Place

Sydney, Australia

Breeds mainly June to Oct Nothing for about 6 months of year Cape Town, South Africa

Octomeris angulosa

False Bay, South Africa

Peak breeding in Aug to Jan Egg masses present most of year except July

Breeding season (200 to 230)×100 –

Late stage eggs in Oct After release empty until June, then peaks in Oct egg 212×133

Size of young (L×B), µm

97

References

Wisely & Blick, 1964

St I 180×100

Sandison, 1954

Achituv & Wortzlavski, 1983

I nauplii (Barker, 1976) of these two species seem to be smaller than in the natural populations (Foster, 1967). Catophragmus polymerus at Sydney has peak breeding seasons from June to September and January to April (Wisley & Blick, 1964). In South Africa Octomeris angulosa at Cape Town contains late stage embryos in eggs in October; there are none from February to June (Sandison, 1954). In False Bay Achituv & Wortzlaviski (1983) found egg masses most of the year except in July; peak breeding was from August to January. Coronuloidea Most information concerns the family Tetraclitidae (Table XIV) which is intertidal and found in tropical and warm-temperate regions. It would, therefore, be expected that Tetraclita species might breed throughout the year. Breeding seasons, however, vary and are not always continuous. For instance T. squamosa rufotincta in the Gulf of Elat has a distinct breeding cycle with a peak in November to December; only 0.5% of the population may be found with egg lamellae at other times of the year (Achituv, 1979; Barnes & Achituv, 1981). According to Lewis (1960) T. s. stalactifera breeds throughout the year in Barbados whereas in Costa Rica T. stalactifera contains lamellae with eyed embryos in October to November (Villalobos, 1980). T. squamosa japonica breeds most of the year with eyed embryos present from June to September in Shimizu Harbour, Japan, (Kosaka & Ishibashi, 1979). Egan & Anderson (1988) record Tesseropora rosea as having a clearly defined breeding season in summer and early winter in New South Wales while Denley & Underwood (1979) say there is much local variability on the same shores. In Tetraclitella purpurascens and Austrobalanus imperator there are peaks of breeding in New South Wales in winter to early summer in the former species and in late autumn to early winter in the latter with some gravid animals present throughout the year in both species (Egan & Anderson, 1988). Villalobos (1979b) suggests that Tetraclita rubescens at Santa Barbara, California, which reaches sexual maturity at the end of its second year may produce offspring one year when there is no somatic growth and then grow the next year, so that in one year energy is allocated to development of gonads rather than growth. T. squamosa rufotincta at Elat becomes mature during its second year (Achituv, 1979) as does T.s. japonica in Japan (Kosaka & Ishibashi, 1979).

98

MARGARET BARNES

Size of late stage eggs and/or stage I nauplii are generally about 240– 310 µm in length. There are three notable exceptions, T.s. rufotincta, T. (Tesseropora) pacifica, and T. divisa with eggs of 508, 527, and 840 µm long (Barnes & Achituv, 1981; Crisp, 1986; Nilsson-Cantell, 1921), respectively. These three species all produce lecithotrophic larvae. The larvae of the first two species pass through the usual six moults to the cyprid stage but development in T. divisa is abbreviated to four stages brooded within the mantle cavity; the cyprid is released from the mantle cavity but has a restricted larval life (Anderson, 1986). The number of eggs produced per brood is recorded in two species in each case depending on the size of the parent. Tetraclitella karandei from Bombay waters contains about 5900 eggs per brood (Gaonkar & Karande, 1980) and Tetraclita squamosa japonica in Japan has 13 000 to 71 000 (Kosaka & TABLE XIV Tetraclita, Tetraditella and Tesseropora spacies: summary of available literature on breeding seasons, size and number of young. cyp=cyprid; St I=stage I nauplius Species

Place

Breeding season

Size of young (L×B), µm

Number of eggs per brood

References

Tetraclita divisa

–

egg 840

–

Nilsson-Cantell, 1921, 1978

Malay Archipelago Tetraclita japonica

–

According to Nilsson-Cantell this is only known operculate to develop to cyprid within the mantle cavity of adult St I 620

20 to 30

Hiro, 1939

egg 300

13 000 to 71 000

Kosaka & Ishibashi, 1979

Tetraclita karandei

Bombay, India

Breeds most of year, eyed embryos in eggs June to Sept –

St I 240

5900

Tetraclita (Tesseropora) pacifica Tetraclita purpurascens (=Tetraditella purpurascens) Leigh, New Zealand

–

–

egg 527×387

–

Karande, 1974b; Karande & Thomas, 1976; Gaonkar & Karande, 1980 Crisp, 1986

Sydney, Australia

–

–

–

Foster, 1967; Harker, 1976

Tetraclita (Tesseropora) rosea

Sydney, Australia

A few nauplii throughout year, peak settlement Oct to Dec St I cultured (290 to 310)× (130 to 140) St I 470×200 Breeds Feb to Mar, releases in May. Mantle

–

–

Shimizu Harbour, Japan

–

Wisely & Blick, 1964; Denley & Underwood, 1979

Wisely & Blick, 1964

EGG PRODUCTION IN CIRRIPEDES

Species

Place

Tesseropora rosea

Sydney, Cape Banks Australia

Tetraclita rubescens

Santa Barbara, California

Tetraclita serrata

Cape Town, South Africa

South Africa

Breeds July to Oct California

Tetraclita squamosa rubescens Tetraclita squamosa rufotincta

Gulf of Elat

Tetraclita squamosa stalactifera Tetraclita stalactifera

Barbados

Costa Rica

Gonads Apr to Aug, eggs Oct to Nov, releases Nov to Dec

Gulf of California

Breeding season

Size of young (L×B), µm

Number of eggs per brood

References

–

–

–

–

Denley & Underwood, 1979; Otway & Underwood, 1987 Villalobos, 1979a, 1980

St I 240×170

–

–

Griffiths, 1979

Breeds June to Sept 2–3 broods a year Breeds Nov to Dec. One brood, may be two per yr Breeds throughout year

egg 340×195

10 770 to 54 090 Hines, 1974, depending on age 1978, 1979

St I 509×336

–

Achituv, 1979; Barnes & Achituv, 1981

–

–

Lewis, 1960

Breeds in summer, peaks in Aug, ends in Nov –

–

–

Malusa, 1986

–

Villalobos, 1980

cavity empty for 6 months of year Some breeding all year but 80% in Apr. Local variability Reproduces every 2 yr egg lamellae Feb to Nov Ripe gonads Oct to Feb none in summer, settles in May St I 274×154

99

Bokenham, 1938; Sandison, 1954

Ishibashi, 1979). There is no indication as to how many broods are produced per year. In contrast T. divisa produces 20–30 eggs per adult according to Hiro (1939); this lower number is to be expected because of the much larger egg size. So far it does not seem to have been possible to bring about fertilisation of ova in Tetraclita species in the laboratory even though the animals have been apparently in a suitable condition. Hines (1978) failed to get T. squamosa rubescens to fertilise although natural populations were fertilising at that time. Achituv & Barnes (1978a) and Barnes & Barnes (unpubl.) tried variable culture conditions with T. s. rufotincta but although well advanced ova and spermatozoa were produced there was no copulation. Members of the Coronulidae include Chelonibia, which according to Crisp (1983) should be spelt Chelonobia, and Platylepas species. Chelonibia testudinaria is found in tropical and temperate seas attached to turtles and according to Kolosváry (1947) in the Adriatic it has one brood a year. Pillai (1958) gives the size of stage I nauplius as wide from an individual found at Kerala, India. Platylepas species are

100

MARGARET BARNES

associated with sea snakes, dugongs, and turtles. Zann & Harker (1978) compared P. ophiophilus and P. hexastylos from northern Queensland, Australia. The length of life of the barnacles depends very much on the frequency of moulting of the host. P. ophiophilus is mature at a diameter of 2.5 mm and produces only one brood of 450–525 eggs of length 120 µm in December. The hosts, sea snakes, moult frequently and although adult barnacles can continue to live on the cast-off skin existence is then precarious, for example due to risk of burial in the sediment. P. hexastylos is mature at 6 mm diameter and eggs the same size as those of P. ophiophilus are found in May and July. Some animals can have up to 6500 eggs per brood, others only 20–100; the average seems to be about 1470–3100. Egg production is probably continuous and because the hosts, turtles, moult less often than sea snakes the barnacles can produce more than one brood in a lifetime. Bathylasma corolliforme is found in Antarctic seas down to about 1400 m. According to Dayton, Newman & Oliver (1982) larvae are continuously “dispersed into the McMurdo Sound region”. The stage I nauplius is relatively large, about 430 µm long, and is capable of a planktonic life. It is probable that there is a pool of larvae throughout the Ross Sea. Balanoidea Species of the genus Elminius are found in Australia, New Zealand, South America and Europe. There is one recorded appearance of E. modestus in South Africa (Sandison, 1950) but it has not been found again since then. It has generally been thought that the European species introduced from Australasia during World War II (Bishop, 1947) was E. modestus but after examining larvae of the Australasian species and comparing them with larval descriptions of European species, Egan & Anderson (1985) have pointed out some anomalies which cast doubt on this. In this review the European species has been called E. modestus as in all the published work from Europe. Foster (1982) has split what was known as E. modestus from eastern Australia into three species, E. modestus, E. covertus, and Hexaminius popeiana. Elminius kingii occurs exclusively in South America. TABLE XV Elminius and Hexaminius species: breeding seasons and size of young. St I=stage I nauplius; o.d.=oven dry Species

Place

Breeding season

Size of young (L×B), µm

References

E. covertus

Sydney, Australia

–

E. kingii

Chile

–

Egan & Anderson, 1985 Stefoni & Contreras, 1979; Arenas, 1982

E. modestus

Sydney, Australia

Auckland, New Zealand Leigh, N.Z.

Egg lamellae present every month except Apr Breed all year

Eyed embryos for 7 months of year –

St I cultured 230×140 egg 250×175 St I 250×140 (120 to 140 also quoted) –

New Zealand

–

St I 360×150 St I cultured (210 to 230)× (100 to 110)

Moore, 1944

Foster, 1967; Luckens, 1970, 1976 Barker, 1976

Wisely & Blick, 1964

EGG PRODUCTION IN CIRRIPEDES

Species

Place

Breeding season

Size of young (L×B), µm

Europe ien general

1800 eggs per 1 mg o.d. body wt –

–

Breeds at 7–1 8°C in May to Oct. Main period July to Sept Main peak July to Aug Some from May to Sept

–

Barnes & Barnes, 1968 Barnes & Barnes, 1965 Harms, 1984

Menai Bridge, North Wales

Nauplii all year

ova 100 to 150 egg 190×93

E. plicatus

New Zealand

–

Auckland, N.Z.

Egg lamellae every month except Apr Some breeding all year –

–

Pontevedra, Spain Heligoland, Germany

Southern England

Leigh, N.Z. Hexaminius popeiana

egg 192

St I (240 to 260) × (110 to 140)

St I 490×200 –

Knight-Jones & Waugh, 1949; Stubbings & Houghton, 1964 Crisp, 1954; Wisely, 1960; Crisp & Patel, 1961 St I cultured (300 to 320)× (120 to 140) Moore, 1944

101

References

Barker, 1976

Foster, 1967; Luckens, 1970, 1976 St I cultured 200×110 Egan & Anderson, 1985

In general, where food is adequate and sea temperatures above 6°C, breeding seasons seem to be continuous, with occasional peaks, for Elminius species (Knight-Jones & Waugh, 1949; Ralph & Hurley, 1952; Skerman, 1958b, 1959; Wisely, 1960; Crisp & Patel, 1961; Wisely & Blick, 1964; Foster, 1967; Luckens, 1970, 1976). Egan & Anderson (1985) found, however, seasonal breeding in E. covertus with peaks in June to August (winter) and September to November (spring). In Europe breeding becomes seasonal at the northern limits of the distribution of E. modestus where sea temperatures drop below 6°C such as in Heligoland (Harms, 1986). Some authors give sizes of ripe eggs or stage I nauplii and these are summarised in Table XV. The only egg counts seem to be those reported by Barnes & Barnes (1968). They found that there was no significant difference in the numbers of eggs produced by E. modestus per given oven dry weight of body throughout the whole range of distribution in Europe—from Scotland to Portugal. The value of 1800 eggs per 1 mg oven dry body weight was obtained from the common regression of egg number against oven dry body weight for all localities and refers to a moderately sized animal. The coral-inhabiting cirripedes of the genus Pyrgoma may be found as far north as the English Channel and southern Irish Sea although they are usually much further south. Hoek (1913) found P. jedani at Station Jedan, containing 60–80 eggs of size 270 µm long×140 µm wide. P. anglicum breeds in summer at the northern limit of its range of distribution producing stage I nauplii of length 280–321 µm. These are the same size as those of Acasta spongites (293– 320 µm) which also breeds in summer in the north (Moyse, 1961). According to Kolosváry (1947) A. spongites in the Adriatic produces one brood a year and cyprids are only found in the spring.

102

MARGARET BARNES

Species of the genus Balanus apparently represent several diverse origins and much has been done recently to separate these. Several species have now been separated into a new family, the Archaeobalanidae while others still remain in the Balanidae (Newman & Ross, 1976, 1977; see also Foster, 1978). For the present purpose the original names used by the authors have been used. The vast volume of data on Balanus species is summarised as far as possible in Table XVI. The differences depend mainly on whether the species is boreo-arctic, warm-temperate or tropical. Those in the north breed at lower temperatures than those further south. In many cases they only produce one brood a year with a large number of eggs compared with southern species that produce more numerous broods containing smaller eggs. The implications of these differences and the general conditions governing breeding and egg production are included in the following sections. FACTORS AFFECTING BREEDING In an introduction to reproductive rhythms in marine invertebrates Barnes (1975) discussed synchrony, and the loss of it, in the reproductive state. He reviewed the effect of temperature, light, and food on population synchrony with particular reference to cirripedes. Experimental work on the conditions affecting breeding has been carried out on several species, such as B. amphitrite var. denticulata (Patel & Crisp, 1960b); B. balanoides (Crisp, 1959, 1964b; Crisp & Clegg, 1960; Barnes, 1963; Barnes & Barnes, 1967; Crisp & Patel, 1969); B. balanus (Crisp & Patel, 1969); TABLE XVI Balanus and related species: summary of some of the literature on breeding seasons, size and number of young. St I=stage I nauplius; o.d.=oven dry Species

Place

Breeding and notes

Size of young (L×B), µm

Number of eggs per brood

References

Balanus alatus B. aligcola

Sulu Sea Table Bay, South Africa

egg 240 St I 180×98

numerous –

Hoek, 1913 Bokenham, 1938; Millard, 1952; Sandison, 1954

B. amaryllis

Bombay, India

St I 240

–

B. a. euamaryllis

Bombay, India

June Year round breeding, peaks in Sept to Dec and Mar to June. 25% with ova all year Breeds all year except Oct –

St I 240

–

B. amphitrite

Genoa, Tyrrhenian Sea, Vado, Ligure Bay

St I 270×140

–

Marseilles

Reproduces May to Nov. Probably up to 8 broods a year

17–18°C temp min for fertilisation Optimum 22–23° C. Settles Mar to Nov –

Wagh, 1965; Karande, 1974a Karande & Thomas, 1976 Relini, 1968; Relini & Giordano, 1969; Geraci & Romairone, 1986

–

Le Reste, 1965

EGG PRODUCTION IN CIRRIPEDES

Species

Place

Breeding and notes

Size of young (L×B), µm

Goa, India

–

3380 to 18 700 in Harkantra et al., adults 3.8 to 9.5 1977 mm diam

–

–

Paul, 1942

–

–

Weiss, 1948b

Europe

Egg lamellae all year but best in pre-and postmonsoon when salinity is high Matures in 16 days Breeds at above 20°C, peaks in Mar, May, June, Oct. No breeding in winter at less than 20°C –

–

B. a. albicostatus

Japan

–

4800 per 3 mg o.d. body wt St I 240×140

Barnes & Barnes, 1968 –

B.a. amphitrite

Bombay, India

–

St I cultured 225

18 700 4607 per 1.3 mg o.d. body wt

Cochin, India

–

–

Andaman Is., Indian Ocean

St I cultured 180×120 –

–

Sydney, Australia

Breeds all year, particularly in Jan to May 25–68% with – ripe eggs from June to Jan. Main release Nov to Mar

Kalyanasundara m & Ganti, 1975 Karande, 1978

–

Wisely & Blick, 1964

Species

Place

Breeding and notes

Size of young (L×B), µm

Number of eggs per brood

References

B. a. communis

Cochin, India

–

250–300

Pillay & Nair, 1972

Bombay, India

Breeds in warmer months but not in May if temp is above 33°C. Peaks usually in

Two peaks of breeding in Dec and May. No breeding in July and Aug St I cultured 180

–

Wagh, 1965; Karande, 1967; Karande & Thomas, 1971

Madras, India Florida, USA

Number of eggs per brood

103

References

Ishida & Yasugi, 1937 Karande, 1973, 1974c; Karande & Thomas, 1976; Gaonkar & Karande, 1980; Rege, Joshi & Karande, 1980

104

MARGARET BARNES

Species

Kerala backwaters, India

Adriatic Sea B. a. cirratus

Port Jackson, Australia

B. a. denticulala

Place Mar to Apr and Oct Breeds all year but max in dry season Dec to Apr. Breeding reduced in May to Aug if salinity is reduced to 6– 7%o Only one brood a year Auckland, New Zealand

Seasonality in breeding, never more than 50% at a time. Min in June to Aug, increase in Sept to Feb. A few breeding Apr to May Smallest gravid animal 6 mm diam Cape Town, South Africa

Swansea, UK

More than one brood a year

North Carolina, USA B. a. hawaiiensis

–

Bombay, India

Breeds in warmer months but not above 33°C. Peaks in Mar to Apr and in Oct Breeds all year. Max in Feb to July

Manila Bay, Philippines

Japan

Breeding and notes

Size of young (L×B), µm

Number of eggs per brood

St I 210×150

–

Pillai, 1958; Karande & Palekar, 1963; Pillay & Nair, 1970

–

–

Kolosváry, 1947

Settles Nov to Mar. Probably mature and breeding within 2 months St I 220

–

–

–

Egan & Anderson, 1986

Ripe eggs in Sept. No development during winter ova 120 egg 150×90

–

St I 210×140

–

St I (160–240) × (140–200) –

–

Crisp, 1954; Patel & Crisp, 1960b Costlow & Bookhout, 1957 –

–

St I cultured 190×100 –

–

–

Karande, 1967

Rosell, 1976

References

Skerman, 1959

Sandison, 1954

Hudinaga & Kasahara, 1942

EGG PRODUCTION IN CIRRIPEDES

Species

Place

Breeding and notes

Size of young (L×B), µm

Number of eggs per brood

References

B. a. variegatus

Bombay, India

Breeds all year including during monsoons

–

35 000

B. balanoides

Bangor, UK

Fertilises Nov. Releases in spring 1 brood a year

ova 210 egg 305×160 St I 34×189

–

Karande & Palekar, 1963; Gaonkar & Karande, 1980 Crisp, 1954, 1962, 1964b, 1968a

Great Britain

–

–

Northern Europe (plankton) Port Erin, UK

–

St I 370

–

5000 to 10 000 in Moore, 1935; Barnes & mature adults; 400–8000 per l.5 Barnes, 1968 mg o.d. body wt Hoek, 1909

Fertilises in Nov

–

Scotland

Fertilises Nov. One brood a year

St I (340–350) ×220

2500–4000, max 13 000 400–8000 per 1. 5 mg o.d. body wt

Herdla, Norway

Fertilises end Oct

–

–

Hammerfest, Norway Spitsbergen

–

–

Fertilises Sept

–

1000 per 1.5 mg o.d. body wt –

Fertilises end Aug to Oct depending on place Nauplii released end May-June –

–

–

–

–

Zevina, 1963a

–

Arnold, 1977

– –

egg 392×186 egg 338×180

4200 in 2nd yr animal, diam – –

Crisp. 1968a Crisp, 1968a

–

egg 341×178

–

Crisp, 1964a

–

–

400–7160 per 1. 5 mg o.d. body wt

Barnes & Barnes, 1968

Scotland

Fertilises about Feb. One brood a year

St I 370×210

–

Greenland

White Sea, Russia New Brunswick, Canada Newfoundland Woods Hole, USA New England, USA Europe—Sweden to Spain excl Great Britain B. balanus

Moore, 1935 Pyefinch, 1948; Barnes & Barnes, 1965, 1968 Runnstrom, 1924–1925 Barnes & Barnes, 1968 FeylingHanssen, 1953 Høpner Petersen, 1966

Barnes & Costlow, 1961;

105

106

MARGARET BARNES

Barnes & Barnes, 1965 Bangor, UK

–

B. calceolns

B. crenatus

South of Salawati Is. Puget Sound, Wash., USA Scotland

Bangor, UK

–

Sea of Japan

Settles in spring and autumn, a few in summer

Species

Place White Sea, Russia Breeds continuously depending on food and temp Breeds Mar to Nov Woods Hole and Long Is., USA

B. cariosus

Puget Sound, Wash., USA

San Francisco Bay, USA B. eburneus

North Carolina, USA

–

Florida, USA

Seasonal breeding, low in mid winter, peak in Oct Breeds May to Oct. Nauplii present June onwards, settles mainly in summer

Genoa, Vada Ligure Bay, Po estuary, Tarante Med. Sea

ova 225 egg 307×168 Collected in Aug

50 000

Crisp, 1954, 1962

egg 170

–

Hoek, 1913

One brood a year in early spring Main breeding Apr to May, but also later in year ; a few broods per year ova 170 egg 237×120 St I cultured 263×124 St I 285×146

–

–

St I 280×162

–

Branscomb & Vedder, 1982 Pyefinch, 1948, 1949; Barnes & Barnes, 1965

–

Crisp, 1954

–

Ovsyannikova & Korn, 1984a,b

Breeding and notes

Size of young (L×B), µm

Number of eggs per brood

References

Nauplii present

–

–

Zevina, 1963a

–

–

Branscomb & Vedder, 1982

egg 190×120 St I 220×120 Breeds all summer, 60 days to maturity. First release in June St I (190 to 230) (160 to 180)

–

Herz, 1933

–

–

–

Costlow & Bookhout, 1957; Barnes & Barnes, 1965 Weiss, 1948b

–

–

St I 315×151

–

Relini, 1968; Relini & Giordano, 1969; Lepore, Scisciola & Gherardi, 1979; Relini & Fasciana, 1982;

Grave, 1933; Landau, Finney & d’Agostino, 1979

EGG PRODUCTION IN CIRRIPEDES

Species

Place

Breeding and notes

Size of young (L×B), µm

Adriatic Sea Black Sea and Sea of Azov Caspian Sea

One brood a year Breeds only at end of summer Egg lamellae in June-July Breeds July and Sept at 21–29°C. Releases Aug to Sept La Jolla and Catalania Is. Calif., USA

– –

– –

–

–

ova 130×80

–

Yasuda, 1970; Honma & Nakajima, 1973

Breeds continuously. Has a complemental male 2 major broods a year, in May and Sept. May be more. No settlement from Nov to May

–

–

Molenock & Gomez, 1972; Gomez, 1974

egg 220×125 St I 244 ×153

2000 to 12 000 depending on size of adult

Johnson & Miller, 1935; Barnes & Barnes, 1956, 1965; Wu & Levings, 1979; Branscomb & Vedder, 1982

Uchiura Bay and Sado Is., Japan

B. galeatus

Number of eggs per brood Geraci & Romairone, 1986 Kolosváry, 1947 Zevina, 1957, 1963b Zevina, 1957

107

B. glandula

British Columbia and Puget Sound, USA

Morro Bay, Calif., USA

6 broods a year in eggs 245×175 winter and spring, limited by heat Great Britain Fertilises early Jan, releases early Mar. One brood a year

3417 to 30 090 depending on age of adult

Hines, 1974, 1978, 1979

ova 265 egg 385×205 St I 442×232

100 000

Friday Harbor, USA Sydney, Australia

–

St I 265×129

–

Eyed embryos for 11 months of year. Some release in Mar to Apr

–

–

B. hameri

B. hesperius B. imperator

B. improvisus

West coast Sweden

Fertilises May to June, more than 60% with egg lamellae in June to Sept, decreases to 10% in Nov to

–

–

References

Moore, 1935; Crisp, 1954, 1962; Patel & Crisp, 1960a; Barnes & Barnes, 1965 Barnes & Barnes, 1959c, 1965 Wisely & Blick, 1964

Blom, 1965

108

MARGARET BARNES

Cuxhaven and Kiel, Germany

Dec. 2–3 broods a year St I (160 to 200) × (80 to 100)

Mediterranean Sea

Breeds May to Oct or Nov until temp is less than 10°C Female gonads present all year. Egg lamellae May to Oct until temp is less than 10°C Two breeding seasons, June to July and Sept Breeds May to Dec

Great Britain

–

ova 123 egg 161×93 St I 195×111

–

Northern waters

–

egg 120×90 St I 190 –

–

–

–

Relini, 1980; Relini, Matricardi & Diviacco, 1980; Relini & Fasciana, 1982 Crisp, 1954; Jones & Crisp, 1954; Barnes & Barnes, 1965 Hoek, 1909; Krüger, 1940 Zevina 1957, 1963b Zevina, 1957

–

–

Yasuda, 1970

–

–

Weiss, 1948a,b

– Early stage embryos in Sept, settles only in spring –

St I 200×135 –

– –

Karande, 1979 Millard, 1952; Sandison, 1954

St I 266×157

–

Breeds all year, probably 23 to 33 broods Breeds in dry season when

–

5900 to 74 400 from adults of 11 to 28 mm diam –

Barnes & Barnes, 1959d, 1965 Hurley, 1973

Netherlands

Arcachon, France

Black Sea and Sea Breeds all year of Azov Caspian Sea Egg lamellae in June to July Uchiura Bay, Breeds early Mar Japan to late June at 14– 2°C. Peak in June at 18–22°C Florida, USA Breeds in pulses rather than continuously all year B. kondakovi E.coast India B. maxillaris Cape Town, South Africa

B. nubilus B. pacificus

B. pallidus stutsburi

Friday Harbor, Wash., USA Baja California and S.Calif. Lagos, Nigeria

–

Buchholz, 1951; Kühl, 1966

–

–

Breeman, 1934

–

–

Dessenoix, 1962

–

–

–

–

Sandison, 1967

EGG PRODUCTION IN CIRRIPEDES

B. perforatus S.W.Great Britain

Species

Arcachon, France

Marseilles, Med. Sea

Genoa, Vado, Ligure Bay, Taranto

B. reticulatus

B. rostratus

B. terebratus B. tintinnabulum tintinnabulum

B. trigonus

salinity is high. 3 to 4 broods a year – ova 160 egg 221×115 St I 279×130

Northern waters Rarely fertilises before mid June, releases same month, settles late July to Sept

St I 280 –

Hoek, 1909

Place

Breeding and notes

Adriatic

One brood a year, – cyprids in winter – –

Dessenoix, 1962

–

–

Le Reste, 1965

St I 402×160

–

egg 250×150

Relini, 1968; Relini & Giordano, 1969; Lepore, Sciscioli & Gherardi, 1979; Geraci & Romairone, 1982, 1986 Depends on temp Yan & Chen, 1980; Cai Rusing and size adult. & Huang More at higher Zongguo, 1981 temp and bigger animals 10 000–15 000 in Korn, 1985 adults 2 cm diam. 100 000 to 500 000 at 5–6 cm diam – Hoek, 1913

–

–

Wagh, 1965

St I 375×171

–

Relini, 1968; Relini & Giordano, 1969;

Breeds in summer, May to June and Aug are main peaks Three periods of reproduction, summer, autumn, winter at 13–24° C. Probably 10 broods a year Larvae in June to Sept Mainly settles in summer Max fertilisation at 22–25°C and release in Aug

East and South China Sea

Breeding begins mid Apr at above 17 to 18°C, ends Dec. Peak May to Oct Peter the Great One brood a year. Bay, Sea of Japan Fertilises in Sept at 17–18°C. About 2 weeks incubation Kei Is. Egg lamellae in Dec Bombay, India Breeds during SW monsoon in July to Sept and Mar to Apr Genoa, Vado, Settles June to Ligure Bay Oct

Size of young (L×B), µm

– Norris & Crisp, 1953; Barnes & Barnes, 1965

109

St I cultured (256 to 275) × (143 to 187)

ova 120–150

Number of eggs per brood

References

–

Kolosváry, 1947

110

MARGARET BARNES

Species

Place

Breeding and notes

Size of young (L×B), µm

Number of eggs per brood

References Geraci & Romairone, 1986

Sydney, Australia

Leigh, New Zealand Table Bay, South Africa

Uchiura Bay,

Eyed nauplii for 7 months, releases Jan and July Summer breeding, settles Nov to Mar Sporadic breeding all year, mainly late summer (Millard). Once a year (Sandison) Breeds May to July and

–

–

Wisely & Blick, 1964

St I cultured (210 to 230)×100 St I 370×190 St I 240×140

–

Skerman, 1959; Foster, 1967; Harker, 1976 Millard, 1952; Sandison, 1954

–

–

Yasuda, 1970

Sept to Nov at 17–19°C. Peaks late May to late June at 18–24°C egg 170×90 –

– Breeds all year, max in spring and autumn. Probably many broods a year Bombay and – Indian harbours

– –

Krüger, 1940 Werner, 1967

St I 240

–

Breeds all year, peaks in Jan to May Some seasonality in breeding. Never more than 50% with egg lamellae, Min in June to Aug increasing to peak in Sept to Feb Smallest gravid animal 7 mm diam

–

–

Karande, 1978

St I 200

–

Egan & Anderson, 1986

Japan

– Florida, USA

B. variegatus

Andaman Is.

Port Jackson, Australia

–

Karande, 1974c; Karande & Thomas, 1976; Gaonkar & Karande, 1980

EGG PRODUCTION IN CIRRIPEDES

B.v. cirratus

Leigh, New Zealand Chile

Nauplii from June to Apr Breeds Apr to June at 14–22°C. Max in June at 18–22°C Predominately winter breeding –

Chile

–

B.M. rosa

Tanabe Bay, Japan

B.M. volcano

Tanabe Bay, Japan Sydney, Australia

Breeds Mar to May with temp rising Breeds July to Oct at max temp Some breeding all year but main peaks in late autumn and early spring Breeds all year but peaks in spring and autumn. Reaches max size of 8 mm in 35 days. Can produce larvae 30 days after settling Reproduces all year with max in Nov to Dec and July to Aug –

B. venustus

B. vestitus B. Austrobalanus flosculus B. Megabalanus psittacus

Austromegabala nus nigrescens

Sydney, Australia Uchiura Bay, Japan

Notomegabalanu s algicola

South Africa

Solidobalanus hesperius

Sea of Japan

S.h. hesperius

Vladivostok

–

–

–

–

Wisely & Blick, 1964 Yasuda, 1970

St I 550×200

–

Foster, 1967

egg 280×160 St I 330×140 –

–

Stefoni & Contreras, 1979 Stefoni & Contreras, 1979

111

–

egg 150×(70 to 100) St I 290×140 –

–

–

Yamaguchi, 1973

St I 270×150

–

Egan & Anderson, 1987

–

–

Branch & Griffiths, 1988

–

–

Ovsyannikova & Levin, 1982

St I cultured 177×85

–

Korn & Ovsyannikova, 1981

Yamaguchi, 1973

B. crenatus (Crisp & Patel, 1969); B. eburneus (Landau, Finney & d’Agostino, 1979); B. glandula (Hines, 1978); B. perforatus (Patel & Crisp, 1960b); Chthamalus fissus (Hines, 1978); Elminius modestus (Patel & Crisp, 1960b); Pollicipes polymerus (Hilgard, 1960); Verruca stroemia (Barnes & Barnes, 1975). Chthamalus stellatus breeds earlier at a lower than higher tidal level (Crisp, 1950) and is controlled by temperature whereas Balanus balanoides breeds first at higher and later at lower levels (Crisp, 1959) and may be controlled by light (Barnes, 1963) and possibly air temperature (Barnes & Barnes, 1976). It seems, therefore, that breeding in intertidal species begins at the shore level having the longest exposure to the changing controlling factors.

112

MARGARET BARNES

Temperature A temperature shock has been suggested as a possible agent synchronising the fertilisation in Balanus balanoides over a considerable section of the eastern North American coast. This may be the final stage of long-term temperature changes which have ensured that all the gonads are available when conditions for fertilisation are favourable. The timing of such conditions may vary slightly from year to year so a shock device is essential to ensure that full advantage is taken of them each year. In general reproductive processes are sensitive to temperature. Some cirripedes will breed over a wide temperature range whereas others will not breed until a certain critical temperature is reached; they are eurythermic or stenothermic, respectively. This critical temperature may be above or below that of their natural environment for most of the year. Boreo-arctic species have an upper temperature barrier and warm-water or tropical species have a lower temperature barrier. Temperature may, therefore, set limits to reproduction and thus become ecologically important in the latitudinal distribution of cirripedes. The effect of temperature may also enable two closely related species to exist in the same region without being in competition. In Tanabe Bay, Japan, for example, Yamaguchi (1973) gives the breeding season of Balanus (Megabalanus) rosa as March to May when sea-water temperature is rising and that of B. (M.) volcano in July to October when temperatures are falling. In those species with a lower temperature barrier to breeding, and living at the northern limit of their distribution, once this barrier is crossed the whole population is released and may fertilise synchronously. Such is the case in B. perforatus at Arcachon, France, released from a low critical temperature in May when about 90% of the mature population fertilises. As more broods are produced throughout the summer the synchrony is lost, because of individual variation between animals, until in October breeding ceases as the temperature barrier is once again imposed (Barnes & Barnes, unpubl. obs.). Landau, Finney & d’Agostino (1979) found that by keeping B. eburneus at 8°C (or below) they could prevent fertilisation for up to 12 months; when the low temperature barrier was removed fertilisation took place. In northern latitudes where animals have an upper temperature barrier temperature, food, and light show marked seasonal variations that are repeated more or less regularly. It is in these regions that cirripedes tend to have restricted breeding cycles which in extreme cases are reduced to one per year. This breeding is timed so that nauplii are released at the most favourable time when food is available for them (Barnes, 1957, 1962; Barnes, Barnes & Finlayson, 1963). B. balanus has a reproductive cycle very similar to that of B. balanoides (Barnes, Barnes & Finlayson, 1963) but the timing is different even though both species release nauplii about March to April. Crisp (1954) considers that a gradual fall in sea-water temperature in the autumn and early winter initiates the final ripening of the gonads in a sublittoral species such as B. balanus. Copulation then takes place at the appropriate time irrespective of any further external stimulus. This may also be the case in other sublittoral species such as B. hameri. B. balanoides needs the shock device because of its intertidal habitat and in the far north fertilisation must take place before the shores become ice-bound. If B. balanoides is maintained above 10°C and in continuous light fertilisation can be prevented indefinitely; when the temperature is reduced to below 10°C and the light to no more than 12 hours in 24 fertilisation takes place in about four weeks (Barnes, 1963; Tighe-ford, 1967). In more tropical regions where temperature, light, and possibly food are more constant or only fluctuate slightly around mean values, breeding might be expected to be asynchronous in a population. Within the optimum temperature range and in relatively stable conditions there is some evidence of seasonal breeding periods superimposed on a general continuous low level of reproduction, for example in Elminius modestus and E. plicatus in New Zealand (Moore, 1944; Foster, 1967) and Chthamalus anisopoma in Baja California (Malusa, 1986). Hines (1978) found temperature to be the controlling factor in the breeding of Balanus glandula in California.

EGG PRODUCTION IN CIRRIPEDES

113

Latitude The effect of latitude on the onset of breeding is difficult to separate from the effect of temperature and it has been most extensively studied on Balanus balanoides in both Europe and North America. As mentioned above boreoarctic cirripedes, whether intertidal or sublittoral, with their southern limits in northern Europe normally have one main breeding season a year, for example, Balanus balanoides (Moore, 1935; Pyefinch, 1948), B. balanus (Barnes & Barnes, 1954), and B. hameri (Moore, 1935; Crisp, 1962). Crisp (1959) gives dates of fertilisation for population of B. balanoides in Europe ranging from 50° to 80°N. There is a trend from north to south with fertilisation in Spitsbergen (Feyling-Hanssen, 1953) being two to three months earlier than in the southwest of Britain. There is also a difference of some days between the east and west coasts of Britain at similar latitudes—probably related to temperature. In eastern Canada at St Andrews, New Brunswick, fertilisation of B. balanoides begins in October (Bousfield, 1954) whereas in the Hudson Strait region it begins in September (Bousfield, 1955). In this region B. balanus also probably fertilises in September (Bousfield, 1955). In Frobisher Bay and Cumberland Sound there may be only one annual brood in B. crenatus as it is near its northern limit (Bousfield, 1955); further south there are two. In Europe, B. crenatus (Pyefinch, 1948) and Verruca stroemia (Barnes & Stone, 1973; Stone & Barnes, 1974; Barnes & Barnes, 1975) have one major breeding period followed by several lesser ones during the season. Warm-water species with their northern limits of distribution in Europe may have more than one breeding cycle in the warmer months but the number is less than in the same species further south. For example, Balanus improvisus produces two to three broods a season on the west coast of Sweden (Blom & Nyholm, 1961; Blom, 1965) but five to six in Florida (Weiss, 1948a). In their normal habitats warm-water species may breed continuously as long as conditions are favourable, for example B. amphitrite in southern India (Daniel, 1958; Pillai, 1958). Breeding in this species is seasonal in the warmer months in temperate regions (Hudinaga & Kasahara, 1942; Sandison, 1954; Costlow & Bookhout, 1958; Wisely & Blick, 1964; Egan & Anderson, 1986). Chamaesipho columna in the region of Sydney, Australia breeds for about six months of the year with release of nauplii from June to October (Wisely & Blick, 1964) whereas in warmer waters at Leigh in New Zealand there is some breeding throughout the year (Moore, 1944; Luckens, 1970, 1976). Light The uniformity of the date of fertilisation of Balanus balanoides in the same area each year suggests that reduction in day length may be a controlling influence (Crisp, 1959). At the same latitude this is the one factor that remains constant from year to year. This explains why animals in shaded situations as well as those on the upper shore, in which light is reduced except when the opercular valves are open under water, fertilise first. Older animals also have more opaque shells (and hence less light penetration) than young ones; older animals fertilise first. The inhibition of fertilisation in B. balanoides by keeping them in continuous light has already been mentioned above (Barnes, 1963). Light probably becomes less important in species from southern regions. In central California, Hines (1978) found that photoperiod had no effect on the reproduction of Chthamalus fissus and Balanus glandula. Feeding Orientation of adult cirripedes, particularly in those species that feed by holding their cirral net in the backwash of water, can affect their ability to feed efficiently. In adverse feeding situations Otway & Underwood (1987) found that there was no reduction in the number of Tesseropora rosea carrying egg lamellae nor in

114

MARGARET BARNES

the weight of egg lamellae produced. There was, however, a reduction in body weight of the parents presumably caused by some starvation due to non-efficient feeding. This agrees with the results of Page (1983) on Pollicipes polymerus. Cimberg (1981) found a higher percentage of P. polymerus breeding at higher than lower tidal levels although the maximum brooding activity was the same at both levels. Because of the way this animal feeds those higher on the shore may be better stimulated to feed. Hines (1978) found food to be the dominating factor in the breeding of Chthamalus fissus in California. Barnes & Barnes (1967, 1975) have shown that the time at which egg lamellae are produced can be controlled by feeding on Balanus balanoides and Verruca stroemia. In the latter species it is poor nutrient conditions that cause breeding to cease in October at Oban in Scotland. As a result, during the subsequent months all the population reaches the same reproductive state and the first fertilisation in the spring is synchronous with 100% of the adults containing egg lamellae. Subsequent broods are not so synchronous— the second has 70% of the population with egg lamellae at the same time. Later broods are even less synchronous so that by June or July breeding is asynchronous until it ceases in October. At this time the temperature is still above that at which reproduction takes place earlier in the year (Barnes & Stone, 1973). By feeding V. stroemia continously Barnes & Barnes (1975) showed that breeding could be maintained asynchronously throughout the year. In a cirripede producing more than one brood a year the time between the laying down of egg masses is important. A longer time due to low food reserves would mean a reduction in the number of broods per year. Hines (1978) has shown that a delay of up to five days between broods in Balanus glandula can reduce the number of broods from six to five in the six months of the breeding season. As far as energy budgets are concerned Wu & Levings (1978) have shown that in this species the production of eggs takes second place after respiration followed by production of shell and body tissue. Crisp & Lewis (1984) have listed the factors affecting the cold tolerance of B. balanoides as reduced metabolism, reduced temperature, and reduced day length—all factors associated with the onset of fertilisation in this species (Crisp & Ritz, 1967). This resistance to cold protects both the parent and the developing egg lamellae and is vital in northern regions. Increased feeding in the spring is sufficient to end the cold tolerant state (Cook & Lewis, 1971). Tooke & Holland (1985) and Tooke, Holland & Gabbott (1985) have compared cold tolerance in B. balanoides and Elminius modestus. Moulting Patel & Crisp (1961) found that in Balanus crenatus and Elminius modestus the moulting phase had no detectable influence on successful copulation. In Balanus balanoides, however, animals which had recently moulted appeared to copulate more readily than those that had moulted more than seven days previously. The moulting cycle in all species is affected by the presence of egg lamellae as a period of reproductive anecdysis follows copulation (Barnes, 1962). The length of this depends on the species. Moulting while egg lamellae are present in the mantle cavity may be a disadvantage although it does occur in boreo-arctic species which have long incubation times. In warm-water species which breed continuously the danger is reduced if the incubation time falls within an intermoult period. This has been found to be so for B. amphitrite, B. crenatus, B. perforatus, Chthamalus stellatus, Lepas anatifera, and Verruca stroemia (Patel & Crisp, 1961). Hatching in these species removes the inhibition to further breeding since oviposition never takes place while egg lamellae are pesent. It also removes any inhibition to moulting—so that the animals moult and are inseminated again. Thus, while able to copulate at any part of the moulting cycle, warmwater cirripedes probably do so most frequently immediately after a post-hatching ecdysis.

EGG PRODUCTION IN CIRRIPEDES

115

Age and Size The size reached by adult barnacles depends on the species and therefore the size at which maturity is reached can only be compared intraspecifically. In warm-water species producing several broods a year those settling early in the year will reach maturity and will themselves reproduce during that year. The age of Chthamalus species when first reproductively active is variable. C. anisopoma in the Gulf of California may be reproductively mature at an age of about six weeks (Malusa, 1986). C. fissus in California first reproduces about two months after settlement (Hines, 1978). C. stellatus in Europe and C. fragilis on the Atlantic coast of North America are about nine months old before they first reproduce (Connell, 1961; Wethey, 1983). The much larger warm-water Tetraclita does not reproduce until much older. T. stalactifera in the Gulf of California does not reproduce until it is about 20 months old (Malusa, 1986). Villalobos (1980) found that T. panamensis in Costa Rica did not reproduce until it was two years old, nor did T. squamosa rubescens in California (Hines, 1978) and T. s. rufotincta in the Gulf of Elat (Achituv & Barnes, 1978a). The age when first reproductive in Balanus species can vary from one to two months as, for example, in B. amphitrite saltonensis (Linsley & Carpelan, 1961), B. trigonus (Werner, 1967), B. pacifica (Hurley, 1973), B. improvisus (Breeman, 1934; Blom, 1965), and B. eburneus (Grave, 1933) to one to two or even three years as, for example, in B. balanoides (Runnström, 1924–1925; Barnes, Barnes & Finlayson, 1963), B. balanus (Barnes & Barnes, 1954; Crisp, 1954), B. hameri (Moore, 1935), and B. restrains (Korn, 1985). In B. balanoides populations of first year adults fertilise later than those consisting of older animals no matter what the tidal level (Crisp, 1959). Age can also affect the numbers of eggs produced by an adult. In the boreoarctic B. balanoides has only one breeding season and animals can reach maturity at the end of their first season of growth. In the early weeks after settlement food resources are used for somatic rather than gonadal growth and so only limited gonadal material is available for reproduction at the end of the first year when adult basal diameter is 5–6 mm. Such animals of about 7–9 mm basal diameter produce only 200–600 eggs whereas second year animals of only 6 mm basal diameter produce about 4200 eggs in Passamaquoddy Bay, Canada. Thus, the adults may be reproductively mature at the end of the first year but the number of eggs produced is only about 10% of the number produced by older animals of a similar size (Arnold, 1977). Crisp (1954) found that the weight of egg masses and number of eggs compared with the weight of the adult was less in small (younger) animals than in larger (older) ones. There is also some indication of this in Elminius modestus (Crisp & Patel, 1961). Age as well as size is therefore important in egg production. Crowding If breeding is connected with food supply then it seems that crowding of the adults, which could cause competition for a limited food supply, might be expected to reduce breeding in some way. Crisp & Davies (1955) and Crisp & Patel (1961) found that the onset of breeding in E. modestus was delayed in slower growing crowded populations by about four weeks. It was also found that in such conditions the male reproductive organs developed at a smaller size and egg lamellae were found in smaller animals than in less crowded situations. Chthamalus dalli and Balanus balanoides produce more eggs per somatic weight when crowded (Wethey, 1984). Wu (1981) found however, that in B. glandula the incidence of cross fertilisation was not increased by decreasing the distance between mature adults. The energy partition of individual barnacles depended on the degree of crowding and this affected the energy channelled into egg production. Uncrowded B. glandula with adequate energy transfer more into egg production than do crowded individuals (Wu, Levings &

116

MARGARET BARNES

Randall, 1977; Wu, 1980; Wethey, 1984). In their first year B. glandula transfer about 50% of their energy production into egg production; this ensures the propagation of the species and counteracts the high mortality that the adults suffer in their first year after settlement (Wu & Levings, 1978). If the whole population of B. glangula is considered, as distinct from individual animals, then in crowded populations in which the animals may have become elongated the egg production per unit area is higher than in uncrowded areas even though the animals in those areas may be contiguous (Wu, 1980). It should be mentioned that this may not be of great ecological value as such over-crowded populations tend to be self-destructive (Barnes & Powell, 1950). Crisp (1964a) found reduced fecundity in high population densities of B. balanoides. Seaweed cover Crisp (1959) found that in B. balanoides local topography had an effect on the percentage of fertilised animals. Fertilisation takes place first in animals at high levels on the shore and becomes progressively later by about 12 days down the shore. Fertilisation was earlier on the underside of boulders, in crevices, and under clumps of seaweed than in open situations although the differences of about three days were not so great as those caused by tidal level. Jernakoff (1985) found that there was no reduction in the weight of egg lamellae produced by Tesseropora rosea overgrown by seaweed compared with those free of cover in the Cape Banks Marine Scientific Research Area of Australia. As this species can contain egg lamellae at all times of the year the effect of cover on the date of fertilision was masked. Salinity The effect of salinity on the breeding season is particularly important in estuarine habitats where freshwater run-off due to river discharge or monsoons is excessive at certain times of the year. In Indian estuaries and harbours Balanus amphitrite amphitrite and B. a. hawaiiensis can breed throughout the year except during the monsoon months when salinity is low (Paul, 1942; Daniel, 1954; Ganapati, Lakshmana Rao & Nagabushanam, 1958; Pillai, 1958; John, 1964; Karande, 1967; Pillay & Nair, 1972; Fernando & Ramamoorthi, 1975). B. variegatus also breeds throughout the year except at very low salinities during the monsoon when the water is also very turbid. B. pallidus in the Vellar estuary does not breed during the summer and early monsoon when salinity is high; there is a slight decrease in breeding during the height of the monsoon at extremely low salinities. This agrees with Sandison (1966, 1967) who reports B. pallidus in Lagos Harbour, Nigeria, as being sensitive to both low and high salinities. In a West Indian mangrove swamp egg masses were found in most adult B. eburneus all the year. There was no indication that the number of eggs was reduced during periods of low salinity but there was evidence that the survival of the hatched nauplii depended on the salinity experienced by the parents during the pre-liberation period (Bacon, 1971). In the Po estuary, Italy, breeding of B. eburneus is more influenced by increase of fresh water than is that of B. improvisus (Relini, 1980; Relini, Matricardi & Diviacco, 1980; Relini & Fasciana, 1982). In B. amphitrite communis in Kerala backwaters in India embryos can be held in the egg lamellae in the mantle cavity of the adult during the rainy season until the salinity becomes more favourable (Daniel, 1958; Pillai, 1958); this was confirmed experimentally. Kühl (1966) found that at Cuxhaven, Germany, B. improvisus would not breed in salinities below 5%o which agrees with the increase of this species in the Baltic when salinity increases to 6%o in some years. Chthamalus malayensis and C. withersi do not breed during the low salinity periods of the monsoons in July to September in India (Karande, 1967).

EGG PRODUCTION IN CIRRIPEDES

117

Bergen (1968) found that egg lamellae of Balanus glandula were less tolerant than adults to reduced salinity. They are, however, protected from low salinity because the adults close their opercular valves in 50% sea water. Crisp & Costlow (1963) investigated the effect of salinity on the in vitro development of egg lamellae in B. eburneus, B. amphitrite amphitrite, and Chelonobia patula; at salinities between 15 and 25%o and between 40 and 60%o development of egg lamellae was delayed and hatching was impaired. Pollution A direct effect of pollution on egg production was found by Rege, Joshi & Karande (1980) in Balanus amphitrite amphitrite at Bombay, India. The time of the breeding season was not affected but in an animal of 1.3 mg oven dry body weight the number of eggs produced was reduced from 4607 in an unpolluted control area to 3458 and 2710 in a moderately and highly polluted area, respectively. Wu & Levings (1980) also found reduced egg production in B. glandula due to bleached kraft pulp mill effluent in British Columbia. Crisp (1959) found that B. balanoides in stagnant or polluted water fertilised earlier than might be expected because of their lower metabolic rate compared with animals on wave-exposed shores. Parasites Johnson (1958) found a marine fungus, Lagenidium chthamalophilum parasitic on the eggs of Chthamalus fragilis var. denticulata. Early stage eggs may be completely destroyed leaving a mass of egg cases filled with fungus mycelium. If, however, the egg lamellae are more mature and the embryos differentiated then some embryos escape infection and hatch normally. Lagenidium callinectes has been found in egg lamellae of Chelonibia patula by Johnson & Bonner (1960). Echiniscoides sp. feeds on egg lamellae of Chthamalus malayensis (Karande, 1967). Peltogaster paguri, a rhizocephalan cirripede, may be rendered sterile by the presence of a female Liriopsis pygmaea (Reinhard, 1942b). Parasitic castration caused by association with gregarian protozoans and an isopod Hemioniscus balani has been reported by several workers (Henry, 1938; Barnes, 1953; Sandison, 1954; Crisp, 1968b). Species susceptible to such infection include Balanus algicola, B. amphitrite amphitrite, B. balanoides, B. balanus, B. glandula, B. hameri, B. improvisus, Chthamalus dalli, C. dentatus, and Eliminius modestus (Crisp, 1968b). Although Sandison (1954) found as many as seven female Hemioniscus in one Balanus algicola she found no apparent deleterious effect. This is unexpected and is not in agreement with Crisp (1968b). Barnes (1953) found that release of nauplii was delayed in a B. balanus heavily parasitised by a protozoan but the nauplii themselves were unaffected when teased out into sea water. RELEASE OF EMBRYOS The actual hatching of cirripede eggs is not within the scope of this article. What is of interest here is the timing of the release of nauplii. The terms ‘nauplii’ and ‘release of embryos’ are preferred to the commonly used ‘spawning’ because in many invertebrates ‘spawning’ refers, as it should, to the release of gametes. The use of the word ‘spawning’ to describe the release of embryos in cirripedes has led to confusion and some misunderstanding. The time of release can vary within the same species depending on its habitat and geographical distribution. At Leigh in New Zealand Luckens (1970) found that Chamaesipho brunnea released nauplii during periods of spring tides, stormy weather, and rough seas. Although eggs may have been ready to hatch release was withheld during neap tides and in calm weather.

118

MARGARET BARNES

More often release is tied to the presence of plankton in the sea water thus ensuring that the nauplius stages have adequate food supplies. Thus, in Chthamalus challengeri near the northern limits of its distribution at Hakodate, Japan, release is in March to April at the time of a diatom outburst in that region (Iwaki, 1975). Korn (1985) reports a similar finding for Balanus rostratus in Peter the Great Bay, Sea of Japan. Fertilisation of the population is in the latter part of September and October, incubation lasts about two weeks and nauplii are found in the plankton in late October to December during an autumn bloom of Skeletonema costatum. Most information is, as usual, available for Balanus balanoides. In this species there is a long time between fertilisation and release of nauplii and the embryos are ready to hatch long before there is a diatom outburst. This readiness is obvious from the in vitro development of egg lamellae as found by Barnes & Barnes (1963). In nature there must, therefore, be some inhibiting factor preventing release until external conditions are favourable. This inhibition is removed once the adult animals start to feed in the spring. In European waters there is a diatom bloom in the spring and the release of nauplii is initiated to coincide with this (Runnström, 1924–25; Barnes, 1957). Delay of a few weeks in the outburst can cause a similar delay in the release of the barnacle nauplii (Barnes, 1956, 1957, 1962). At Woods Hole, USA fertilisation is slightly earlier than in Europe and the embryos are ready and are released in late autumn because here there is abundant planktonic food available throughout the winter (Barnes, 1959; Barnes & Barnes, 1959a). In regions where the intertidal region of the shore is not permanently frozen during the winter B. balanoides nauplii are released first from animals on the upper shore followed by those on the lower shore a little later. In areas where the intertidal area is permanently frozen for several months each winter the animals at the upper levels will remain frozen longer than those at the lower levels. The development of the egg lamellae on the upper shore may, therefore, be arrested, and more so in colder than in less cold winters, because the parents are frozen. The onset of spring feeding by the adults will also be delayed compared with those on the lower shore. In such cases the animals on the lower shore release their nauplii first. Such conditions are found in the Murman region of Russia (Rzepishevsky 1959, 1962). Release here also coincides with the phytoplankton increase in spring. Shore and sea ice can also affect the reproductive timing in B. balanoides in Greenland and Spitsbergen. In Greenland animals fertilise from August to the end of October depending on the place and release begins in March and may continue until August. In some places it may not even start until July and in extreme conditions some nauplii may not be released at all in that year. Embryonic development has to be completed before the formation of the ice foot and there can be no release until it has broken away (Høpner Petersen, 1966). A whole generation can also be lost at Spitsbergen if severe local conditions prevent release of nauplii so long that there is no time to complete the life cycle and fertilise before the onset of the next winter (Feyling-Hansen, 1953). EGGS The number of eggs produced, their size and shape as well as their chemical composition are all of interest. Chemical composition Several studies have been made of the biochemistry of cirripede eggs: B. balanus and B. balanoides (Barnes, 1965; Dawson & Barnes, 1966; Barnes & Evens, 1967; Barnes & Blackstock, 1975); Chthamalus stellatus (or probably C. montagui) (Achituv & Barnes, 1976); Balanus perforatus, Pollicipes cornucopia, and Tetraclita squamosa rufotincta (Achituv & Barnes, 1978b; Gilboa-Garber, Achituv & Mizrahi, 1983);

EGG PRODUCTION IN CIRRIPEDES

119

Chthamalus dentatus and Octomeris angulosa (Achituv & Wortzlavski, 1983); Pollicipes polymerus (Eastman, 1968). The oxygen uptake of the egg lamellae of Balanus balanoides and Pollicipes polymerus has been discussed by Barnes & Barnes (1959b). Lucas & Crisp (1987) were principally concerned with energy metabolism of Balanus balanoides eggs during embryogenesis. Detailed consideration of these papers is, however, not within the scope of this article. Number of eggs Barnes & Barnes (1968) defined some of the terms used in discussing reproduction in cirripedes. Egg number denotes the total number of eggs in the egg lamellae carried by an adult at any one time. In crustaceans this is a function of the size of the parent. Comparisons should therefore be made between animals of the same size. From graphs relating egg number to size of parent a “standard” animal can be arbitrarily selected for the purposes of comparison. The number of eggs produced by such a standard animal can be used to compare environmental conditions within a restricted habitat or over widely separated regions. The term fecundity is often used loosely but in the case of cirripedes it is best defined as the number of eggs produced per standard animal per unit time when comparing the same species or, when comparing different species, as the number of eggs produced per unit time by a given increment of body tissue. Other conditions being constant this number will be inversely proportional to egg size. Thus, per unit body weight increment the fecundity of Chthamalus stellatus is several times greater than that of Balanus balanoides. When populations are concerned the age structure must, amongst other factors, be considered so that the population fecundity of Chthamalus stellatus compared with that of Balanus balanoides and possibly B. perforatus and Elminius modestus at equal population densities is lower because of its smaller adult size (Barnes & Barnes, 1968). In both Pollicipes polymerus and Chthamalus fissus the reproductive effort increases with lower tidal level. Usually, however, most C. fissus are found above this level because of poor survival at lower levels. The higher reproductive effort of which this species is capable at low levels may be to offset the high mortality (Page, 1984). In central California Hines (1974) found that the weight specific fecundity of C. fissus was higher than in Balanus glandula or in Tetraclita squamosa in that order, but that fecundity over an individual’s life is highest in T. squamosa followed by Balanus glandula and then Chthamalus fissus, yet population fecundity was highest in C. fissus followed by Balanus glandula and then Tetraclita squamosa. Of the three species Balanus glandula was only highest in population reproductive productivity followed by Tetraclita squamosa and then Chthamalus fissus. Using the volume (V) of the egg just prior to hatching or that of the stage I nauplius and the number (N) of eggs produced per 50 µg oven dry body weight, Barnes & Barnes (1968) used N×V values for a range of species from different habitats to compare the relative weights of eggs produced per standard increment of body weight. This emphasised the contrast between boreo-arctic species with N×V values between 1500 and 3500 and those of less than 500 for all other species tested. In order to compare what Barnes & Barnes (1968) called the “metabolic efficiency of egg production” but what Hines (1978) prefers to call “reproductive output” or “reproductive effort” the product N×V×B should be used, where B is the number of broods produced by an animal per year. Barnes & Barnes (1968) thought that metabolic efficiency of a wide variety of species was similar. From the data collected in the intervening years it is now known that reproductive effort can vary within a species in different habitats and also between species (Hines, 1978; Page, 1984). Chthamalus fissus has a high reproductive effort and a rapid response to available food. Pollicipes polymerus, on the other hand, reacts more slowly, perhaps because of its specialised feeding

120

MARGARET BARNES

behaviour (Barnes & Reese, 1959, 1960), and has a lower reproductive effort (Page, 1984). It behaves more like Tetraclita squamosa rubescens (Hines, 1978). These last two species channel more energy into a large adult before reaching maturity and so have to use more energy to maintain that size at the expense of reproduction. Cirripedes incubate their eggs in the mantle cavity and so it is obvious that the volume of this cavity must regulate the number of eggs that can be accommodated. Animals with the same weight of soft tissue but thick shells will have a smaller space for egg lamellae than thinner walled animals of the same basal diameter. Some of the mantle cavity in both cases will be occupied by the prosoma and muscles. Species that remain small can only increase the number of eggs produced by reducing the egg size or by increasing the number of broods produced in a lifetime. The number of eggs produced may be influenced by latitude as well as the position of the adults in the intertidal zone (Crisp, 1959; Barnes & Barnes, 1965, 1968; Cimberg, 1981; Lewis & Chia, 1981). In the East and South China Sea Cai Rusing & Huang Zongguo (1981) found that the number of eggs in Balanus reticulatus depended on temperature as well as on shell diameter. Arnold (1977) said that B. balanoides in Passamaquoddy Bay, Canada, produces more eggs than the same species in Europe giving figures of 4200 for an animal of 6 mm diameter and 19 000 for one of 19 mm diameter. No mention was made of the oven dry weight of these animals. The figures quoted by Barnes & Barnes (1968) were adjusted to 1.5 mg oven dry body weight so that they were all comparable. An animal of about 6 mm basal diameter would have this body weight and so the figure of 5500–8000 for Millport, Scotland is greater than that given by (1977) and 4200 for animals further north at Corpach, Scotland is about the same as Arnold’s. Egg number in this species varies with shore level, from place to place, and from year to year. There is also considerable individual variation and so care is needed when making comparisons. Optimal egg numbers seem to be reached, where conditions are fully marine but where there is some protection, even at the southern limits of its distribution in Spain (Barnes & Barnes, 1968). B. pacificus is found on small objects on sandy bottoms in California and Baja California and matures in two to three months at a basal diameter of 10 mm. Egg lamellae of a single brood contain on the average 15 000 eggs and are ready for release in 8 to 11 days. Hurley (1973) calculates that this barnacle is capable of producing 23 to 33 broods a year; at an average of 28 this amounts to 420 000 eggs per year; no egg size is given. This enormous annual egg production is no doubt related to the rather hazardous existence of the adult and ensures that at least a few cyprids may find a suitable settling site. Egg size and shape The size and shape of cirripede eggs has recently been reviewed by Crisp (in press) and little can be added to his summary. He lists the volume of a series of cirripede eggs including and extending those given by Barnes & Barnes (1968). The egg increases in size during development of the contained embryo and when comparing egg sizes in the considerable literature available care must be taken to use data referring to the same stage of development. This is not always possible as the stage may not be given. Sometimes only the size of the first nauplius stage is available and so that rough comparisons can be made this may be assumed to reflect closely that of the final egg size before hatching (Groom, 1894; Barnes & Barnes, 1968). Apart from the stage of development of the embryo, temperature, genetic differences, individual variation, and variations within an egg lamella can affect the size of the egg within a species. Variations within an egg lamella are really differences in stage of development as eggs at the outer edges of large lamellae, having more access to oxygen, tend to be further advanced, and hence larger, than those in the centre.

EGG PRODUCTION IN CIRRIPEDES

121

Crisp (1959) and Barnes & Barnes (1965) have discussed the influence of temperature on egg size and have recorded differences due to latitude in B. balanoides. Barnes & Barnes found the largest (396 µm long) eggs of this species at Murmansk and the smallest (263–314 µm long) in Great Britain. Within these extremes the relationship of egg size and latitude is not close. Animals from places with severe winters (during which the embryos are developing) have large eggs irrespective of latitude. For example, eggs at Murmansk (69°00’N) were 396 µm long and those at St Johns, Newfoundland (47°34’N) were 386 µm long. Sizes (348 µm long) at St Andrews on the Canadian eastern coast (45°05’N) were similar to those (356 µm long) at Disko, Greenland (69°20’N). These places all have severe winters compared with Great Britain. In areas with less severe winters, after taking any local variations into account, there is little connection between latitude and egg size. For example, at Pornic, France (47°07’N) the egg length was 282 µm which comes within the range found in Great Britain some 10 degrees of latitude further north. It seems that eggs of B. balanoides are largest where winters are severe and summers relatively cold irrespective of latitude. Latitude has not been shown to have a striking effect on the size of eggs of Chthamalus stellatus (Barnes & Barnes, 1965) in Europe; sizes range from 158–186 µm long at latitudes from 36 to 48°N. At the extreme northern limit of this species in Fair Isle (59°32’N) Powell (1954) quotes a length of 230 µm. The size of C. dalli eggs is given by Barnes & Barnes (1965) as 183 µm long on the Pacific coast of North America (about 40°N) and by Korn & Ovsyannikova (1979) as 218 µm long at Vladivostok (43 °N) at about the same latitude. Given adequate food in controlled experiments Patel & Crisp (1960b) found that temperature affected egg size in several warm-water species. Under these conditions the animals were breeding continuously and an increase in temperature caused a decrease in egg size. Lucas & Crisp (1987) explained this as an increase in the metabolic rate of eggs during development resulting in extra consumption of food reserves. Crisp (in press) regards the egg’ size of Balanus balanoides as being genetically controlled in the different races of this species but taking the animals at Boothbay Harbor, Maine, “as typical of the American E.coast race” seems to neglect the obvious differences found in egg size and rate of embryonic development found by Barnes & Barnes (1965, 1976) for this species further south on the American coast. Barnes & Barnes (1965) found that within limits larger B. balanoides (up to about 10 mm basal diameter) produced larger eggs. Crisp (in press) has shown that it is the relative amount of tissue available for egg production that determines the egg size. At Boothbay Harbor small animals in probably poor nutrient conditions had smaller eggs than larger animals but at Menai Bridge, North Wales, in good nutrient conditions egg size was the same throughout the size range of adults. Eggs of boreo-arctic species are large compared with practically all others; there are a few exceptions that will be mentioned later. This large size is mainly an adaptation to boreo-arctic conditions and the survival value of producing a large nauplius capable of dealing with the available planktonic food (Barnes & Barnes, 1965). Crisp (in press) suggests that it may also be related to greater dispersal of larvae having a long planktonic life and the time needed to find suitable substrata for settlement. Moyse & Knight-Jones (1967) say that the smaller eggs in warm-water species are related to shorter planktonic lives and probably the quick turnover of food supplies by adults producing numerous broods per year. Although many cirripede nauplii are pelagic and planktotrophic some are pelagic and lecithotrophic while some are merely lecithotrophic and develop to the cyprid stage within the mantle cavity of the parent. Animals producing lecithotrophic nauplii usually produce broods of fewer than usual eggs of a larger size than normal. The rhizocephalans are an exception. The eggs are small and numerous. Cyprids of these parasitic cirripedes do not, however, need excessive reserves of yolk because as soon as a suitable host is found for settlement then ample nutrient is available for further development. Many of the acrothoracicans also have lecithotrophic larvae but the eggs are much bigger and fewer in number than in the rhizocephalans.

122

MARGARET BARNES

In the Thoracica several species produce lecithotrophic larvae. In the Lepadomorpha Ibla quadrivalvis (Anderson, 1965), I. cumingi (Karande, 1974a; Klepal, 1985) and I. idiotica (Batham, 1945a) are all known to have large eggs and only relatively few per brood. Anelasma squalicola (Darwin, 1851; Hoek, 1909, Krüger, 1940) has a very large egg of about 500 µm length, much larger than in other Heteralepadoidea (about 200 µm in length). All the deep-water Scalpellum and Acroscalpellum species on which there is adequate information form lecithotrophic larvae. Pollicipes (Mitella) spinosus (Batham, 1946) has a very large egg (690 µm long) immediately before release of larvae compared with about 250 µm length in planktotrophic P. polymerus and P. cornucopia. P. spinosus also produces far fewer eggs per brood. In the Balanomorpha there are striking examples of lecithotrophy in the Coronuloidea in Tetraclita species. Three species produce lecithotrophic larvae while the rest, as far as is known are planktotrophic. The three exceptional species are T.squamosa rufotincta (Achituv & Barnes, 1978a; Barnes & Achituv, 1981), T. (Tesseropora)pacifica (Crisp, 1986), and T. divisa (Nilsson-Cantell, 1921; Anderson, 1986). The volume of the egg of T. squamosa rufotincta when ready to hatch is ml compared with ml for T. squamosa, ml for T. rubescens, and ml for T. serrata. It also greatly exceeds that of boreo-arctic species such as B. balanoides and B. balanus with volumes of and ml, respectively (Barnes & Achituv, 1981). The shape of a cirripede egg has been dealt with fully by Crisp (in press). Suffice it to say here that the general shape is ovoid with the variable short diameter being longer at one end than the other giving a tapering shape. Eggs producing lecithotrophic larvae are much more globular than those giving planktotrophic larvae presumably due to the yolk reserves carried in the former right through to the cyprid stage and maybe beyond. The increased amount of yolk retards the rate of embryonic development (Anderson, 1965). At 23°C Ibla quadrivalvis eggs take about 17 days to hatching and at 14– 15°C those of Pollicipes spinosus take about 32 days. In contrast Balanus perforatus and B. eburneus eggs produce planktotrophic larvae and take about 8 and 11 days, respectively at 15°C. CONCLUSIONS In order to exist an animal must reproduce itself successfully. Models of reproductive methods in marine benthic invertebrates proposed by Vance (1973a,b) have caused some controversy (Underwood, 1974; Vance, 1974). Steele (1977) and Strathmann (1977) have attempted to resolve the arguments. In nature, however, things are always more complicated and interrelated; the situation in cirripedes is no exception. Steele & Steele (1975) report that in crustaceans egg size and duration of embryonic development are correlated. This is often the case in cirripedes. At high latitudes with low environmental temperatures optimal conditions for larval life and survival of young adult barnacles is short. The time of releasing larvae is critical and must be timed to coincide with the spring outburst of phytoplankton. Only a single brood can be produced each year because a second brood cannot be reared successfully as young would be released when conditions were unfavourable. In the far north ice-bound shores limit the time available for fertilisation and embryonic development. Large eggs ensure that sufficient nutrient is available during the long period waiting for optimal conditions for release when the ice thaws. Smaller eggs are produced at low latitudes where temperatures are higher and favourable conditions last longer. Embryonic development is quicker and there is no delay in release of nauplii except in a few cases, for example where there is a sudden drastic reduction in salinity. Many broods can be produced in quick succession providing adequate food is available and other environmental conditions remain favourable. The production of many broods increases the chance of reproductive success especially in cases where a whole generation may fail or where there is excessive predation of the planktonic stages. Such prédation can

EGG PRODUCTION IN CIRRIPEDES

123

be avoided by incubation to the cyprid stage within the mantle cavity or the reduced necessity of a long planktonic life because the adults provide the eggs with sufficient nutrient to carry the young through until metamorphosis, that is, the larvae are lecithotrophic. This mode of life seems to have been developed independently in several groups of cirripedes and almost always with some advantage to the species concerned or as an adaptation to the local environment. It usually means the production of much larger eggs to accommodate the large amount of stored yolk. As has been said above this is not necessary in rhizocephalans. Lecithotrophy is no doubt an advantage in deep-water species such as Acroscalpellum and Scalpellum. In these species an abbreviated nauplius development within the mantle cavity is also an advantage. The cyprid when released remains in the vicinity of the parent group; restricted dispersal in a habitat with few suitable substrata for settlement is advantageous. At first it is not why the three species of Tetraclita should behave differently from conspecifics which produce the usual planktotrophic larvae. Achituv & Barnes (1978a) thought that the large nauplii of T. squamosa rufotincta might be an adaptation to feeding on large particles such as may be the case in boreoarctic species. Barnes & Achituv (1981) found, however, that the nauplii did not feed and compared them with other non-feeding nauplii. This species and T. (Tesseropora) pacifica (Crisp, 1986) live in the vicinity of coral reefs and where the water is very and nutrients are limited. The lecithotrophic nauplii seem, therefore, to be an adaptation to environmental conditions. That this is extended further to an abbreviate nauplius development in T. divisa may be because planktonic dispersion would be a disadvantage in this insular and cave-living species (Crisp, 1986). According to Anderson (1986) the cyprid of this last species contains no yolk reserves and immediate settlement in the vicinity of the parent stock is favoured. Although this species is insular it is also circumtropical, thus dispersal must be by adults attached to floating objects. Rapid settlement of newly released cyprids will then ensure quick colonisation of any new sites. In Anelasma squalicola the abbreviate nauplius development ensures that cyprids when released are in the immediate vicinity of a suitable substratum, in this case members of the same shoal of fish. Lecithotrophy in Ibla species presents a problem unless it is again connected with lack of nutrients in the water such as with I. cumingi at Elat. Moyse (1987) has suggested that it may be habit carried over from an earlier deep-sea existence. From the foregoing it is evident that in cirripedes, egg production, egg sizes and numbers as well as number of broods produced and the breeding cycles and seasons cannot be divorced from the life style and general ecology of the animal concerned. ACKNOWLEDGEMENTS It is a pleasure to acknowledge the help of many colleagues, particularly Univ. Doz. Dr W.Klepal, in the preparation of this review. Miss E.Walton and Miss R.Gow helped in tracing some of the references and ProfessorD. J.Crisp supplied me with a proof of his paper on the shape and size of cirripede eggs. I appreciate the care taken by Mrs M.Fletcher in the typing of the manuscript. REFERENCES Achituv, Y., 1979. Israel J. Zool., 28, 54 only. Achituv, Y., 1986. Crustaceana, 51, 259–269. Achituv, Y. & Barnes, H., 1976. J. Exp. Mar. Biol. Ecol, 22, 263–267. Achituv, Y. & Barnes, H., 1978a. J. Exp. Mar. Biol. Ecol., 31, 315–324. Achituv, Y. & Barnes, H., 1978b. J. Exp. Mar. Biol. Ecol, 32, 171–176.

124

MARGARET BARNES

Achituv, Y. & Klepal, W., 1981. P.S. Z.N. I. Mar. Ecol., 2, 295–305. Achituv, Y. & Wortzlavski, A., 1983. J. Exp. Mar. Biol. Ecol., 69, 137–144. Anderson, D.T., 1965. Aust. J. Zool., 13, 1–15. Andersen, D.T., 1986. Proc. Linn. Soc. N.S.W., 109, 107–116. Arenas, J.N., 1982. Stud. Neotrop. Fauna Environ., 17, 61–67. Arnold, D.C., 1977. J.Fish. Res. Board Can., 34, 273–275. Aurivillius, C.W.S., 1894. Kongl. Svensk. Vetensk-Akad. Handl, 26(7), 107pp. Bacon, P.R., 1971. J. Exp. Mar. Biol. Ecol., 6, 187–198. Bainbridge, V. & Roskell, J., 1966. In, Some Contemporary Studies in Marine Science, edited by H.Barnes, Alien & Unwin, London, pp. 67–81. Barker, M.F., 1976. N. Z. J. Mar. Freshwater Res., 10, 139–158. Barnard, K.H., 1924. Ann. S. Afr. Mus., 20, 1–103. Barnard, K.H., 1925. Union of S.Africa, Fisheries and Mar. Biological Survey. Rept. No. 4, part 6, pp. 1–5. Barnes, H., 1953. Nature (London), 172, 128 only. Barnes, H., 1956. J. Anim. Ecol., 25, 72–84. Barnes, H., 1957. Année Biol., 33, 67–85. Barnes, H., 1958. J. Mar. Biol. Assoc. U.K., 37, 427–33. Barnes, H., 1959. In, Marine Boring and Fouling Organisms, edited by D.L.Ray, Friday Harbor Symposia, University Washington Press, Seattle, pp. 234–245. Barnes, H., 1962a. Limnol. Oceanogr., 7, 462–473. Barnes, H., 1962b. Crustaceana, 4, 118–122. Barnes, H., 1963. J. Mar. Biol. Assoc. U.K., 43, 717–727. Barnes, H., 1965. J. Mar. Biol. Assoc. U.K., 45, 321–339. Barnes, H., 1972. J. Exp. Mar. Biol. Ecol., 8, 89–100. Barnes, H., 1975. Pubbl. Staz. Zool. Napoli, 39 (Suppl.), 8–25. Barnes, H. & Barnes, M., 1954. Oikos, 5, 63–76. Barnes, H. & Barnes, M., 1956. Pacif. Sci., 10, 415–22. Barnes, H. & Barnes, M., 1958. Ecology, 39, 550 only. Barnes, H. & Barnes, M., 1959a. Oikos, 10, 1–18. Barnes, H. & Barnes, M., 1959b. Kieler Meeresforsch., 15, 242–251. Barnes, H. & Barnes, M., 1959c. Can. J. Zool., 37, 237–244. Barnes, H. & Barnes, M., 1959d. Can. J. Zool., 37, 15–23. Barnes, H. & Barnes, M., 1963. Vidensk. Medd. Dan. Naturhist. Foren. Khobenhavn, 125, 93–100. Barnes, H. & Barnes, M., 1965. J. Anim. Ecol., 34, 391–402. Barnes, H. & Barnes, M., 1966. In, Some Contemporary Studies in Marine Science, edited by H.Barnes, Alien & Unwin, London, pp. 83–105. Barnes, H. & Barnes, M., 1967. J. Exp. Mar. Biol. Ecol., 1, 1–6. Barnes, H. & Barnes, M., 1968. J. Exp. Mar. Biol. Ecol., 2, 135–153. Barnes, H. & Barnes, M., 1975. J. Exp. Mar. Biol. Ecol., 19, 227–232. Barnes, H. & Barnes, M., 1976. J. Exp. Mar. Biol. Ecol., 24, 251–269. Barnes, H., Barnes, M. & Finlayson, D.M., 1963. J. Mar. Biol. Assoc. U.K., 43, 185–221. Barnes, H., Barnes, M. & Klepal, W., 1977. J. Exp. Mar. Biol. Ecol., 27, 195–218. Barnes, H. & Blackstock, J., 1975. J. Exp. Mar. Biol. Ecol., 19, 59–79. Barnes, H. & Costlow Jr, J.D., 1961. J. Mar. Biol. Assoc. U.K., 41, 59–68. Barnes, H. & Crisp, D.J., 1956. J. Mar. Biol. Assoc. U.K., 35, 631–639. Barnes, H. & Evens, R., 1967. J. Mar. Biol. Assoc. U.K., 47, 171–180. Barnes, H., Klepal, W. & Munn, E.A., 1971. J. Exp. Mar. Biol. Ecol., 1, 173–196. Barnes, H. & Powell, H.T., 1950. J. Anim. Ecol., 19, 175–179. Barnes, H. & Reese, E.S., 1959. Proc. Zool. Soc. London, 132, 569–585.

EGG PRODUCTION IN CIRRIPEDES

125

Barnes, H. & Reese, E.S., 1960. J. Anim. Ecol., 29, 169–185. Barnes, H. & Stone, R.L., 1972. J. Exp. Mar. Biol. Ecol., 9, 303–309. Barnes, H. & Stone, R.L., 1973. J. Exp. Mar. Biol. Ecol., 12, 279–297. Barnes, M. & Achituv, Y., 1981. J. Exp. Mar. Biol. Ecol., 54, 149–165. Bassindale, R., 1936. Proc. Zool. Soc. London, 1936, 57–75. Bate, C.S., 1869. Ann. Mag. Nat. Hist., Ser. 4, 3, 465–166. Batham, E.J., 1945a. Trans. R. Soc. N.Z., 75, 347–356. Batham, E.J., 1945b. Trans. R. Soc. N.Z., 74, 359–374. Batham, E.J., 1946. Trans. R. Soc. N.Z., 75, 405–418. Batham, E.J. & Tomlinson, J.T., 1965. Trans. R. Soc. N.Z. Zool., 7, No. 9, 141– 154. Bergen, M., 1968. Crustaceana, 15, 229–234. Berndt, W., 1903. Z. Wiss Zool., 74, 396–457. Bigelow, M.A., 1896. Anat. Anz., 12, 263–269. Bigelow, M.A., 1902. Bull. Mus. Comp. Zool., Harv. Univ., 40(2), 61–144. Bishop, M.W.H., 1945. Nature (London), 159, 501 only. Blom, S.-E., 1965. Zool. Bidr. Uppsala, 37, 59–76. Blom, S.-E. & Nyholm, K.-G., 1961. Zool. Bidr. Uppsala, 33, 149–155. Bocquet-Védrine, J., 1961. Cah. Biol. Mar., 2, 455–593. Bocquet-Védrine, J., 1964. Zool. Meded. Deel XXXIX, Feestbundel H. Boschma, pp. 1–11. Bocquet-Védrine, J., 1968. Arch. Zool. Exp. Gén., 109, 257–267. Bocquet-Védrine, J., 1972a. C.R.Acad. Sci., Paris, Ser. D, 275, 2145–2148. Bocquet-Védrine J., 1972b. Cah. Biol. Mar., 13, 615–626. Bocquet-Védrine, J., 1979. Crustaceana, 37, 153–164. Bocquet-Védrine, J. & Pochon-Masson, J., 1969. Arch. Zool. Exp. Gén., 110, 595– 616. Bokenham, N.A.H., 1938. Ann. Natal Mus., 9, 47–81. Boschma, H., 1931. Siboga-Expeditie, Monographie, No. XXXI bis, 67 pp. Bousfield, E.L., 1954. Ann. Rep. Nat. Mus. Canada for 1952–53, Bull. No. 132, 112– 154. Bousfield, E.L., 1955. J. Fish. Res. Board Can., 12, 762–767. Branch, G.M. & Griffiths, C.L., 1988. Oceanogr. Mar. Biol. Annu. Rev., 26, 395– 486. Branscomb, E.S. & Vedder, K., 1982. Crustaceana, 42, 83–95. Breeman, L. van, 1934. Zool. Anz., Suppl., 105, 247–257. Broch, H., 1922. Natur. Foren. Medd., No. 73/74, 215–358. Broch, H., 1924. Vid.-Selsk. Skr. I, Mat.-Naturv. Kl 1924, No. 17, 121 pp. Broch, H., 1927. In, Handbuch der Zoologie von W. Kukenthal, Vol. 3, Walter de Gruyter & Co., Berlin, pp. 503–552. Buchholz, H. von, 1951. Kiel. Meeresforsch., 8, 49–57. Cai Rusing & Huang Zongguo, 1981. Acta Zool. Sinica, 27, 274–280 (in Chinese, English summary). Callan, H.G., 1941. Nature (London), 148, 258 only. Cheung, P.J. & Nigrelli, R.F., 1969. Am. Zool., 9, 502 only. Cimberg, R.L., 1981. Biol. Bull. (Woods Hole, Mass.), 160, 31–42. Codreanu, R., 1959. C.R.Acad. ScL, Paris, 248, 3739–3741. Coker, R.E., 1901. Bull. U.S.Fish. Comm., 21, 399–412. Colón-Urban, R., Cheung, P.J., Ruggieri, G.D. & Nigrelli, R.F., 1979. Int. J.Invert. Reprod., 1, 245–252. Connell, J.H., 1961. Ecology, 42, 710–723. Cook, P.A. & Lewis, A.H., 1971. Mar. Biol., 9, 26–30. Costlow Jr, J.D. & Bookhout, C.G., 1957. Biol. Bull. (Woods Hole, Mass.), 112, 313–324. Costlow Jr, J.D. & Bookhout, C.G., 1958. Biol. Bull. (Woods Hole, Mass.), 114, 284–295. Crisp, D.J., 1950. Nature (London), 166, 311–312. Crisp, D.J., 1954. J. Mar. Biol. Assoc. U.K., 33, 473–96. Crisp, D.J., 1959. Oikos, 1, 275–289.

126

MARGARET BARNES

Crisp, D.J., 1962. Crustaceana, 4, 123–130. Crisp, D.J., 1964a. In, Grazing in Terrestrial and Marine Environments, edited by D.J.Crisp, Brit. Ecol. Symp. No. 4, pp. 251–264. Crisp, D.J., 1964b. J. Mar. Biol. Assoc. U.K., 44, 33–15. Crisp, D.J., 1968a. J. Fish. Res. Board Can., 25, 2633–2641. Crisp, D.J., 1968b. J. Fish. Res. Board Can. 25, 1161–1167. Crisp, D.J., 1983. Mar. Biol. Lett., 4, 281–294. Crisp, D.J., 1986. In, Indian Ocean. Biology of Benthic Marine Organisms, edited by M.-F. Thompson et al., A.A.Balkema, Rotterdam, pp. 69–84. Crisp, D.J., in press. In, Contributions in Marine Science, Dr S.Z.Qasim Sastyabdapurti Felicitation Volume, pp 1–26. Crisp, D.J. & Clegg, D.J., 1960. Oikos, 11, 265–275. Crisp, D.J. & Costlow Jr, J.D., 1963. Oikos, 14, 22–34. Crisp, D.J. & Davies, P.A., 1955. J. Mar. Biol. Assoc. U.K., 34, 357–380. Crisp, D.J. & Lewis, A.H., 1984. J. Mar. Biol. Assoc. U.K., 64, 125–145. Crisp, D.J. & Patel, B., 1961. Limnol. Oceanogr., 6, 105–115. Crisp, D.J. & Patel, B., 1969. Mar. Biol., 2, 283–295. Crisp, D.J. & Ritz, D.A., 1967. Helgol. Wiss. Meeresunters., 15, 98–115. Dalley, R., 1984. Crustaceana, 46, 39–54. Daniel, A., 1954. J. Madras Univ., Ser. B, 24, 189–212. Daniel, A., 1958. J. Madras Univ., Ser. B, 28, 23–17. Darwin, C., 1851. A Monograph on the Sub-Class Cirripedia, with Figures of all the Species. The Lepadidae; or, Pedunculate cirripedes. The Ray Society, London, 400 pp. Darwin, C., 1854. A Monograph on the Sub-Class Cirripedia with Figures of all the Species. The Balanidae (or sessile cirripedes), The Verrucidae etc. etc. etc. The Ray Society, London, 685 pp. Darwin, C., 1873. Nature (London), 8, 431–432. Dawson, R.M.C. & Barnes, H., 1966. J. Mar. Biol. Assoc. U.K., 46, 249–261. Day, J.H., 1935. Q.J.Microsc. Sci., 77, 549–583. Dayton, P.K., Newman, W.A. & Oliver, J., 1982. J. Biogeography, 9, 95–109. Delage, Y., 1884. Arch. Zool. Exp., 2nd Ser., 2, 417–738. Denley, E.J. & Underwood, A.J., 1979. J. Exp. Mar. Biol. Ecol., 36, 269–293. Dessenoix, C., 1962. Thesis, Inst. Biol. Mar d’Arcachon, 97 pp. Dineen, J.F., 1987. Biol. Bull. (Woods Hole, Mass.), 172, 284–298. Eastman, R.C., 1968. Exp. Cell Res., 51, 323–329. Egan, E.A. & Anderson D.T., 1985. Aust. J. Mar. Freshwater Res., 36, 383–04. Egan, E.A. & Anderson, D.T., 1986. Crustaceana, 51, 188–207. Egan, E.A. & Anderson, D.T., 1987. Aust. J.Mar. Freshwater Res., 38, 511–522. Egan, E.A. & Anderson, D.T., 1988. J. Nat. Hist., 22, 1379–1405. Elston, R.A., Wilkinson, M.T. & Burge, R., 1985. Aquaculture, 49, 359–361. Evans, F., 1958. Nature (London), 182, 1245–1246. Fernando, A.S. & Ramamoorthi, K., 1975. Bull. Dept. Mar. Sci., Univ. Cochin, 7, 721–732. Feyling-Hanssen, R.W., 1953. Norsk Polarinstitutt, Skr., Nr. 98, 644 pp. Foster, B.A., 1967. Tane, 13, 33–42. Foster, B.A., 1978. N.Z.Oceanogr. Inst. Mem., No. 69, 160 pp. Foster, B.A., 1982. Proc. Linn. Soc. N.S.W., 106, 21–32. Foster, B.A., 1983. Aust. Mus. Mem., 18, 133–139. Foxon, G.E.H., 1940. J. Mar. Biol. Assoc. U.K., 24, 253–264. Fyhn, U.E.H. & Costlow, J.D., 1975. Biol. Bull. (Woods Hole, Mass.), 149, 316– 330. Ganapati, P.N., Lakshmano Rao, M.V. & Nagabhushanam, R., 1958. Andra Univ. Mem. Oceanogr. , Ser. 62, 2, 193–209.

EGG PRODUCTION IN CIRRIPEDES

127

Gaonkar, S.N. & Karande, A.A., 1980. J. Bombay Nat. Hist. Soc., 76, 305–310. Genthe, K.W., 1905. Zool. Jahrb., Abt J.Morph., 21, 181–200. George, A.I., 1959. J. Zool. Soc. India, 11, 171–204. George, M.J., 1949. Proc. Ind. Acad. Sci., Sect. B, 30, 207–214. Geraci, S. & Romairone, V., 1982. P. S. Z. N. I. Mar. Ecol., 3, 225–232. Geraci, S. & Romairone, V., 1986. P. S. Z. N. I. Mar. Ecol., 1, 151–164. Ghiselin, M.T., 1969. Q. Rev. Biol., 44, 189–208. Ghiselin, M.T., 1974. The Economy of Nature and the Evolution of Sex. University of California Press, Berkeley, California, 346 pp. Gilboa-Garber, N., Achituv, Y. & Mizrahi, L., 1983. Mar. Biol. Lett., 4, 37–46. Gomez, E.D., 1974. Dissertation Abstr. Int. (B), 34, 3902 only. Gomez, E.D., 1975. Crustaceana, 28, 105–107. Gomez, E.D., Faulkner, D.J., Newman, W.A. & Ireland, C., 1973. Science, 179, 813–814. Grave, B.H., 1933. Biol. Bull. (Woods Hole, Mass.), 65, 375–386. Griffiths, R.J.I., 1979. Trans. R. Soc. S. Afr., 44, 97–111. Groom, Th. T., 1894. Phil. Trans. R.Soc. Ser. B, 185, 119–232. Grygier, M.J., 1982. Int. J.Invert. Reprod., 4, 323–332. Grygier, M.J., 1983a. Crustacean Phytogeny, edited by F.R.Schram, A.A.Balkema, Rotterdam, pp. 73–104. Grygier, M.J., 1983b. J.Nat. Hist., 17, 213–239. Grygier, M.J., 1983c. J. Plankton Res., 5, 197–202. Grygier, M.J., 1987. J. Crust. Biol., 7, 87–104. Grygier, M.J. & Newman, W.A., 1985. Trans. San Diego Soc. Nat. Hist., 21, 1–22. Hallberg, E., Elofsson, R. & Grygier, M.J., 1985. Sarsia, 70, 167–171. Hancock, A., 1849. Ann. Mag. Nat. Hist., Ser. 2, 4, 305–314. Harkantra, S.N., Nair, A., Ansari, Z.A. & , A.H., 1977. Mahasagar, 10, 165–168. Harker, B.M., 1975. M.Sc. thesis, James Cook University of North Queensland, Australia. Harms, J., 1984. Helgol. Meeresunters., 38, 123–134. Harms, J., 1986. Helgol. Meeresunters., 40, 355–376. Heath, J.R., 1971. J. Exp. Mar. Biol. Ecol., 6, 15–22. Heath, J.R. & Barnes, H., 1970. J. Exp. Mar. Biol. Ecol., 5, 199–233. Henry, D.P., 1938. Arch. f.Protistenkund., 90, 414–431. Henry, D.P. & McLaughlin, P.A., 1965. Nature (London), 207, 1107–1108. Henry, D.P. & McLaughlin, P.A., 1967. Crustaceana, 12, 43–58. Herz, L.E., 1933. Biol. Bull. (Woods Hole, Mass), 64, 432–442. Hilgard, G.H., 1960. Biol Bull. (Woods Hole, Mass.), 119, 169–188. Hines, A. H., 1974. Am. ZooL, 14, 1266 only. Hines, A. H., 1978. Biol. Bull. (Woods Hole, Mass.), 154, 262–281. Hines, A. H., 1979. In, Reproductive Ecology of Marine Invertebrates, edited by S.E. Stancyk, Belle W.Baruch Library in Marine Science, No. 9, University of South Carolina Press, Columbia, pp. 213–234. Hines, J. A., 1966. Cirripedologist’s Newsletter, 2(3), 2–3. Hiro, F., 1939. Mem. Coll. Sci. Kyoto Imp. Univ., Ser. B, 15, 245–284. Høeg, J. & Lützen, J., 1985. Marine Invertebrates of Scandinavia, No. 6, 92 pp. Høeg, J. T., 1982. J. Exp. Mar. Biol. Ecol, 58, 87–125. Høeg, J. T., 1984. J. Exp. Mar. Biol. Ecol, 16, 145–156. Høeg, J. T., 1985a. Acta Zool. (Stockh.), 66, 1–45. Høeg, J. T., 1985b. J. Exp. Mar. Biol Ecol, 89, 221–235. Høeg, J. T., 1987. Phil. Trans. R. Soc. Land. Ser. B, 317, 47–63. Høeg, J. T. & Ritchie, L. E. (deceased), 1985. J. Exp. Mar. Biol. Ecol., 87, 1–11. Hoek, P.P.C., 1883. Rep. Sci. Res. Voyage H.M.S. Challenger 1873–76, Zool., 8, part III, 169 pp.

128

MARGARET BARNES

Hoek, P.P.C., 1884. Rep. Sci. Res. Voyage H.M.S. Challenger 1873–76, Zool, 10, part III, 47 pp. Hoek, P.P.C., 1888. Rep. Sci. Res. Voyage H.M.S. Challenger 1873–76, Zool., 24, 919–926. Hoek, P.P.C., 1906. Akad. Wet. Amsterdam, Proc., 8, 659–662. Hoek, P.P.C., 1907. Siboga-Expeditie, 31, Part A, 1–128. Hoek, P.P.C., 1909. Nordisches Plankton, Zoologischer Teil, 4, VIII, 265–332. Hoek, P.P.C., 1913. Siboga-Expeditie, 31, Part B, 129–170. Honma, Y. & Nakajima, J., 1973. Bull. Jap. Soc. Scient. Fish., 39, 357–365. Høpner Petersen, G., 1966. Medd. Grønl, 159(12), 114 pp. Hudinaga, M. & Kasahara, H., 1942. Zool. Mag., 54, 108–118. Hui, E. & Moyse, J., 1984. J. Mar. Biol Assoc. U.K., 64, 91–97. Hurley, A. C., 1973. Limnol Oceanogr., 18, 386–393. Ichikawa, A. & Yanagimachi, R., 1958. Annot. Zool. Jpn., 31, 82–96. Ichikawa, A. & Yanagimachi, R., 1960. Annot. Zool. Jpn., 33, 42–56. Ishida, S. & Yasugi, R., 1937. Bot. Zool. Tokyo, 5(9), 1659–1666 (in Japanese). Iwaki, T., 1975. Bull. Fac. Fish. Hokkaido Univ., 26, 1–10. Jeffries, W. B. & Voris, H. K., 1979. Crustaceana, 37, 123–132. Jeffries, W. B. & Voris, H. K., 1983. Fieldiana (Zool), N.S., No. 16, 10 pp. Jernakoff, P., 1985. J. Exp. Mar. Biol. Ecol, 94, 89–97. John, P.A., 1964. Fish. Tech., 1, 189–201. Johnson, M.W. & Miller, R.C., 1935. Univ. Wash. Publ. Oceanogr., 2(1), 1–18. Johnson, Jr, T.W., 1958. Biol. Bull. (Woods Hole, Mass.), 114, 205–214. Johnson, Jr, T.W. & Bonner, R.R., 1960. J. Elisha Mitchell Sci. Soc., 76, 147–149. Jones, L.W.G. & Crisp, D.J., 1954. Proc. Zool. Soc. London, 123, 765–780. Kalyanasundaram, N. & Ganti, S.S., 1975. Bull. Dept. Mar. Science, Univ. of Cochin, 7, 761–767. Kamens, T.C., 1981. MS thesis, University of Delaware, 130 pp. Karande, A.A., 1967. Proc. Symp. on Crustacea, Ernakulam Jan. 12–15, 1965, Part IV, pp. 1245–1253. Karande, A.A., 1973. Proc. Ind. Acad. Sci, Sect. B, 77, 56–63. Karande, A.A., 1974a. Ind. J. Mar. Sci., 3, 173–177. Karande, A.A., 1974b. Biol. Bull. (Woods Hole, Mass.), 146, 249–257. Karande, A.A., 1974c. Crustaceana, 26, 229–232. Karande, A.A., 1978. Ind. J. Mar. Sci., 7, 39–43. Karande, A.A., 1979. Proc. Ind Acad. Sci., Sect. B, 88, 73–83. Karande, A.A. & Palekar, V.C., 1963. J. Defence Sci., 13, 130–137. Karande, A.A. & Thomas, M.K., 1971. Curr. Sci., No. 5, 109–110. Karande, A.A. & Thomas, M.K., 1976. Proc. Ind. Acad. Sci., 83B, (5), 210–219. Kaufmann, R., 1962. Naturwissenschaften, Jahrgang 49, Heft 10, 238–239. Kaufmann, R., 1965. Zeit. Morph. Okol. Tiere, 55, 161–232. Kaufmann, R., 1971. Mitt. Inst. Columbo-Alemán Invest, Cient., 5, 73–85. Klepal, W., 1985. P. S. Z.N. I. Mar. Ecol., 6, 47–119. Klepal, W., 1987. Oceanogr. Mar. Biol. Annu. Rev., 25, 285–351. Klepal, W. & Barnes, H., 1975. J. Exp. Mar. Biol. Ecol., 17, 269–396. Klepal, W. & Barnes, H., 1977. J. Exp. Mar. Biol. Ecol., 27, 261–287. Klepal, W., Barnes, H. & Barnes, M., 1975. In, Ninth European Marine Biology Symposium, edited by H.Barnes, Aberdeen University Press, Aberdeen, pp. 275– 286. Klepal, W., Barnes, H. & Barnes, M., 1977. J. Exp. Mar. Biol. Ecol., 27, 289–304. Knight-Jones, E.W. & Waugh, G.D., 1949. J. Mar. Biol. Assoc. U.K., 28, 413–428. Kolosváry, G.von, 1947. Ann. Hist. Nat. Mus. Hung., 40, 1–88. Korn, O.M., 1985. Biol. Morya, Vladivostok, 1985, No. 3, 36–43. Korn, O.M. & Kolotukhina, N.K., 1983. Biol. Morya, Vladivostok, 1983, No. 2, 31–38 (in Russian, English summary).

EGG PRODUCTION IN CIRRIPEDES

129

Korn, O.M. & Ovsyannikova, I.I., 1979. Mar. Biol. Vladivostok, 5, 60–69 (in Russian, English summary). Korn, O.M. & Ovsyannikova, I.I., 1981. Zool. Zh., 60(10), 1472–1479 (in Russian, English summary). Kosaka, M. & Ishibashi, I., 1979. J. Fac. Mar. Sci. Technol. Tokai Univ., 12, 191– 207 (in Japanese, English abstract). Krüger, P., 1902. Arch. Mikroskop. Anat., 96, 355–386. Krüger, P., 1927. Die Tierwelt der Nord- und Ostsee, Lieferung. VIII, Teil X, Crustacea, d, 40 pp. Krüger, P., 1940. In, Dr H.G.Bronns Klassen und Ordnungen des Tierreichs. Fünfter Band. Abt. 1, Buch 3, Teil III, 560 pp. Kühl, H., 1966. Cirripedologist’s Newsletter, 2(2), 7–8. Kühnert, L., 1934. Zeit. Morph. Okol. Tiere, 29, 45–78. Landau, M., 1976. Crustaceana, 30, 105–106. Landau, M., Finney, C.M. & D’Agostino, A., 1979. Crustaceana, 37, 241–246. Lang, W.H., 1976. Biol Bull. (Woods Hole, Mass.), 150, 255–267. Lepore, E., Sciscioli, M. & Gherardi, M., 1979. Mem. Biol. Mar. Oceanogrqfia, N.S. 9, 115–122. Le Reste, L., 1965. Bull. Recl. Trav. Stn. Mar. Endoume, Bull. 38, Fasc. 54, 33–121. Leuckart, R., 1859. Ann. Mag. Nat. Hist., 4, 422–429. Lewis, C.A., 1975. Mar. Biol., 32, 141–153. Lewis, C.A. & Chia, F.-S., 1981. Can. J. Zool., 59, 893–901. Lewis, J.B., 1960. Can. J. Zool., 38, 391–35. Linsley, R.H. & Carpelan, L.H., 1961. In, The Ecology of the Salton Sea, California, in Relation to the Sportfishery, edited by B.W.Walker, Calif. Dept. Fish and Game, Fish Bull. No. 113, pp. 43–47. Lucas, M.I. & Crisp, D.J., 1987. J. Mar. Biol. Assoc. U.K., 67, 27–54. Luckens, P.A., 1968. Bull. Mar. Biol. Stn. Asamushi, Tohoku Univ., 13(2), 75–82. Luckens, P.A., 1969. Bull. Mar. Biol. Stn. Asamushi, Tohoku Univ., 13(4), 251–254. Luckens, P.A., 1970. N. Z. J. Mar. Freshwater Res., 4, 497–514. Luckens, P.A., 1976. Mem. N. Z. Oceanogr. Inst., No. 70, 64 pp. Lützen, J., 1981. J. Exp. Mar. Biol. Ecol., 50, 231–254. Lützen, J., 1984. Sarsia, 69, 91–106. Malusa, J.R., 1986. Biol. Bull. (Woods Hole, Mass.), 170, 409–428. McLaughlin, P.A. & Henry, D.P., 1972. Crustaceana, 22, 13–30. McMurrich, J.P., 1917. Proc. Trans. R. Soc. Can. Ser. 3, 11 Sect. 4, 47–61. Millard, N., 1952. Trans. R. Soc. S. Afr., 33, 415–445. Molenock, J. & Gomez, E.D., 1972. Crustaceana, 23, 100–108. Moore, H.B., 1935. J. Mar. Biol. Assoc. U.K., 20, 57–63. Moore, L.B., 1944. Trans. R. Soc. N. Z., 73, Part 4, 315–334. Moyse, J., 1961. Proc. Zool. Soc. London, 137, 371–392. Moyse, J., 1987. In, Barnacle Biology, edited by A.J.Southward, A.A.Balkema, Rotterdam, pp. 329–362. Moyse, J. & Knight-Jones, E.W., 1967. Proc. Symp. Crustacea, Ernakulam, Part II, 595–611. Müller, F., 1862. Archiv. für Naturgeschichte, 28, 1–9. Munn, E.A. & Barnes, H., 1970a. J. Exp. Mar. Biol. Ecol., 4, 261–286. Munn, E.A. & Barnes, H., 1970b. J. Exp. Cell Res., 60, 277–284. Newman, W.A., 1971. J. Zool., 165, 423–29. Newman, W.A., 1974. J. Mar. Biol. Assoc. U.K., 54, 437–56. Newman, W.A., 1980. Téthys, 9, 379–398. Newman, W.A., 1982. In, The Biology of Crustacea, Vol. 1, edited by L.G.Abele, Academic Press, New York, pp. 197–221. Newman, W.A., 1987. In, Barnacle Biology, edited by A.J.Southward, A.A. Balkema, Rotterdam, pp. 3–12. Newman, W.A. & Ross, A., 1971. Antarctic Res. Ser., 14, 257 pp. Newman, W.A. & Ross, A., 1976. San Diego Soc. Nat. Hist., Mem., 9, 108 pp. Newman, W.A. & Ross, A., 1977. Crustaceana, 32, 102 only.

130

MARGARET BARNES

Newman, W.A. & Tomlinson, J.T., 1974. Crustaceana, 27, 204–208. Newman, W.A., Zullo, V.A. & Withers, T.H., 1969. In, Treatise on Invertebrate Paleontology. R. Arthropoda, 4, Vol. 1, pp. R206-R295. Nilsson-Cantell, C.-A., 1921. Zool. Bidr. Uppsala, 7, 75–395. Nilsson-Cantell, C.-A., 1930a. ‘Discovery’ Rept., 2, 223–260. Nilsson-Cantell, C.-A., 1930b. Res. Sci. Voyage aux Indes Orientales Néerl. de LL. AA. RR. le Prince et 1a Princesse Léopold de Belgique, 3, (3) 24 pp. Nilsson-Cantell, C.-A., 1932. Ark. Zool., 23, part 4A, No. 18, 12 pp. Nilsson-Cantell, C.-A., 1950. Rept. Swedish Deep-Sea Exped., Zoology, 2(17), 215– 220. Nilsson-Cantell, C.-A., 1978. Cirripedia Thoracica and Acrothoracica. Marine Invertebrates of Scandinavia, No. 5, Universitetsforlaget, Oslo, Norway, 133 pp. Noll, F.C., 1875. Zeit. Wiss. Zool., 25, 114–128. Norris, E. & Crisp, D.J., 1953. Proc. Zool. Soc. London, 123, 393–409. Orton, J.H., 1936. Ann. Mag. Nat. Hist., Ser. 10, 17, 617–625. Otway, N.M. & Underwood, A.J., 1987. J. Exp. Mar. Biol. Ecol., 105, 85–106. Ovsyannikova, I.I. & Korn, O.M., 1984a. Soviet J. Mar. Biol., 10, 275–280. Ovsyannikova, I.I. & Korn, O.M., 1984b. Mar. Biol. Vladivostok, No. 5, 34–40. Ovsyannikova, I.I. & Levin, V.S., 1982. Mar. Biol. Vladivostok, No. 4, 44–51. Page, H.M., 1983. J. Exp. Mar. Biol. Ecol., 69, 189–202. Page, H.M., 1984. J. Exp. Mar. Biol. Ecol., 74, 259–272. Page, H.M., 1986. Mar. Ecol. Prog. Ser., 29, 157–164. Patel, B. & Crisp, D.J., 1960a. Physiol. Zool., 33, 104–119. Patel, B. & Crisp, D.J., 1960b. J. Mar. Biol. Assoc. U.K., 39, 667–680. Patel, B. & Crisp, D.J., 1961. Crustaceana, 2, 89–107. Patel, B.A., 1959. J. Mar. Biol. Assoc. V.K., 38, 589–597. Paul, M.D., 1942. Proc. Ind. Acad. Sci., Ser. B, 15, 1–42. Pawlik, J.R., 1987. J. Crust. Biol., 7, 265–273. Pillai, N.K., 1958. Bull. Res. Inst. Univ. Kerala, Ser. C, 6, 117–130. Pillay, K.K. & Nair, N.B., 1970. J. Kerala Acad. Biol., 2, 22–32 (English summary). Pillay, K.K. & Nair, N.B., 1972. Ind. J. Mar. Sci., 1, 8–16. Pilsbry, H.A., 1907. Proc. Acad. Nat. Sci. Philad., 59, 408–16. Pope, E.C., 1965. Proc. Linn. Soc. N. S. W., 90, 10–77. Potts, F.A., 1915. Papers from the Dept. of Mar. Biol., Carnegie Inst. of Washington, 8, 1–32. Powell, H.T., 1954. Nature (London), 173, 688 only. Pyefinch, K. A, 1948. J. Mar. Biol. Assoc. U.K., 27, 464–503. Pyefinch, K.A., 1949. Proc. Zool. Soc. London, 118, 916–923. Ralph, P.M. & Hurley, D.E., 1952. Zool. Publ. Victoria Univ. Coll., No. 19, 22 pp. Rasmussen, T., 1980. Int. Congr. Mar. Corrosion and Fouling, No. 5, 37–44. Rege, M.S., Joshi, S.S. & Karande, A.A., 1980. Ind. J. Mar. Sci., 9, 15–18. Reinhard, E.G., 1942a. J. Morph., 70, 389–402.. Reinhard, E.G., 1942b. Biol. Bull. (Woods Hole, Mass.), 83, 401–15. Reinhard, E.G., 1944. J. Wash. Acad. Sci., 34, 49–58. Reinhard, E.G., 1946. J. Wash. Acad. Sci., 36, 127–131. Reinhard, E.G. & Evans, J.T., 1951. J. Morph., 89, 59–66. Reinhard, E.G. & Stewart, T.C., 1956. Proc. Helminth Soc. Wash., 23, 162–168. Reischman, P.G., 1959. Proc. Ned. Akad. Wet., Amsterdam, 62C, 409–435. Relini, G., 1968. Bull. Mus. Ist. Biol. Univ. Genova, 36, 185–190. Relini, G., 1980. XXVIIe Congr.-Assemblé pléniere, Cagliari, Oct. 1980, 2 pp. Relini, G., 1983. Rapp. P.-V. Réun. Commn. Int. Explor. Scient. Mer Médit., 28(3), 273–275.

EGG PRODUCTION IN CIRRIPEDES

131

Relini, G. & Fasciana, C., 1982. Bull. Mus. Ist. Biol. Univ. Genova, 50, 306–312. Relini, G. & Giordano, E., 1969. Nat.-Soc. It. Sc. Nat., Mus. Civ. St. Nat. e Acquario Civ, Milano, 60, 251–258. Relini, G., Matricardi, G. & Diviacco, G., 1980. 5th Int. Congr. Mar. Corrosion and Fouling, pp. 45–60. Ren, X., 1983. Oceanologia Limnol. Sin., 14, 74–87 (in Chinese, English summary). Rigo, L., 1941. Atti Ist. Veneto, Cl. Sci. Mat. Nat., 100, 639–651. Ritchie, L.E. & Høeg, J.T., 1981. J. Crust. Biol., 1, 334–347. Resell, N.C., 1967. Nat. App. Sci. Bull, 20, 277–319. Rosell, N.C., 1976. Int. Congr. Mar. Corrosion and Fouling No. 4, 457–459. Ross, A., 1975. Crustaceana, 28, 17–20. Ross, A. & Newman, W.A., 1969. In, Distribution of Selected Groups of Marine Invertebrates in Waters South of 35°S Latitude. Antarctic Map Folio Series, Folio 11, Am. Geog. Soc., pp. 30–32. Runnström, S., 1924–1925. Bergens Mus. Årbok Naturv. Raekke, No. 5, 46 pp. Rzepishevsky, I.K., 1959. In, Internationalis Oceanography Congress Preprints, edited by M.Sears, Am. Advmt. Sci., Washington, D.C., pp. 247–249. Rzepishevsky, I.K., 1962. Int. Rev. Gesamt. Hydrobiol., 47, 471–479. Sandison, E.E., 1950. Nature (London), 165, 79–80. Sandison, E.E., 1954. Trans. R. Soc. S. Afr., 34, 69–101. Sandison, E.E., 1966. J.Anim. Ecol., 35, 363–378. Sandison, E.E., 1967. Crustaceana, 13, 161–174. Sastry, A.N., 1983. In, The Biology of Crustacea, Vol. 8, edited by F.J.Vernberg and W.B.Vernberg, Academic Press, New York, pp. 179–270. Sewell, R.B.S., 1926. Rec. Ind. Mus., 28, 269–330. Shirase, S. & Yanagimachi, R., 1957. Zool. Mag. Tokyo, 32, 253–257 (in Japanese, English summary). Skerman, T.M., 1958a. N. Z. J. Sci, 1, 402–411. Skerman, T.M., 1958b. N. Z. J. Sci., 1, 224–257. Skerman, T.M., 1959. N. Z. J. Sci., 2, 57–94. Smith, G., 1906. Fauna et Flora des Golfes von Neapel, 29, 123 pp. Smith, G. & the late Weldon, W.F. R., 1920. In, The Cambridge Natural History, Vol. IV, edited by S.F.Harmer and A.E.Shipley, Macmillan & Co., Ltd., London, pp. 79–109. Smyth, M.J., 1986. J. Crust. Biol., 6, 143–157. Southward, A.J., 1976. J. Mar. Biol. Assoc. U.K., 56, 1007–1028. Southward, A.J. & Southward, E., 1967. Arctic (J. Arctic Inst. N. Am. Arctic), 20, 8–20. Spivey, H.R., 1979. Bull. Mar. Sci., 29, 497–508. Stebbing, T.R.R. & Fowler, G.H., 1904. Trans. Linn. Soc. Land. Ser. 2, Zool, 10, 13–54. Steele, D.H., 1977. Am. Nat., 111, 371–372. Steele, D.H. & Steele, V.J., 1975. Int. Rev. Gesamt. Hydrobiol., 60, 711–715. Stefoni, D.L. & Countreras, G.T., 1979. Acta Zool. Lilloana, 35, 547–561 (in Spanish, English Summary). Stewart, F.H., 1911. Mem. Ind. Mus., 3, 33–51. Stone, R.L. & Barnes, H., 1973. J. Exp. Mar. Biol. Ecol., 12, 167–185. Stone, R.L. & Barnes, H., 1974. J. Exp. Mar. Biol. Ecol., 14, 27–36. Strathmann, R.R., 1977. Am. Nat., 111, 373–376. Straughan, D., 1971. In, Biological and Oceanographical Survey of the Santa Barbara Channel Oil Spill 1969–1970, Vol. I. Biology and Bacteriology, edited by D. Straughan, Allan Hancock Foundation, University of California, pp. 223–244. Stubbings, H.G., 1936. John Murray Expedition 1933–34, Sci. Res., 4(1), 70 pp. Stubbings, H.G., 1961. Res. Sci. Camp, de la ‘Calypso’, Fasc. V, 179–192. Stubbings, H.G., 1965. Bull. Inst. Fr. Afr. Noire, 27, Ser. A, No. 3, 876–907. Stubbings, H.G. & Haughton, D.R., 1964. Int. Rev. Gesamt. Hydrobiol., 49, 233–279. Svane, I., 1986. Mar. Biol., 90, 249–253.

132

MARGARET BARNES

Tenerelli, V., 1958. Boll. Accad. Gioenia Sci. Nat. Catania, Ser. IV, 4, 263–280. Tenerelli, V., 1959. Boll. Accad. Gioenia Sci. Nat. Catania, Ser. IV, 5, 93–107. Thomer, E., 1967. Z. Morph. Ökol. Tiere, 60, 176–226. Thorner, E. & Ankel, W.E., 1966. Natur und Museum, 96(6), 209–220. Tighe-Ford, D.J., 1967. Nature (London), 216, 920–921. Tomlinson, J., 1966. J. Theor. Biol., 11, 54–58. Tomlinson, J., 1967. Bull. Natl. Mus. Singapore, No. 33, 101–105. Tomlinson, J.T., 1953. J. Wash. Acad. Sci., 43, 373–381. Tomlinson, J.T., 1955. J. Morph., 96, 97–121. Tomlinson, J.T., 1960. Bull. Inst. Fr. Afr. Noire, 22, Ser. A, 402–410. Tomlinson, J.T., 1969. U.S. Natl. Mus. Bull., No. 296, 162 pp. Tomlinson, J.T., 1973. Wasmann J. Biol., 31, 263–288. Tooke, N.E. & Holland, D.L., 1985. J. Exp. Mar. Biol. Ecol., 87, 241–253. Tooke, N.E., Holland, D.L. & Gabbott, P.A., 1985. J. Exp. Mar. Biol. Ecol., 87, 255–269. Tubb, J.A., 1946. Rec. Aust. Mus. Sydney, 21, 383–385. Turquier, Y., 1963. Pacif. Sci., 17, 299–301. Turquier, Y., 1967a. Cah. Biol. Mar., 8, 75–87. Turquier, Y., 1967b. Arch. Zool. Exp. Gén., 108, 33–47. Turquier, Y., 1967c. Arch. Zool. Exp. Gén, 108, 111–137. Turquier, Y., 1968. Arch. Zool. Exp. Gén., 109, 113–122. Turquier, Y., 1971. Arch. Zool. Exp. Gén., 112, 301–348. Turquier, Y., 1972a. Arch. Zool. Exp. Gén., 113, 499–551. Turquier, Y., 1972b. Arch. Zool. Exp. Gén., 113, 167–196. Turquier, Y., 1977. Arch. Zool. Exp. Gén., 118, 133–154. Turquier, Y, 1985a. Bull. Soc. Zool. Fr., 110, 151–168. Turquier, Y., 1985b. Bull. Soc. Zool. Fr., 110, 169–189. Turquier, Y. & Pochon-Masson, J., 1969. Arch. Zool. Exp. Gén., 110, 453–70. Underwood, A.J., 1974. Am. Nat., 108, 874–878. Utinomi, H., 1950. Mem. Coll. Sci. Univ. Kyoto, Ser. B, 19, No. 3, Art. 18, 8 pp. Utinomi, H., 1960. Publ. Seto Mar. Biol. Lab., 8, 223–279. Utinomi, H., 1961. Publ. Seto Mar. Biol. Lab., 9, 413–446. Utinomi, H., 1964. Publ. Seto Mar. Biol. Lab., 12, 117–132. Utinomi, H., 1967. Publ. Seto Mar. Biol. Lab., 15, 117–120. Vance, R.R., 1973a. Am. Nat., 107, 339–352. Vance, R.R., 1973b. Am. Nat., 107, 353–361. Vance, R.R., 1974. Am. Nat., 108, 879–880. Van Kampen, P.N. & Boschma, H., 1925. Siboga-Expeditie, Monographie, No. XXXI bis, 62 pp+plates. Veillet, A., 1943a. Bull. Inst. Océanogr. Monaco, No. 841, 4 pp. Veillet, A., 1943b. Bull. Inst. Océanogr. Monaco, No. 845, 4 pp. Veillet, A., 1945. Ann. Inst. Océanogr. Monaco, 22, 193–341. Veillet, A., 1956. Bull. Séances Soc. Sci. Nancy, 15, 1–3. Veillet, A., 1961. C. R. Acad. Sci. Paris, 253, 3087–3088. Veillet, A., 1962. C. R. Acad. Sci. Paris, 254, 176–177. Villalobos, C.R., 1979a. Rev. Biol. Trop., 27, 279–293. Villalobos, C.R., 1979b. Rev. Biol. Trop., 27, 293–300. Villalobos, C.R., 1980. Rev. Biol. Trop., 28, 193–201. Visscher, J.P, 1938. Biol. Bull. (Woods Hole, Mass.), 75, 363 only. Wagh, A.B., 1965. Ph.D. thesis, University of Bombay, India, 338 pp. Walker, G., 1977a. J. Mar. Biol. Assoc. U.K., 57, 969–972.

EGG PRODUCTION IN CIRRIPEDES

Walker, G., 1977b. Experientia, 33, 1603–1604. Walker, G, 1980. Phil. Trans. R. Soc. London, Ser. B, 291, 147–162. Walker, G., 1985. J. Exp. Mar. Biol. Ecol, 93, 131–145. Walker, G., 1987. J. Exp. Mar. Biol. Ecol., 106, 151–163. Walley, L.J., 1965. J. Mar. Biol. Assoc. U.K., 45, 115–128. Walley, L.J., White, F. & Brander, K.M, 1971. J. Mar. Biol. Assoc. U.K., 51, 489–494. Weiss, C.M., 1948a. Biol. Bull. (Woods Hole, Mass.), 94, 236–243. Weiss, C.M., 1948b. Ecology, 29, 153–172. Wells, H.W. & Tomlinson, J.T., 1966. Quart. J. Florida Acad. Sci., 29, 27–37. Werner, Jr, W.E., 1967. Bull. Mar. Sci., 17, 64–84. Wethey, D.S., 1983. Biol. Bull. (Woods Hole, Mass.), 165, 330–341. Wethey, D.S., 1984. Can. J. Zool., 62, 1788–1795. White, F., 1970. Mar. Biol., 5, 29–34. Willemöes-Suhm, R.von, 1876. Phil. Trans. R. Soc. London, 166, 131–154. Wisely, B., 1960. Aust. J. Mar. Freshwater Res., 11, 42–54. Wisely, B. & Blick, R.A.P., 1964. Aust. J. Mar. Freshwater Res., 15, 162–171. Witschi, E., 1935. Biol. Bull. (Woods Hole, Mass.), 68, 263–267. Wu, R.S.S., 1980. Can. J. Zool., 58, 559–566. Wu, R.S.S., 1981. Can. J. Zool., 59, 890–892. Wu, R.S.S. & Levings, C.D., 1978. Mar. Biol, 45, 225–235. Wu, R.S.S. & Levings, C.D., 1979. Mar. Biol., 54, 83–89. Wu, R.S.S. & Levings, C.D., 1980. Mar. Pollut. Bull., 11, 11–15. Wu, R.S.S., Levings, C.D. & Randall, D.J., 1977. Can. J. Zool., 55, 643–647. Yamaguchi, T., 1973. Publ. Seto Mar. Biol. Lab., 21, 115–140. Yan, W. & Chen, X., 1980. Nanhai Stud. Mar. Sinica, 1, 125–134 (in Chinese, English abstract). Yanagimachi, R., 1960. Bull. Mar. Biol. Stn. Asamushi, 10, 109–110. Yanagimachi, R., 1961a. Biol. Bull. (Woods Hole, Mass.), 120, 272–283. Yanagimachi, R., 1961b. Crustaceana, 2, 183–186. Yanagimachi, R. & Fujimaki, N., 1967. Annot. Zool. Jpn., 40, 98–104. Yasuda, T., 1970. Bull. Jap. Soc. Scient. Fish, 36, 1007–1016. Zann, L.P., 1975. In, The Biology of Sea Snakes, edited by W.A. Dunson, University Park Press, Baltimore, Maryland, pp. 267–286. Zann, L.P. & Harker, B.M., 1978. Crustaceana, 35, 206–214. Zevina, G.B., 1957. C. R. Acad. Sci. URSS, Doklady N.S., 113, 450–453. Zevina, G.B., 1963a. Trudy Inst. Okeanol. U.S.S.R., Akad. Nauk, 70, 52–70 (in Russian, English summary). Zevina G.B., 1963b. Trudy Inst. Okeanol. U.S.S.R., Acad. Nauk, 70, 72–75. Zevina, G.B., 1972. Crustaceana, 22, 39–63. Zevina, G.B., 1974. Crustaceana, 27, 209–215. Zevina, G.B., 1987. Zool. Zh., 66, 1304–1313. Zevina, G.B., 1988. Zool. Zh., 67, 31–0. Zevina, G.B. & Poliakova, I.W., 1986. Zool. Zh., 65, 1420–1423. Zullo, V.A., 1979. NOAA Technical Report NMFS (Nat. Mar. Fish. Services) Circular, No. 425, 27 pp.

133

Oceanogr. Mar. Biol. Annu. Rev., 1989, 27, 167–272 Margaret Barnes, Ed. Aberdeen University Press

BIOLOGY OF MARINE HERBIVOROUS FISHES* MICHAEL H.HORN Department of Biological Science, California State University, Fullerton, California 92634, and Ocean Studies Institute, Long Beach, California 90840, USA

ABSTRACT iF s hes that feed on benthic algae occur as solitary individuals, members of roving groups or defenders of territories containing their algal food. They may be broadly classified as browsers or grazers , the latter if sediment is ingested during feeding. nO tropical re efs, the Acanthuridae, Scaridae, Siganidae, and oP macen tridae have abundant and ecologically important herbivorous members. In temperate waters the Aplodactylidae, G irellidae, dO acidae, and Stichaeidae contain or consist solely of plant-eating species. The K yphosidae and Sparidae contain herbivorous species in both temperate and tropical waters. uN merous herbivorous fishes select relatively tender and palatable algae, and apparently are deterred by combinations of morphological and chemical defences found in many seaweeds. N evertheless, some species regularly consume tough and chemically defended seaweeds. H erbivorous fishes can assimilate algal material, but few growth studies have been done. M ost herbivores have longer guts than carnivores and relatively high ingestion rates and fast gut transit times. Several species, however, feed more intermittently and retain food longer. H erbivorous fishes apparently do not produce cellulases to break down plant cell walls but gain access to the contents by lysing the cells in a highly acidic stomach, grinding the food in a muscular stomach or pharyngeal mill or harbouring microbes that ferment the food in a hindgut caecum. eH rbi vorous and detritivorous fishes maintain large populations on their low protein diets and have evolved several specialisations similar to those of terrestrial herbivores to cope with the low nitrogen content of their diets. rF e e-ranging herbivorous fishes, especially surgeonfishes and parrotfishes, markedly affect coral reef ecosystems by contributing to erosion and sedimention as well as affecting the abundance and composition of benthic algal communities. Because of their mobility these fishes decrease the spatial variability of herbivore effects on community structure but often appear to have smaller local effects on seaweed communities than grazing s ea urchins. Territorial damselfishes by defending and maintaining algal patches strongly affect the abundance, diversity, and productivity of seaweed communities on coral reefs. iL mited evidence suggests that herbivorous fishes might also influence community structure of temperate algae.

BIOLOGY OF MARINE HERBIVOROUS FISHES

135

H erbivores constitute a small proportion of marine fishes and most are eP rci formes, a large and almost certainly polyphyletic assemblage of teleosts. They are largely confined to coastal habitats between 40°N and 40°S and are so mewhat more diverse in southern rather than northern temperate habitats. Inexplicably, herbivorous fishes are more species-rich in tropical than in temperate waters. aL ti tudinal differences in abundance (density) are less obvious, but too few data are available to draw any conclusions about reasons or consequ ences.

INTRODUCTION Fish herbivory in the sea is an important ecological subject because the trophic interactions at the base of food webs, the energetic requirements and physiological capabilities of fishes and the role of fishes in the flow of energy through marine communities are involved. A diet of seaweeds or seagrasses confers on fishes a set of basic physiological problems that has made fish herbivory an intriguing and provocative topic for a continuing series of investigators. Like other herbivorous animals, plant-eating fishes face seemingly formidable difficulties concerning the quality of, and access to, the nutrients available in the seaweeds they consume. Animals have much greater nitrogen requirements than plants and also use nitrogen less efficiently (Mattson, 1980). Therefore, herbivores must consume relatively large quantities of food and assimilate it efficiently in order to meet their nitrogen (and energy) requirements (Mattson, 1980; Crawley, 1983). These challenges are hard to meet partly because plants, as stationary organisms, defend themselves against herbivore attack in a variety of ways and partly because plant nutrients are components of, or are contained within, relatively indigestible cell walls. Several major questions surround herbivory by fishes in the marine environment, and consideration of these questions forms the framework of this review. They are as follows. (1) How do herbivorous fishes select their diets from an array of seaweeds (and seagrasses) that vary in availability, nutritional quality, thallus (leaf) toughness, secondary chemistry, and digestibility? (2) How do these fishes obtain an adequate diet from a food source relatively low in nitrogen and, hence, protein? In other words, what morphological and physiological specialisations allow herbivorous fishes to gain access to and assimilate the nutrients locked inside plant cell walls? (3) What ecological impacts do herbivorous fishes have on seaweed populations and on nearshore marine communities in general? (4) Why are there so few strictly herbivorous fishes in temperate and polar latitudes as compared with tropical latitudes? To consider these questions, the review is divided into four main sections: (1) Food and feeding; (2) Digestion and digestive mechanisms; (3) Ecological impacts; and (4) Distribution, diversity and abundance. The protein requirements of herbivorous fishes and the evolutionary responses of herbivorous fishes to nitrogen shortage form two additional sections that complete the review. A few definitions and restrictions are necessary at the outset to provide the boundaries for the review. Coverage is limited to those species of marine fishes that consume benthic plants: macroalgae (seaweeds), diatoms or seagrasses. Fishes that feed on phytoplankton are not included except where their mention contributes in a comparative way to the herbivores of interest. Emphasis is placed on fishes that appear to be dependent on seaweeds, diatoms or seagrasses for most of their energy and nutritional requirements during at least part of their life histories. This focus is qualitative and imprecise, however, because fishes in general are often catholic and notoriously opportunistic in their food habits. Nevertheless, it is important to try to

*Contribution No. 61 from the Ocean Studies Institute

136

MICHAEL H.HORN

limit discussion to those fishes feeding primarily on the plant end of the food spectrum; otherwise, the review grades into an appraisal of omnivores, which are common in many coastal marine habitats (Bakus, 1969). What topics have been emphasised to date in studies of marine herbivorous fishes? The literature on marine fish herbivory is far less voluminous than that on terrestrial herbivores, especially insects and mammals, but also less so than that on marine invertebrate herbivores, particularly sea urchins and gastropod molluscs. Among the reasons for the latter disparity, aside from the abundance and ecological importance of these two invertebrate groups, is that fishes are more mobile and therefore more difficult to study. The advent of SCUBA and related diving gear and techniques, however, has led to rapid developments in the in situ study of coastal fishes, especially on coral reefs (Sale, 1980), where herbivorous fishes are particularly abundant and diverse. Studies of fish communities in tropical waters (e.g. Hiatt & Strasburg, 1960; Randall, 1967; Hobson, 1974) included accounts of the feeding behaviour and food habits of the herbivorous components and recognised that although herbivorous fishes made up a minor proportion of the total diversity, they were often among the most abundant species. In temperate waters, true herbivores have been viewed as rare or non-existent in both earlier (e.g. Quast, 1968) and more recent (e.g. Moreno & Jara, 1984) studies of temperate-zone fish assemblages even though several species were found to contain substantial amounts of macroalgae in their stomachs. The notion has prevailed until very recently that even fishes whose guts were packed with seaweeds were not digesting the algal material but actually gaining their nutrition from the epibionts on the plant surfaces (e.g. Wheeler, 1980). Only in recent years has it become apparent that year-round herbivores occur in the temperate waters of both the northern (e.g. Horn, Murray & Edwards, 1982) and southern (e.g. Russell, 1983) hemispheres and that fishes from both regions can digest macroalgae (e.g. Edwards & Horn, 1982; Anderson, 1987). It is of interest that herbivores in the southern hemisphere, especially in the waters around New Zealand and southern Australia, are more diverse than those in the northern hemisphere (e.g. see Burchmore et al., 1980; Stephens & Zerba, 1981) and make up large proportions of the total fish biomass of rocky reef communities (e.g. Russell, 1983). In the last decade, an increasing number of studies have shown that herbivorous fishes in tropical waters, particularly on coral reefs, have profound influences on individual seaweed populations and community structure (e.g. Hixon, 1986). Damselfishes, in particular, have been shown in a continuing series of papers to have a marked effect on the abundance and species composition of seaweeds in their territories (e.g. Hixon, 1983). Also in tropical waters, the morphological and chemical defences of seaweeds have been intensively studied of late and shown to affect strongly the food selection of certain herbivorous fishes (e.g. Hay & Fenical, 1988). Greater recognition has been given in recent years to the differences in herbivore diversity in temperate compared with tropical latitudes (e.g. Choat, 1982; Gaines & Lubchenco, 1982; Thayer et al., 1984), but little progress has been made in providing and testing explanatory hypotheses. Although anatomical specialisations of herbivorous fishes have long been recognised (e.g. Suyehiro, 1942; Al-Hussaini, 1947) and their digestive physiology generally summarised more recently (e.g. Kapoor, Smit & Verighina, 1975; Lobel, 1981; Pandian & Vivekanandan, 1985), much remains to be known about the physiological and biochemical mechanisms by which fishes break down and absorb seaweed material. The synthesis of cellulolytic enzymes by fishes remains unproven (see Lewis & Peters, 1984; Urquhart, 1984), but a microbial fermentation system is known to occur in at least two herbivorous fishes (Rimmer & Wiebe, 1987). Greater knowledge of the digestive physiology of herbivorous fishes may be a key to understanding the differences in diversity and abundance of these species in tropical and temperate habitats (Clements & Bellwood, 1988). To my knowledge, this review represents the first attempt to provide a comprehensive account of marine fish herbivory on a worldwide basis. Previously published review articles on herbivorous fishes have

BIOLOGY OF MARINE HERBIVOROUS FISHES

137

concentrated either on their ecological roles (Ogden & Lobel, 1978; Hixon, 1983) or their digestive mechanisms (Lobel, 1981). Many other reviews that were focused on larger or peripheral topics such as plant-herbivore interactions (e.g. Lubchenco & Gaines, 1981; Gaines & Lubchenco, 1982; Hay & Fenical, 1988), fish feeding (e.g. Hyatt, 1979; Choat, 1982; Hixon, 1986) and fish digestion and energetics (e.g. Kapoor et al., 1975; Brett & Groves, 1979; Fange & Grove, 1979; Pandian & Vivekanandan, 1985) have included information on herbivorous fishes. FOOD AND FEEDING Even though most herbivorous fishes consume a variety of seaweeds, they still must select from dozens to hundreds of the different plant species available to meet their nutritional requirements. How is this accomplished and what are the outcomes of the process? In the words of Howe & Westley (1988), few questions in ecology are of more fundamental and practical interest than “why do different herbivores eat different plants?” Food choice is still poorly known for fish herbivores but deserves further attention because, for example, preferences often determine (Lubchenco & Gaines, 1981) the role of the animal in community organisation. This section considers the feeding behaviour of herbivorous fishes, their food habits and preferences and the factors influencing diet selection. FEEDING BEHAVIOUR Fishes that eat algae usually have short, blunt snouts with closely set teeth that form a cropping edge (Ogden & Lobel, 1978). In their extreme form, the teeth are fused to form a beak as in the parrotfishes (Scaridae), most of which bite into the inorganic substratum and obtain algal mat and endolithic algae (Ogden & Lobel, 1978). Odacids (Odacidae) use the beak to bite pieces out of the thalli or reproductive structures of laminarian and fucoid algae (Clements, 1985; Clements & Bellwood, 1988). Herbivorous fishes can be classified as either grazers or browsers (Hiatt & Strasburg, 1960; Jones, 1968; Backus, 1969; Ogden & Lobel, 1978). Grazers pick up inorganic substratum while feeding by scraping or sucking, whereas browsers bite or tear more upright macroalgae and rarely ingest any inorganic material (Jones, 1968). Grazers tend to feed non-selectively because their algal food is small and closely affixed to the substratum, whereas browsers take whole or parts of individually recognisable seaweeds and, therefore, are more selective (Lobel, 1981; Choat, 1982). In tropical reef communities, herbivorous fishes are daytime feeders and are among the species that seek shelter within the reef as twilight progresses and remain inactive at night (e.g. Earle, 1972). Travel between foraging areas and slumber sites occurs in both temperate (Meekan, 1986) and tropical species (Fishelson, Montgomery & Myrberg, 1987) and seems to be prevalent in habitats where food supply and hiding places are spatially separated (Fishelson et al., 1987). During the day, herbivorous fishes, especially those associated with coral reef habitats, forage in one of three ways: (1) territorial defence; (2) group foraging, or (3) individual home ranges (Ogden & Lobel, 1978). Territorial defence in tropical species The most obvious holders of feeding territories on coral reefs are herbivorous damselfishes (Pomacentridae) (Hixon, 1983). Their territories are readily identified as patches of algae frequently of a particular colour and consistency. These fishes have important effects on the reef itself and on other herbivorous fishes in

138

MICHAEL H.HORN

their habitat. The behaviour and ecology of territorial damselfishes have been the subject of intensive study over the past 15 years. Other tropical reef fishes including blenniids (Nursall, 1977), scarids and acanthurids (Foster, 1985a) are also known to hold feeding territories. These territories vary in time, space and intensity with which they are defended. Some overlap and are shared between unrelated species; others are defined by the morphological limitations of the holder. This range of territorial situations provides a glimpse into the complexities of biological interactions in coral reef communities. The following examples illustrate this spectrum of territorial relationships. The parrotfish Scarus croicensis holds territories that appear to serve the dual functions of feeding and reproduction (Ogden & Buckman, 1973; Buckman & Ogden, 1973) and defends against both conspecifics and other benthic feeding fishes (Robertson, Sweatman, Fletcher & Cleland, 1976). In the San Bias Islands (Panama), however, some individuals of this species hold permanent feeding territories while others form feeding schools (Robertson et al., 1976). In St Croix, US Virgin Islands, S. croicensis does not hold territories (Buckman & Ogden, 1973), possibly because of a more dispersed food supply (Ogden & Lobel, 1978). The Caribbean blenny Ophioblennius atlanticus fiercely defends small feeding territories on exposed surfaces of coral rock against conspecifics but weakly against other species (Nursall, 1977, 1981). Its territory often lies within the larger territory of the damselfish Eupomacentrus dorsopunicans and, although these two species overlap in diet, they interact little with each other. Nursall credits this lack of interaction to the benthic habits, non-aggressive behaviour and time-minimised use of the algal food resource by the blenny. A third species, the damselfish Microspathodon chrysurus, uses the same zone as a foraging area and together with the more aggressive Eupomacentrus dorsopunicans protects the space including the territory of the blenny against roving grazers and other potential invaders. On the Great Barrier Reef, another blenny-damselfish pair has been shown to occupy overlapping territories and share the algal turf food resource with little mutual interference (Roberts, 1987). The damselfish Pomacentrus flavicauda and the blenny Salarias fasciatus had similar diets except that the former fed to some degree on plankton. Removal experiments by Roberts (1987) failed to provide evidence of exploitation competition between these species or of changes in levels of interference competition by the damselfish after blenny removal. Damselfish were unable to exclude blennies from their territories because the blennies took refuge in holes. The feeding activities of the blennies appeared not to be detrimental to Pomacentrus flavicauda, at least in the short-term. This result is intriguing given the estimates in other studies (Hatcher, 1981; Walker, 1984) that blennies seem to exert the greatest pressure among grazers on algal turf in areas dominated by territorial fishes. Some damselfish populations, however, may not be at the carrying capacity of the habitat (Robertson, Hoffman & Sheldon, 1981) and, in certain cases, limited more by recruitment than resource availability (Wellington & Victor, 1985). A three-species system involving the symbiotic sharing of feeding territories and algal food has been described for coral reef fishes at Aldabra in the Indian Ocean (Robertson & Polunin, 1981). A damselfish (Stegastes fasciolatus) occurs in the feeding areas of two much larger surgeonfishes (Acanthurus lineatus and A. leucosternon). All three defend feeding areas against conspecifics and other fishes with similar diets but show little aggression toward each other. Like the blennies in the above examples, the damselfish provides little of the interspecific defence and cannot be excluded by the surgeonfishes because it can take refuge in holes. The authors argue that the cost to the surgeonfishes of having the damselfish in their feeding areas is small because the damselfish has a low biomass density, a slightly different diet and contributes even if in a minor way to the defence of the shared feeding territories. Long-term experiments are required, as they are in the study by Roberts (1987) described above, to sort out the costs and benefits of

BIOLOGY OF MARINE HERBIVOROUS FISHES

139

the relationship to each species. Such a relationship may represent a mechanism by which some of the numerous coral reef fish species can coexist with little apparent competition on the same potentially limiting resources. In another study at Aldabra involving the same two acanthurids, Robertson, Polunin & Leighton (1979) claimed that morphological limitations dictated differences in territoriality and food habits between these two species. They reasoned that A. leucosternon, with its smaller size, smaller mouth, more ovoid body with larger median fins and more truncate tail, has greater flexibility in its feeding behaviour than A. lineatus and, therefore, can feed more efficiently on smaller, more sparsely distributed algae and small growths in crevices. A. leucosternon then, exploits food unavailable to other species but is restricted to poorer quality habitat by more dominant species. A. lineatus, on the other hand, must eat food that has to be defended against many other species, and it forms large colonies in which individual fish defend small territories containing thick algal mats. Like Jones (1968), these authors maintain that the several acanthurid species on a given reef coexist through resource partitioning and interspecific dominance hierarchies. Territorial defence in temperate species In general, few of the herbivorous fishes in temperate waters appear to defend feeding territories. Many temperate species seem to be roving browsers or grazers (see Table I) or relatively inactive and secretive species such as the TABLE I Names, feeding types and diets of herbivorous species in 14 shallow marine fish communities; species were included if stomach contents were more than 50% plant material by volume, mass or, in some cases, frequency of occurrence; if no quantitative data were available, herbivore designation was based on information from other sources; see Table VIII for numbers and proportions of herbivorous fishes in these communities Family—species

Feeding type

Main dietary items

France (Mediterranean, rocky littoral), Gibson (1968) Blenniidae Blennius sanguinolentus Grazer Filamentous and foliose algae California (rocky intertidal), Grossman (1986) Cottidae Clinocottus globiceps Browser Foliose green and red algae, sea anemones California (rocky intertidal), L.G.Allen & M.H.Horn, unpubl. data Stichaeidae Cebidichthys violaceus Browser Foliose green and red algae Xiphister mucosus Browser Foliose green and red algae New Zealand (rocky subtidal), Russell (1983) Aplodactylidae Aplodactylus arctidens Grazer Turf-forming red, brown and green algae Girellidae Browser/grazer Turf-forming red algae and large brown algae Girella tricuspidataa Kyphosidae Kyphosus sydneyanus Browser Large brown algae Odacidae

140

MICHAEL H.HORN

Family—species

Feeding type

Odax pullus Browser Pomacentridae Parma alboscapularis Browser/grazer Australia (rocky subtidal), Burchmore et al. (1985)b Acanthuridae Acanthurus xanthopterus Grazer Prionurus microlepidotus ? Aplodactylidae Crinodus lophodon Grazer Girellidae Girella elevata Browser/grazer Girella tricuspidata Browser/grazer Odacidae Olisthops cyanomelas Browser Pomacentridae Mechanichthys immaculatus Browser/grazer? Parma microlepis Browser/grazer? Parma oligolepis Browser/grazer? Parma polylepis Browser/grazer? Parma unifasciata Browser/grazer? Siganidae Browser Siganus spinus Family—species

Main dietary items Large brown algae Turf-forming red algae, foliose green algae, invertebrates

Algae Algae Algae Algae Algae Algae Algae Algae Algae Algae Algae Algae

Feeding type

California (rocky subtidal), Stephens & Zerba (1981) Girellidae Girella nigricans Browser/grazer Scorpididae Medialuna californiensis Browser/grazer California (kelp bed), Quast (19681) Girellidae Girella nigricans Browser/grazer Kyphosidae Hermosilla azurea Browser/grazer Scorpididae Medialuna californiensis Browser/grazer Mexico (rocky intertidal), Thomson & Lehner (1976) Girellidae Girella simplicidens Browser/grazer Kyphosidae

Main dietary items

Algae Algae

Red, green and brown algae; invertebrates Red, green and brown algae; invertebrates Red, green and brown algae; invertebrates

Algae, invertebrates

BIOLOGY OF MARINE HERBIVOROUS FISHES

Family—species

Feeding type

Main dietary items

Hermosilla azurea Browser/grazer Algae, invertebrates Mugilidae Mugil curema Grazer Diatoms, blue-green algae, detritus, mud and silt Pomacentridae Eupomacentrus rectifraenum Grazer Algae South Africa (rocky littoral), Berry et al. (1982) Acanthuridae Acanthurus triostegus Browser Filamentous algae Acanthurus lineatus Browser Filamentous algae Mugilidae Valamugil buchanani Grazer Detritus, diatoms Scorpididae Neoscorpis lithophilus Browser Foliose green and red algae Sparidae Sarpa salpa Browser Foliose green and red algae Kermadec Islands (rocky subtidal), Schiel, Kingsford & Choat (1986) Aplodactylidae Aplodactylus etheridgi Grazer Turf-forming red, brown and green algae Girellidae Browser/grazer Algae Girella cyaneaa Browser/grazer Algae Girella fimbriatusa Kyphosidae Kyphosus fuscus Browser Algae Pomacentridae Parma alboscapularis Grazer/browser Turf-forming red algae, foliose green algae, invertebrates Parma polylepis (=P.kermadecensis) Grazer/browser Algae, invertebrates Grazer Algae, detritus Stegastes fasciolatus Family—species

Feeding type

Hawaii (coral reef), Jones (1968) Acanthuridae Acanthurus achilles Browser Acanthurus glaucopareius Browser Acanthurus guttatus Browser Acanthurus leucopareius Browser Acanthurus nigrofuscus Browser Acanthurus sandvicensis Browser Acanthurus nigroris Browser/grazer Acanthurus dussumieri Grazer/browser Acanthurus mata Grazer

Main dietary items

Filamentous algae Filamentous algae Filamentous algae Filamentous algae Filamentous algae Filamentous algae Filamentous algae, diatoms and detritus Diatoms, detritus and filamentous algae Diatoms and detritus

141

142

MICHAEL H.HORN

Family—species

Feeding type

Main dietary items

Acanthurus olivaceus Grazer Diatoms and detritus Acanthurus xanthopterus Grazer Diatoms and detritus Ctenochaetus hawaiiensis Grazer Diatoms and detritus Ctenochaetus strigosus Grazer Diatoms and detritus Naso brevirostris Browser Foliose algae Naso lituratus Browser Foliose algae Naso unicornis Browser Foliose algae Zebrasoma flavescens Browser Filamentous algae Zebrasoma veliferum Browser Filamentous algae Hawaii (coral reef), Hobson (1974) Balistidae Melichthys niger Browser/grazer Foliose algae, coralline algae Blenniidae Cirripectus variolosus Grazer Filamentous algae and detritus Canthigasteridae (Tetraodontidae) Canthigaster amboinensis Grazer Coralline algae, filamentous algae Chaetodontidae (Pomacanthidae) Centropyge potteri Grazer Filamentous algae and detritus Kyphosidae Kyphosus cinerascens Browser Benthic algae, drifting algal fragments Monacanthidae (Balistidae) Cantherines sandwichiensis Grazer Filamentous and coralline algae Pomacentridae Abudefduf sindonis Grazer Algae, including diatoms, detritus Abudefduf sordidus Grazer Algae, including diatoms, detritus Pomacentrus jenkinsi Grazer Algae, including diatoms, detritus Scaridae Scarus rubroviolaceus Grazer Algal scrapings from rock surfaces, calcareous powder, sand Scarus sordidus Grazer Algal scrapings from dead coral, calcareous powder, sand Scarus taeniurus Grazer Algal scrapings from rock surfaces, calcareous powder, sand Puerto Rico/Virgin Islands (coral reef), Randall (1967) Acanthuridae Acanthurus bahianus Grazer Algae and detritus, seagrasses Acanthurus chirurgus Grazer Algae and detritus, seagrasses Browser Algae and detritus, seagrasses Acanthurus coeruleus Family—species Balistidae Melichthys niger Blenniidae Blennius cristatus

Feeding type

Main dietary items

Browser/grazer

Algae, invertebrates, seagrasses

Grazer

Algae and detritus

BIOLOGY OF MARINE HERBIVOROUS FISHES

Family—species

Feeding type

Main dietary items

Blennius marmoreus Grazer Algae and detritus, invertebrates Entomacrodus nigricans Grazer Algae and detritus Ophioblennius atlanticus Grazer Algae and detritus Chaetodontidae (Pomacanthidae) Centropyge argi Grazer/browser Algae and detritus Gobiidae Gnatholepis thompsoni Grazer Algae and detritus, invertebrates Coryphopterus glaucofraenum Grazer Algae and detritus, invertebrates Hemiramphidae Hemiramphus brasiliensis Browser Seagrasses, fishes Kyphosidae Kyphosus incisor Browser Brown algae Kyphosus sectatrix Browser Brown and red algae Monacanthidae (Balistidae) Alutera schoepfi Browser Seagrasses and algae Mugilidae Mugil curema Grazer Diatoms, blue-green algae, detritus, mud and silt Pomacentridae Abudefduf taurus Browser Brown, green and red algae Microspathodon chrysurus Browser/grazer Filamentous red and blue-green algae, silt Pomacentrus fuscus Grazer Algae and detritus, invertebrates, silt Scaridae Scarus coelestinus Grazer Algae, inorganic sediment Scarus croicensis Grazer Algae, inorganic sediment Scarus guacamaia Grazer Algae, seagrasses, inorganic sediment Scarus taeniopterus Grazer Algae, seagrasses, inorganic sediment Scarus vetula Grazer Algae, inorganic sediment Sparisoma aurofrenatum Grazer Algae, inorganic sediment Sparisoma chrysopterum Grazer Algae, seagrasses, inorganic sediment Sparisoma rubripinne Grazer Algae, seagrasses, inorganic sediment Sparisoma radians Browser Seagrasses, algae Sparisoma viride Grazer Algae, inorganic sediment Sparidae Archosargus rhomboidalis Browser/grazer Seagrasses, algae Diplodus caudimacula Grazer Algae, invertebrates, sand Marshall Islands (coral reef), Hiatt & Strasburg (1960) Acanthuridae Acanthurus achilles Browser Filamentous algae Acanthurus aliala Browser Filamentous and foliose algae Acanthurus gahhm Grazer Filamentous algae, sand

143

144

MICHAEL H.HORN

Family—species

Feeding type

Main dietary items

Acanthurus guttatus Acanthurus lineatus Acanthurus mata Acanthurus nigroris Acanthurus olivaceus Acanthurus triostegus

Browser/grazer Browser Grazer Browser/grazer Grazer Browser

Foliose and filamentous algae Filamentous algae Filamentous algae, sand Foliose and filamentous algae Calcareous, filamentous and foliose algae, calcareous powder Filamentous algae

Family—species

Feeding type

Acanthurus xanthopterus Grazer Ctenochaetus striatus Grazer Naso lituratus Browser Naso unicornis Browser Zebrasoma veliferum Browser Balistidae Rhinecanthus rectangulus Grazer Rhinecanthus aculeatus Grazer/browser Melichthys vidua Grazer Blenniidae Cirripectus variolosus Grazer Cirripectus sebae Grazer Exallias brevis Grazer Istiblennius coronatus Grazer Istiblennius paulus Grazer Canthigasteridae (Tetraodontidae) Canthigaster solandri Grazer Chaetodontidae Chaetodon ephippium Browser Chaetodon reticulatus Grazer Kyphosidae Kyphosus cinerascens Browser Mugilidae Crenimugil crenilabis Grazer Neomyxus chavtali Grazer Pomacanthidae Centropyge flavissimus Browser/grazer Pomacentridae Abudefduf amabilis Grazer/browser Abudefduf biocellatus Browser/grazer Abudefduf dicki Browser/grazer Abudefduf glaucus Grazer

Main dietary items Filamentous algae, sand Algal scrapings and filaments, calcareous powder Foliose and filamentous brown algae Foliose brown algae Filamentous and foliose algae Algal scrapings and filaments, calcareous powder Algal scrapings, foliose algae, coralline algae Filamentous and foliose algae, sand, calcareous powder Filamentous algae, detritus Filamentous algae, detritus Filamentous algae, detritus Filamentous algae, detritus, sand Algal scrapings and filaments, detritus, sand Foliose and filamentous algae, calcareous powder Coral polyps, filamentous algae Filamentous algae Foliose and filamentous algae Detritus, blue-green algae, diatoms Diatoms, desmids, filamentous algae Foliose and filamentous algae Filamentous and foliose algae, foraminiferans Foliose and filamentous algae, invertebrates, fish Filamentous algae, fish, detritus and sand Filamentous algae, invertebrates, fish

BIOLOGY OF MARINE HERBIVOROUS FISHES

Family—species

Feeding type

Main dietary items

Abudefduf lacrymatus Abudefduf saxatilis Abudefduf septemfasciatus Abudefduf sordidus Pomacentrus albofasciatus Pomacentrus jenkinsi Pomacentrus nigricans Pomacentrus vaiuli

Browser Browser Browser Grazer Grazer Browser/grazer Grazer Browser

Filamentous algae, foraminiferans, invertebrates Filamentous and foliose algae, invertebrates Foliose algae, invertebrates Algal scrapings and invertebrates Algal filaments and scrapings, sand Foliose and filamentous algae Filamentous and foliose algae, algal scrapings Filamentous algae, fish, invertebrates

Family—species

Feeding type

Main dietary items

Scaridae Cryptotomus spinidens Scarus bicolorc

Grazer Grazer

Scarus sordidusc

Grazer

Scarus spp.c (7 unidentified species)

Grazers

Filamentous algae, calcareous powder Coral polyps, filamentous algae, calcareous powder Coral polyps, filamentous algae, calcareous powder Coral polyps, filamentous algae, calcareous powder

Siganidae Siganus rostratus Tanzania (coral reef), Talbot (1965) Acanthuridae Acanthurus bariene Acanthurus bicommatus Acanthurus fuliginosus Acanthurus leucosternon Acanthurus lineolatus Ctenochaetus striatus Ctenochaetus strigosus Naso brevirostris Naso lituratus Zebrasoma scopas Zebrasoma veliferum Pomacanthidae Apolemichthys trimaculatus Centropyge bispinosus Scaridaec Calotomus spinidens Scarus aeruginosus Scarus africanus

Browser

Algal fronds and filaments,

Grazer? Grazer? Grazer? Grazer? Grazer? Grazer? Grazer? Browser? Browser? Browser? Browser?

Algae Algae Algae Algae Algae Algae Algae Algae Algae Algae Algae

Grazer? Grazer?

Algae Algae

Grazer? Grazer? Grazer?

Algae Algae Algae

145

146

MICHAEL H.HORN

Family—species

Feeding type

Main dietary items

Scarus bipallidus Scarus forsteri Scarus globiceps Scarus guttatus Scarus harid Scarus javanicus Scarus microrhinos Scarus niger Scarus pectoralis Scarus scaber Scarus sordidus Scarus vermiculatus Siganidae Siganus oramin Siganus stellatus

Grazer? Grazer? Grazer? Grazer? Grazer? Grazer? Grazer? Grazer? Grazer? Grazer? Grazer? Grazer?

Algae Algae Algae Algae Algae Algae Algae Algae Algae Algae Algae Algae

Browser? Browser?

Algae Algae

a

This species was placed in the Kyphosidae by the author (s). Designation of herbivorous species was aided by M.J.Kingsford (pers. comm.). c All scarids were classified as coral feeders by Talbot (1965); 9 scarids were classified as omnivores by Hiatt & Strasburg (1960). b

aplodactylids (Clements, 1985) and stichaeids (Ralston & Horn, 1986). Exceptions to this generality include mainly those species that belong to primarily tropical families such as the Pomacentridae and Blenniidae. As examples, the Australian damselfish Parma microlepis (Moran & Sale, 1977), the New Zealand damselfish Parma alboscapularis (Choat & Ayling, 1987) are territorial, and the Mediterranean blenny Parablennius sanguinolentus loosely defends a feeding territory (see Taborsky & Limberger, 1980). Another pomacentrid, Hypsypops rubicunda, which occurs in waters off southern California, maintains and defends a red algal turf nest but relies mainly on animal food (Quast, 1968; Clarke, 1970; Foster, 1972). Group foraging Foraging groups are common in tropical waters especially among members of the Acanthuridae and Scaridae (Ogden & Lobel, 1978). These groups may be made up of one to several species. If multispecific, the aggregations often consist of several herbivorous species and a few carnivorous ones that apparently consume the invertebrates and smaller fishes disturbed by the feeding activities of the foraging group (Ogden & Buckman, 1973; Hobson, 1974; Alevizon, 1976; Robertson et al., 1976). Although participation in these groups may confer any of several advantages on the individual species, the major benefit seems to be to overwhelm the defences of territorial damselfishes and thereby gain access to the algal patches maintained within damselfish territories. Robertson et al. (1976) and Foster (1985a,b) argue that territorial damselfishes, as abundant and aggressive defenders of space on tropical reefs, promote the widespread habit of group foraging among non-territorial herbivores. This behaviour may be likened to the pack hunting of carnivores (Foster, 1985a,b). The occurrence of aggregations seems to increase (and incidence of solitary individuals decrease) with increase in density of territorial herbivores (Barlow, 1974a; Doherty, 1983). The biting rate of individual participants increases with group size, apparently because individuals in large

BIOLOGY OF MARINE HERBIVOROUS FISHES

147

groups are attacked less frequently by damselfish (Foster, 1985a). A damselfish at first attacks the invaders, but soon the attack rate per fish in the aggregation declines and the damselfish waits passively for the group to move away. In the parrotfish Scarus croicensis schooling is a mechanism for circumventing the territoriality of competitors (Robertson et al., 1976). Some individuals hold permanent territories while others form feeding schools. Territorial parrotfish inhibit the feeding of non-territorial conspecifics, but both types of individuals are subordinate to the omnivorous, territorial damselfish Eupomacentrus planifrons. Nonterritorial Scarus croicensis in schools feed at higher rates and are attacked by territory holders less often than are non-schooling, non-territorial individuals. Non-schooling S. croicensis seem to make up that portion of the population unable to acquire feeding territories, probably because of the aggressiveness of Eupomacentrus planifrons. Schooling, therefore, helps maintain the coexistence of territorial and nonterritorial individuals. Group foraging as a means to obtain algal food inside damselfish territories is expected to be rare in temperate waters because herbivorous fishes in general, and herbivorous, territorial damselfishes in particular, are rare at temperate latitudes (see Quast, 1968; Choat & Ayling, 1987). Sites of such activity may, however, exist in southern Australia and New Zealand where territorial pomacentrids and girellids, which feed as roving browsers or grazers, co-occur (Burchmore et al., 1985; Choat & Ayling, 1987) and overlap somewhat in diet (Russell, 1983). Home ranges Home ranges of coral reef fishes appear to differ from feeding territories in at least three ways (Ogden & Lobel, 1978): (1) home ranges do not increase the productivity of the food resources even if stronger defences are applied; (2) home ranges are defended only against closely related species and possibly competitors for shelter or spawning sites; and (3) home range behaviour develops where the food resource is neither limiting nor widely used by other species. A variety of reef fishes are known to limit their movements to specific areas of the reef where their feeding is concentrated (Reese, 1973; Sale, 1977; Bouchon-Navaro & Harmelin-Vivien, 1981; Russ, 1984a). To some extent, fishes using home ranges are intermediate between species such as certain scarids and acanthurids that feed over relatively wide areas in mixed-species schools and species such as many pomacentrids and some acanthurids that actively defend feeding territories. Species using home ranges may belong to these same families or even the same species. Local-scale heterogeneity in reef structure may cause site-associated variations in feeding behaviour and territorial defence such that a fish may defend a territory at one site but only occupy a home range at another (see Choat & Bellwood, 1985). Little information is apparently available on home range use by temperate herbivorous fishes. The highly restricted areas used by the stichaeid Cebidichthys violaceus in a rocky intertidal habitat may indicate that this fish occupies a home range, but more data are required to test this possibility (Ralston & Horn, 1986). Comparative feeding behaviour of three herbivorous fishes Recent studies on a tropical parrotfish (Scaridae), a temperate odacid (Odacidae) and a temperate aplodactylid (Aplodactylidae) (Clements, 1985; Clements & Bellwood, 1988) provide an opportunity to compare the feeding behaviours and mechanisms of three distinctly different herbivores. Scarus rubroviolaceus, a tropical grazer This parrotfish and the odacid share the morphological features of (1) a fused dental structure, with the premaxillary and dentary bones forming a beak, (2) an opposable

148

MICHAEL H.HORN

pharyngeal jaw, and (3) the lack of a gastric stomach (Clements & Bellwood, 1988). The Scaridae and the Odacidae along with the Labridae form a monophyletic assemblage of pharyngognathous acanthopterygian teleosts (Liem & Greenwood, 1981). S. rubroviolaceus occurs from eastern Africa to Baja California and exhibits a typical parrotfish feeding behaviour of scraping algae from the surface of rocks and coral rubble (Rosenblatt & Hobson, 1969). S. rubroviolaceus is described (Clements & Bellwood, 1988) to feed in an oblique head-down position, scraping the surface of the turf-covered substratum. Each bite produces a pair of narrow parallel scrapes marked by dislodged algae, but substratum scarring occurs only occasionally. Bites, about 15–20 per minute, are grouped into short feeding bouts with little or no searching behaviour within or between bouts. The jaw morphology and feeding behaviour of S. rubroviolaceus indicate a strong scraping bite but a weak sucking action. The pharyngeal apparatus appears to be capable of a great deal of movement, and a powerful grinding action is produced by the broad, flat dentigerous surfaces of the pharyngeal bones. Once ingested, algae are triturated by the action of the pharyngeal jaws, The intestinal contents consist of finely ground algal material and large amounts of calcium carbonate particles, attesting to the grinding efficiency of the pharyngeal apparatus. The gut is modified with unusual intestinal sacculae, which may increase the holding capacity and also retention times by reducing laminar flow rates. Odax pullus, a temperature browser Odacids are temperate fishes endemic to Australia and New Zealand, and O. pullus is found only in New Zealand waters (Gomon & Paxton, 1985). This species displays a variety of feeding methods depending on the food type and appears to be a more selective feeder than the parrotfish; the main foods are primarily laminarian (i.e. Ecklonia radiata) and fucalean (i.e. Carpophyllum spp.) brown seaweeds (Clements, 1985; Clements & Bellwood, 1988). When the fish feeds on Ecklonia radiata, the oral surface is applied to the lamina and held there by opercular suction, and a disc of algal tissue is excised. The bite is usually taken from the tips and centres of the secondary laminae where the sori are located in the reproductive season. When feeding on Carpophyllum spp., Odax pullus removes the reproductive bunches and thalli with a bite and a sideways flick of the head. Bites, two or fewer per minute, are grouped into bouts, which are punctuated by searching behaviour. The feeding rates are highest during the first two hours after dawn. Like Scarus rubroviolaceus, Odax pullus is a strictly diurnal feeder and, like numerous tropical reef species (Fishelson, Montgomery & Myrberg, 1987) and some other temperate fishes, especially of tropical affinities (Ebeling & Bray, 1976), seeks shelter at night. After each feeding bout, the fish hangs in a tail-down position and pharyngeal mastication occurs, as evidenced by the raising and lowering of the hyoid apparatus. The jaw morphology and feeding behaviour of O. pullus indicate a relatively weak bite but a strong sucking action. The sharp, finely serrated oral jaws provide a cutting action that is enhanced by the narrow cutting edges of the dentigerous surfaces of the pharyngeal apparatus. Once ingested, algae are shredded into small pieces but left uncrushed as they pass into the intestine. No carbonate material is found in the gut because O. pullus feeds on large upright seaweeds and not on turf or encrusting algae. The simple intestine is shorter than would be expected for a herbivore and, overall, how O. pullus digests seaweeds is an intriguing question. Aplodactylus arctidens, a temperate browser Aplodactylids are temperate fishes found only in coastal waters of Peru, Chile, New Zealand and Australia (Nelson, 1984), and A. arctidens is confined to the latter two regions (Clements, 1985). A. arctidens is apparently a crepuscular species, preferring to feed during the early morning and late evening (Doak, 1978). The feeding behaviour of this species as described here follows that of Clements (1985). A. arctidens is a sluggish, negatively buoyant fish that feeds primarily on red turfing algae growing on rocky reefs. Its relatively weak jaws, weakly implanted jaw teeth and lack of opposable pharyngeal jaws limit this fish to a diet of relatively tender, low growing seaweeds. Larger fish, however, eat a higher proportion of tougher, foliose algae. Clements (1985) calls A. arctidens a non-

BIOLOGY OF MARINE HERBIVOROUS FISHES

149

selective grazer, more like the parrotfish than the highly selective odacid in this comparison. Unlike either the parrotfish or the odacid, this fish does not grind nor shred its food, but it does have a distinct, muscular stomach. The three fishes clearly must have different mechanisms for digesting algae, a topic to be addressed in the next section (p. 197). DIETS AND FOOD PREFERENCES The food habits of many herbivorous fishes indicate that collectively these fishes consume a great variety of the plant material in the sea (Table I). Although red and green macroalgae seem to predominate in the diets of the majority of herbivorous fishes, virtually all types of algae from diatoms to large kelps are eaten as well as several species of seagrasses. At first glance, the food habits of plant-eating fishes might seem to be merely representative of the widely reputed versatility and opportunism exhibited by fishes in general. Closer examination, however, reveals a much more complex, and still incompletely understood, picture of dietary patterns in these fishes. Herbivorous species are constrained morphologically, for example, by the relatively small gapes that characterise most of them (Ogden & Lobel, 1978) and by differences in their jaw teeth and pharyngeal apparatus (e.g. Clements & Bellwood, 1988). Although most herbivorous fishes are diurnally active (Ogden & Lobel, 1978), they may variously feed throughout the day (Earl, 1972), primarily during twilight hours (Doak, 1978), or reach peaks of feeding intensity early in the morning (Meekan, 1986) or late in the afternoon (Taborsky & Limberger, 1980). Some species, especially those in temperate waters (Horn, Murray & Edwards, 1982; Horn, 1983), often shift their diets with the season, and many species in both temperate and tropical latitudes show ontogenetic changes in their food habits (Montgomery, 1977; Barton, 1982; Horn, Murray & Edwards, 1982; Clements, 1985; Meekan, 1986). These latter species commonly begin life as carnivores then become herbivores as adults. Still another source of variation is that some herbivorous fishes have different diets in different parts of their geographic range (e.g. Odum, 1970; Collins, 1981), perhaps mainly a result of changes in food availability but possibly because of other factors as well. Also, a number of seaweed species, including seemingly palatable forms, are avoided by herbivorous fishes (Montgomery & Gerking, 1980). Relatively few studies have been completed in which: (1) the availability of the algal resource was known so that food selectivity in nature could be quantitatively determined, or (2) fishes were presented with an array of seaweeds, either in the field or laboratory, so that food preferences could be established. Emerging studies of herbivorous fishes, however, including those on functional morphology and digestive physiology and on their responses to seaweed defences, are beginning to shape a clearer understanding of the complexity in dietary patterns and to provide a better answer to the question of why do herbivores eat what they eat. The purpose of this portion of the food and feeding section is first to summarise information on the diets, food selectivity and dietary preferences of the herbivorous members of several tropical and temperate fish families and then to evaluate the factors that appear to influence diet choice in herbivorous fishes. A compilation of the diets and feeding modes of the herbivorous species encountered in several major studies of tropical and temperate fish communities is presented in Table I. Tropical families Acanthuridae Acanthurids (surgeonfishes) comprise one of the most abundant and diverse fish families on tropical reefs. Although a few species are zooplankton feeders, most are either grazing or browsing herbivores (Jones, 1968). In a study of 20 species of Hawaiian and Johnston Island acanthurids, Jones

150

MICHAEL H.HORN

(1968) found that two were planktivores, feeding on zooplankton, six were grazers of diatoms and detritus and 12 were browsers, three feeding on foliose brown algae and nine on smaller, filamentous red and, to a lesser extent, green algae. Five major algal divisions, Cyanophyta (blue-greens), Chrysophyta (diatoms), Chlorophyta (greens), Rhodophyta (reds) and Phaeophyta (browns) were represented in the diets of these 18 herbivorous surgeonfishes. Although they consumed a wide variety of algae, the 12 species of browsers still fed selectively on a rather small range of seaweeds. Of the 160 algal genera available to browsing acanthurids in Hawaii, Jones (1968) found only 40 in the fish stomachs he examined. If he omitted rarely occurring genera, the number was reduced to 32 or 20% of those species available in the habitat. More specifically, 2 of 24 (8%) available blue-greens, 9 of 27 (33%) greens, 15 of 97 (15%) reds and 6 of 16 (38%) browns were eaten. A further indication of dietary selectivity was that one of the browsers of large, foliose algae, Naso lituratus, apparently preferred to eat brown algae in the genus Pocockiella because the genus was found in all 37 fish examined and made up the entire stomach contents of 25 of these specimens. In a laboratory study of one of the above surgeonfishes, Acanthurus trio stegus sandvicensis (=A. sandvicensis), Randall (1961a) showed that this fish preferred the red alga Polysiphonia sp. and the green alga Enteromorpha sp. over a variety of other algae even though numerous other filamentous species and diatoms were found in stomach contents. Randall (1961a) believed that many of the algae, including the blue-greens, were consumed by the fish incidentally according to their abundance in the habitat. Recent field studies, including algal transplant experiments (Lewis, 1985), observations of tagged fishes (Wolf, 1985), preference tests (Lewis, 1986) and experiments to assess the effectiveness of algal secondary compounds in deterring herbivores (Paul, 1987; Hay et al., 1988a; Wylie & Paul, 1988) have all shown acanthurids to be selective feeders. Scaridae Scarids (parrotfishes) are abundant and diverse residents of tropical reefs (Hiatt & Strasburg, 1960; Randall, 1967; Hobson, 1974), and some species also occur in adjacent seagrass beds (Randall, 1967; Ogden, 1976). Although earlier studies (Al-Hussaini, 1947; Gohar & Latif, 1959; Hiatt & Strasburg, 1960; Talbot, 1965) considered parrotfishes to be primarily coral feeders, subsequent investigations (Randall, 1967; Hobson, 1974; Smith & Paulson, 1974) cast doubt on this conclusion, and these fishes are now classified as virtually exclusive herbivores (Russ, 1984b; Randall, 1986). With their fused jaw teeth and pharyngeal mill they are able to graze on a variety of algae on reef surfaces (Randall, 1967; Lobel, 1981) and to browse on seagrasses (Randall, 1967; Ogden 1976). Although most parrotfishes are grazers (Ogden, 1976), Sparisoma radians is a seagrass browser and a characteristic member of Caribbean seagrass communities (Thayer et al., 1984). Two relatively recent studies on S. radians constitute almost all the experimental work on parrotfish food preferences except for some more general field experiments on food choice in tropical herbivores that included parrotfishes (see below). The first of these studies showed that the seagrass Thalassia testudinum was the most abundant item in the field diet and, with epiphytes, the top-ranking plant species in laboratory preference tests (Lobel & Ogden, 1981). Nevertheless, fish fed a mixed diet including all seagrasses and algae in the preference hierarchy provided the greatest certainty of survival. Lobel & Ogden speculated that Sparisoma radians eats a variety of plants to maintain a balanced diet. In a subsequent study, Targett, Targett, Vrolijk & Ogden (1986) showed that the three least preferred food genera in Ogden & Lobel’s experiments, the green algae Caulerpa, Halimeda and Penicillis, contain terpenoid secondary compounds that deter feeding by Sparisoma radians. Coating Thalassia blades with crude extracts of Halimeda incrassata reduced the fish’s bite rate on, and consumption of, the seagrass to the same level as that for H. incrassata. The results of this study indicate that secondary metabolites can play an important role in food selection by herbivorous fishes. Feeding observations corroborate this dietary selectivity in that Sparisoma radians prefers

BIOLOGY OF MARINE HERBIVOROUS FISHES

151

the epiphytised tips of Thalassia, which are lower in protein but also lower in phenolic compounds than younger, basal leaves (Thayer et al., 1984). Recent field studies including algal transplant experiments (Lewis, 1985, 1986), observations on tagged fishes (Wolf, 1985) and experiments to assess the role of algal secondary compounds in deterring herbivores (Hay et al., 1988a) have all shown that scarids are selective feeders, but often with different responses than acanthurids. Siganidae Siganids (rabbitfishes) are Indo-Pacific herbivores that feed as grazers on filamentous algae or as browsers on larger, foliose seaweeds (Hiatt & Strasburg, 1960; Lundberg & Lipkin, 1979). They are locally abundant around tropical reefs (Tsuda & Bryan, 1973), and some species are harvested (Bryan, 1975) and cultured (Bryan, 1975; Avila, 1986) for human consumption. Bryan (1975) reported that the foliose alga Gelidiopsis intricata and the filamentous algae Boodlea composita, Sphacelaria tribuloides, and Centroceras clavulatum were the most abundant items in the stomachs of Siganus spinus captured in waters around Guam. In a study of three siganids in the northern Red Sea, Lundberg & Lipkin (1979) found that these fishes ate green, red and brown algae in different proportions and also consumed small quantities of blue-green algae and seagrasses. Their conclusion that all three species fed selectively was based on comparisons of stomach contents with availability of the seaweeds and seagrasses in the habitat. Several other studies have shown that siganids feed selectively in nature or prefer certain algae in the laboratory. Tsuda & Bryan (1973) observed that the green algae Enteromorpha, Caulerpa, Boodlea, and Cladophoropsis were the first seaweeds to disappear from the reef flats of Guam following the invasion of huge schools of Siganus restrains and S. spinus. Their preference experiments showed that the algae chosen were also grazed heavily in the field. Other preference studies on siganids (Westernhagen, 1974; Bryan, 1975) have shown that Enteromorpha is a highly preferred algal type. Enteromorpha, however, was only of intermediate importance in the siganid stomachs examined by Bryan (1975). This discrepancy could be a result of differential digestibility or local availability of the alga at the time the fish were captured (Ogden & Lobel, 1978). The latter notion is corroborated by the importance of Enteromorpha in the diets of two rabbitfishes in a Red Sea location where the alga was abundant in the habitat (Lundberg & Lipkin, 1979). In multiple choice feeding trials with 50–60 seaweed species, juvenile Siganus rostratus and S. spinus rejected 26 species whereas only 12 were never consumed by adults (Tsuda & Bryan, 1975). Bryan (1975) claimed that the adults but not the juveniles were physically able to ingest the larger, tougher algal species and that this limitation was a factor in the high mortalities of starving juveniles recorded by Tsuda & Bryan (1975) after an invasion of the juveniles caused the depletion of filamentous algae from the reef flats of Guam. Siganids were among the coral reef herbivores whose foraging was often deterred by secondary compounds extracted from chemically defended seaweeds and applied to the surface of otherwise palatable and preferred Enteromorpha plants in field experiments on Guam (Paul, 1987; Paul et al., 1987; Paul & Van Alstyne, 1988). It is of interest that elatol, a terpenoid compound extracted from the red alga Laurencia obtusa, has been shown in field experiment to deter parrotfish feeding in the Caribbean (Hay, Fenical & Gustafson, 1987b). L. obtusa, however, along with two other species of Laurencia pre-dominated in the diets of Siganus rivulatus in the Red Sea (Lundberg & Lipkin, 1979). These findings suggest that different herbivores react differently to seaweed secondary chemicals, an increasingly recognised source of variation in recent field experiments. Pomacentridae Pomacentrids (damselfishes) are widespread, conspicuous and ecologically important fishes on tropical reefs around the world (Emery & Thresher, 1980). They are solitary or form small to dense aggregations depending upon the species, and they range in food habits from strict carnivores to omnivores to almost exclusive herbivores, with most maintaining an omnivorous diet (Hiatt & Strasburg, 1960). The territorial species, which are usually omnivores or herbivores, exert profound and well-

152

MICHAEL H.HORN

documented effects on the social organisation of other herbivorous reef fishes and on the structure of algal communities in tropical waters. Some pomacentrids also occur in warm temperate waters where a few are herbivores (Russell, 1983) and others are omnivores (Quast, 1968; Moran & Sale, 1977) or planktivores (Quast, 1968). The herbivorous damselfishes are reef-dwellers and feed primarily on turfing red and green algae. Much attention has been given to their diets because of the impacts their defence of a feeding territory have on other fish herbivores and on the algal populations within their territories. Two studies that focused on feeding ecology in particular serve to illustrate the diets and food selectivity in herbivorous pomacentrids. Montgomery (1980b), despite acknowledged small sample sizes, found that two co-occurring species in the southern Gulf of California, Eupomacentrus rectifraenum and Micro-spathodon dorsalis, have somewhat similar diets in that both consume highly productive, foliose red and green algae but ignore brown and calcareous seaweeds. The two species, however, differ in feeding behaviour. The smaller Eupomacentrus rectifraenum feeds intensively and selectively on a defended algal mat, choosing primarily the green alga Ulva and secondarily red algae, especially Gracilaria. Montgomery argued that the recruitment rate and productivity of Ulva must be high to persist under the heavy grazing pressure. In contrast, Microspathodon dorsalis feeds less intensively and non-selectively on its defended algal mat, foraging without close inspection of the substratum by tearing a large circular patch of algae from the mat. Its diet and the mat inside its territory are dominated by the red alga Polysiphonia, not because this alga is selected, but presumably because it is the only alga with a growth rate sufficient to persist under the influence of heavy, nonselective grazing pressure. Thus, maintenance of the species composition in the algal mats of both fishes requires high primary productivity and rapid growth rates of the dominant species. In the second study, Lassuy (1984) showed that Stegastes lividus in Guam shifted from omnivory as juveniles to herbivory as adults. Both ontogenetic stages were classed as browsers. Red algae, particularly those in the genera Polysiphonia, Gelidiopsis and Ceramium, were the main plant components of the diet of all size classes. Enteromorpha and Cladophora were the most common green algae in the stomachs examined. Brown algae, mainly Ectocarpus and Sphacelaria, and blue-green algae were recorded in smaller quantities; the proportion of the latter type in the diets decreased with size of the fish. Tropical-temperate families Kyphosidae Kyphosids (rudderfishes) and their relatives do not classify easily into either phylogenetic arrangements or latitudinal categories. This family is sometimes considered to be made up of three subfamilies, Kyphosinae, Girellinae and Scorpidinae, but in other cases these taxa are raised to family status (Nelson, 1984). The three taxa are treated here as separate families. All three families contain herbivorous species, but discussion is limited in this section to the kyphosids and girellids because, unlike the scorpidids, essentially all their members are herbivorous and because their herbivorous habits are better known than those of the plant-eating scorpidids. Kyphosids occur in both tropical and temperate waters (Nelson, 1984) and are therefore considered in this tropical-temperate subsection, whereas girellids are temperate species and thus included in that subsection. Kyphosids, in particular the genus Kyphosus, occur worldwide in tropical to temperate (mainly warm temperate) seas (see Nelson, 1984). They browse in small to large schools on red and brown algae (Hiatt & Strasburg, 1960; Randall, 1967; Hobson, 1974), and are also known to feed near the surface on drifting seaweeds in tropical (Randall, 1967; Hobson, 1974; Littler, Taylor & Littler, 1983b) and temperate (Rimmer, 1986) waters. They consume cleanly-bitten pieces of algae and pass the material with little or no trituration into the stomach (Russell, 1983; Rimmer, 1986). Kyphosus sydneyanus, a south temperate

BIOLOGY OF MARINE HERBIVOROUS FISHES

153

species, feeds mainly on brown algae in both New Zealand (Russell, 1983) and Australian (Rimmer & Wiebe, 1987) waters, whereas K. cornelii, another Australian species, feeds chiefly on red algae (Rimmer & Wiebe, 1987). Several species of Kyphosus seem to prefer brown algae, a notion reinforced in an anecdotal way by the practice (see Titcomb, 1972) of some Hawaiian fishers who bait their hooks with brown seaweeds, especially Sargassum, to catch Hawaiian rudderfishes as well as the acanthurid, Naso unicornis, another brown algae eater. The food and feeding ecology of Hermosilla azurea, a warm temperate species occurring in waters off southern California and northern Mexico, is poorly known but appears to consume a variety of red, green and brown seaweeds (Quast, 1968). Kyphosids have complex digestive tracts, and the two south temperate species mentioned above have recently been shown to have a microbial fermentation system in their hindguts (Rimmer & Wiebe, 1987). Sparidae Sparids (porgies) are active fishes commonly associated with reefs, seagrass beds and other coastal benthic habitats in tropical to warm temperate regions around the world (Smith & Smith, 1986). As a group, sparids have extremely varied diets, and many undergo complex, ontogenetic changes in food habits that are often correlated with age-related changes in dentition and gut morphology (Christensen, 1978; Stoner, 1980; Stoner & Livingston, 1984). Porgies are of interest in the context of this review because many are omnivorous, consuming both animal and plant material during their lives. The use of plant material varies with age, season and location but is usually greater in adults (Adams, 1976; Christensen, 1978; Stoner, 1980; Livingston, 1982; Gerking, 1984; Ogburn, 1984; Stoner & Livingston, 1984). Sparids variously graze on filamentous algae or browse on larger, foliose algae or seagrasses (Randall, 1967; Christensen, 1978; Stoner, 1980; Gerking, 1984; Ogburn, 1984; Stoner & Livingston, 1984). In Lagodon rhomboides (Stoner, 1980) and Sarpa salpa (Christensen, 1978) canine teeth are replaced in older fish by incisors, which are apparently better suited for taking bites of the plant material that becomes a larger part of the diet with age in these fishes. The powerful molariform teeth in the sides of the jaws of many sparids may not be used to crush algal cell walls because Ogburn (1984) found only whole and completely intact algae in the stomachs of Archosargus probatocephalus. Sarpa salpa, a species found in the Mediterranean and along western and southern African coasts (Smith & Smith, 1986), is probably as an adult as close to being a complete herbivore as any of the sparids, even though animals still supplement the red and green algae that make up the bulk of the diet. Not only does the gut lengthen and the teeth become incisiform with age in S.salpa (Christensen, 1978), but the fish has been shown to have relatively high assimilation efficiencies when fed the green alga Ulva lactuca (Gerking, 1984). Nevertheless, the fish lost weight on a strict Ulva diet in the laboratory (Gerking, 1984), as did another sparid, Archosargus rhomboidalis, when fed only a single algal species (Vaughan, 1978). These results suggest that the two species require animal food and perhaps accurately reflect the omnivorous habits of most sparids. Recent field and laboratory experiments have shown that the feeding selectivity of two sparids, Dlplodus holbrooki and Lagodon rhomboides, is influenced by unpalatable (i.e. not consumed, probably because of taste or texture) seaweeds and seaweed secondary compounds. Both fishes were found to feed at significantly higher rates on the palatable red algae Hypnea and Spyridia occurring alone than on these algae attached to the unpalatable brown seaweed Sargassum (Hay, 1986) and to be deterred from feeding by a secondary chemical extracted from Dictyota dichotoma, a brown alga little preferred by these two fishes (Hay, Duffy, Pfister & Fenical, 1987a; Hay, Renaud & Fenical, 1988b).

154

MICHAEL H.HORN

Temperate families The food and feeding behaviour of herbivorous members of two south temperate fish families, the Aplodactylidae and the Odacidae, were described in an earlier part of this section of the review to illustrate nonselective grazing and selective browsing and are not included here. Girellidae Girellids (nibblers) are widespread in the warm temperate waters of the north Pacific and south Pacific (Nelson, 1984) where they feed in small to large schools as grazers or browsers on rocky substrata or in kelp bed habitats (Quast, 1968; Russell, 1983). The close-set, cuspate incisors of girellids are movable and seem to allow these fishes to be efficient grazers of rock surfaces (Norris & Prescott, 1959), perhaps more so than the related kyphosids, which have fixed but otherwise similar teeth (Thomson, Findley & Kerstitch, 1979). Girellids, however, are also able to browse on foliose algae including kelps, for they remove clean bites from a variety of seaweeds and pass the material with little or no trituration into the stomach (Bell, Burchmore & Pollard, 1980; Russell, 1983). Most girellids consume at least some invertebrate prey, either intentionally or incidentally, but the bulk of the diet, at least in adults, is usually made up of macroalgae (Williams & Williams, 1955; Quast, 1968; Bell et al., 1980; Russell, 1983). Nevertheless, the mixed diets of girellids led to the impression (Williams & Williams, 1955; Quast, 1968) that these fishes gained most of their nutrition from animal material despite the large amount of algae in their guts. Only recently has it been demonstrated (Anderson, 1987) that a girellid, the Australian Girella tricuspidata, can assimilate nutrients, including those in the cell wall, from algal material. Much remains to be learned about feeding and digestion in girellids and the related kyphosids and scorpidids, especially in north temperate waters. Stichaeidae Stichaeids (pricklebacks) are elongate, bottom-dwelling fishes that occur in cold temperate, primarily inshore waters of the Northern Hemisphere (Eschmeyer, Herald & Hammann, 1983). Of the 23 stichaeid species recognised for the eastern North Pacific (Eschmeyer et al., 1983), only two, Cebidichthys violaceus and Xiphister mucosus, are considered to be herbivores (Barton, 1982; Horn, Murray & Edwards, 1982). Once these two species shift from carnivory to herbivory (as small juveniles of about 45 mm standard length), they feed selectively in the rocky intertidal zone on red and green algae and avoid encrusting, calcareous and brown seaweeds (Horn et al., 1982). Laboratory experiments have shown that Cebidichthys violaceus prefers three species of red algae that have the highest protein contents among the eight species making up the bulk of the diet. This result combined with rather consistently high assimilation efficiencies shows that the preferred species provide the greatest amount of protein per bite to the fish (Horn & Neighbors, 1984). These three algae, however, are annuals and available mainly during the summer (Horn et al., 1982). Diet selection from a seasonally changing array of seaweeds in the habitat results in both C.violaceus and Xiphister mucosus consuming carbohydrate-rich, protein-poor algae in the winter and protein-rich, carbohydrate-poor algae during the rest of the year (Horn, Neighbors & Murray, 1986). The small amounts of animal material in the gut of fish beyond the dietary transition phase (Montgomery, 1977; Horn et al., 1982; Miller & Marshall, 1987), the highly acidic stomach that apparently is effective in releasing nutrients from algal cells (Urquhart, 1984) and the assimilation of radioactively labelled algal material (Horn, Neighbors, Rosenberg & Murray, 1985) demonstrate that Cebidichthys violaceus is a herbivore. Although less intensively studied, Xiphister mucosus seems to be equally dependent on algal foods. These two stichaeids appear to be among the very few year-round herbivorous fishes in cold temperate waters. They may be matched in diet and thermal environment (approximately 8–15°C annual range of sea surface temperatures) only by the aplodactylids and odacids in southern New Zealand waters (see Ayling & Cox, 1982).

BIOLOGY OF MARINE HERBIVOROUS FISHES

155

FACTORS AFFECTING DIET SELECTION What are the factors that account for the food selectivity seen among herbivorous fishes and are these predictive for unstudied species and plant-eating fishes in general? The question is difficult and complex even when stated in the simpler terms of why herbivorous fishes eat what they eat. Morphological and behavioural constraints (e.g. mouth size, tooth type, swimming mode and buoyancy) certainly limit the dietary range. For instance, the blades of a large kelp may be too large or too tough for a fish to bite effectively or too far into the water column for a negatively buoyant, benthic fish to acquire safely or economically. Also, availability of potential dietary items may vary in time and space because of, for example, abiotic factors, algal life history patterns or the feeding actions of other herbivores. This variation obviously influences food habits: absent plants are not eaten, and very rare ones usually contribute little to the total diet. At the same time, however, extremely abundant seaweeds do not necessarily predominate in herbivorous fish diets. Although fish morphology and food availability are important influences on fish herbivore diets, the selectivity exhibited by plant-eating fishes seems to be governed by more than just these factors. For example, as many as 15 or 20 or even dozens of algal species have been recorded from the stomach contents of a herbivorous fish such as Cebidichthys violaceus (e.g. Horn et al., 1982; Miller & Marshall, 1987), but only a few species make up the bulk of the diet, and these are not consumed in proportion to their abundance in the habitat (Horn et al., 1982). Moreover, only two or three species of brown algae constitute almost the entire diet of Odax pullus even though many more seaweeds are available in the habitat (Russell, 1983; Clements, 1985; Meekan, 1986; Clements & Bellwood, 1988). Clearly, selectivity is practised by herbivorous fishes. But what drives this selectivity besides the feeding abilities of the fishes and the availability of the plants? Are diets more the result of positive or of negative selection (i.e. avoidance)? The answers to these questions are explored below in the context of other factors that may influence diet selection. Emphasis is placed on the chemical and morphological features of the seaweeds and seagrasses that are potential food sources for herbivorous fishes. Food quality Herbivorous fishes might be expected to choose seaweeds high in nutritional quality, i.e., high in energy and especially protein, given the low levels of this nutrient in macroalgae. The cues that would allow fishes to make such choices remain obscure (see Stephens & Krebs, 1986), but if food quality as defined here were correlated with other features of the algae such as colour, texture, toughness or taste then fishes could have evolved to make such an association. In any case, is nutritional quality predictive of food choice regardless of the cues? The answer seems to be yes, at least in part. Optimal foraging theory including models of optimal diets is represented by a large and growing body of literature (Schoener, 1971; Pyke, Pulliam & Charnov, 1977; Hughes, 1980; Pyke, 1984) but remains controversial (Gray, 1987; Pierce & Ollason, 1987; Schoener, 1987; Stearns & Schmid-Hempel, 1987). This body of thought and conflict has stimulated research on the diet and feeding behaviour of a variety of animals. The basic assumption of foraging theory is that energy is the currency to be optimised leading to the energy maximisation premise (Townsend & Hughes, 1981). Most tests of the model, including the optimal diet component, have been consistent with the theory (Schoener, 1987); the theory was, however, developed with carnivores in mind (Stenseth & Hansson, 1979), and few studies have been done on marine herbivores (Hughes, 1980). Although Hughes (1980) concluded that the diets of marine herbivorous browsers largely meet the expectations of optimal diet theory, he mentioned few studies, and none was on fishes. Hughes

156

MICHAEL H.HORN

further concluded that foraging behaviour could not be adequately understood solely on the basis of the energy maximisation premise. In one of the few studies to examine the diets of herbivorous fishes in the context of optimal diet theory, Horn (1983) found the diets of the stichaeid fishes Cebidichthys violaceus and Xiphister mucosus to be partially consistent with the three most important predictions (Schluter, 1981) of the model. (1) The prediction that at high food densities a forager should concentrate solely on the energetically most valuable items was incompletely met by these two fish species. Cebidichthys violaceus and Xiphister mucosus increased their consumption of energy-rich annual seaweeds during periods (summer and fall) of high food abundance, but still continued to take a mixed diet. (2) The prediction that abundance of lower-valued foods does not determine their inclusion in the diet was largely upheld by the feeding habits of these two intertidal fishes. The probability of an item being consumed apparently depends upon its abundance as well as its chemical composition. (3) The prediction that foragers will generalise as food abundance declines was largely met by the two fishes because their diets broadened noticeably during periods (especially winter) of reduced food supply. An overriding outcome of this and other studies indicates that optimal diet models cannot be based solely on energy maximisation but should also include nutrient constraints in order to account more completely for the seasonally fluctuating, mixed diets of these fishes and other generalist herbivores. Foliage-eating herbivores (generalists) face nutrient constraints because foliage often varies greatly in abundance and in chemical composition, both within species on a seasonal basis and between species (Westoby 1974, 1978). According to Westoby (1974, 1978) variety in the diets of such generalist herbivores might best be explained by sampling and, hence, diet mixing by the animal in order to keep pace with seasonal changes and thereby maintain a nutritionally adequate diet. The food preferences and seasonal diets of herbivorous stichaeids are relevant in the context of diet mixing. The seaweed species most highly preferred by Cebidichthys violaceus and Xiphister mucosus are the annual red algae Microcladia coulteri, Porphyra perforata and Smithora naiadum (Horn, et al., 1982), which have the highest protein contents of the seaweeds in the diets of these fishes (Horn & Neighbors, 1984). Thus, these two fishes prefer the seaweeds that yield the most protein per bite. Furthermore, in the winter when these preferred annuals are scarce or absent, Cebidichthys violaceus and Xiphister mucosus feed selectively on those perennial seaweeds, especially the red alga Iridaea cordata var. cordata (as I. flaccida), that provide them with greater carbohydrate and, therefore, greater energy return than would the other algae contributing to the bulk of the diet (Horn et al., 1986). These results mean that the two stichaeids consume carbohydrate-rich, protein-poor algae in the winter and protein-rich, carbohydrate-poor algae during the rest of the year. Thus, contrary to an expectation of optimal diet theory, energy intake was increased by selectivity only during the winter, whereas protein intake was increased at the expense of energy intake at all other times of the year (Horn et al., 1986). The recently documented dietary shift with season in a tropical surgeonfish, Acanthurus nigrofuscus, appears to be tied to the nutritional requirements of the fish during the reproductive cycle (Fishelson, Montgomery & Myrberg, 1987) and may be another example in support of Westoby’s (1974, 1978) proposition of seasonal diet mixing by herbivores to meet nutritional requirements. At this point, it is pertinent to describe the use of protein (P) and energy (E) contents together in ratio form (the P/E ratio) as an indicator of food quality and, in turn, of fish growth. In a series of publications on detritivorous and herbivorous freshwater fishes in the family Cichlidae, Bowen (1979, 1982, 1984b) has demonstrated the application of the P/E ratio. The P/E ratio as used by Bowen is defined as the amount of assimilable protein per unit of assimilable energy. P/E ratios required for maintenance and for maximum growth appear (Russell-Hunter, 1970) to be remarkably uniform across a wide range of animal taxa. Bowen (1984b) cited these ratios as being about 4 mg digestible protein per kJ of digestible energy for maintenance and 20– 30 mg of protein per kJ for maximum growth (Fig 1). He found growth of goldfish (Carassius auratus)

BIOLOGY OF MARINE HERBIVOROUS FISHES

157

Fig1.—Predicted bimonthly growth rates of the stichaeid fishes Cebidichthys violaceus (months listed above curve) and Xiphister mucosus (months listed below curve) as a function of the amount of assimilable protein relative to the amount of assimilable energy (P/E ratio) obtained from consumption of eight species of algae; theoretical growth curve for different P/E values from Bowen (1982); figure slightly modified from Horn, Neighbors & Murray (1986).

between maintenance and maximum growth levels to be directly proportional to protein concentration in the diet (Bowen, 1984b). Growth rates predicted by the Bowen (1982) model for Cebidichthys violaceus and Xiphister mucosus (Fig 1) based on P/E ratios calculated from assimilation of protein and energy by the fishes from major dietary species are, as expected, lowest and near maintenance levels during the winter and increase but do not reach maximum growth levels at other times of the year (Horn et al., 1986). Such growth predictions can provide the basis for actual growth experiments using diets with modified P/E ratios. Comparison of the potential food quality among different algal groups offers important but somewhat limited clues to dietary selectivity in herbivorous fishes. Montgomery & Gerking (1980) analysed the ash, energy and nutrient contents of 16 species to assess the food quality of fleshy red, green, brown and calcareous red algae. On the basis of ash, calories, total protein, and total lipid contents, they concluded first that fleshy algae should be superior to calcareous algae as foods for fishes and secondly that green algae should be superior to brown algae and brown algae superior to red algae. When Montgomery & Gerking considered the digestibility of storage and extracellular carbohydrates, they predicted that green and red algae should be superior to brown algae as food for fishes. Montgomery & Gerking then cited two Gulf of California damselfishes, Eupomacentrus rectifraenum and Microspathodon dorsalis, as examples of herbivores that feed as predicted by algal biochemistry in that they eat red and green algae, assimilate nutrients from both types and ignore brown and calcareous algae. As mentioned above, most herbivorous fishes probably do seem to eat mainly red and green algae, and to this extent, these predictions are valuable. Caution, however, is required in expanding the model because of within-group variation in chemical composition of red and green algae (see Horn et al., 1986). Moreover, several different types of herbivorous fishes, including kyphosids, odacids and certain acanthurids and siganids, feed selectively on brown algae, which also exhibit (Steinberg, 1985; Ragan & Glombitza, 1986; Estes & Steinberg, 1988) large withingroup variation in chemical composition, especially secondary metabolites. The discussion of food quality and diet prediction to this point indicates that energy content alone does not account for herbivore diets and that energy and nutrient contents together are better predictors but still deficient. It is important to emphasise that data showing that stichaeids prefer the seaweeds of highest protein content (Horn & Neighbors, 1984) and that these fishes feed selectively for greatest protein or carbohydrate return (Horn et al., 1986) distinguish only among dietary species. Non-dietary seaweeds were not studied. So the question is not just which species are preferred among a larger set that constitutes the total diet, but also which species are avoided and why or under what circumstances. The answers require greater understanding of seaweed defences, a subject that has received increasing attention in recent years, and of fish digestive mechanisms, a subject that has yet to receive adequate study. These topics are taken up in sequence below. Seaweed defences Traits of seaweeds that may deter herbivores include low digestibility, calcified or otherwise tough (leathery or rubbery) thalli and toxicity resulting from the synthesis and sequestering of secondary compounds. The evidence that these characteristics defend against marine herbivores, especially fishes, is summarised below.

158

MICHAEL H.HORN

Low digestibility Differential digestibility of carbohydrates, which comprise most of the bulk and available energy in macroalgae, has been cited (Montgomery & Gerking, 1980) as a factor influencing, or at least correlated with, dietary patterns in herbivorous fishes. Fishes in general have carbohydrases capable of breaking alpha-linked but not beta-linked polymers of monosaccharides (Fange & Grove, 1979). All plants contain beta-linked polymers (cellulose and structurally similar compounds) in their cell walls (Mackie & Preston, 1974), and presumably, therefore, these compounds cannot be digested enzymatically by fishes. A major difference among plant groups occurs in the linkages of their polysaccharide storage products. Green and red algae and seagrasses produce alpha-linked storage compounds (Craigie, 1974). Thus, digestibility apparently does not explain why, for example, the stichaeids Cebidichthys violaceus and Xiphister mucosus do not consume (Horn et al., 1982) red algae such as Endocladia muricata or Prionitis lanceolata and the seagrass Phyllospadix scouleri. Brown seaweeds, however, contain beta-linked polysaccharides (Craigie, 1974), so digestibility (or lack of it) may help to explain why these algae are avoided by the stichaeids (Horn et al., 1982) and also by certain pomacentrids (Montgomery & Gerking, 1980). In support of the apparent indigestibility of brown algae, Cebidichthys violaceus did not assimilate carbon from either Macrocystis integrifolia or Fucus distichus when force-fed these two brown seaweeds in the laboratory (Horn et al., 1985). As has already been mentioned, however, a variety of fishes such as certain acanthurids (Jones, 1968), siganids (Lundberg & Lipkin, 1979), kyphosids (Russell, 1983; Rimmer & Wiebe, 1987) and odacids (Russell, 1983; Clements, 1985; Meekan, 1986; Clements & Bellwood, 1988) regularly eat brown seaweeds as a major portion of their diets and almost certainly are able to digest these algae. There are distinct differences in digestive capabilities among herbivorous fishes, therefore, and further study should be undertaken on the digestion of various algal carbohydrates by herbivorous fishes. Calcification and toughness Calcified thalli appear to be effective in deterring herbivorous invertebrates and fishes (Paine & Vadas, 1969; Littler, 1976; Montgomery & Gerking, 1980; Lewis, 1985; Paul & Hay, 1986). The strong negative correlations between ash and both energy and nutrient content in calcareous seaweeds suggest that avoidance of these species would be of selective value for most herbivorous fishes (Montgomery & Gerking, 1980). The diets of only a few herbivorous fishes appear to contain substantial amounts of calcareous algae (see Table I, p. 173). Corallina officinalis, a heavily calcified red alga, fits the pattern of avoidance in that it is not eaten by Cebidichthys violaceus or Xiphister mucosus (Horn, Murray & Edwards, 1982). Steneck (1983) has argued that diversity of herbivores and the intensity of their grazing activities have escalated over geological time and that the conspicuousness and frequent dominance of coralline algae on hard substrata in shallow marine communities is due in part to their ability to withstand and adapt to deep grazing (i.e. excavating) herbivores such as parrotfishes. Toughness refers to the resistance to breaking or tearing of the thallus or leaf (if on a seagrass) when pulled or bitten. It is a morphological characteristic of seaweeds and seagrasses that can deter feeding by marine herbivores such as fishes and sea urbins (Littler, Taylor & Littler, 1983) and molluscs (Steneck & Watling, 1982). Toughness alone, however, is frequently not sufficient to explain the dietary selectivity of herbivorous fishes but is a trait that often appears to be part of a combination of morphological and chemical defences present in seaweeds (see p. 197). Secondary compounds Many marine plants produce secondary metabolites, and increasing numbers of these compounds are now known to have a role in deterring herbivorous invertebrates and fishes (see Ragan & Glombitza, 1986; Hay & Fenical, 1988). Deterrence of herbivorous fishes by a variety of these chemicals, either as extracts or purified compounds, has been demonstrated in numerous recent field or laboratory studies (Paul, Sun & Fenical, 1982; Paul & Fenical, 1983; Sun, Paul & Fenical, 1983; Hay, 1986; Paul & Hay, 1986; Targett et al., 1986; Hay et al., 1987a; Hay, Fenical & Gustafson, 1987b; Paul, 1987; Paul et al., 1987; Hay et al., 1988a; Hay, Renaud & Fenical, 1988b; Paul & Van Alstyne, 1988;

BIOLOGY OF MARINE HERBIVOROUS FISHES

159

Wylie & Paul, 1988). Although the incidence of defensive chemicals appears to be more common in seaweeds of herbivore-rich tropical and subtropical waters (Hay & Fenical, 1988), some temperate algae produce secondary metabolites that deter herbivorous invertebrates (Steinberg, 1984, 1985) and fishes (Hay, 1986; Hay et al., 1987a). Brown seaweeds in north temperate seas have been shown (Steinberg, 1985) to produce phlorotannins in high concentrations (order Fucales) or in low concentrations (order Laminariales). Fucalean and other brown seaweeds in tropical habitats all appear to be low in phlorotannins (Steinberg, 1986). Interestingly, common species of fucoid algae in temperate Australia and New Zealand appear to have phlorotannin concentrations two to three times those of phlorotannin-rich species in California waters (Estes & Steinberg, 1988; Steinberg, 1988). This striking difference intensifies the mystery of how the southern hemisphere odacid and kyphosid fishes, whose diets often are made up mostly of these fucoids, tolerate and digest such seemingly heavily defended algae. In California, the girellid Girella nigricans and the scorpidid Medialuna californiensis are known (Quast, 1968; Harris, Ebeling, Laur & Rowley, 1984) to eat varying quantities of the phlorotannin-poor laminarian kelps, but no fish in the northern hemisphere is known to consume fucoid algae. The rapid progress that has been made in collecting, handling, purifying and applying the often unstable secondary compounds from seaweeds (Norris & Fenical, 1985) has fuelled the recent surge of studies on the deterrent effects of these chemicals on fishes and other marine herbivores. An especially useful and rewarding technique has been to coat pieces of a palatable seaweed with an extract or pure metabolite of another alga known or expected to be unpalatable. The coated piece is then offered along with pieces of the same seaweed coated only with the solvent (control) to a particular herbivore species in the laboratory (e.g. Targett et al., 1986; Hay et al., 1987a; Wylie & Paul, 1988), or to populations of several different herbivores (e.g. parrotfishes, surgeonfishes and rabbitfishes) in the wild (e.g. Hay et al., 1987b; Paul & Van Alstyne, 1988). Experimental and control pieces are then either counted or weighed after short-term (minutes to hours) feeding trials to calculate the deterrent effect of the extract or metabolite on the herbivores in question. This technique has been limited to non-polar extracts and metabolites (e.g. terpenoids) because, being lipid soluble, they adhere to the algal surface after the solvent evaporates and are not quickly lost to sea water. Polar secondary compounds, especially phlorotannins, have not been studied with this technique because their separation and identification is more complex (Norris & Fenical, 1985) and, being water soluble, they are subject to greater loss in sea water. Recent experiments involving secondary compounds show that seaweeds vary in their production of these metabolites and that different herbivores respond differently to the same compounds. Some of these findings are as follows. (1) Palatable seaweeds are eaten less when growing on, or in association with, herbivore-resistant species (Hay, 1986; Littler et al., 1986). (2) Different metabolites from the same seaweed confer different levels of resistance against different herbivores (Hay et al., 1987a; Paul, 1987; Paul et al., 1987; Hay et al., 1988a; Paul & Van Alstyne, 1988). (3) Extracts from some populations of the green alga Halimeda in areas of high herbivore activity appear to be more effective deterrents against herbivorous fishes than extracts from populations of the same species in habitats subject to lower rates of herbivory (Paul & Van Alstyne, 1988). (4) The secondary compounds of the brown alga Dictyota dichotoma that deter feeding by herbivorous fishes do not deter, or may even increase, feeding by small, sedentary grazers such as amphipods and polychaetes (Hay et al., 1987a; Hay et al., 1988b). Hay and co-workers suggested that the small herbivores that live on plants are more resistant to chemical defences than are large, mobile herbivores like fishes that move among many plants and that prédation and herbivory by fishes may be major factors selecting for amphipods that can live on and eat seaweeds unpalatable to fishes. These last statements might be interpreted as evidence that plant-herbivore relationships in the sea are the products of coevolution. Hay & Fenical (1988), however, argue that fundamental differences in marine and

160

MICHAEL H.HORN

terrestrial communities make the potential for coevolution less likely in the marine environment. The major difference between the herbivore communities in the two environments is the manner in which juveniles locate their host plants. The adults of many specialised insects are not tied to the food plant but search widely and lay eggs on appropriate hosts. In contrast, herbivorous fishes and sea urchins have planktonic larval stages that spend time developing in the pelagic community before returning to coastal habitats where they settle, metamorphose into juveniles and begin eating seaweeds either immediately as small juveniles or perhaps in most cases, especially for fishes, as larger juveniles. Although some larvae settle selectively in response to general habitat cues (Marliave, 1977; Butman, 1987), they are unlikely to respond only to inconspicuous seaweed hosts because these would be rarely encountered. Moreover, larvae approaching shore are subject to intense prédation (Gaines & Roughgarden, 1987) and unlikely to have the opportunity of multiple entries to find a rare host seaweed. These factors should select for generalist feeders that can use whatever seaweeds they encounter as they join coastal communities. Such constraints would not appear to promote coevolution between herbivores and seaweeds and may explain why specialists appear to be rarer in marine than terrestrial communities. Hay & Fenical (1988) further reason that the potential for coevolution between seaweeds and small sedentary grazers such as amphipods is limited because of the minor impact of these grazers on plant fitness relative to the major effects of herbivorous fishes and sea urchins. These reservations about coevolution between seaweeds and herbivores reinforces the cautionary note sounded by Meekan (1986) that the chemical and morphological traits of seaweeds may have evolved in response to a variety of biotic and abiotic influences, not just herbivorous animals. Despite all the recent investigations of seaweed secondary metabolites, including pharmacological assays, the physiological effects of these compounds on herbivores remain virtually unknown (Hay & Fenical, 1988). When Hay et al. (1987a) fed pinfish (Diplodus holbrooki) a diet containing 1 % pachydictyol-A from the brown alga Dictyota dichotoma, the fish grew half as rapidly as control fish over a three-week period. The physiological basis for this reduced growth, however, was not investigated. Fishes such as odacids and kyphosids that consume large amounts of brown algae would be important subjects for the study of the physiological effects of secondary compounds on herbivorous fishes. Lobel (1981) has suggested that the neutral or weakly alkaline environment of the alimentary tracts of some fishes may dissociate the proteintannin complexes formed by the digestion of algae with high phlorotannin concentrations. Interestingly, the gut pH of the stomachless Odax pullus is alkaline (Clements & Bellwood, 1988), whereas the stomach pH is strongly acidic and the intestine weakly acidic to alkaline in Kyphosus cornelii and K. sydneyanus (Rimmer & Wiebe, 1987). This area of fish herbivory requires further study and promises to yield exciting discoveries. Combined defences In a study of herbivory on a Caribbean reef, Littler, Taylor & Littler (1983b) presented evidence that seaweeds grouped according to a functional-form model (Littler & Littler, 1980) increasingly resisted fish and sea urchin herbivory in the following order of six morphological types: sheet, filamentous, coarsely-branched, thick-leathery, jointed-calcareous and crustose groups. A combination of chemical and morphological defences (toxicity, low calorific content, calcified or otherwise tough thalli) appears to be present in the most resistant functional-form groups (Littler et al., 1983b). Moreover, several recent studies (Hay, 1984a; Lewis, 1985; Wolf, 1985; Paul & Hay, 1986; Hay et al., 1988a; Paul & Van Alstyne, 1988) show the difficulty of distinguishing between the effects of these two types of deterrents in tropical algae. For example, Lewis (1985) found that alleged defences were prevalent in some algal species that are nevertheless highly susceptible to fish grazing. She found that the brown algae Sargassum polyceratium and Turbinaria turbinata, which are tough and leathery and contain phlorotannins, and another brown alga, Padina jamaicensis, which is lightly calcified, to be highly preferred by parrotfishes in the genus Sparisoma but avoided by the surgeonfishes Acanthurus bahianus and A. coeruleus, both with weaker jaw structure and dentition than the parrotfishes. Both Lewis

BIOLOGY OF MARINE HERBIVOROUS FISHES

161

(1985) and Wolf (1985) stressed that algal defences must be interpreted with respect to individual herbivores. Multiple defences might be expected among seaweeds in habitats of high herbivore diversity such as coral reefs where no single defence is apt to be effective against all the herbivores that encounter these plants (Paul & Hay, 1986). These combinations of deterrents may undergo dramatic temporal shifts in tropical algae in response to the diel activity patterns of herbivores. According to Hay et al. (1988a) new growth in the green alga Halimeda is produced at night when herbivorous reef fishes are inactive, and although chemically defended, is susceptible to damage by herbivores. Older plant portions (more than 48 hours old), which are exposed to daytime herbivory, have lower concentrations of the terpenoid feeding deterrent but are lower in food value and more highly calcified, thus relatively resistant to herbivory. Hay et al. (1988a) speculate that coordinating diel patterns of growth with rapidly mobilisable defences could be an important component of the ecology of not only Halimeda but other seaweeds and marine phytoplankton as well. Seaweeds in temperate waters, predictably, may lack the combined defences of algae in tropical habitats because of the lower intensity of herbivory, especially by fishes, in higher latitudes. Little is known, however, about seaweed deterrence against fish herbivores in temperate habitats. Comparison of the defensive traits of algae in north temperate latitudes with those in south temperate waters, especially around Australia and New Zealand where brown algae are known (Estes & Steinberg, 1988; Steinberg, 1988) to have high concentrations of certain secondary compounds, should make a rewarding study. DIGESTION AND DIGESTIVE MECHANISMS EVIDENCE FOR DIGESTION AND ASSIMILATION OF ALGAL MATERIAL The ability of herbivorous fishes to digest and assimilate algal material is apparent and convincing for many species, but the most compelling evidence for herbivory (i.e. growth on an algal diet) is limited to only a few species. The kinds of evidence, in order of increasing confirmation of herbivory, are as follows: (1) diets largely of seaweeds; (2) specialised morphology of the alimentary canal; (3) assimilation of seaweed compounds; and (4) growth on a seaweed diet. These topics are discussed in turn below. Diets largely of seaweeds This evidence is usually obtained from observations of feeding behaviour or analyses of gut contents of fishes or both. Many fishes in tropical waters (e.g. Hiatt & Strasburg, 1960; Randall, 1967; Jones, 1968; Hobson, 1974) and fewer in temperate habitats (e.g. Quast, 1968; Gibson, 1968; Horn, Murray & Edwards, 19c2; Russell, 1983) have been shown to have such diets (see Table I, p. 173). But whether browser or grazer, a plant-eating fish is always going to ingest some animal material simply because many species of small invertebrates live in close association with seaweeds. Herbivorous fishes do not appear to winnow out the animals before swallowing morsels of algal foods, so it seems that no marine fishes are herbivores in the strictest sense. Does the ingested animal material contribute significantly to the fish’s nutrition? This question remains largely unanswered, but because the digestive enzymes of carnivorous and herbivorous fishes are about the same, differing mainly in concentration (Kapoor, Smit & Verighina, 1975; Fänge & Grove, 1979), the animal matter probably contributes to the total assimilated material. In fishes that seem to be unequivocal herbivores the proportion of animal material can be exceedingly small. Proportions of animal material in the guts of both tropical (e.g. Randall, 1967) and temperate (e.g. Russell, 1983) species may be less than 1 % by volume. Diets of the stichaeids Cebidichthys violaceus and Xiphister mucosus are

162

MICHAEL H.HORN

usually less than 2% animal material by dry weight (Horn et al., 1982), an amount that does not change the P/E ratio of the total diet for the two species and therefore does not alter their predicted growth rates (Horn, Neighbors & Murray, 1986; see Fig. 1, p. 192). The diet of X. mucosus, however, occasionally contains as much as 8% animal material, which increases the P/E ratios of assimilated matter by 20–50% but still does not change the predicted growth pattern of the fish (Horn et al., 1986; see Fig 1). Specialised morphology of the alimentary canal Further evidence for herbivory is apparent in the morphological specialisations of the digestive tract of fishes whose diets are made up largely of algal material. The beak-like jaws and pharyngeal mill of parrotfishes, one of the most abundant groups of tropical herbivorous fishes (Randall, 1967, 1986), allow these fishes to scrape epilithic algae from rock substrata and dead coral and grind the material into a mass of fine particles (Randall, 1967; Clements & Bellwood, 1988) presumably available for enzymatic action and subsequent assimilation. Other plant-eating fishes such as acanthurids (Jones, 1968), girellids (Norris & Prescott, 1959) and odacids (Clements & Bellwood, 1988) have jaw teeth that appear to be specialised for a herbivorous diet. The stomachs of several species of grazing herbivorous fishes, especially the mugilids (Thomson, 1954) and certain acanthurids (Jones, 1968), are muscular and function as gizzards to grind filamentous algae and diatoms into fine particles apparently ready for chemical breakdown and subsequent absorption. The intestine or entire gut of herbivorous fishes is usually longer than that of their noncarnivorous relatives (e.g. Al-Hussaini, 1947; Barton, 1982; Goldschmid, Kotrschal & Wirtz, 1984; Hallacher & Roberts, 1985). Finally, the hindguts of at least two species of kyphosids are enlarged and specialised as fermentation chambers (Rimmer & Wiebe, 1987). These morphological specialisations are discussed in greater detail under Digestive Mechanisms (see p. 207). Assimilation of seaweed compounds Assimilation efficiencies obtained by both direct (i.e. measurement of food consumed and faeces produced) and indirect (i.e. use of an indigestible marker) methods for a variety, but still a relatively small number, of herbivorous fishes (Table II) indicate strongly that these fishes are digesting and absorbing carbohydrate, lipid and protein from their seaweed diets. Herbivores are expected to have low assimilation efficiencies because of the relatively large amounts of indigestible plant material contained in their diets (Kapoor et al., 1975; Brett & Groves, 1979; Knights, 1985), and this expectation is often met as the values in Table II reveal. Nevertheless, these efficiencies are highly variable as a whole because they range from values below 10% for lipid or carbohydrate assimilation in certain fishes (Montgomery & Gerking, 1980; Horn et al., 1986) to values for protein assimilation in the 95–98% range (Horn et al., 1986). These latter efficiencies match those recorded for carnivorous fishes (Kapoor et al., 1975). Within- and between-species differences in assimilation efficiencies among herbivorous fishes are driven by an enormous range of factors including temperature, type of food eaten, age of the fish, recent history of feeding, gut morphology and transit time of food through the gut. Two examples illustrate intraspecific variation. (1) Nitrogen assimilation efficiency in the damselfish Stegastes lividus increases from 36% in juveniles to 79% in adults, in accord with the ontogenetic shift from omnivory to herbivory in this species (Lassuy, 1984). (2) Periods of food deprivation cause reduced assimilation efficiencies for carbon, nitrogen, lipid and protein in the girellid Girella tricuspidata, with a 4-day deprivation period having a greater effect than a 24-h deprivation period (Anderson, 1988).

BIOLOGY OF MARINE HERBIVOROUS FISHES

163

Because of the differences in the direct and indirect methods of determining assimilation efficiencies and because of criticisms of the latter method, it is appropriate at this point to compare the two techniques. Most studies of assimilation efficiency in fishes including herbivores have used an indirect method of determination because of the difficulty of collecting faeces quantitatively (Talbot, 1985). The procedural advantage of the indirect method is that nutrient assimilation can be calculated without measuring food intake or faeces output. An indicator or marker substance is added to the diet, or an integral, naturally occurring component of the diet is used as an indicator or marker. Ideally, indicators should be completely indigestible and evacuated at the same rate as the other gut contents. Internal indicators, i.e., naturally occurring constituents of the diet, are preferable because assimilation of TABLE II Assimilation efficiencies of some marine herbivorous fishes Family—species Food item

Assimilation efficiency

Method

Temperature °C Reference

Acanthuridae Acanthurus guttatus

Natural diet, filamentous algae Natural diet, diatoms & detritus? Natural diet, filamentous algae Natural diet, diatoms & detritus? 37%

Total

11%

Indirect (ash marker)

28–29

Charlock, 1983

Total

16%

Indirect (ash marker)

28–29

Chartock, 1983

Total

14%

Indirect (ash marker)

28–29

Charlock, 1983

Total

20%

Indirect (ash marker)

?

Nelson & Wilkins, 1988

Enteromorpha sp. (green alga)

% 14C recovered in fish tissue of that administered in whole or fractionated alga (force fed): 1.4 4.2

Carbon-14

20

Anderson, 1987

Acanthurus olivaceus Acanthurus triostegus Ctenochaetus striatus Nitrogen Girellidae Girella tricuspidata

Control Cell walls (16 h after feeding) Cell walls (5 d after feeding) Protoplasts (16 h after feeding) Whole alga (16 h after feeding)

6.8 9.8 15.0

164

MICHAEL H.HORN

Girella tricuspidata Lipid Total nitrogen Protein nitrogen Total carbon (decline in efficiencies with food deprivation) Hemiramphidae Hyporhamphus melanochir Heterozostera tasmanica

Enteromorpha intestinalis (green alga) −20 to 55% 53 to 83% 42 to 79%

Whole alga

18 to 78%

Direct

20

Anderson, 1988

Organic matter

38%

Indirect (lignin marker)

17

Klumpp & Nichols, 1983

Carbon

8%

Carbon-14 uptake

?

Conacher, Lanzing & Larkum, 1979

22%

Carbon-14 uptake

?

Nitrogen

85%

Direct

19 & 28

Menzel, 1959

24%

Indirect (ash marker)

?

Montgomery & Gerking, 1980

26 to 82%

Natural seagrass diet, Protein

Lipid Energy Monacanthidae (Balistidae) Monacanthus Microdictyon chinensis umbilicatum (green alga) Posidonia Carbon australis (seagrass) Pomacanthidae Holacanthus Enteromorpha bermudensis salina & Monostroma Carbohydrate oxysperma (green algae) Energy 78% Pomacentridae Eupomacentrus Natural diet, rectifraenum green & red algae Protein 88% Lipid 56% 2% Carbohydrate

50% 76% 28%

72%

Total

BIOLOGY OF MARINE HERBIVOROUS FISHES

165

Family— species

Food item

Assimilation efficiency

Method

Temperature ° C

Reference

Microspathodo n dorsalis

Natural diet, Polysiphonia (red alga) 57% 46% 37% Enteromorpha clathrata (green alga)

Total

Indirect (ash marker)

?

Montgomery & Gerking, 1980

Total 29% (juveniles)

Indirect (ash marker)

28

Lassuy, 1984

Natural algal turf diet

Nitrogen 92% (winter)

Indirect (ash marker)

?

Polunin, 1988

Direct & Indirect?

?

Lobel & Ogden, 1981

Carbon-14 uptake

?

Bryan, 1975

Direct

25

Vaughan, 1978

Protein Lipid Carbohydrate Stegastes lividus Total 72% (adults) Nitrogen 36% (juveniles) Nitrogen 79% (adults) Plectroglyphid odon lacrymatus

20%

77% (summer) Carbon 78% (winter) 58% (summer) Phosphorus 61% (summer) Scaridae Sparisoma radians Siganidae Siganus spinus

Carbon 6 to 39% (adults) Sparidae Archosargus rhomboidalis Enteromorpha flexuosa (green alga) Polysiphonia subtilissima (red alga)

Seven different algae & seagrasses

Total (depending on item eaten)

Enteromorpha compressa (green alga)

Carbon 9 to 60% (juveniles)

Total

48%

Total

58%

7 to 81%

166

MICHAEL H.HORN

Sarpa salpa Total (ash free dry weight) Protein Energy Stichaeidae Cebidichthys violaceus Protein Lipid Carbohydrate (depending upon alga fed) Cebidichthys violaceus Nitrogen (depending upon alga fed) Cebidichthys violaceus

Four algae, epibionts reduced (2 browns, 1 green, 1 red) Four algae (2 browns, 1 green, 1 red) Nitrogen (depending upon alga fed) Cebidichthys violaceus Carbohydrate (depending upon alga fed) Xiphister mucosus

Ulva lactuca (green alga) 59%

Total (dry weight)

61%

Direct

19

Gerking, 1984

Four dietary algae (2 greens, 2 reds) 43 to 81% 21 to 44% 46 to 62%

Total

31 to 52%

Indirect (ash marker)

15

Edwards & Horn, 1982

Eight dietary algae (2 greens, 6 reds) 71 to 84%

Protein

77 to 95%

Indirect (ash marker)

15

Horn & Neighbors, 1984

Three algae, epibionts not reduced (1 brown, 1 green, 1 red) Carbon (depending upon alga fed)

Carbon (depending upon alga fed)

0 to 37%

Carbon-14 uptake

15

Horn et al., 1985

1 to 13%

Carbon-14 uptake

Carbon

−14 to 68%

Indirect (ash marker)

Eight dietary algae (2 greens, 6 reds) 18 to 83%

Lipid

0 to 59%

Indirect (ash marker)

15

Horn, Neighbors & Murray, 1986

Eight dietary algae (2 greens, 6 reds)

Protein

69 to 98%

Indirect (ash marker)

15

Horn, Neighbors & Murray, 1986

81% 65%

0 to 80%

BIOLOGY OF MARINE HERBIVOROUS FISHES

Lipid Carbohydrate (depending upon alga fed)

167

6 to 78% 11 to 74%

natural foods can be determined in both laboratory and field situations (Talbot, 1985). A large number of the assimilation studies on herbivorous fishes have used the ash component of the food as an internal indicator (Montgomery & Gerking, 1980; Edwards & Horn, 1982; Horn & Neighbors, 1984; Lassuy, 1984; Horn et al., 1985, 1986) stemming from its original use in assimilation studies with zooplankton (Conover, 1966). Any removal of ash from the food upon passing through the gut would result in under-estimation of assimilation efficiency. In certain studies (Buddington, 1980) some of the ash has been shown to be absorbed and, as a result, its use as an indicator or marker has been criticised (Bjorndal, 1985). Several other natural dietary substances including hydrolysis-resistant organic matter (Buddington, 1980; De Silva & Perera, 1983), hydrolysis-resistant ash (De Silva & Perera, 1983), lignin (Klumpp & Nichols, 1983) and acid-insoluble ash (Atkinson, Hilton & Slinger, 1984) have been used with varying degrees of success, but none seems to be without limitations. Both Gerking (1984) and Anderson (1988) used direct methods, i.e., quantitative measurement of food consumed and faeces produced, to determine assimilation efficiencies and obtained values generally similar to those estimated by indirect methods. Even more convincing evidence for assimilation of algal compounds by fishes is provided by data showing that radioactively labelled (14C) algal material is absorbed from the digestive tract into the fish’s body. Such evidence is available for at least four species of marine herbivorous fishes (Table II). The study on the stichaeid Cebidichthys violaceus (Horn et al., 1985) was the first to demonstrate uptake of radioactively labelled algal material by a cold-temperate marine fish. Moreover, this work showed that the fish could assimilate 14C from macroalgae largely free of epibionts (diatoms and bacteria) and from a nondietary brown alga (Macrocystis) as well as dietary species. Andersen’s (1987) study was the first to show that a herbivorous fish can assimilate the cell wall components of a seaweed although the mechanism remains unknown. Growth on a seaweed diet Finally, the most compelling evidence for complete herbivory would be growth and maintenance of good health by fish on a strictly algal diet. Few such growth experiments have been done, and less than definitive results have been obtained in the studies completed. When Menzel (1959) fed Monostroma to the predominantly herbivorous angelfish Holacanthus bermudensis at 28cC, the fish gained some weight, but when the fish were given Enteromorpha at 19°C, they lost weight. The weight gain at 28°C was attributed to deposition of lipids not proteins. Menzel expressed doubt that Holacanthus bermudensis could grow on algal matter alone unless consumption rates were much higher in nature than those seen in captivity. Similarly, weight losses were recorded for the sparids Archosargus rhomboidalis (Vaughan, 1978) and Sarpa salpa (Gerking, 1984) when fed unialgal diets in the laboratory. Much of the understanding of growth of marine fishes on strictly herbivorous diets is based on the results of these three studies, but they are misleading for two reasons. First, the fish used in each case was more an omnivore than a strict herbivore and, secondly, unialgal diets were used, which are unnatural and probably do not provide a sufficiently balanced diet for growth. Other studies suggest that fishes survive better on multispecies diets than on unialgal diets. Although they did not measure growth, Lobel & Ogden (1981) found that survival in the strictly herbivorous parrotfish Sparisoma radians was higher on a mixed plant diet than on any single

168

MICHAEL H.HORN

dietary item, including the most preferred species. Fry of Chanos chanos, a euryhaline herbivore and widely cultured fish in tropical Asia, also survive better on a combined diet of two kinds of freshwater algae than on either one alone (Pantastico, Baldia & Reyes, 1986). The maintenance ration calculated by Gerking (1984) for Sarpa salpa apparently represents the first and only such estimate for a marine herbivorous fish. Clearly, more studies of growth in marine herbivorous fishes are needed, not only to provide a greater understanding of their digestive capabilities but also to derive estimates of food consumption in nature. DIGESTIVE ENZYMES The major emphasis in enzymatic studies of herbivorous fishes has been on the search for cellulolytic enzymes, while research on the action of the basic digestive enzymes in these fishes has largely been neglected. These two topics are discussed in turn below. Assays for endogenous celluloses One of the most lingering and troublesome questions in fish herbivory is whether any fishes produce a cellulase to break down the cell walls of the plants they consume. Numerous papers have been published on the subject with generally mixed and inconclusive results. Broad surveys have shown that cellulase activity is unrelated to food habits and produced by the intestinal microflora (Stickney & Shumway, 1974; Stickney, 1975) or related either to the amount of highly processed plant detritus in the gut (Prejs & Blaszczyk, 1977) or a diet of invertebrates that contain cellulase or a cellulolytic micro-flora (Niederholzer & Hofer, 1979; Lindsay & Harris, 1980). Cellulase activity has been reported in the intestinal tracts of the sparid Lagodon rhomboides (Weinstein, Heck, Giebel & Gates, 1982), and the clupeid Brevoortia tyrannus (Lewis & Peters, 1984), but the authors in each case refrain from claiming the existence of an endogenous cellulase system for their fish. Neither Klumpp & Nichols (1983) studying the south temperate hemirhamphid Hyporhamphus melanochir nor Urquhart (1984) studying the north temperate stichaeid Cebidichthys violaceus found significant cellulase activity in the guts of these plant-eating fishes. Unlike Klumpp & Nichols (1983) and most other investigators who have measured the degradation of methyl—or carboxymethyl cellulose by an enzyme extract, Urquhart (1984) assayed for glucose production from native, crystalline cellulose when the latter was subjected to homogenates of gut tissue and gut contents prepared from specimens of C. violaceus. He reasoned that native crystalline cellulose was a more appropriate substrate for the assay because it and not methyl—or carboxymethyl-cellulose occurs in plant cell walls (Mackie & Preston, 1974). Future studies should attempt to standardise the assay techniques and probably should use crystalline cellulose as a substrate. Other enzymes The uncertainty surrounding cellulolytic enzymes is paralleled to a certain degree for the other digestive enzymes known to occur in herbivorous fishes. Plant-eating fishes appear to produce the complement of digestive enzymes found in other fishes although in different concentrations (Kapoor et al., 1975; Fange & Grove, 1979), but a general lack of knowledge persists about digestive enzymes and their action in fish herbivores, especially among marine species. A few general statements can be made, based on the reviews of Kapoor et al. (1975) and Fange & Grove (1979), but even here inconsistencies expose the fragmentary knowledge of enzymology in these herbivores. The gastric protease, pepsin, has optimal proteolytic activity around pH2 but often has another maximum between pH3 and pH4. Pepsin is probably found in all fishes

BIOLOGY OF MARINE HERBIVOROUS FISHES

169

except stomachless species. Apart from the appropriate pH of the stomach contents, proteolytic gastric digestion is enhanced by a high pepsin concentration, high temperatures and intense stomach motility. Other enzymes such as lipases, trypsin or its analogs, and several carbohydrases including amylase are produced either in the pancreas or intestine and require a neutral or slightly alkaline environment for activity. Enzyme production seems to be correlated with the composition of the diet, in particular, that carbohydrases are produced in larger amounts in herbivores and proteases in greater quantities in carnivores. For example, amylase activity is much higher in the freshwater herbivore Tilapia than in Perca, a carnivore (Fish, 1960), and proteolytic activity is weak in the herbivorous surgeonfish Acanthurus triostegus sandvicensis (Randall, 1961a). Moreover, stomachless fishes, which lack a pepsin, tend to be herbivores or omnivores (Kapoor et al., 1975) and have nearly neutral or slightly alkaline guts (Klumpp & Nichols, 1983), whereas carnivorous fishes, according to Kapoor et al. (1975), possess a true stomach in which peptic digestion takes place. This statement helps to explain the weakly acidic to slightly alkaline digestive tracts of such stomachless herbivores as odacids (Clements & Bellwood, 1988), scarids (Gohar & Latif, 1961b; Smith & Paulson, 1974; Lobel, 1981) and hemiramphids (Klumpp & Nichols, 1983) because acidic guts in these fishes would impede the action of intestinal enzymes. The Labridae, however, one of the largest marine fish families (Nelson, 1984), is composed of species almost all of which are carnivorous, yet they too are stomachless (Suyehiro, 1942; Al-Hussaini, 1947; Gohar & Latif, 1959) and have alkaline digestive tracts (Gohar & Latif, 1961b). All carnivorous fishes, therefore, cannot be characterised as having true stomachs and peptic digestion. One of the most detailed series of studies on fish digestive enzymes is that by Gohar & Latif who described the morphology (Gohar & Latif, 1959) and histology (Gohar & Latif, 196la) as well as trie carbohydrases (Gohar & Latif, 1961b), lipases (Gohar & Latif, 1961c) and proteolytic enzymes (Gohar & Latif, 1963) of the digestive tracts of some labrid and scarid fishes. These investigators found that the scarids had much stronger amylase activity and somewhat greater lipase and protease activities than the labrids, a pattern not completely in accord with the hypothesis of a close correlation between enzyme production and diet. Gohar & Latif considered the labrids they studied to be carnivores, especially adept at crushing mollusc shells in their pharyngeal jaws, and the scarids to be coral-feeders despite describing several structural features of the digestive systems (grinding pharyngeal teeth, long intestine, large gut mucosal area) of these fishes that suggest herbivorous habits. DIGESTIVE MECHANISMS How is the nutrition from seaweeds obtained? In other words, what mechanisms are found among herbivorous fishes that allow them access to the nutrients locked inside algal cells? The emphasis here is on the morphological and physiological specialisations of the herbivore gut. The available information is placed into an expanded and somewhat modified version of Lobel’s (1981) framework of alimentary canal types based on the relationship between their morphology and physiology and the digestibility of different kinds of algae. Preceding this framework are discussions of gut length and the transit time of algal food through the digestive tract, two measurable gut characteristics important for understanding the complexities of food processing in herbivorous fishes. Following the discussion of alimentary canal types, the digestive mechanisms of herbivorous fishes are analysed in the light of broader conceptual or comparative studies involving both herbivorous and non-herbivorous animals. The subjects of these studies are animal guts as chemical reactors, intestinal nutrient transport in different vertebrates, protein requirements of fishes and the evolutionary responses of herbivores to plant nitrogen content.

170

MICHAEL H.HORN

Gut length One of the most consistent and widely recognised features of herbivorous fishes is that with few exceptions their digestive tracts are longer than those of non-herbivorous species of equivalent size. This pattern is seen in several fish families exhibiting a variety of feeding modes including the Blenniidae (Goldschmid, Kotrschal & Wirtz, 1984; Kotrschal & Thomson, 1986), Gobiidae (Geevarghese, 1983), Pomacentridae (Emery, 1973) and Stichaeidae (Barton, 1982). Increased gut length is also reported in those species that represent the only taxa that resort to occasional or seasonal herbivory in otherwise carnivorous fish families such as the Myctophidae (Robison, 1984) and Scorpaenidae (Hallacher & Roberts, 1985). Most herbivorous fishes begin life as carnivores or omnivores (White, 1985) before adopting a stricter plant diet as adults. Not surprisingly, increased gut length accompanies the ontogenetic shift to herbivory as has been documented for a variety of plant-eating taxa including odacids (Clements, 1985), pomacentrids (Emery, 1973; Lassuy, 1984), sparids (Christensen, 1978; Ogburn, 1984; Stoner & Livingston, 1984) and stichaeids (Montgomery, 1977; Barton, 1982). Relative gut length, however, does not always continue to increase with size, but may level off or even decline slightly when the fish reaches a certain length (Emery, 1973; Montgomery, 1977; Lassuy, 1984; Ogburn, 1984; Stoner & Livingston, 1984) and may reflect a critical period in the dietary shift to herbivory. Gut length is a convergent feature of herbivorous fishes in that taxonomically unrelated plant-eating fishes share the characteristic of a relatively long digestive tract. Al-Hussaini (1947) in a still frequently cited study arranged 60 species of Red Sea fishes into four categories based on their food habits and relative gut lengths (defined as the distance from the posterior end of the pharynx to the anus divided by the distance from the tip of the snout to the insertion of the caudal fin). Plankton-feeders had the shortest relative gut lengths (0.5–0.7), carnivores the next (0.6–2.4), then omnivores (1.3–4.2) and herbivores the longest relative gut lengths (3.7–6.0). Attempts to place species other than those studied by Al-Hussaini into this framework can be misleading, and, for several reasons, precautions are required in deriving the trophic status of fishes from their gut lengths (see Table III). First, many fishes now known to be rather strict herbivores have shorter guts than those in Al-Hussaini’s herbivore category including parrotfishes, which he considered to be coral-feeders. Stomachless herbivorous fishes, in particular, have relatively short guts. For example, Odax pullus has a relative gut length of around 1.5 and Scarus rubroviolaceus one of about 2.0 (Clements & Bellwood, 1988). A more extreme example is the hemiramphid Hyporhamphus melanochir, which has a relative gut length of 0.5 (Robertson & Klumpp, 1983). This species is an omnivore but consumes large quantities of plant material, especially seagrasses, which it can digest with moderate efficiency (Klumpp & Nichols, 1983). Other herbivorous fishes with true stomachs but relative gut lengths less than 3.0 include the stichaeids Cebidichthys violaceus and Xiphister mucosus (Barton, 1982) and the aplodactylid Aplodactylus arctidens (Clements, 1985). A second precaution is that gut length can vary with nutritional status of the fish. Montgomery & Pollak (1988) showed that the relative gut length of the surgeonfish Acanthurus nigrofuscus decreases by 30–50% during periods of starvation of only several hours to about two days. The mullet Mugil cephalus appears to grow a longer gut when consuming a diet relatively high in detritus compared to algal material (Odum, 1970) although Collins (1981) found the reverse pattern, and the cyprinid Cyprinus carpio has a longer intestine when fed a diet containing increased amounts of glucose (Buddington, 1987). A third precaution is that the surface area of the intestinal mucosa per body weight or volume may actually decrease with size in some herbivorous fishes (Al-Hussaini, 1949; Gohar & Latif, 1959; Montgomery, 1977). These reported declines in absorptive surface area may be related to the lower metabolic rates of larger fish, but further studies using additional specimens and other species are required before definitive conclusions can be drawn about this relationship.

BIOLOGY OF MARINE HERBIVOROUS FISHES

171

What is the importance of increased gut length in herbivorous fishes? The pervasiveness of this trait among plant-eating fishes suggests that it is important for the digestive processes of these fishes. Sibly & Calow (1986) argued on theoretical grounds that animals eating relatively poor quality food should have larger digestive chambers, other factors being equal. The net rate of obtaining energy is a product of (1) the weight of material in the digestive tract and (2) the amount of energy yielded from the material for a given retention period. The second quantity in the product is lower for a poorer quality food. So, for a given nutritional requirement, the animal must carry a greater weight of digesta (the first quantity in the product above). This relationship, according to Sibly & Calow (1986), provides an explanation for the widespread observation that animals sustaining themselves on poorer quality food have larger guts. The food processing actions of herbivorous fishes with a long, coiled or looped intestine seem to be consistent with this model in that the large quantities of high fibre, low nutrient food they ingest are distributed along the extensive surface area of the intestine to maximise TABLE III Relative gut lengths of some marine herbivorous fishes Family—species

Feeding type

Relative gut length

Acanthuridae Acanthurus achilles Acanthurus glaucopareius Acanthurus guttatus Acanthurus leucopareius Acanthurus nigrofuscus Acanthurus sandvicensis Acanthurus nigroris Acanthurus dussumieri Acanthurus mata

Browser Browser Browser Browser Browser Browser Browser/grazer Grazer/browser Grazer

5.2 4.9 4.6 6.9 4.0 5.8 4.4 3.0 3.8

Dimensions of ratio

Reference Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968

Family—species

Feeding type

Relative gut length Dimensions of ratio Reference

Acanthurus olivaceus Ctenochaetus hawaiiensis Ctenochaetus strigosus Naso brevirostris Naso lituratus Naso unicornis Zebrasoma flavescens Zebrasoma veliferum Acanthurus nigrofuscus

Grazer Grazer

3.4 3.5

Jones, 1968 Jones, 1968

Grazer Browser Browser Browser Browser Browser Browser

3.5 2.2 3.2 3:2 3.7 3.7 3.0 (0800–1600 h)

Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968 Jones, 1968 Montgomery & Pollak, 1988

2.3 (1800–0600 h) 1.4 (starved 48h) 1.9 (starved and fed)

172

MICHAEL H.HORN

Aplodactylidae Aplodactylus arctidens Blenniidae Salarias fasciatus Blennius sanguinolentus Entomacrodus chiostictus Ophioblennius steindachneri Girellidae Girella elevata Girella tricuspidata Gobiidae Boleophthalmus pectinorostris Awous stamineus Pseudogobius javanicus Family—species Hemiramphidae Hyporhamphus melanochir Kyphosidae Kyphosus tahmel Kyphosus cornelii Kyphosus sydneyanus Mugilidae Mugil auratus Aldrichetta forsteri Liza argentea Liza dussumieri Mugil cephalus

Grazer

1.6–1.8

Clements, 1985

Grazer Grazer Grazer Grazer

4.1 2.4 2.5 4.8

Al-Hussaini, 1947 Goldschmid, Kotrschal & Wirtz, 1984 Kotrschal & Thomson, 1986 Kortschal & Thomson, 1986

Browser/grazer 2.1 (adults) Browser/grazer

1.7 (juveniles)

Bell, Burchmore & Pollard, 1980

1.9–2.9

Anderson, 1986

Grazer? Grazer? Grazer?

6.0 1.5 1.3

Suyehiro, 1942 Geevarghese, 1983 Geevarghese, 1983

Feeding type Relative gut length Dimensions of ratio Reference Browser

0.5

Robertson & Klumpp, 1983

Browser? Browser Browser

4.1 4.6 4.0

Al-Hussaini, 1947 Rimmer & Wiebe, 1987 Rimmer & Wiebe, 1987

Grazer Grazer Grazer Grazer Grazer

3.7 5.5–5.7 5.5–5.7 5.5–5.7 5.5–5.7

Al-Hussaini, 1947 Thomson, 1954 Thomson, 1954 Thomson, 1954 Thomson, 1954

Mugil georgii Myxus elongatus Mugil cephalus

Grazer Grazer Grazer

Liza dumerilii Liza richardsonii Liza tricuspidens Mugil cephalus

Grazer Grazer Grazer Grazer

Odacidae

5.5–5.7 5.5–5.7 3.2–5.5 (depending on diet) 2.4–2.5 2.5–2.7 2.1–2.5 4.5–5.2 (slightly longer in smaller fish)

Thomson, 1954 Thomson, 1954 Odum, 1970 Marais, 1980 Marais, 1980 Marais, 1980 Marais, 1980

BIOLOGY OF MARINE HERBIVOROUS FISHES

Odax pullus Pomacentridae Stegastes lividus Plectroglyphidodon lacrymatus Scaridae Pseudoscarus ghobbam

Browser

1.5

Clements & Bellwood, 1988

Browser Grazer

4.4 8.5–10.0

Lassuy, 1984 Polunin, 1988

Grazer

1.4–1.6

Gohar & Latif, 1959

173

Family—species

Feeding type Relative gut length

Dimensions of ratio Reference

Pseudoscarus harid Scarus rubroviolaceus Siganidae Siganus fuscescens Siganus spinus Sparidae Sarpa salpa 1.4 (juveniles) 2.7 (adults) 2.3 (adults) Archosargus probatocephalus 2.0–2.1 (adults) Stichaeidae Cebidichthys violaceus Xiphister mucosus

Grazer Grazer

1.6–1.9 2.1

Gohar & Latif, 1959 Clements & Bellwood, 1988

Browser Browser

2.5–3.0 3.5–4.0

Suyehiro, 1942 Bryan, 1975

Browser

0.9 (small juveniles)

Christensen, 1978

Browser

Gerking, 1984 1.7 (juveniles)

Ogburn, 1984

Browser Browser

1.1 0.8

Barton, 1982 Barton, 1982

digestion and assimilation (see Buddington, Chen & Diamond, 1987). Intestinal surface area is, of course, constrained by the size of the body cavity and the size and shape of the fish (Montgomery, 1977). Kotrschal & Thomson (1986) speculated that, although the smallest reef fishes (e.g. chaenopsids) should be grazers like certain blennies (Nursall, 1981) because such a feeding habit minimises handling time and thereby reduces the threat of predators, these species are not herbivores probably because gut length is constrained by the small body size. Another possible function of a long intestine in herbivorous fishes is reabsorption of proteolytic enzymes thereby conserving protein, which is often in short supply for plant-eating animals. Hofer & Shiemer (1981) found evidence for active reabsorption of these enzymes in some herbivorous and omnivorous freshwater fishes. The efficiency of reabsorption reached an optimum at relative gut lengths of 2.5–3.0. This avenue of research might be usefully extended to marine species. Gut transit times Gut transit time, or gut retention time, is the interval of time required for a particle of food to pass through the digestive tract and the undigested portion voided as faeces. Gastric evacuation time is a component of this interval, and both are often expressed as rates (Fänge & Grove, 1979). Methods used to measure the rate of movement of food through the digestive tract include serial killing of individuals in an experimental laboratory population, monitoring of food translocation by X-radiography of an ingested opaque marker and

174

MICHAEL H.HORN

recording the time of appearance of coloured faeces after a dye has been incorporated into the food (Fange & Grove, 1979). Gut transit time is a critical component in food processing by fishes because it influences assimilation efficiencies and helps determine food consumption rates. The appetite offish appears to return as the digestive tract empties (Fange & Grove, 1979; Grove & Crawford, 1980). Gut transit time (or rate) varies with temperature, meal size, food type, fish size, method of feeding and feeding history of the fish (Fange & Grove, 1979). The frequently cited scenario that herbivorous fishes consume large quantities of plant material by continuous feeding, pass the material rapidly through the gut and assimilate it with moderate efficiency (Ogden & Lobel, 1978; Brett & Groves, 1979; Gerking, 1984; Pandian & Vivekanandan, 1985; Buddington et al., 1987) is more fully understood in the context of herbivore diversity and the exceptions to the pattern. Herbivores do appear to have shorter gut transit times although the data are sparse. The summaries of gastric evacuation and total gut emptying times compiled by Fänge & Grove (1979) are heavily dominated by carnivorous species, and almost all of the few herbivorous species are freshwater forms, both sets of data reflecting the historical bias in species chosen for study. Although the values are strongly temperature dependent, the total gut emptying times range from 10 to 158 h for carnivores and are mostly less than 10 h for herbivores. Gut retention times for the few marine herbivores that have been studied are mostly tropical species, and these times are also below 10 h (Table IV). This mode of food processing, which involves the passing of ingesta across the large surface area of an extended intestine, apparently results in a more favourable cost/benefit ratio for herbivores than the more carnivore-like TABLE IV Time required for food to pass through the gut of some marine herbivorous fishes Transit time, h Family—species

Stomach Entire gut

Acanthuridae Acanthurus triostegus sandvicensis Hemiramphidae Hyporhamphus melanochir Kyphosidae Kyphosus sydneyanus Mugilidae Mugil cephalus Liza dumerilii Liza falcipinnis Pomacentridae Stegastes lividus Plectroglyphidodon lacrymatus Scaridae Scarus gibbus Scarus jonesi

Temperature, °C Reference

2

?

Randall, 1961a

4.4 (seagrass diet) 8.3 (crustacean diet)

17

Klumpp & Nichols, 1983

21

23

Rimmer & Wiebe, 1987

2–6

20–26 ? ?

Odum, 1970 Payne, 1978 Payne, 1978

4.5 (1 juvenile) 9.7, 9.9 (2 adults) 5–6

28

Lassuy, 1984

?

Polunin & Koike, 1987

4–6 4–6

? ?

Smith & Paulson, 1974 Smith & Paulson, 1974

3 3

BIOLOGY OF MARINE HERBIVOROUS FISHES

175

Transit time, h Family—species

Stomach Entire gut

Siganidae Siganus spinus Sparidae Sarpa salpa Stichaeidae Cebidichthys violaceus

38

Temperature, °C Reference

1.5 (juveniles) 2–3 (adults)

?

Bryan, 1975

Several

19

Gerking, 1984

53

15

Urquhart, 1984

actions of intermittent consumption, longer retention times and higher assimilation efficiencies. Studies on feeding and digestion in the hemiramphid Hyporhamphus melanochir (Robertson & Klumpp, 1983; Klumpp & Nichols, 1983) are instructive for this comparison. This warm temperate Australian fish feeds on seagrass during the daytime but switches to crustaceans at night. It is a continuous feeder, and the seagrass passes through the short tubular gut twice as fast as the crustacean food (4.4 compared with 8.3h; see Table IV). The crustaceans are available only at night when they emerge into the water column; thus, seagrass is much more available during the day than the crustaceans. The dilution of food quality by inorganic material has been shown to increase ingestion and egestion rates in fishes, i.e., to decrease gut emptying times (Rozin & Mayer, 1964; Lee & Putnam, 1973). Similarly, other studies, but on carnivores with stomachs, have shown that gastric emptying times are faster in fish on low energy diets than those consuming high energy meals (Jobling, 1980, 1987). Klumpp & Nichols (1983) argued, therefore, that the indigestible fibre in seagrass may have the same effect on gut transit time in H. melanochir. Thus, rapid passage of the seagrass material through the fish’s short gut apparently results in the low assimilation efficiencies (only 50% for protein) reported by these authors. Night-time consumption of crustaceans then becomes, according to Klumpp & Nichol’s calculations, a necessary source of protein for the fish. H. melanochir and other stomachless herbivores such as the odacids and scarids may be limited to rapid gut transit times because of their short, simple guts although it is difficult to sort out the effects of this factor from other influences. Clements & Bellwood (1988) speculated that the intestinal folds of Scarus rubroviolaceus may slow laminar flow rates and thereby increase food retention times. Exceptions to the above scenario are instructive and help illustrate the diversity of feeding modes and food processing among marine herbivorous fishes. The cold temperate stichaeid Cebidichthys violaceus has a gut transit time of greater than 50 hours (Urquhart, 1984; see Table IV), possibly the longest reported for a herbivorous fish, and moderately high assimilation efficiencies (Edwards & Horn, 1982; Horn & Neighbors, 1984; Horn, Neighbors & Murray, 1986). This sluggish fish is an intermittent feeder, apparently consuming algae in its rocky intertidal habitat only on the incoming tide and then becoming inactive for most of the diel cycle (Ralston & Horn, 1986). The time and energy required to digest and assimilate its seaweed food at low temperatures may limit the fish’s scope for activity. The warm temperate rudderfish Kyphosus sydneyanus also appears to be an intermittent feeder and to have a relatively long gut transit time, but for reasons different from those of Cebidichthys violaceus. Kyphosus sydneyanus has been shown to have a microbial fermentation system in its hindgut (Rimmer & Wiebe, 1987). The gut retention time of 21 h (Rimmer & Wiebe, 1987; see Table IV) and feeding activity limited apparently to crepuscular periods (Russell, 1983) seem to be traits appropriate to allow time for a batch of seaweed food to be broken down by a microbial colony in the fish’s caecum.

176

MICHAEL H.HORN

Fig 2.—Digestive tracts of four species of fishes, each representing one of four types of alimentary canals in marine herbivorous fishes; gut lengths but not other dimensions are drawn to the same scales as the fishes; salient features of the four canal types are lettered consecutively; a=upper and lower jaw teeth of Acanthurus nigrofuscus; b=stomach of A. nigrofuscus, consisting of a thin-walled, distensible cardiac region and a thin-walled but more muscular pyloric region; c=close-set, sieve-like gill rakers of Mugil cephalus; d=thick-walled, muscular pyloric region (shaded) of the stomach of M.cephalus; e=fused upper and lower jaw teeth of Scarus rubroviolaceus; I=lateral view of the upper and lower pharyngeal bones of S. rubroviolaceus; g=hindgut caecum of Kyphosus sydneyanus, separated by valves (shaded areas) from the intestine and the rectum; pyloric caeca, located at the junction of the stomach and intestine, range in number in these four fishes from none in Scarus rubroviolaceus, to 2 in Mugil cephalus, to 5 in Acanthurus nigrofuscus to an uncounted mass in Kyphosus sydneyanus; information for drawings: Acanthurus nigrofuscus (Jones, 1968); Mugil cephalus (Thomson, 1966); Scarus rubroviolaceus (Clements & Bellwood, 1988); Kyphosus sydneyanus (Rimmer & Wiebe, 1987); see Table V for further details and additional representative species of the four alimentary canal types.

TYPES OF ALIMENTARY CANALS A central question in fish herbivory is how fishes gain access to the nutrients inside plant cell walls. Based on the apparent absence of endogenous cellulolytic enzymes and a fermentative microflora in herbivorous fishes, Lobel (1981) claimed that fish can digest seaweeds only if breakage of the cell occurs so that the fish’s regular complement of enzymes can react with the cell contents. He proposed that such breakage can occur by either (1) lysis, resulting from acidic stomach secretions, or (2) mechanical action, resulting mainly from trituration in a pharyngeal mill or gizzard-like stomach. Based on Lobel’s (1981) findings and those of more recent studies, four types of alimentary canals are discussed below. Highly acidic stomach (Type I) Herbivorous fishes with thin-walled stomachs and no trituration mechanism other than the jaw bite appear to use highly acidic stomach fluids to lyse algal cell walls (Fig 2; Table V). Acid lysis as a means of releasing plant cell contents was first suggested by Fish (1960) and has since been demonstrated by Moriarty (1973) and Bowen (1976) for freshwater cichlids and by Lobel (1981) and Urquhart (1984) in laboratory tests simulating the acidic gastric pH environment of several herbivorous marine fishes. Gastric pH levels drop to as low as 1.3–1.5 in the cichlids and are apparently sufficient to lyse the cells of bluegreen algae (Moriarty, 1973) and detrital bacteria (Bowen, 1976) eaten by these fishes. Lobel (1981) found that the stomach pH of certain acanthurids, pomacanthids and pomacentrids ranged from 2.4 to 4.3 (Table V), acidities seemingly as effective as trituration in releasing the cell contents of some algal species. He used spectrophotometry to measure the concentrations of cell contents released into the medium by the acid treatments and noted that algae were differentially susceptible to acid lysis. This finding indicates that an alga might be available but not digestible to a particular consumer. Urquhart (1984) obtained stomach pH values of 2.2–2.5 (Table V) for the stichaeid Cebidichthys violaceus, and his laboratory simulations TABLE V Characteristics of four types of alimentary canals in marine herbivorous fishes with a list of representative species and their gut pH values; first three types based on information in Lobel (1981); see Fig 2 for illustrations of these alimentary canal types Representative fishes Family Species Stomach pH Intestinal pH Reference I. Digestive mechanism: primarily chemical; acid lysis in stomach releases cell contents Gut morphology: thin-walled stomach; long intestine Feeding type: mainly browsing

BIOLOGY OF MARINE HERBIVOROUS FISHES

177

Representative fishes Acanthuridae Acanthurus lineatus Acanthurus triostegus Zebrasoma rostratum Acanthurus nigrofuscus

Chanidae Pomacanthidae Centropyge flavissimus Pomacentridae Eupomacentrus planifrons Sparidae

Acanthurus glaucopareius 4.2 4.3

4.2

7.3

3.2 7.0–7.2

3.7–6.2 (pyloric) (depending on time of collection & feeding state) Chanos chanos

(pyloric caeca)

Eupomacentrus nigricans 2.4

Lobel, 1981 Lobel, 1981 Lobel, 1981

2.9–5.7 (cardiac)

Apolemichthys xanthopunctatus 3.1

Lobel, 1981

1.9–6.6 (depending on diet) 3.0

Montgomery & Pollak, 1988

Lobel, 1981 7.6

Lobel, 1981

Lobel, 1981 2.7

6.9

5.6

Lobel, 1981

Lobel, 1981

Archosargus 2.0 Ogburn, 1984 probatocephalus Stichaeidae Cebidichthys 3.0 7.3–7.9 Edwards & Horn, violaceus 1982 2.2–2.5 Urquhart, 1984 II. Digestive mechanism: primarily mechanical; food ground in thick-walled, gizzard-like stomach; sediment ingested to aid in grinding process Other gut morphology: long intestine Feeding type: mainly grazing Acanthuridae Acanthurus bleekeri 6.7 Lobel, 1981 Acanthurus gahhm 7.4 Lobel, 1981 Lobel, 1981 Acanthurus guttatusa 6.3 Acanthurus mata 6.5 Lobel, 1981 Acanthurus olivaceus 6.7 Lobel, 1981 Acanthurus 7.0 Lobel, 1981 xanthopterus Ctenochaetus 7.1 Lobel, 1981 striatus 7.3 Lobel, 1981 Ctenochaetus cyanoguttatus Aplodactylidae

Aplodactylus arctidensb

Clements, 1985

178

MICHAEL H.HORN

Mugildae

Crenimugil 7.9 6.8 Lobel, 1981 crenilabis Liza dumerilii 7.0–8.5 8.5 Payne, 1978 Liza falcipinnis 2.0–5.0 8.5 Payne, 1978 Mugil cephalus 3.5–7.0 Moriarty, 1976 8.5 8.5 Payne, 1978 7.2 Lobel, 1981 Mugil curema 8.5 8.5 Payne, 1978 III. Digestive mechanism: primarily mechanical; food ground in pharyngeal mill; sediment ingested by grazers Gut morphology: no stomach; moderately long intestine Feeding type: grazing or browsing Hemiramphidae Hyporhamphus 6.5–7.0 Klumpp & Nichols, melanochir 1983 Odacidae Odax pullus >8.1 Clements & Bellwood, 1988 Gohar & Latif, 1961b Scaridae Pseudoscarus 6.1–6.8 (duodenum) ghobbam 6.4–7.9 (posterior limbs) Pseudoscarus harid 6.1–6.8 (duodenum) Gohar & Latif, 1961b 6.1–7.8 (posterior limbs) Scarus gibbus 6.4 (pyloric caecum) Smith & Paulson, 6.4–6.5 (intestine) 1974 7.5 (rectum) Smith & Paulson, Scarus jonesi 6.8 (pyloric caecum) 1974 6.9–7.5 (intestine) 8.2 (rectum) (feeding) 7.2 (pyloric caecum) 7.5–7.6 (intestine) (nonfeeding) Sparisoma radians 8.4 (anterior Lobel, 1981 intestine) 8.6 (posterior intestine) IV. Digestive mechanism: primarily chemical; hindgut caecum for fermentation; acidic stomach but cell lysis not apparent Other gut morphology: well-defined stomach; long intestine Feeding type: bro owsing Kyphosidae Kyphosus cornelii 2.9–3.9 5.7–7.5 (intestine) Rimmer & Wiebe, 6.1–6.2 (caecum) 1987 6.4 (rectum) 2.8–3.0 6.4–8.2 (intestine) Rimmer & Wiebe, Kyphosus 6.3–6.7 (caecum) 1987 sydneyanus 6.5–7.0 (rectum)

a b

This species is described as a browser with a thin-walled stomach by Jones (1968). This species is reported not to ingest sediment (Clements, 1985).

BIOLOGY OF MARINE HERBIVOROUS FISHES

179

showed that acid lysis of the cell walls of the green alga Ulva lobata significantly increased the concentration of protein and carbohydrate in the medium compared to control solutions. Carbohydrates leached into acidic solutions simulating the stomach fluids, whereas proteins leached into solutions simulating the slightly alkaline environment of the intestine after the alga had been treated with acidic solutions simulating those in the stomach. The exact mechanism by which low pH causes cell lysis is poorly understood. Increased acidity appears to weaken the hydrogen bonds between the cellulose units in terrestrial plant cell walls and, in effect, to loosen or expand the cell walls (Wilkins, 1984), which presumably would allow the cell contents to leach out through the cell wall. A similar mechanism may operate in seaweeds. As mentioned above, marine algae are differentially affected by acidic conditions (Lobel, 1981). These differences may result from differences in cell wall structure among the various groups of seaweeds. Gut pH in fishes is influenced by a variety of factors and conditions. Lobel (1981) found that fishes with thin-walled (non-gizzard-like) stomachs had more acidic stomach contents (mean pH of 3.4) than fishes with thick-walled (gizzard-like) stomachs (mean pH of 7.0). Moreover, a thin-walled stomach increases in acidity as it is filled and becomes distended (Randall, 196la; Smit, 1967, 1968; Moriarty, 1973; Norris, Noms & Windell, 1973; Kapoor, Smit & Verighina, 1975; Lobel, 1981). Assimilation efficiency is positively correlated with stomach fullness and acidity in the cichlid Tilapia nilotica (Moriarty, 1973). Stomach pH also can vary with the type of food eaten. For example, Lobel (1981) reported a gastric pH of 1. 9 in the milkfish, Chanos chanos, when it had been feeding on green algae and a pH of 6.6 when the fish had been eating invertebrates. The stomach pH of carnivorous fishes, although not as acidic as in herbivores, may enable them to digest algae to a certain degree or to take advantage of the filling effect of ingested seaweeds to lower acidity and thereby enhance digestion of animal prey. Lobel (1981) has even proposed that gastric acidity in such fishes may preadapt them to herbivory. Only moderate character evolution, such as an increase in gut length or a change in jaw tooth shape or number, might allow increased herbivory in species belonging to otherwise carnivorous families. For example, relatively long guts and increased consumption of algae distinguish Ceratoscopelus warmingii from the other, carnivorous myctophid fishes (Robison, 1984) and Sebastes mystinus from the other, carnivorous scorpaenid fishes in the genus Sebastes (Hallacher & Roberts, 1985). Also, some carnivorous fishes are known to consume seaweeds when animal prey is scarce. Zooplanktivorous damselfishes in the genera Amblyglyphidodon, Chromis and Pomacentrus eat drifting algae where zooplankton is in low abundance (Hobson & Chess, 1978), and the scorpaenid Sebastes mystinus, which feeds primarily on zooplankton during the upwelling season, consumes large quantities of macroalgae during the non-upwelling season (Hallacher & Roberts, 1985). In summary, acid lysis appears to be one of the principal mechanisms whereby certain herbivorous fishes gain access to the nutrients inside algal cell walls. The alimentary canals of these fishes are perhaps the least morphologically specialised of the plant-eating marine species. Algal types suitable as food for fishes with this kind of alimentary canal should be primarily green and red algae with large cell sizes (Lobel, 1981). Lobel (1981) asserted that fishes with acidic stomachs are browsers rather than grazers so as to minimise ingesting sand while feeding and thereby prevent the rapid buffering of stomach contents by calcium carbonate material. All 12 species of acanthurids recognised as browsers by Jones (1968) have thinwalled stomachs. Other fishes in this category include the chanid, pomacentrids, pomacanthids, sparids and stichaeids (Table V). Siganids have thin-walled stomachs and long intestines (Bryan, 1975) and may belong in this category, but gastric pH apparently has not been measured in these fishes. Herbivorous blenniids have long, coiled intestines and anterior intestinal swellings but lack a true stomach (Al-Hussaini, 1947; Goldschmid, Kotrschal & Wirtz, 1984). These fishes may also belong in this category but, again, digestive

180

MICHAEL H.HORN

tract acidity apparently has not been recorded. It is of interest that the luderick Girella tricuspidata is the only fish studied to date that has the ability to assimilate components of algal cell walls (Anderson, 1987). The highly acidic (pH as low as 2) and muscular stomach of this fish (Anderson, 1986) probably contributes to this ability. In having both of these traits, G. tricuspidata is intermediate between herbivorous fishes with thin-walled, highly acidic stomachs and those with muscular, less acidic stomachs discussed below. Whether the pharyngeal apparatus of G. tricuspidata, mentioned by Kaufman & Liem (1982), aids in rupturing cell walls remains unknown. Gizzard-like stomach (Type II) A thick-walled, gizzard-like stomach represents one of the two mechanical means by which herbivorous fishes rupture algal cell walls (Lobel, 1981). The proposed function of such a heavily muscularised stomach is trituration of bacteria, blue-green algae, diatoms, and macroalgae, especially filamentous green and red forms, ingested with sand or other inorganic material. Mugilids and certain acanthurids are the main groups of herbivorous fishes with this type of alimentary canal (Fig 2; Table V). The mullet Mugil cephalus typically feeds either by sucking up the surface layer of mud or by grazing on submerged rock or plant surfaces (Odum, 1970). Sand or other sediment particles may comprise more than 50% of the stomach contents (by weight) of M. cephalus (Collins, 1981). Grazing surgeonfishes in the genus Acanthurus can commonly be seen moving in schools over the bottom picking up mouthfuls of sand (Jones, 1968). That the gizzard-like stomach is a grinding organ is based upon observations of the condition of food before and after passage through the stomach (Al-Hussaini, 1947; Pillay, 1953; Jones, 1968; Odum, 1968, 1970; Payne, 1978). Mullets (Mugilidae), perhaps the best known group of fishes with a gizzard-like stomach, seem to rely solely (Payne, 1978) on the muscular, grinding action of the pyloric portion of the stomach and the abrasion of sand grains to lyse the cells of bacteria and blue-green algae. Payne cites methods in microbiology that tend to corroborate the effectiveness of the mullet grinding mechanism for lysing cells. Wet attrition milling with the use of abrasives is known to lyse bacterial cells, and the resultant particle size can be as small as 0. 02 mm (Hughes, Wimpenny & Lloyd, 1971). Such milling is said to be particularly effective when the abrasive is continually added as would be the case in feeding mullet. Furthermore, the uniform size of the sand grains ingested (Odum, 1968) may increase the milling efficiency (Payne, 1978). Changes in permeability resulting from the formation of cracks in the cell envelope during the grinding process may be sufficient to lyse the cells (Hughes et al., 1971). The stomach contents of mullets become enveloped in a stiff mucous membrane that Payne (1978) claims is a protective device for the stomach epithelium during the grinding process. Herbivorous fishes with thick-walled stomachs appear to have slightly acidic to slightly alkaline stomach fluids (Table V). Although more highly acidic pH values (3.5–4.5) have been recorded for the mullets Liza falcipinnis and Mugil cephalus, such acidities are thought to be insufficient to cause cell lysis (Moriarty, 1976; Payne, 1978). Gizzard-like stomachs have been reported to lack acid-secreting cells and proteolytic enzymes (Ishida, 1935, 1936), but these findings are controversial (Lobel, 1981) because pepsin-producing glands apparently have been found in a species of mullet (Pillay, 1953) and moderately low pH levels have been recorded for two other mullets as noted above. Fishes with gizzard-like stomachs tend to have shorter guts than species with thin-walled stomachs, at least within the Acanthuridae. Among the herbivorous surgeonfishes studied by Jones (1968), four species with thick-walled stomachs had a mean relative gut length of 3.6 (± 0.2SD), whereas 12 species with thin-walled stomachs had a mean length of 4.3 (± 1.3SD) Intestinal length in mullets may be greater than in those fishes

BIOLOGY OF MARINE HERBIVOROUS FISHES

181

with thin-walled stomachs (see Fig 2, p. 219), but gut length in mullets varies according to diet although the pattern is not clear. Odum (1970) found that a population of Mugil cephalus feeding chiefly on plant detritus in a salt marsh had longer intestines than a population consuming mainly diatoms in a seagrass bed. Odum reasoned that a longer intestine is required to digest and assimilate detritus than diatoms. In contrast, Collins (1981) reported that M.cephalus populations with a greater proportion of algae (mostly diatoms) in their diets had longer intestines than those ingesting mainly detritus. The basis for this discrepancy between intestinal length and diet data is unknown but might result from variations in importance of detrital amino acids or differential digestibility of algae in the diets of mullets from different localities (Collins, 1981). In summary, herbivorous fishes with thick-walled stomachs consume algae and microorganisms along with quantities of sand or other sedimentary material, which apparently aids in the rupture of cells during the grinding process in the stomach. Fishes in this category are considered to be grazers (Lobel, 1981), and they fit Jones’ (1968) definition of the term in that inorganic material is ingested in the feeding process. In addition to the mullets and certain surgeonfishes, aplodactylids may also belong to this category (see Table V, p.220). Clements (1985) reported that Aplodactylus arctidens has a muscular stomach and tentatively classified the species among Lobel’s (1981) fishes with gizzard-like stomachs. Clements noted, however, that gut pH data are lacking for A.arctidens and that, although the species is a non-selective grazer, it does not swallow sand to aid in food preparation. Pharyngeal mill (Type III) A specialised pharyngeal apparatus represents the second of the two mechanical means by which herbivorous fishes rupture algal cell walls (Lobel, 1981). In addition to the pharyngeal mill, fishes with this type of alimentary canal have slightly acidic to alkaline digestive tracts and no stomach (Fig 2; Table V). These alimentary canal features are found in scarids (Gohar & Latif, 1961b; Smith & Paulson, 1974; Lobel, 1981; Clements & Bellwood, 1988), odacids (Clements & Bellwood, 1988) and hemiramphids (Klumpp & Nichols, 1983). Scarids are reported to lack acid-secreting cells (Gohar & Latif, 196la), so the slight acidity of their intestines may result from bile secretions (Gohar & Latif, 1961b). The relative gut lengths of fishes with pharyngeal mills are generally shorter than either those with highly acidic, thin-walled stomachs or gizzard-like stomachs (Lobel, 1981; see Fig 2). Scarids are the best known of the herbivorous fishes with a pharyngeal mill. As already indicated, they use their fused jaw teeth either to graze algae from reef surfaces (Randall, 1967; Ogden, 1977; Lobel, 1981; Clements & Bellwood, 1988) or to browse on seagrasses (Randall, 1967; Ogden, 1976; Thayer et al., 1984), and then use their powerful pharyngeal apparatus to grind the food into small particles before it is passed into the intestine (Randall, 1967; Ogden, 1977; Clements & Bellwood, 1988). The pharyngeal mill in scarids is similar to that in cichlids, which like the scarids are members of a monophyletic assemblage of pharyngognathous perciform fishes (Liem & Greenwood, 1981; Kaufman & Liem, 1982), but cichlids have retained a thin-walled acidic stomach (Lobel, 1981). The inorganic material ingested by parrotfishes is mostly calcium carbonate, whereas that which might be swallowed by the freshwater cichlids while feeding is inert silica (Lobel, 1981). Odacids are labroid fishes closely related to scarids. Both taxa have been included in the Labridae (Liem & Greenwood, 1981; Kaufman & Liem, 1982) based primarily on osteological and myological characters associated with the feeding apparatus. Richards & Leis (1984), however, advocated separate family status for labrids, scarids and odacids based on early life history characters. In either case, odacids have fused jaw teeth like scarids and a true diarthrosis in the pharyngeal jaws (Kaufman & Liem, 1982) characteristic of the labroid lineage. Odax pullus, the only herbivorous odacid in which food and feeding have been studied in

182

MICHAEL H.HORN

detail, browses on brown macroalgae and shreds but does not grind its food in the pharyngeal apparatus (Clements, 1985; Clements & Bellwood, 1988). Its gut is relatively short and simple, and apparently no sedimentary material is ingested. The alkaline fluids (pH>8.1) of the intestine may dissociate protein-tannin complexes (Lobel, 1981) and allow O. pullus to feed mainly (Clements, 1985; Meekan, 1986; Clements & Bellwood, 1988) on the brown seaweeds known (Estes & Steinberg, 1988) to have high phlorotannin concentrations. Hemiramphids are included in this category because, although they are mostly omnivorous, some species eat seagrasses as a major portion of the diet (Thomson, 1959; Randall, 1967; Coetzee, 1981; Robertson & Klumpp, 1983) and triturate the ingested plant material in a pharyngeal mill (Klumpp & Nichols, 1983). The family is defined taxonomically by the third pair of upper pharyngeal bones being ankylosed into a plate (Collette, McGowen, Parin & Mito, 1984). This plate is part of the pharyngeal apparatus, which is similar in function to that of the unrelated labroid fishes but different in structure. In Hyporhamphus melanochir, the gut is extremely short, a stomach is lacking and the intestinal pH ranges from slightly acidic to neutral (Klumpp & Nichols, 1983). Hindgut fermentation chamber (Type IV) A fourth type of alimentary canal found in herbivorous fishes was not described by Lobel (1981), but he recognised the possibility that intestinal microorganisms might be found in certain species with the ability to digest plant cell walls. Since the publication of Lobel’s study, compelling evidence has been provided (Rimmer & Wiebe, 1987) that two species of kyphosids in warm temperate to subtropical Australian waters contain a hindgut microflora that can fermentatively digest seaweed material ingested by these fishes. The two species, Kyphosus cornelii and K. sydneyanus, are large (60– 80 cm in lenght), strictly herbivorous fishes with long, coiled intestines and thin-walled caecum-like pouches near the posterior end of the intestine (Fig 2; Table V). The capacity of these pouches is about 1.5–2 times the stomach volume. The pouches were found to be well vascularised and separated by valves from the adjacent parts of the gut. In K. sydneyanus the caecal pouch is single-lobed, whereas in K. cornelii the pouch contains two blind, lateral sacs connected by valves to the median lobe. A short, vascularised rectum is separated by a distal valve from the caecum. Food material taken from caecal pouches contained an abundant and diverse microflora of bacteria and ciliated and flagellated protozoans, which were undetectable in the anterior part of the gut. Rimmer & Wiebe (1987) claimed they found conclusive evidence of fermentation by recording the presence of volatile fatty acids (VFAs) in the material sampled from the caecal pouch of these two fishes and the rectum of K. sydneyanus. (VFAs are the assimilable, anaerobic degradation products following hydrolysis of polysaccharides (e.g. cellulose) to their constituent sugars by the microfloral symbionts in the gut; see Smith & Douglas, 1987). The VFA concentrations were slightly lower than those reported for the rumen contents of sheep and cattle. Several basic requirements must be met in order for a vertebrate to maintain an efficient gut microflora (Bjorndal, 1987). These requirements include (1) constant, preferably elevated, body temperature; (2) constant food supply; (3) slow passage of digesta to allow sufficient time for microbial reproduction; (4) anaerobic conditions; (5) control of gut pH; and (6) removal of fermentation waste products. Most of these requirements seem to be readily met by the two kyphosid fishes. The staples of their diets, red and brown seaweeds, are abundant in their habitat. Gut transit time is a relatively slow 21 h, anaerobic conditions most likely prevail in the caecal pouches and the caecal pH values measured by Rimmer & Wiebe were in a fairly narrow range— 6.1–6.2 for K. cornelii and 6.3–6.7 for K. sydneyanus. Samples from the rectum contained many more lysed bacteria than did anterior regions of the gut, perhaps indicative of waste product removal.

BIOLOGY OF MARINE HERBIVOROUS FISHES

183

The temperature requirement seems to be the most restrictive because these fishes live at latitudes where temperatures would be expected to fluctuate to some extent with the season. Water temperatures, however, varied only 4°C (22–26°C) over the year at Rimmer & Wiebe’s (1987) collection site in coastal waters of Western Australia (Rimmer, 1986). Apparently this temperature range is sufficiently narrow to allow effective microbial fermentation in the digestive tracts of K. cornelii and K. sydneyanus. The digestive tracts of the two kyphosids do not appear to fit readily into any of the three other types of alimentary canals discussed above and first described by Lobel (1981). Pieces of algae are bitten off cleanly and swallowed intact without trituration. Although the stomach fluids are strongly acidic in both K. cornelli (pH 2.9–3.9) and K. sydneyanus (pH 2.8–3.0), lysis of algal cells was not apparent to Rimmer & Wiebe (1987). Algae were softened but retained their natural colour and intact appearance before passing from the stomach to the intestine. K. cornelii is known (Rimmer, 1986) to have a specialised gut by the time it settles on to the reef as a juvenile and to have a resident microflora at sizes less than 40 mm fork length. The fish’s gut lengthens rapidly during the transition from an omnivorous to a strictly herbivorous diet. Kyphosids, especially species in the genus Kyphosus, occur worldwide in warm temperate to tropical waters and frequently eat brown seaweeds that are considered to be (Littler, Taylor & Littler, 1983b; Estes & Steinberg, 1988) among the most chemically defended of marine plants. Comparison of digestive mechanisms among kyphosid taxa and investigation of their apparent tolerance to seaweed defences should be highly rewarding. Symbiotic microorganisms have also been found in the guts of surgeonfishes, especially Acanthurus nigrofuscus, in the Red Sea (Fishelson, Montgomery & Myrberg, 1985). The symbionts included bacteria, flagellates and an undescribed protozoan that attains extremely high densities in the gut. Fishelson and coworkers found no evidence of fermentation, but reasoned that the location of the protozoan near the gut lining and not in the food bolus meant that it is not involved in primary digestion. Two to three days of starvation in A. nigrofuscus caused the elimination of the symbiont, and subsequent feeding by starved fish that were released and observed foraging with other surgeonfishes did not re-establish the protozoan in the gut (Montgomery & Pollak, 1988). Small, unidentified flagellates have been found in the intestinal contents of Mugil cephalus by Odum (1970) who thought they probably served to break down the cellulose walls of plant detritus particles. When the flagellates disappeared from the gut after the fish was starved for 1–2 days, Odum concluded that they were digested by the fish. He quoted earlier speculation that the intestinal tracts of many fishes become sterile during periods when no food is ingested. Further study is required to understand the role of microbial symbionts in the guts of mugilids, acanthurids and other fishes. Conclusions on the types of alimentary canals The four kinds of alimentary tracts discussed above provide a general framework for most of the different groups of herbivorous fishes in the sea. Much more information, however, needs to be obtained on all of these groups, and further research will undoubtedly lead to modification and greater resolution of this basic model of digestive mechanisms. Gut pH seems certain to continue to be an important parameter of herbivore alimentary canals, both as an instantaneous signal of food processing conditions and as a predictive indicator of digestive function. Techniques should be devised to monitor continuously the pH in the different parts of the herbivore digestive tract without causing undue stress to the fish. Recent work showing that gut pH and even intestinal length are subject to short-term and local variations (Montgomery & Pollak, 1988) emphasises the dynamic nature of fish digestive systems and underscores the need for more detailed studies on the physiology and biochemistry of digestion in herbivorous fishes.

184

MICHAEL H.HORN

HERBIVORE GUTS AS CHEMICAL REACTORS One approach that holds promise for providing the needed greater resolution and precision on the functioning of alimentary canals in herbivorous fishes is that of treating the gut as chemical reactor. This avenue has recently been taken for animal guts in general and for deposit feeding invertebrates and mammals with fermentative digestion in particular (Penry & Jumars, 1986, 1987). The purpose here is not to provide a new analysis of the digestive tracts of herbivorous fishes in the context of chemical reactor models but rather to describe briefly the general classes of reactors and suggest which herbivore digestive mechanisms fit each of these classes. These tentative pairings might then serve as springboards for future research. An animal’s net rate of energy and nutrient gain are determined by foraging and digestion, two stages of a single process (Penry & Jumars, 1986, 1987). Optimality models have stressed foraging but neglected digestion, although, as Penry and Jumars indicate, digestive parameters are increasingly being taken into account (e.g. Milton, 1981; Taghon, 1981; Troyer, 1984; Horn, Neighbors & Murray, 1986). Penry & Jumars (1986) suggested that principles of chemical reactor theory can be used to formulate optimisation constraints in a general theory of digestion and then followed those principles (Penry & Jumars, 1987) to develop explicit models of digestion. These authors used performance equations and kinetic models of both enzymatic catalysis and fermentative digestion to reveal functional relationships among initial concentrations of the limiting food component, gut volume, gut retention time and digestive reaction kinetics. An optimality approach that combines foraging and digestion seems especially relevant to herbivore studies because food quality becomes a composite of a number of different measurable factors including concentration of limiting components, susceptibility to degradation by the animal’s enzymes or microbes, cost of producing enzymes or maintaining microbes and costs of adding new body tissues from breakdown products of particular foods. Three ideal theoretical models form the basis of all chemical reactor design: (1) batch reactors; (2) plugflow reactors; and (3) continuous-flow stirred-tank reactors. Batch reactors are filled with reactants, continuously stirred during the reaction, and then emptied of products after a given reaction period. In plugflow reactors, reactants enter continuously and products exit continuously with no mixing along the flow path. In continuous-flow stirred-tank reactors, reactants enter continuously and products continuously exit a stirred vessel. The guts of animals eating discrete meals can be classified as batch reactors and those of animals that eat continuously as either plug-flow or continuous-flow stirred-tank reactors. Herbivorous fishes with thin-walled acidic stomachs or with gizzard-like stomachs seem to fit best the continuous-flow stirred-tank model. In both of these groups ingestion and egestion rates are high in most cases, and the food is subjected either to acid lysis or trituration in the stomach (the stirred tank) before being passed into the rather long intestine. An apparent exception is the stichaeid Cebidichthys violaceus, which has a thin-walled acidic stomach. This fish seems to represent more closely the batch reactor model because of its intermittent feeding schedule, long gut retention times and generally sluggish lifestyle. Although Penry & Jumars (1987) claimed by their analysis that batch processing is an undesirable digestive mode, they admitted that a variety of animals use such a method, especially those with low metabolic requirements and those that are time minimisers constrained to widely separated foraging intervals. They listed hydras, hydroids, jellyfishes, sea anemones, corals, combjellies, some glycerid polychaetes, brittle stars, starfishes and chaetognaths as animals that can be modelled as having batch-reactor guts. Herbivorous fishes with a pharyngeal mill and no stomach seem to fit the plug-flow reactor model. These fishes have high ingestion rates, relatively short guts and brief gut transit times. Food is apparently neither stirred nor stored on route from the pharyngeal mill to the anus although intestinal peristalsis may provide some stirring action. Penry & Jumars (1987) pointed out that an animal dependent on its own digestive

BIOLOGY OF MARINE HERBIVOROUS FISHES

185

enzymes should function as a plug-flow reactor. The digestive tracts of geese, corophiid amphipods and deposit-feeding polychaetes with simple tubular guts approximate plugflow reactors (Penry & Jumars, 1986). Animals fermenting refractory materials (i.e. plants) should combine the continuous-flow stirred-tank and plug-flow reactors in series (Penry & Jumars, 1987). Foregut fermenters should begin digestion with the continuous-flow stirred-tank type of reactor, whereas hindgut fermenters should operate starting with the plug-flow type. Therefore, the kyphosid fishes with a hindgut fermentation chamber have guts that are best modelled as a combination of a plug-flow reactor and a continuous-flow stirred-tank reactor in sequence. The distinctions between foregut and hindgut fermenters are incomplete in the Penry-Jumars analysis, probably a reflection of the complexity of guts with both catalytic and fermentative modes of digestion. This brief account tentatively shows that all four of the chemical reactor models and combinations are represented among the alimentary types found in herbivorous fishes. A more detailed analysis of herbivore digestive mechanisms in the context of these models should be rewarding. INTESTINAL NUTRIENT TRANSPORT IN HERBIVOROUS FISHES A component of digestive mechanisms in herbivorous fishes not yet considered in this review is that of nutrient uptake across the intestinal wall. Does the uptake rate for different nutrients, in particular, carbohydrate and protein, vary between herbivores and carnivores? First of all, uptake rates in general are lower for fish intestines than those for mammal, bird and reptile intestines (Karasov, Buddington & Diamond, 1985). Moreover, intestinal glucose absorption appears to be highest in herbivores, intermediate in omnivores and lowest in carnivores within several vertebrate classes (Karasov et al., 1985) including fishes (Ferraris & Ahearn, 1983). These findings indicate that the rate of intestinal glucose transport is correlated with the carbohydrate content of the natural diet of each species. More recently, Buddington, Chen & Diamond (1987) studied the ratio of proline to glucose uptake in two carnivores (the salmonid Salmo gairdneri and the percichthyid Morone saxatilis), two omnivores (the ictalurid Ictalurus punctatus and the acipenserid Acipenser transmontanus) and four herbivores (the cyprinids Cyprinus carpio and Ctenopharyngodon idella, the cichlid Tilapia zillii and the stichaeid Cebidichthys violaceus). The sole marine herbivore was C. violaceus, which was treated both as a carnivore and a herbivore depending upon size because it is known (Montgomery, 1977; Barton, 1982; Horn, Murray & Edwards, 1982) to undergo an ontogenetic shift from carnivory to herbivory. The authors fed all eight species on the same manufactured diet and determined the uptake of proline and glucose by an in vitro, everted intestinal sleeve technique. They found that the ratio of proline to glucose uptake decreased from carnivores to omnivores to herbivores and that the intestine’s uptake capacity for glucose, a non-essential nutrient, was much higher in the herbivores than in the carnivores. This result suggests, as do the earlier studies cited above, that glucose transport is matched to the carbohydrate content of the natural diet. Proline uptake, however, varied less among the eight species, apparently because species with different natural diets still have similar protein requirements. Buddington et al. (1987) drew the conclusion that the species differences in uptake are genetic adaptations not phenotypic responses because all the species were fed the same diet. This study seems to have shown real differences in intestinal uptake among fishes of different trophic levels. The results, however, for the only marine herbivore in the study, Cebidichthys violaceus, are problematical because of the experimental conditions used and because of the interpretation of results for different-sized fish. C. violaceus was maintained in the laboratory at 19°C, and the in vitro experiments were performed at 20°C. These temperatures would seldom if ever be experienced by this fish in its cold

186

MICHAEL H.HORN

temperate coastal habitat (see Horn, Murray & Seapy, 1983; Murray & Horn, 1989), and C. violaceus is known to show increased stress and mortality at these temperatures in other laboratory experiments (Riegle, 1976). The size range of C. violaceus examined was approximately 100–370 mm total lenght (= fork lenght in this case). Although Buddington et al. (1987) designated for unexplained reasons fish shorter than 120 mm as carnivores and longer than 120 mm as herbivores, stomach content data (Montgomery, 1977; Barton, 1982; Horn et al., 1982) would indicate that all the fish they examined were herbivores. The diets of fish as small as about 55 mm (total length) appear to consist almost totally of macroalgae (Horn et al., 1982). The ratio of proline to glucose uptake for C. violaceus varied greatly, especially in the 100–120 mm size range, in which the highest values were 3–4 times greater than the lowest values. Perhaps this variation in the ratio of proline to glucose uptake represents, at the level of the gut wall, the rapid shift from a carnivore to a herbivore gut, but it leaves in question how presumably fast-growing juveniles in the size range between 55 mm and 100–120 mm survive as herbivores. The conclusions drawn by Buddington et al. (1987) about fish herbivory seem to be over generalised based on their examination of three freshwater species and one marine species. They state that herbivores distribute their absorptive tissue along the surface of a long thin intestine, whereas carnivores concentrate it in a short thick intestine or allocate much of it to the pyloric caeca. They then suggest that these differences in allocation of absorptive tissue between carnivores and herbivores arise from the herbivore’s very high food ingestion rates and fast intestinal transit times. Buddington and his co-workers reasoned that allocation of absorptive tissue to pyloric caeca is incompatible with fast transit times because food material would have to travel in and out of the blind caeca whereas the material transits the intestine only once. Nutrients would have little time to diffuse down to the deeper absorptive cells of the thick mucosa of the caeca and would thus be inefficiently assimilated. This scenario, however, is too narrow to accommodate the diversity and complexity of herbivore alimentary canals and digestive mechanisms surveyed in the foregoing parts of this review. Just two examples help to illustrate this point. First, all herbivores do not have fast gut transit times. C. violaceus, the marine herbivore studied by Buddington et al. (1987), requires more than 50 h to evacuate its digestive tract at 15°C (Urquhart, 1984), and Kyphosus Sydney anus, a hindgut fermenter, requires about 21 h at 23°C (Rimmer & Wiebe, 1987). Secondly, although herbivores do commonly have only a few (less than 10) pyloric caeca (see Suyehiro, 1942; Al-Hussaini, 1947) as Buddington et al. imply, these caeca are numerous in kyphosids (Al-Hussaini, 1947; see Fig 2, p. 219) and number more than 100 in girellids (Suyehiro, 1942; Bell, Burchmore, & Pollard, 1980; Anderson, 1988). Finally, on a more general point, all herbivores do not have long thin intestines; some species have relatively short, apparently thicker intestines (see Fig 2). As Penry & Jumars (1987) show in their plug-flow model, radial elaboration of absorptive surface area and lengthening of the gut are not equivalent tactics if the gut is considered to be a chemical reactor. A radial increase in the absorptive surface increases the uptake of digestive products at any point along the gut. Both a radial increase and an elongation of the gut increases the surface area, but lengthening increases the total gut volume as well. All other factors being equal, the resulting increase in retention time increases the extent to which food materials are converted to digestive products. Increase in surface area thus affects only the extent of absorption, whereas lengthening the gut influences both the extent of absorption and the extent of reaction. Studies of a wider range of herbivorous fishes are required before broad generalisations about digestive mechanisms in these fishes can be established. PROTEIN REQUIREMENTS OF HERBIVOROUS FISHES Do herbivorous fishes require as much protein in their diet as carnivorous species? This seems to be a reasonable question to ask given the low protein content of plants (Mattson, 1980) and the low rates of

BIOLOGY OF MARINE HERBIVOROUS FISHES

187

amino-acid transport in the intestines of herbivores (Buddington et al., 1987). Fishes, in general, have traditionally been considered to require more dietary protein than other vertebrate animals (Cowey & Sargent, 1979; Millikin, 1982; Pandian & Vivekanandan, 1985; Tacon & Cowey, 1985; Wilson & Halver, 1986). Whereas birds and mammals usually attain maximum growth on diets of 12–25% protein, fishes are reported to require diets with 35–55% protein to achieve maximum growth rate. Pandian & Vivekanandan (1985) argued that the low cost of maintaining body temperature and position in water, the ease with which nitrogenous waste is excreted as ammonia in water and the capacity to digest protein efficiently regardless of feeding habit have led to the high protein requirement and therefore to carnivory in the majority of fishes. They concluded that omnivorous, herbivorous and detritivorous fishes have the same high protein needs as carnivores, thus perpetuating the widely held view that herbivores and detritivores must digest microorganisms attached to their algal and detrital food to satisfy this high protein requirement. This view has recently been challenged by Bowen (1987) who, while not denying that herbivores may use attached microorganisms, cited examples (Dabrowski, 1982; Bowen, 1984a,b) indicating that many fishes that specialise as herbivores and detritivores in freshwater ecosystems maintain very large populations on low protein diets. The same can be said for marine herbivorous fishes on tropical (Odum & Odum, 1955; Hiatt & Strasburg, 1960) and temperate reefs (Russell, 1977, 1983). In a careful reassessment of the protein requirements of fishes and terrestrial homeotherms, Bowen (1987) concluded that the two animal groups differ only in relative protein concentration in the diet required for maximum growth rate. This difference can be explained by the greater energy requirement of birds and mammals and does not reflect absolute differences in protein requirement among these three vertebrate groups. Other measures of protein requirement examined by Bowen suggest that these groups are remarkably similar in their use of protein as a nutritional resource. According to Bowen (1987), two circumstances have led to the conclusion that fishes have high protein requirements (Bowen, 1987). First, the requirement for diets relatively high in protein has been misinterpreted as a requirement for higher absolute amounts of protein. Bowen’s analysis, however, indicates that the difference in dietary requirement of fishes and terrestrial homeotherms is in their need for energy, not protein. To clarify the relationship between protein and energy requirements, nutritionists working with vertebrates other than fishes have used protein, the less variable quantity, as the denominator to form an E/P ratio (e.g. Sell, Hasiak & Owings, 1985) rather than the P/E ratio often used by fish nutritionists such as Bowen (1984b). Second, most studies of digestive physiology and nutrition in fishes have focused almost entirely on carnivorous species, yet the results have often been generalised to fishes of all trophic levels. As more comparative studies similar to that by Buddington et al. (1987) are completed, the digestive and metabolic specialisations of different taxa and trophic levels can be better understood. According to Bowen (1987) the explanation for why many fishes on low protein diets achieve rapid population growth probably resides not in unique nutritional requirements but in food choice and digestive specialisations. Thus, the need for further research on the digestive mechanisms of herbivorous (and detritivorous) fishes is expressed again in this review. EVOLUTIONARY RESPONSES TO NITROGEN SHORTAGE In an article focused primarily on terrestrial herbivores (insects and vertebrates), Mattson (1980) reviewed the evidence that nitrogen is scarce and perhaps a limiting nutrient for many herbivores and that in response to this selection pressure many herbivores have evolved behavioural, morphological, physiological and other adaptations to cope with and use the nitrogen available in their environments. The responses in Mattson’s list are: (1) ability to find and use the most nitrogen-rich plants or plant parts; (2) increased consumption

188

MICHAEL H.HORN

rates; (3) prolonged periods of feeding, digestion and development; (4) specialised alimentary canals and digestive systems that rely on the presence and activity of endosymbionts or ectosymbionts; (5) occasional carnivory (cannibalism and predation); (6) switching among plant parts and plant species; (7) regulation of plant chemistry; and (8) evolution of larger body size. The purposes here are to consider briefly each of these responses in turn for marine herbivorous fishes and thereby provide a concise and somewhat speculative appraisal of the feeding behaviour and digestive mechanisms of these fishes in the context of nitrogen limitation. Some evidence is available to indicate that herbivorous fishes feed on the most nitrogen-rich plants or plant parts. Other factors, however, complicate food choice, as has been discussed earlier in the review. For example, the parrotfish Sparisoma radians feeds preferentially on the upper, older portions of turtlegrass blades with microalgal epiphytes (Lobel & Ogden, 1981). These parts of the plant are lower in nitrogen but also lower in phenolic compounds than the younger, basal leaf parts (Thayer et al., 1984). Many, if not most, plant-eating fishes have higher consumption rates than those of non-herbivorous species. These higher rates are usually accompanied by relatively short gut transit times. Important exceptions to this pattern have been discussed earlier in this review. Prolonged periods of feeding generally go together with increased consumption rates. Again, there are exceptions, as in the intermittently feeding stichaeids and kyphosids. Apparently, prolonged periods of digestion are not common in herbivorous fishes because the majority of species have short gut transit times. Food material is retained in the gut for relatively long periods in cold temperate species (stichaeids) and those with microbial fermentation (kyphosids), but no herbivorous fishes are known to have gut retention times as long as the 5.5- to 7-day periods of iguanid lizards (Nagy & Shoemaker, 1984; Troyer, 1984) and the 13-day period of the gopher tortoise (Bjorndal, 1987). A cellulolytic microflora is found in most, if not all, tropical iguanids and this temperate tortoise (Bjorndal, 1987). Prolonged periods of development, such as the extended life cycles of certain insects like the cicadas (Mattson, 1980), are not obvious features of herbivorous fishes, but many plant-eating fishes are relatively large (see p. 234) and may be long-lived and slow growing. For example, a 600-mm specimen of Kyphosus sydneyanus was estimated to be 49 years old based on counts of otolith growth rings (Ayling & Cox, 1982). Specialised alimentary canals and digestive systems that rely on endosymbionts are known for two species of kyphosid fishes. Microbes are also abundant in the intestines of mullets and certain acanthurids. Direct or indirect rôles of gut microbes in the digestive physiology of herbivorous fishes will probably become increasingly apparent as more species are studied in detail. Ectosymbionts, i.e., microorganisms living on the surfaces of seaweeds and seagrasses, most likely contribute directly to the nutrition of herbivorous fishes that consume these plants, but they also may play an important rôle by breaking down plant material and making it more usable to herbivores. Little definitive work has been done to demonstrate either of these processes for marine herbivorous fishes although mullets have been shown to digest bacteria and detritus and to have a resident microflora. Several lines of evidence suggest that occasional carnivory is a response to nitrogen shortage and therefore important to the nutrition of herbivorous fishes. First of all, carnivory on a small scale is inevitable for virtually all herbivorous fishes whether they are browsers or grazers because small animals encrust or otherwise dwell on the surfaces of foliose seaweeds or occur in close association with filamentous algae or diatoms on the substratum. The contribution of these small amounts of animal material to the nutrition of herbivorous fishes is essentially unknown. Omnivory, in which the diet is frequently composed of substantial proportions of both plant and animal material, is widespread among fishes and perhaps could be considered to be a response to nitrogen shortage but may also serve to dilute toxic materials that are ingested. Conversely, consumption of algal material by largely carnivorous fishes with acidic stomachs

BIOLOGY OF MARINE HERBIVOROUS FISHES

189

might serve as a filler that enhances digestion of animal material by stimulating flow of acid and thus lowering gastric pH. Perhaps the most telling evidence of all that carnivory is a response to nitrogen shortage is that all herbivorous fishes, with a few apparent exceptions such as the gobiescocid Sicyases sanguineus (Cancino & Castilla, 1988), probably begin life as carnivores. Even in Kyphosus cornelii, which apparently acquires its intestinal microflora as a small juvenile, some animal material contributes to the diet at this stage before the fish shifts to strict herbivory (Rimmer, 1986). White (1985) has gone so far as to assert that with few exceptions the young of all herbivores are not herbivores. Rather, they have near total dependence on access to animal or microbial protein for survival and growth. Finally, on a more anecdotal level, herbivorous fishes such as Girella nigricans and Medialuna californiensis (North, 1973), Odax pullus (Clements, 1985) and certain scarids (Bakus, 1969; Hay, 1981a) have been observed to feed on animal material offered to them in nature or captivity. Switching among plant species or plant parts has not been well demonstrated in herbivorous fishes apart from spatial or temporal changes in diet that seem largely driven by availability of preferred foods. In some cases these foods are richer in nitrogen and protein but, again, the reasons for switching are often influenced by a variety of factors. Herbivorous fishes are not known to regulate plant chemistry in a manner parallel to certain insects and ungulate mammals that secrete agents in their saliva that alter plant metabolism or promote plant growth (see Mattson, 1980). Nevertheless, like terrestrial herbivores, fishes, especially those such as territorial damselfishes, can enhance primary productivity presumably by their grazing or browsing activities (Klumpp, McKinnon & Daniel, 1987) and may increase local nitrogen concentrations within their territories through excretion and defaecation (Polunin & Koike, 1987). Circumstantial evidence for the evolution of larger body size in herbivorous fishes is seen in data on diet and maximum body length compiled for three families, Odacidae, Pomacentridae and Stichaeidae, each composed of members with different feeding habits (Table VI). In each case the largest two or three species are herbivores, whereas the smallest species are carnivores. Other vertebrates, including reptiles (Pough, 1973; Rand, 1978), birds (Morton, 1978) and mammals (Eisenberg, 1978), also show such a relationship between body size and dietary mode. Larger body size confers several advantages on the animal, and these may all be necessary on a nitrogenpoor diet (Mattson, 1980). Mattson cites the following advantages of increased body size. (1) Larger body size increases TABLE VI Maximum body lengths of species in three marine fish families with carnivorous, herbivorous and omnivorous (Pomacentridae only) members Family—species

Trophic position

Maximum body length, mm Remarks

Odacidae, Australia and New Zealand. Gomon & Paxton. 1985 Odax pullus Herbivore 750

All sizes are standard lengths except total length for O.pullus from Ayling & Cox (1982). The list of herbivorous species is from M.J.Kingsford (pers. comm.). S.argyrophanes is an extremely slender species. All species in the family are listed.

190

MICHAEL H.HORN

Family—species

Trophic position

Siphonognathus Carnivore argyrophanes Odax cyanoallix Herbivore Odax cyanomelas Herbivore Haletta semifasciata Carnivore Odax acroptilus Herbivore Siphonognathus radiatus Carnivore Neoodax balteatus Herbivore Siphonognathus attenuatus Carnivore Siphonognathus beddomei Carnivore Siphonognathus tanyourus Carnivore Siphonognathus caninus Carnivore Pomacentridae, Florida Keys, Emery, 1973 Abudefduf taurus Herbivore

Maximum body length, mm Remarks 400 350 350 290 240 180 140 120 120 110 100 250

Microspathodon chrysurus Herbivore 200 Abudefduf saxatilis Omnivore 180 Chromis multilineata Carnivore (zooplanktivore) 165 Eupomacentrus fuscus Omnivore 150 Eupomacentrus planifrons Herbivore 125 Eupomacentrus variabilis Omnivore 125 Chromis cyanea Carnivore (zooplanktivore) 125 Eupomacentrus partitus Herbivore 100 Eupomacentrus Omnivore 100 leucostictus Chromis enchrysurus Carnivore (zooplanktivore) 100 Chromis insolatus Carnivore (zooplanktivore) 100 Chromis scotti Carnivore (zooplanktivore) 100 Stichaeidae, Alaska to Baja California, Eschmeyer, Herald & Hammann, 1983 Cebidichthys violaceus Herbivore 760

All sizes are total lengths from Randall (1968) or Robins, Ray & Douglass (1986). All species studied by Emery (1973) are listed except an unidentified species of Eupomacentrus.

All sizes are total lengths, C. violaceus and X. mucosus are known to be herbivores (Horn, Murray & Edwards, 1982; Barton, 1982). All other species are listed as carnivores because of lack of evidence of herbivory. The species listed include all those

BIOLOGY OF MARINE HERBIVOROUS FISHES

Family—species

Trophic position

Xiphister mucosus Lumpenus sagitta Chirolophis decoratus Stichaeopsis sp. Lumpenella longirostris Bryozoichthys marjorius Xiphister atropurpureus Poroclinus rothrocki Stichaeus punctatus Anoplarchus purpurescens Phytichthys chirus

Herbivore Carnivore Carnivore Carnivore Carnivore Carnivore Carnivore Carnivore Carnivore Carnivore Carnivore

191

Maximum body length, mm Remarks reported by Eschmeyer, Herald & Hammann (1983) to occur in the region. 580 510 420 320 310 300 300 250 220 200 200

Family—species

Trophic position

Maximum body length, mm

Plagiogrammus hopkinsii Lumpenus maculatus Chirolophis nugator Anisarchus medius Kasatkia sp. Plectobranchus evides Anoplarchus insignis Allolumpenus hypochromus

Carnivore Carnivore Carnivore Carnivore Carnivore Carnivore Carnivore Carnivore

200 180 150 140 140 130 120 75

Remarks

efficiency of locomotion thus permitting wider foraging for higher volume consumption. (2) Larger animals have lower mass-specific energy requirements and therefore need to extract less energy per unit of ingested material than do smaller animals. (3) Lower respiration rates mean that on equivalent diets larger animals should have higher food conversion efficiencies than smaller species. (4) Larger size may also confer mechanical advantages in foraging because low nitrogen levels usually accompany increased toughness of plant tissues. (5) Finally, larger body size may be necessary for the development of complex digestive systems important for breaking down nitrogen-poor, refractory plant foods. Both Demment & Van Soest (1985) and Penry & Jumars (1987) provide evidence to show that the extent to which refractory plant material is digested increases with body size. All of these advantages seem as applicable to fish herbivores as to other plant-eating vertebrates. ECOLOGICAL IMPACTS OF HERBIVOROUS FISHES Herbivorous fishes by their grazing and browsing activities affect shallow water benthic communities, especially those of tropical reefs, in a variety of ways. Their scrapings and excavations of hard substrata

192

MICHAEL H.HORN

contribute to erosional and sedimentary processes and their intensive cropping of algal stands influences the composition, abundance and spatial patterns of seaweed populations. In addition, territorial defence, especially as exhibited by damselfishes, protects algal stands from other, roving herbivorous species and creates a mosaic of lush algal growth inside territories and heavily cropped zones outside the territories. These ecological impacts in coral reef habitats represent perhaps the most intensively studied topic in all of fish herbivory over the past 15–20 years (see recent reviews by Sale, 1980; Borowitzka, 1981; Lubchenco & Gaines, 1981; Gaines & Lubchenco, 1982; Hixon, 1983, 1986; Chapman, 1986). This section of the present review is divided into tropical and temperate segments and concentrates on tropical habitats, especially coral reefs, where the great majority of the studies have been conducted. The topics discussed are the contribution of herbivorous fishes to bioerosion and sedimentation on coral reefs, the distributions of herbivorous fishes across the reef environment, the impacts of roving (largely non-territorial) fish herbivores on the structure of seaweed communities, the relative importance of grazing by fishes and sea urchins in reef habitats and the effects of territorial damselfishes on the diversity and abundance of algal assemblages and other reef inhabitants. TROPICAL HABITATS Bioerosion and sediment formation Grazers were defined earlier in the review as those fishes (and invertebrates) that in the course of feeding ingest some inorganic material from the substratum. They bite, rasp or scrape off pieces of the substratum and subsequently excrete the material as sediment. The effects of this grazing on the erosion of coral reefs and on the production of carbonate sediments has been documented for fishes (Cloud, 1959; Bardach, 1961; Gygi, 1969; Glynn, 1973; Randall, 1974; Ogden, 1977; Hutchings, 1986; Sammarco, Carleton & Risk, 1986) and for a variety of invertebrates (Scoffin et al., 1980; Hutchings, 1986). Parrotfishes and surgeonfishes are the major bio-eroders and sediment producers among herbivorous fishes (Hiatt & Strasburg, 1960; Randall, 1967, 1974; Gygi, 1969, 1975; Ogden, 1977; Frydl & Stearn, 1978, Sammarco et al., 1986). With their heavy beaklike jaws and pharyngeal mill, parrotfishes break down pieces of dead coral skeleton and calcareous red and green algae into fine sand and silt (Ogden, 1977). They digest the algal tissue and void as faeces the fine, inorganic material, which has been shown to comprise an average of 70% of the dried gut contents of Caribbean scarids (Randall, 1967). Their high ingestion rates and fast gut transit times are typical of many tropical herbivorous fishes and result in the frequent passage (Hiatt & Strasburg, 1960) of faeces as great masses of calcareous powder. Ogden (1977) determined that at Isla Pico Feo in Caribbean waters of Panama Scarus croicensis fed 8 h each day (data from Ogden & Buckman, 1973) and defaecated 50 mg (dry weight) in each of 20 defaecations per day. Based on an estimated population of 3200 at Pico Feo, S. croicensis would deposit 7300 kg/year of sediment on the reef. Assuming that half of the carbonate material deposited consists of sediment already present on the reef, Ogden calculated that the fish produces new carbonate sediment at the rate of 0.49 kg.m−2.yr−1. This estimate, however, was considerably lower than those he calculated for the sea urchins Diadema antillarum (4.6 kg.m−2.yr−1) and Echinometra lucunter (3.9 kg.m−2.yr−1). Although Ogden (1977) admitted that estimates of turnover for both fish and sea urchin grazers were subject to some major sources of error, he concluded that such grazers are important contributors to geological and biological processes on coral reefs and therefore to the physical structure of reef ecosystems. The surgeonfishes that ingest sand are grazers on filamentous algae, diatoms and detritus (Jones, 1968). These fishes have thick-walled, muscular stomachs in which they presumably use the inorganic material to

BIOLOGY OF MARINE HERBIVOROUS FISHES

193

grind the food into finer particles to make more of the cell contents available for digestion (Hiatt & Strasburg, 1960; Randall, 1967; Jones, 1968). Surgeonfishes in the genus Ctenochaetus are sediment feeders and have long, flexible teeth with spatulate tips, gill rakers bearing many very fine teeth and bristles and a muscular, gizzard-like stomach (Jones, 1968). The teeth on the upper and lower jaws function as opposing brooms, and the fishes appear to use them to sweep and suck up fine material from the reef. The flexibility of the teeth apparently allows the fish to scrape irregular rock surfaces and comb attached algal thalli to remove adhering detritus and diatoms (Jones, 1968). A member of the genus Ctenochaetus, C. striatus, recently has been reported to reduce the sizes of ingested carbonate particles on the reef flat at Moorea, French Polynesia (Nelson & Wilkins, 1988). Mean particle size in stomach samples of this fish were smaller than those in the feeding area, and particle sizes in the rectal contents were smaller than those in stomach samples. Nelson & Wilkins hypothesised that grinding of ingested materials by the muscular stomach of C. striatus was responsible for the observed reduction in size of particles as the particles pass through the gut. This view is consistent with that of Randall (1974) who asserted that particle size is undoubtedly reduced in the stomach of surgeonfishes with muscular stomachs. That particle size is smaller in the rectum than in the stomach of C. striatus is the expected result of grinding action in the stomach, but the reduction in particle size from the sediments to the stomach could also occur if the fish selects smaller particles when feeding. Nelson & Wilkins acknowledged this possibility but failed to mention the elaborate, finely divided gill raker apparatus of C. striatus that Jones (1968) proposed would be highly effective in handling the fine particulate matter ingested. Mullets, which also have an average particle size in their stomachs that is smaller than those in the sediments, apparently feed selectively on the smaller, nutrient-rich particles (Odum, 1968; Marais, 1980). Mullets have a pharyngeal filtering device (Ebeling, 1957) that they use to select the very fine particles from sediments of various sizes (Odum, 1968). Whether the surgeonfish C. striatus selectively ingests smaller particles or mechanically reduces particle size in the stomach or both remains an important but unanswered question. Regardless of the process, C. striatus produces smaller particles than those found in ingested sediments and, therefore, probably has a local impact on the particle-size distributions of sediments within reef and lagoon habitats. Nelson & Wilkins (1988) found that C. striatus has relatively low assimilation efficiencies (Table II) for total organic matter (20%) and nitrogen (37%) but cited evidence (Barlow, 1974b) that Ctenochaetus species have high feeding rates, which may compensate for the low assimilation efficiencies. Species in this genus are also known to be among the most abundant fishes on tropical reefs (e.g. Jones, 1968; Bouchon-Navaro & Harmelin-Vivien, 1981; Robertson & Gaines, 1986). Thus, given the high abundance and probable high feeding rate of C. striatus, it is reasonable to predict that the fish significantly influences particle size distributions in coral reef habitats. One consequence of particle size reduction is that finer sediments have more surface area available for microbial colonisation, which enhances sediment productivity (Moriarty et al., 1985; Yamamoto & Lopez, 1985). Impacts of roving herbivorous fishes Free-ranging herbivorous fishes are not uniformly distributed across tropical reefs, and the spatial patterns they form are relevant to a consideration of the impacts of these fishes on algal communities. Spatial distributions on coral reefs Plant-eating fishes are most common at the shallower depths on tropical reefs. Ranges of 0–6 m (Gosline, 1965) and 0–10 m (Bakus, 1969) have been given as the depths of greatest abundance although some species can be found at considerably greater depths. In a survey of deep reefs (30– 180 m) by small submersible at Enewetak, Marshall Islands, Thresher & Colin (1986) found that

194

MICHAEL H.HORN

Fig 3.—Abundances of five guilds of herbivorous fishes in five sections of six different coral reefs on the central Great Barrier Reef; black bars represent midshelf reefs, white bars represent outershelf reefs; suckers and scrapers are types of grazers, and croppers are browsers; suckers of fine sediments include surgeonfishes in the genus Ctenochaetus; sand suckers include the surgeonfishes Acanthurus dussumieri, A.mata and A.nigricauda; scrapers include several species of parrotfishes in the genus Scarus, e.g. S. brevifilis, S. gibbus, and S. rivulatus; large croppers include the surgeonfishes Acanthurus lineatus and Naso lituratus and the rabbitfish Siganus doliatus; small croppers include the surgeonfishes Acanthurus glaucopareius and A. nigrofuscus; from Russ (1984b).

herbivorous fishes were most abundant at the shallowest depth (30 m) of the survey and disappeared by 90 m. Although space limitations or limited recruitment rates have been proposed (Sale, 1980) as more important than food abundance in affecting the distribution and abundance of reef fishes, Thresher & Colin (1986) asserted that the depth distributions of herbivores (and coralivores) are ultimately constrained by their food supply, i.e., by light-dependent organisms. Acanthurids and scarids, the two major groups of large, roving herbivorous fishes associated with coral reefs, have generally different depth distributions. Acanthurids tend to predominate in shallow habitats, whereas scarids occur more prominently in deeper zones. This pattern has prevailed in studies in the Caribbean (Randall, 1963; Lewis & Wainwright, 1985), Red Sea (Bouchon-Navaro & Harmelin-Vivien, 1981) and Indo-Pacific (Bradbury & Goeden, 1974; Jones & Chase, 1975) and, therefore, may be typical of acanthurid and scarid distributions in many reef systems. These depth differences combined with differences in feeding behaviour and food preferences of these two major members of the roving herbivore guild suggest that algal populations in different depth zones are subject to different grazing and browsing pressures. The non-uniform spatial distributions of mobile herbivorous fishes, however, are not explained simply by reef bathymetry, but also by reef topography, local availability of food resources and other physical features and biological interactions (Lewis & Wainwright, 1985). Recent studies on reefs in widely separated geographic regions lend support to this statement. On the central Great Barrier Reef, Russ (1984a) found inshore reefs to have fewer individuals and species of acanthurids and scarids than mid- and outershelf reefs and proposed that changes in food availability across the shelf may be the best explanation for the pattern. The assemblage studied by Russ fed primarily on filamentous and highly productive turf-forming algae that predominate on the mid- and outershelf reefs. Inshore reefs were characterised by larger, probably less palatable seaweeds such as the brown algae Sargassum and Turbinaria. In a related study focusing on five reef zones, Russ (1984b) found distinctive assemblages of herbivorous fishes on reef slopes, reef flats, reef crests, and over sandy areas in both lagoon and back reef zones. Acanthurids and scarids generally had higher numbers of species and individuals on reef crests and in lagoons than on reef flats or reef slopes, whereas siganids were more diverse and abundant in lagoons and back reefs. Different guilds of herbivorous fishes were also distributed differently among the zones (Fig 3): (1) suckers feeding on fine sediments were most abundant near windward and leeward edges of reefs; (2) suckers feeding over sand were most abundant in back reefs and lagoons; and (3) croppers and scrapers were more numerous in shallow zones (reef crest, reef flat, lagoon) than in deep zones (reef slope, back reef). Interspecific interactions and local-scale heterogeneity further complicate spatial distributions of herbivorous fishes as seen in a study (Choat & Bellwood, 1985) on the Great Barrier Reef of the territorial surgeonfish Acanthurus lineatus, one of the large croppers in the studies of Russ (1984a,b) described above. This reef-crest species is strongly site-attached and aggressive toward other herbivorous fishes. Behavioural observations of these fishes at two adjacent sites of high A.lineatus abundances revealed a complex pattern of interactions among species. Large scarids, including Scarus gibbus, a scraper (Russ, 1984b), fed within Acanthurus lineatus territories at one site but were rare at the other. A.nigrofuscus, a small cropper (Russ,

BIOLOGY OF MARINE HERBIVOROUS FISHES

195

1984b) and a consistent target of aggression by A.lineatus, was moderately common at one site but rare at the second site. Ctenochaetus striatus, an abundant surgeonfish and a sucker of fine sediments (Russ, 1984b), was present in high densities at both sites and fed within Acanthurus lineatus territories. Although the degree of aggression by A. lineatus seems to be somewhat related to feeding guild membership of the other species, Choat & Bellwood (1985) emphasised the importance of small but consistent differences in reef structure at each site. They hypothesised that localised differences in the within-habitat component of acanthurid and scarid abundances and distributions reflect site-associated variability in recruitment, postrecruitment mortality or behaviour independent of the activities of A. lineatus. In the Red Sea (Gulf of Aqaba), acanthurids were found to dominate on the reef flat, whereas scarids were more abundant on the outer reef slope (Bouchon-Navaro & Harmelin-Vivien, 1981). In this study the observed complexity in the spatial distributions of fish herbivores was related in large part to differences in social structure of the fishes according to reef biotope. Surgeonfishes formed dense schools in the reef zones nearshore, smaller schools seaward and behaved as solitary individuals on the outer slope. Parrotfishes displayed similar but less marked patterns of schooling versus solitary behaviour in the different reef zones. The occurrence of both schooling and solitary individuals within the same species has been reported in other studies of acanthurids (Randall, 196la; Barlow, 1974b; Vine, 1974) and scarids (Ogden & Buckman, 1973; Barlow, 1975; Robertson et al., 1976). The schools of both acanthurids and scarids generally comprised adults at the Red Sea site (Bouchon-Navaro & Harmelin-Vivien, 1981). Juvenile acanthurids were concentrated in shallow (less than 10 m) waters primarily as solitary individuals, whereas young scarids occurred in the shallows but also as deep as 30 m on the outer reef slopes, either as solitary individuals or in small aggregations. Another complicating factor in the equation for explaining the spatial distributions of fish herbivores on tropical reefs is the risk of predation. Although relatively few data are available, Hay (1985) pointed out that predators could have a strong impact on the spatial patterns of habitat use by herbivorous fishes (and sea urchins) and thus indirectly help form the mosaic pattern of grazing on reefs. Grazing and browsing fishes may avoid deep sections of reefs because of the reduced algal food supply, as discussed above, but also because of the heightened risk of predator attack at decreased light levels (see Hobson, 1972). Herbivorous minnows avoid both shallow areas of streams where they be attacked by aerial or terrestrial predators (Power, 1984) and areas near large predatory fishes (Power & Matthews, 1983). Similar behaviour by marine herbivorous fishes is suggested by the results of several studies. For example, herbivorous fishes may avoid structurally simple reef flats (Hay, 1984a), deep sand plains (Earle, 1972; Dahl, 1973; Hay, 1981b; Hay, Colburn & Downing, 1983) and shallow sandy lagoons (Randall, 1965; Hay, 198la, 1984b) because they provide few hiding places and, therefore, leave these fishes more vulnerable to predators. On the other hand, Hay (1985) also provided evidence that herbivore feeding rates are higher in more open reef areas than areas near overhangs, which may provide ambush sites for predators. In terms of seaweed distributions, such a predatory threat to herbivores may result in small safe patches for seaweeds to colonise even though they would be rapidly eaten in immediately adjacent areas. In summary, these studies illustrate some of the complex interactions that are involved in determining the spatial distributions of roving herbivorous fishes across reef habitats. They provide an appropriate background for the following discussion of the effects of these herbivores on coral reef algal communities. Effects on algal communities Two striking features of coral reefs are the abundance of animals and the scarcity of conspicuous plants (see Randall, 1967; Earle, 1972; Menge & Lubchenco, 1981). What are the causes of such a low standing biomass of seaweeds and the prominence of small plants in these highly productive tropical habitats? Gaines & Lubchenco (1982) discuss this question at length and identify two main causal factors: low nutrient levels and intensive feeding by herbivorous fishes and sea urchins. They

196

MICHAEL H.HORN

concluded that nutrient levels may set limits on the total size (and thus biomass) of algae, but that herbivory can limit algal size and biomass well below the constraints imposed by low nutrients. Fishes and sea urchins, by their foraging on larger, upright seaweeds, can prevent crustose and small filamentous algae from being overgrown by the erect species. The free-ranging herbivorous fishes, in particular, parrotfishes, surgeonfishes and rabbitfishes, appear to have great overall impacts on reef-inhabiting algal communities. Gaines & Lubchenco (1982) drew attention to their well-developed visual capabilities, large and powerful feeding apparatus and high potential mobility as traits making fishes important tropical herbivores. Roving fishes are thus able to decrease spatial variability in the effects of herbivory on community structure. More specifically, they can reduce between-habitat algal diversity across the range of reef habitats where they occur as compared to areas with less mobile herbivores. Ogden & Lobel (1978) suggest that fishes with these traits and capabilities are similar to herbivorous mammals in a number of ways. For example, these fishes have diversified their diets and probably learn quickly to feed mainly on familiar foods. Nevertheless, they can be expected to remain flexible in order to cope with temporal changes in seaweed availability. They appear to be able to tolerate some algal secondary compounds because they consume certain chemically defended species. Recent studies of food preferences show, however, that herbivorous fishes tend to choose plants species or plant parts with small amounts of secondary metabolites. That these mobile herbivorous fishes are intensive feeders and greatly affect the abundance, diversity and distribution of reef algae has been demonstrated in numerous field experiments (see reviews by Sale, 1980; Borowitzka, 1981; Hixon, 1983, 1986). A few representative studies are discussed here. Two of the earliest experimental studies were by Stephenson & Searles (1960) in the intertidal zone at Heron Island on the Great Barrier Reef and Randall (1961b) in shallow subtidal waters on the island of Hawaii. In both cases algal populations showed rapid and conspicuous increases in standing crop when protected by exclosures from herbivorous fishes. At Heron Island, rabbitfishes were the most important of several species of grazing and browsing fishes on intertidal beach rock, and on Hawaii, surgeonfishes were the main species controlling algal density. Within 48 h after removal of the exclosures rabbitfish toothmarks were visible in all the formerly caged areas. Several other investigations have shown that algal abundances are reduced by foraging fishes (John & Pople, 1973; Wanders, 1977; Montgomery, Gerrodette & Marshall, 1980; Miller, 1982; Hatcher & Larkum, 1983; Lewis, 1986). Grazer-resistant crustose forms have been observed to become dominant on grazed surfaces in many studies (John & Pople, 1973; Wanders, 1977; Brock, 1979; Hixon & Brostoff, 1981, 1985; Lewis, 1986). Miller (1982), however, recorded increased coverage by filamentous blue-green algae and a diatom-bacterial film in a heavily grazed area of an intertidal reef at Enewetak, Marshall Islands. Reduced algal density is common, but algal diversity can also vary as a function of fish browsing and grazing as has been documented in several studies (Stephenson & Searles, 1960; John & Pople, 1973; Day, 1977). Algal diversity varied according to different intensities of scarid grazing pressure (as measured by different fish densities) in experiments conducted by Brock (1979). At low fish densities, algal diversity was low because of dominance by a few species; at intermediate densities, the highest algal species richness was obtained; and at high fish densities, low algal diversity developed. Brock (1979) suggested that parrotfish may serve as keystone species (sensu Paine, 1966) on some Hawaiian reefs. Factors that seem to influence the distributions of herbivorous fishes on coral reefs in turn affect the spatial distributions of algal species. Lubchenco & Gaines (1981) pointed out that some seaweed species normally excluded from many reef slopes by fish grazing often persist on reef flats where shallow water restricts fish foraging (Hay & Goertemiller, 1983), on sand plain areas that lack protection from predators for many herbivorous fishes (Earle, 1972; Dahl, 1973; Hay, 1981b; Hay et al., 1983) or in areas of intense

BIOLOGY OF MARINE HERBIVOROUS FISHES

197

wave action or surge that also limit fish feeding (Van den Hoek, Breeman, Bak & Van Buurt, 1978; Ruyter van Steveninck, Kamermans & Breeman, 1988). The narrow zone of bare sand often seen between patch reefs and seagrass beds in the Caribbean appears to be the result of heavy grazing by parrotfishes and surgeonfishes (Randall, 1965) or by sea urchins (Ogden, Brown & Salesky, 1973) that stay close to the reefs for shelter from predatory fishes. Finally, territorial fishes, especially damselfishes, strongly influence the foraging ranges of the roving herbivorous fishes and the spatial mosaic of seaweed species. The responses of seaweeds to the distributional patterns and feeding activities of herbivorous fishes appear to be more extensive than just variations in spatial distributions and population densities. Transplant experiments have shown that algal species characteristic of low-herbivory spatial refuges are highly susceptible to being eaten by surgeonfishes and parrotfishes (Lewis, 1986). Conversely, algae from habitats of high herbivore feeding intensity seem to be resistant to fish browsing and grazing (Hay, 1981a, 1984a; Littler, Littler & Taylor, 1983a; Lewis, 1985). Lewis (1986) saw these results as consistent with the hypothesis of a trade-off in algal resource allocation between competitive ability and herbivore resistance as has been proposed for other prey assemblages (Lubchenco, 1978; Littler & Littler, 1980; Hay, 1981b; Lubchenco & Gaines, 1981). Morphological plasticity as seen in the brown alga Padina jamaicensis has been interpreted as another kind of response to different herbivore grazing pressures (Lewis, Norris & Searles, 1987). This alga persists in a prostrate, turf morphology under high grazing intensity but responds rapidly (96 h) to reduced herbivory by shifting to an erect foliose morphology. The latter morphology was shown to be consumed preferentially by parrotfishes in transplant experiments. Thus, P.jamaicensis has the phenotypic plasticity enabling it to respond rapidly to temporal or spatial fluctuations in herbivory. Comparative impacts of roving herbivorous fishes and sea urchins Numerous studies have clearly established that mobile herbivorous fishes, especially acanthurids and scarids, and echinoids, especially those in the genus Diadema, exert strong influences on algal communities in coral reef habitats. On the Great Barrier Reef, the fishes appear to be the major grazers (Stephenson & Searles, 1960; Day, 1977; Borowitzka, 1981; Hatcher, 1981; Russ, 1984a,b; Sammarco, 1983). In the Caribbean, both fishes and sea urchins have been recognised as playing important roles as primary consumers (cf. Ogden et al., 1973; Sammarco, Levinton & Ogden, 1974; Ogden, 1976; Sammarco, 1982a,b; Carpenter, 1981, 1983, 1986, 1988; Steneck, 1983; Hay, 1984a,b; Hay & Taylor, 1985; Foster, 1987; Hughes, Reed & Boyle, 1987; Levitan, 1988). On the Great Barrier Reef, Hatcher (1981) estimated that about 60% of the daily algal production was removed by herbivores, with fishes being the only significant consumers. On a Virgin Islands reef, in contrast, Carpenter (1986) reported that of the 97% removal of algal production by herbivores, the urchin Diadema antillarum removed a slightly larger proportion than did the herbivorous fishes (mainly juvenile parrotfishes). The percentage of the primary production entering the grazing food chain on this Caribbean reef is the highest reported for any ecosystem (Carpenter, 1986; see Wiegert & Owen, 1971). The relative magnitude of the grazing impacts of fishes and sea urchins varies with the locality and other conditions. This variation has led to different interpretations and to controversy about the relative influence of these two herbivore groups on Caribbean reefs. Hay (1984a,b) stirred debate (see Hixon, 1985) when he questioned earlier studies (e.g. Ogden et al., 1973; Sammarco, 1982a,b) purporting to show that sea urchins, especially D. antillarum, were the most important grazers on typical coral reefs. Hay proposed that the results obtained at the particular sites of these studies were atypical because they were on reefs heavily overfished by humans, thus resulting in unusually high densities of sea urchins that had been from predation by and perhaps competition with fishes. Using strips of the seagrass Thalassia testudinum as a field

198

MICHAEL H.HORN

bioassay for the intensity of herbivory, Hay found that on seven lightly fished reefs scattered throughout the Caribbean grazing intensity decreased with depth and almost all the herbivory resulted from fish grazing. Sea urchin densities were low on these reefs. In contrast, he found the opposite, traditional pattern on four heavily fished Caribbean reefs. Hay stressed that on reefs affected by human activity many patterns may be recent, having existed only for the past few hundred years, and therefore that positing evolutionary implications from ecological studies conducted on heavily fished reefs is unjustified. These assertions, however, have been criticised (see Hixon, 1985), and the Thalassia bioassay technique has been shown to produce results contradictory to other measurements of herbivory (Steneck, 1983). Hay (1984a) himself admitted that the technique under-estimated browsing by smaller herbivores such as damselfishes and young surgeonfishes because it is difficult for them to bite through the Thalassia blade. The mass mortality of Diadema antillarum, which occurred throughout the Caribbean in 1983–1984 (Lessios, Robertson & Cubit, 1984), provided a unique opportunity to compare the impacts of grazing sea urchings and fishes. Studies following this mass mortality (Hughes et al., 1987; Carpenter, 1988; Levitan, 1988) appear to reinforce the earlier impression that D. antillarum can have a dramatic influence on coral reef habitats, perhaps often greater than that of herbivorous fishes in localised areas. For example, Carpenter (1988) showed that within five days after the die-off at St Croix, Virgin Islands, algal biomass increased by 20%, algal community primary productivity dropped by 37% (per unit area) and 61% (per unit biomass), and algal biomass removed by herbivores declined by 50%. These changes in the algal community were accompanied by increases in the rates of grazing by herbivorous fishes (mostly juvenile scarids), suggesting that exploitative competition for food occurs between Diadema and some herbivorous fish species. The criticisms and controversy surrounding the studies comparing the impacts of fishes and urchins led Hixon (1985) to identify three basic goals for future studies of coral reef communities: (1) make greater use of properly controlled field experiments; (2) undertake long-term field studies; and (3) include several study sites in order to determine the extent of observed patterns at a single site. Spatial and temporal variations in reef habitats and in the relative abundance of the different resident herbivores provide strong support for Hixon’s recommendations. Impacts of territorial damselfishes Unlike the roving herbivorous fishes, territorial pomacentrids closely tend a patch of the reef substratum and maintain a growth of algae in the territory that is suitable as food. As a result of these habits, damselfishes exert a pronounced influence on the structure of benthic algal communities on coral reefs. These fishes are conspicuous and frequently pugnacious residents of shallow water reef habitats. Thus, they are readily accessible for study and have been the focal point of a large number of manipulative field experiments carried out over the past 15 years. Their abundance alone suggests that they play an important role in coral reef communities. For example, as much as 40% to 50% of the surface of certain reef habitats can be under the influence of a single species of territorial damselfish (Sammarco & Williams, 1982; Klumpp, McKinnon & Daniel, 1987). By establishing and defending distinct algal mats, territorial damselfishes affect (1) coral recruitment, growth, and bioerosion; (2) local microfaunal abundance; (3) nitrogen fixation by blue-green algae; and (4), algal abundance and local diversity (Hixon, 1983). Each of these effects are discussed in turn below, followed by an account of recent studies on the effects of these fishes on algal productivity. The effects of territorial damselfishes on coral recruitment, growth and bioerosion are varied and complex. Birkeland (1977) reported that algal mats inhibit settlement by corals, and Potts (1977) speculated

BIOLOGY OF MARINE HERBIVOROUS FISHES

199

that the scarcity of the coral Acropora palifera at a Great Barrier Reef site was caused primarily by the territoriality of the damselfish Dischistodus perspicillatus. By encouraging the growth of filamentous algae, D. perspicillatus presumably inhibits settlement and then suppresses the growth of the young coral survivors. Both Vine (1974) and Lobel (1980) claimed that inhibition of coral growth inside territories could lead to a weakening and eventual collapse of the basal reef network. Moreover, Risk & Sammarco (1982) found that pieces of staghorn coral (Acropora) inside the territories of two damselfish species were significantly more eroded than coral pieces outside territories, with most of the destruction caused by boring sponges and sipunculids. They suggested that bioerosion is accelerated within territories as a result of reduced grazing and predation on the borers by fishes. Irvine (1982) observed that Eupomacentrus planifrons actually kills corals by its biting activity and thereby opens up new substrata for algal colonisation. Wellington (1982) showed that the coral killing activities of E. acapulcoensis reduces or eliminates the massive coral Pavona from shallow waters thus promoting establishment of the branching coral Pocillopora, which it does not attack. In contrast to the foregoing studies that largely report negative effects of territorial damselfishes on corals, other works have revealed positive or neutral effects. Sammarco & Carleton (1981) found that the density of coral spat within damselfish territories was about five times higher than on exposed surfaces and in caged areas at 10–11 m and that diversity of coral spat was greater in these defended areas. And, although territories may be sites of high local mortality of coral spat, these defended areas may also serve as refuges for rare species (Sammarco & Williams, 1982). The net effect would be an overall increase in coral diversity in the community. Finally, in studies on the Great Barrier Reef, external bioerosion (i.e. erosion caused by grazing fishes) was found to be effectively reduced within the borders of damselfish territories (Sammarco, Carleton & Risk, 1986), whereas total rates of internal bioerosion as caused by invertebrate borers was found not to vary significantly with changes in grazing pressure (Sammarco, Risk & Rose, 1987). Regional differences in the abundance and diversity of algae, invertebrates and fishes may account for some of the variation in the effects of territorial fishes on corals. Further research seems necessary to sort out the apparent conflicts in published results. Although relatively few studies have been published on the effects of territorial damselfishes on local microfaunal (invertebrate) abundance, the available data indicate that densities are generally higher inside than outside the defended areas. Lobel (1980) reported that the number of small, motile invertebrates is greater within the territories of two damselfish species, Eupomacentrus planifrons from the Caribbean and E. nigricans from the central Pacific. He concluded that the territories function as refuges for juvenile benthic invertebrates, such as crabs and sea stars, and demersal plankton. Hixon & Brostoff (1982) showed that increased grazing on Hawaiian reefs causes a decrease in the biomass of erect algae and the abundance of associated small invertebrates and an increase in coverage of crustose coralline and prostrate algae. Invertebrate density inside territories of Stegastes fasciolatus was intermediate (34.3 per plate) between that outside of territories (1.7 per plate) and that within fish exclusion cages (48.8 per plate). Zeller (1988) found that densities of invertebrates (crustaceans, mainly copepods and amphipods) were greater on experimental plates inside than outside territories of S. apicalis on the Great Barrier Reef. Territorial damselfishes may indirectly affect nitrogen fixation by blue-green algae on reefs, but the data that are available appear to be in conflict. Lobel (1980) and Hixon & Brostoff (cited in Hixon, 1983) both found more blue-green algae inside than outside territories on Hawaiian reefs. Wilkinson & Sammarco (1983), however, reported that rates of nitrogen fixation at a Great Barrier Reef site were directly proportional to the extent of fish grazing and inversely proportional to total algal biomass. In other words, rates were lowest in caged areas, intermediate in damselfish territories and highest in areas fully exposed to fish grazing. More recently, Ruyter Van Steveninck (1984) found no difference in the amounts of

200

MICHAEL H.HORN

Fig 4.—A, diversity of algae after one year on substrata within fish exclusion cages (left), inside territories of the damselfish Stegastes fasciolatus (centre) and exposed outside territories (right) on Hawaiian coral reefs; vertical and horizontal lines through means represent 95% confidence limits; results support the hypothesis shown in B. B, graph of the intermediate-disturbance hypothesis, showing that a keystone species can enhance local diversity either by increasing predation intensity from point 1 toward point 2 (classic type), or by decreasing overall predation intensity form point 3 toward point 2 (reverse type as exemplified by the damselfish); slightly modified from Hixon & Brostoff (1983).

filamentous blue-green algae inside and outside damselfish territories on a reef in the Florida Keys. These contradictory results may reflect regional differences in the local abundances of blue-green algae. Many studies involving removal or caging experiments have shown that algal biomass is greater inside than outside damselfish territories (Vine, 1974; Brawley& Adey, 1977; Lassuy, 1980; Hixon & Brostoff, 1981, 1982; Williams, 1981; Sammarco, 1983; Ruyter Van Steveninck, 1984; Hinds & Ballantine, 1987; Klumpp et al, 1987; Russ, 1987; Zeller, 1988). Although this effect of territoriality seems to be clearly established, a few studies have provided exceptions. Montgomery (1980a) showed that the non-selective feeding of Microspathodon dorsalis maintains a near monoculture of the red alga Polysiphonia in its territory. Polysiphonia appears to be the only alga that can grow fast enough to persist under the influence of this fish’s intense, nonselective grazing. This alga is more productive than the multispecies algal mat outside the territory but of much lower standing biomass. Also, Foster (1987) found that while Stegastes dorsopunicans can effectively defend its territory against most solitary fishes, it is unable to prevent the sea urchin Diadema antillarum from grazing in its territory. As a result, this damselfish exerts a minor impact on algal biomass in the habitat, unlike Diadema, or the more aggressive damselfish, Stegastes planifrons, which canexclude sea urchins from its territory. Several studies have shown that algal diversity is also higher inside than outside damselfish territories. In an exhaustive study carried out on Hawaiian reefs, Hixon & Brostoff (1981, 1982, 1983) tested the intermediate disturbance hypothesis (Connell, 1978), which proposes that as the intensity or frequency of disturbance progressively increases from zero, the species diversity of the affected community will initially increase then subsequently decrease. Hixon & Brostoff’s results supported the hypothesis in that substrata located within the defended territories of the damselfish S. fasciolatus were subjected to intermediate grazing intensity and, as a result, showed greater algal diversity than substrata either protected within fish exclusion cages or exposed to intense fish grazing outside territories (Fig 4). Similar experimental studies in Guam (Lassuy, 1980), Australia (Sammarco, 1983) and Puerto Rico (Hinds & Ballantine, 1987) have all shown that algal diversity is greatest inside damselfish territories. Hixon & Brostoff (1983) labelled S. fasciolatus a keystone species (sensu Paine, 1966) in reverse because it maintains high diversity of algae by decreasing rather than increasing the overall predation intensity on the algae. Also, Williams (1980) has referred to a damselfish, Eupomacentrus planifrons, on Jamaican reefs as a non-carnivorous keystone species because it acts as a selective predator by its stronger response to the more invasive (Diadema antillarum) of two competing sea urchin species (D. antillarum and Echinometra viridis) in its territory. Thus, Eupomacentrus planifrons apparently stabilises competitive interactions and reduces the potential competitive exclusion of sea urchin species, actions primarily attributed to predatory keystone species. Territorial damselfishes, however, do not always conform to the intermediate disturbance hypothesis. Grazing intensity can be greater inside than outside territories, as in those species that actively weed certain species from the defended algal mats (Lassuy, 1980; Irvine, 1982). Moreover, as mentioned on page 247, the Gulf of California damselfish Microspathodon dorsalis maintains an almost pure stand of an algal species in its territory (Montgomery, 1980a). A model (Fig 5) relating the intensity of non-selective grazing, like that of M. dorsalis, to productivity of algal assemblages has been proposed by Montgomery (1980a), who considers the model a special case and extension of the disturbance-dependent models of Connell &

BIOLOGY OF MARINE HERBIVOROUS FISHES

201

Fig 5.—Graphical model of the relationship between nonselective grazing by territorial damselfishes and algal productivity, both represented in the same units (e.g. biomass removed or produced per unit time); increased levels of grazing intensity (represented by higher Roman numerals) would progressively remove all but the most productive or all the algal assemblages (represented by lower case letters), whereas decreased grazing intensity would allow succession to occur toward the least productive assemblages; from Montgomery (1980a).

Slatyer (1977). Grazing, which is expressed in the same units as productivity, is seen as a function of the fish’s requirements and largely independent of the composition of algal communities. Grazing at the lowest intensity (I in Fig 5) will not lead to the terminal algal assemblage (i) because the grazing rate is too small to remove a significant amount of the production. The next level of grazing (II) would remove assemblages with lower productivity than the grazing rate (f—i), whereas grazing at level III would eliminate all but the most productive assemblage (a). Montgomery argues that this is the situation that occurs in territories of M. dorsalis and various other damselfishes. Even higher grazing rates would remove all erect algae and result in the overgrazing commonly observed, particularly on coral reefs (e.g. Randall 1961b; Earle, 1972). Reduction in grazing intensity would allow succession of algal assemblages to proceed. Finally, at least two studies have reported that damselfish territories are zones of high algal productivity relative to surrounding epilithic algal communities. As mentioned above, Montgomery (1980a) found that M. dorsalis by its non-selective grazing maintains a near monoculture of the red alga Polysiphonia in its territory. On a per unit biomass basis, this alga was 34– 47 times as productive as the algal mat outside the territory. More recently, Klumpp et al. (1987) have shown that the algal communities inside the territories of four species of damselfishes on the coral reefs of the Great Barrier Reef and Papua New Guinea were 1.5– 3.4 times more productive (per unit biomass) than algae growing outside the territories. These authors offered three possible explanations for the high primary productivity of damselfish territories: (1) “weeding” activities of the damselfishes promote more highly productive algal species or forms; (2) cropping action of the damselfishes, as distinct from other fish grazers, keeps algae in an exponential phase of growth; and (3) the damselfishes “fertilise” the algae with their waste products. Weeding has been reported for three damselfish species. Lassuy (1980) found that both Eupomacentrus lividus and Hemiglyphidodon plagiometopon remove the larger, tougher, and presumably less palatable and less productive seaweeds from their territories leaving the more preferred, lower ash content forms for consumption. Irvine (1982) reported that Eupomacentrus planifrons weeds out less preferred algae, which opens up space for settlement and growth of preferred species. Preferred algae were Polysiphonia spp. and diatoms, whereas those removed were primarily filamentous green algae and the brown alga Dictyota bartayresii. Whether the preferred algae are more productive was not determined. Damselfishes feed predominantly upon the algae in their territories (Klumpp et al., 1987) and, in terms of the rate of removal of algal tissue, may feed more intensively than the herbivorous fishes outside their territories (Russ, 1987). Whether such feeding maintains the algae in an exponential growth phase and that by roving herbivores outside the territories does not is apparently yet to be discovered. Temporal and spatial focusing of the release of nitrogenous wastes occurs in Plectroglyphidodon lacrymatus, a territorial damselfish on coral reefs of Papua New Guinea (Polunin & Koike, 1987; Polunin, 1988). Polunin & Koike found that most defaecation by this fish occurred at a single site in each territory and was far more important than excretion in the generation of nitrogen by the fish. Nearly all of the nitrogen defaecated, and perhaps one-third of that excreted, was transferred initially to the reef framework below the fish’s territory. Polunin (1988) showed that P. lacrymatus ingests 91% of the nitrogen incorporated into the algal community and thus spatially focuses the transfer of nitrogen within its territory. All three of these explanations are intriguing and deserve further study. Nevertheless, it is still possible that damselfish territories are highly productive because the

202

MICHAEL H.HORN

fishes are somehow able to establish their territories in areas more suitable for algal growth (see Eakin, 1987; Zeller, 1988). TEMPERATE HABITATS On hard substrata in temperate waters the dominant herbivores are molluscs, echinoderms and crustaceans (Gaines & Lubchenco, 1982). In California waters, for example, gastropod molluscs (especially limpets and trochids) appear to be the major intertidal herbivores and sea urchins the principal subtidal herbivores (Foster & Schiel, 1985). Knowledge is meagre at best about the impacts of plant-eating fishes on seaweed and seagrass communities in temperate waters (Choat, 1982; Gaines & Lubchenco, 1982; Thayer et al., 1984; Hixon, 1986; Castilla & Paine, 1987; Santelices, 1987), but the prevailing notion is that fishes may have negligible effects on benthic algal abundance and diversity (Limbaugh, 1955; Bakus, 1964, 1969; Earle, 1972; Ogden & Lobel, 1978; Lubchenco & Gaines, 1981; Menge & Lubchenco, 1981; Hixon, 1986). This impression has emerged for at least two reasons: (1) herbivorous fishes have been perceived as rare in kind and numbers in temperate waters and (2) individual algae and algal standing stocks are often large in temperate latitudes and, based on general observations, seem little affected by grazing or browsing fishes. The observations leading to this opinion have been made largely in north temperate regions, especially the eastern North Pacific. The small amount of evidence, however, that does suggest that plant-eating fishes may have some effect on temperate-zone seaweeds has until recently come from these same north temperate regions, mostly from California coastal waters. Here, three species, the girellid Girella nigricans, the kyphosid Hermosilla azurea and the scorpidid Medialuna californiensis, are known to consume a variety of seaweeds, including Macrocystis, the giant kelp (Quast, 1968). Girella nigricans and Medialuna californiensis are the most common of the three species in southern California. Their herbivorous tendencies are well known because they eat young Macrocystis plants that have been transplanted as part of kelp restoration (North, 1972, 1973) and artificial reef (Grant, Wilson, Grover & Togstad, 1982) projects. Gill nets have been used to reduce the number of Girella nigricans and Medialuna californiensis around kelp transplant areas (North, 1973), which attests to the seriously detrimental effects of their grazing on young Macrocystis sporophytes. The potential or actual impacts of these two fishes have been recognised in other studies in California kelp beds. Three of these studies are summarised as follows. (1) Leighton (1971) attributed the reduction in fleshy green and red algae relative to a caged treatment to the action of herbivorous fishes. He reached this conclusion largely because the experimental block was placed in a location where invertebrate herbivores were absent and because the green alga Ulva occurred in appreciable amounts only in the guts of Girella nigricans among those of six fish species examined in an attempt to find the responsible grazing agent. (2) Foster (1975) found increased growth in the red algae Gigartina spp. in cage treatments at a giant kelp forest site. Although Foster observed that Girella nigricans and Medialuna californiensis were common in the area and that control plants bore evidence of structural damage from fish feeding, he presented no data to link the two observations. An alternative explanation, which possibly diminishes the role of herbivorous fishes in affecting seaweed biomass, stems from work by Kennelly (1983) in a kelp forest habitat near Sydney, Australia. He found that exclusion of fishes from settlement plates and the natural substratum resulted in less not more algal cover during colonisation. Kennelly hypothesised that exclusion of carnivorous fishes led to an increase in the abundance of small invertebrate grazers, especially amphipods, that caused reduced algal stands inside cages. Perhaps similar grazers avoided Gigartina and ate other algae inside the cages used by Foster, or perhaps they modified algal abundance inside the cages by eating spores

BIOLOGY OF MARINE HERBIVOROUS FISHES

203

or germlings. The other algae could also have been outcompeted for space by sessile animals in the low light conditions of the habitat (Foster, 1975). Kennelly, however, presented no data or observations on herbivorous fishes, several species of which are known to occur at least in the general vicinity of his study area (see Bell, Burchmore & Pollard, 1980; Burchmore et al., 1985). (3) In one of the only studies designed to determine the impacts of temperate herbivorous fishes on seaweed populations, Harris et al. (1984) found that Girella nigricans and Medialuna californiensis can be important grazers on young Macrocystis plants on a small scale. These workers manipulated early colonising, fast-growing algal stands after a severe storm had denuded a southern California kelp bed habitat to test for the role of filamentous brown algae in helping sporophytes survive fish grazing. Grazing affected 59% of small plants (less than 10 cm tall) concealed in a turf of the filamentous algae on new surfaces, whereas 94% of the sporophytes amidst shorter turf on old surfaces had fish bites. These results indicate that young kelp plants growing within stands of filamentous algae receive some protection from herbivorous fishes. Larger plants were unaffected, perhaps because of a size refuge. The grazing resulted in a small-scale change in the dispersion pattern of juvenile giant kelp on the reef. Harris et al. (1984), however, did not mention the feeding behaviour of the fishes or report their abundances in the habitat. Several other southern California fish species occasionally consume small amounts of algae, including two species of surfperches (Embiotocidae) (Bray & Ebeling, 1975) and the labrid Oxyjulis californica (Bray & Ebeling, 1975; Bernstein & Jung, 1979). The latter species, although not a herbivore, can apparently have an impact on Macrocystis by removing blade portions encrusted with a bryozoan, one of its preferred foods (Bernstein & Jung, 1979). Algae had the third highest utilisation index among food categories of kelp bed fishes recognised by Quast (1968), which he attributed to be mostly a result of accidental ingestion with animal prey. A recent study, on a North Carolina rock jetty, also suggests that temperate fishes can have an impact on seaweed abundance. Hay (1986) reported that palatable seaweeds such as Hypnea musciformis, Chondria spp., Gracilaria spp., Enteromorpha spp. and Ulva spp. that are common during most of the year on the jetty decrease sharply in abundance during midsummer when algae-eating fishes such as the sparids Diplodus holbrooki and Lagodon rhomboides and that monacanthid Monacanthus hispidus are most prevalent. In the spring and late fall when these fishes are rare or are feeding on invertebrates instead of algae, most of the palatable seaweeds are growing on primary substratum. In midsummer, on the other hand, when the fishes are numerous and feeding on algae, the palatable seaweeds are mostly epiphytes on unpalatable brown algae such as Sargassum and Padina. These tantalising accounts in the North Pacific and North Atlantic offer hints of the importance of herbivorous fishes in temperate regions, but in virtually all cases the studies from which they emerged were focused on other objectives. I am aware of no study in north temperate waters that was designed to investigate rigorously the impacts of herbivorous fishes on algal populations or communities. In the southern hemisphere, especially in the temperate waters around New Zealand and southern Australia, recent studies show signs that herbivorous fishes may affect algal communities. First of all, herbivorous fishes are surprisingly diverse in both the Australian (Coleman, 1980; Burchmore et al., 1985) and New Zealand (Doak, 1978; Ayling & Cox, 1982; Russell, 1983; Choat & Ayling, 1987) regions. Moreover, species such as Girella tricuspidata are highly abundant and can constitute as much as 51% of the total fish biomass in dense kelp cover in New Zealand coastal waters (Russell, 1977). The zone of short algal turf on shallow rocky reefs in northeastern New Zealand may be at least partially maintained by browsing and grazing fishes (Russell, 1983). Neither of the two studies of which I am aware that provide data on the impacts of herbivorous fishes on algal abundance, however, show evidence of an important influence by fishes. Conacher, Lanzing &

204

MICHAEL H.HORN

Larkum (1979) concluded that browsing by the monacanthid Monacanthus chinensis had little impact on algal or seagrass production rates in Botany Bay, New South Wales, and Meekan (1986) found that browsing by the odacid Odax pullus imparted no more than a minor cost to the brown alga Ecklonia radiata in shallow waters off northern New Zealand. Meekan’s (1986) study is one of the most detailed yet completed on the relationship between a herbivorous fish and its major food plant. He found that browsing damage on Ecklonia radiata was closely correlated with abundance patterns of Odax pullus. Most browsing activity was concentrated on adult plants and was selective within plants. Generally, only the secondary laminae of Ecklonia radiata were consumed, and a higher proportion of reproductive tissue than vegetative tissue of these laminae was eaten by Odax pullus. Meekan cited evidence that reproductive tissue consumed by the fish may survive digestion. Browsing damage on Ecklonia radiata varied spatially between reef types, localities, transects, sites, plants within a site, and, temporally, on individual plants. Moreover, browsing damage experiments and monitoring of individuals over a three-month period suggested that Odax pullus kills few plants and that the plants rapidly recover from either fish browsing or artificially imposed damage. A longer-term study that would assess the fish’s impact on other seaweeds in its diet (e.g. the fucoid algae Carpophyllum spp.) should be rewarding.

DISTRIBUTION, DIVERSITY AND ABUNDANCE OF HERBIVOROUS FISHES GENERAL PATTERNS OF DISTRIBUTION AND DIVERSITY Herbivorous fishes are largely confined to shallow-water coastal habitats. They are limited in their vertical and horizontal distributions because they depend directly on light-requiring organisms for food. As mentioned in the previous section, tropical species occur as deep as 90 m (Thresher & Colin, 1986) but mainly live in the shallowest 10 m of inshore waters (Bakus, 1969). Temperate species show essentially the same, if not shallower, bathymétric pattern (see Quast, 1968; Russell, 1983; Burchmore et al., 1985). Offshore, herbivorous fishes are restricted because the substratum becomes too deep for benthic plants to survive. Moreover, the browsing and grazing fishes considered in this review are not equipped morphologically to feed on phytoplankton in the water column. High seas herbivores are exceedingly rare, apparently because phytoplankton populations are too sparse in the open ocean to support filter-feeding fishes. The main phytoplanktivorous fishes in the sea are filter-feeding clupeoids such as menhaden and some species of anchovies and sardines, and they are confined to coastal regions in various parts of the world (see Blaxter & Hunter, 1982). Like other herbivorous fishes, these clupeoids begin life as carnivores then through early ontogeny develop an elaborate set of gill rakers, epibranchial organs and a longer gut as the diet becomes increasingly made up of phytoplankton. An unusual case of open ocean herbivory is seen in the myctophid Ceratoscopelus warmingii, a dominant midwater fish in the oligotrophic North Pacific Gyre (Robison, 1984). This fish feeds on the dense diatom mats (Rhizosolenia spp.) that form in the upper 30 m of the water column. The intestine of Ceratoscopelus warmingii is longer than that of other myctophids, but Robison noted that the fish is only an occasional herbivore and feeds otherwise on zooplankton. Browsing and grazing herbivorous fishes make up only a small proportion of the total diversity of coastal fishes. About 19 families contain almost all of the herbivorous fishes in the sea (Table VII). Thus, herbivore diversity is concentrated in only about 5% of the 409 recognised (Nelson, 1984) families of teleostean fishes. Moreover, 15 of the 19 families belong to the Perciformes (Table VII), the largest traditionally

BIOLOGY OF MARINE HERBIVOROUS FISHES

205

recognised order of teleosts, but one more recently viewed (Lauder & Liem, 1983; Johnson, 1984) as a polyphyletic assemblage. At the species level, it is difficult to estimate with any reasonable degree of accuracy the total number of marine herbivorous fishes because for most of the families the proportions of herbivorous fishes cannot be determined with available information (see Table VII). For example, little is known about the feeding habits of many species in the large, diverse families such as the Blenniidae and Gobiidae. LATITUDINAL PATTERNS Diversity One of the most striking and widely recognised (Bakus, 1964, 1969; Medd, 1970; Earle, 1972; Ogden & Lobel, 1978; Choat, 1982; Gaines & Lubchenco, TABLE VII Principal teleostean fish families containing marine herbivorous species ; order, suborder, family distribution and total species from Nelson (1984) Order

Suborder

Family

Family distribution

Total no. species

Estimated no. herbivorous speciesb

Gonorynchiform es Atheriniformes Perciformes

Chanoidei

Chanidae

Tropical

1

1

Exocoetoidei Percoidei

Hemiramphidae Sparidae

Tropical Tropical/ temperate Temperate Tropical/ temperate Tropical/ temperate Tropical Tropical/ temperate Temperate Tropical/ temperate Temperate Tropical Temperate Tropical/ temperate Tropical/ temperate Tropical Tropical

80a 100

? ?

20 10

20 10

15

?

74 235

? ?

5 70a

5 70

12 68 60 3o1a

5 68 ≥2 ?

1500a

?

76 25

7 25

Girellidae Kyphosidae Scorpididae Pomacanthidae Pomacentridae

Mugiloidei Labroidei

Aplodactylidae Mugilidae

Zoarcoidei Blennioidei

Odacidae Scaridae Stichaeidae Blenniidae

Gobioidei

Gobiidae

Acanthuroidei

Acanthuridae Siganidae

206

MICHAEL H.HORN

Order

Suborder

Family

Family distribution

Total no. species

Estimated no. herbivorous speciesb

Tetraodontiform es

Balistoidei

Balistidae

Tropical/ temperate Tropical

135

?

Tetraodontoidei Tetraodontidae 118a ? include freshwater species. sEstimates are included where all species are expected to be herbivores with two exceptions. Total number of species in the Odacidae is from Gomon & Paxton (1985) and the number of herbivorous odacids from Choat & Ayling (1987) and M.J. Kingsford (pers. comm.). The minimum number of herbivorous stichaeids is based on Horn, Murray & Edwards (1982). aTotals

1982; Estes & Steinberg, 1988) yet enigmatic distributional patterns in coastal marine waters is the rarity of strictly herbivorous fish species in temperate latitudes as compared to the much larger numbers of such species in tropical waters. Bakus (1969) estimated that coral reef fish communities are composed of 22% herbivores, 69% carnivores and 9% omnivores, whereas rocky reef and kelp bed fishes in southern California are roughly 33% carnivores and 67% omnivores. In other words, he recognised no temperate herbivores in this warm temperate region of the eastern North Pacific. Although Bakus under-estimated temperate-zone herbivory, a survey of 19 studies of shallow marine fish communities (Table VIII) shows that the occurrence of herbivores is inversely correlated with latitude in terms of both the absolute numbers and the proportions of herbivorous species in the community. Furthermore, the Table shows that the fish communities at the highest latitudes contain no herbivores. Thus, strictly TABLE VIII Numbers and proportions of herbivores in shallow marine fish communities at different latitudes; see Table I for list of herbivorous species in those communities where they occur Locality (habitat)

Latitude Total no. species studied Herbivores Reference

No. vores

%

Antarctica (rocky subtidal) Scotland (rocky intertidal) Chile (kelp bed) Auckland Islands (rocky subtidal) France (Atlantic, rocky intertidal) France (Mediterranean, rocky littoral) California (rocky intertidal) California (rocky intertidal)

61°S 56°N 55°S 51°S

9 15 18 8

0 0 0 0

0 0 0 0

49°N

13

0

0

Targett, 1981 R.N.Gibson, unpubl. data Moreno&Jara, 1984 Kingsford, Schiel & Battershill, 1988 Gibson, 1972

42°N

19

1

5

Gibson, 1968

38°N 36°N

15 28

1 2

7 7

New Zealand (rocky subtidal) Australia (rocky subtidal) California (rocky subtidal) California (kelp bed)

36°S 34°S 33°N 32°N

50 102 46 45

5 12 2 3

10 12 4 7

Grossman, 1986 L.G.Allen & M.H.Horn, unpubl. data Russell, 1983 Burchmore et al., 1985 Stephens & Zerba, 1981 Quasi, 1968

BIOLOGY OF MARINE HERBIVOROUS FISHES

207

Locality (habitat)

Latitude Total no. species studied Herbivores Reference

No. vores

%

Mexico (rocky intertidal) South Africa (rocky littoral) Kermadec Islands (rocky subtidal) Hawaii (coral reef) Puerto Rico/Virgin Islands (coral reef) Marshall Islands (coral reef) Tanzania (coral reef)

31°N 30°S 29°S

25 66 45

4 5 7

16 8 16

Thomson & Lehner, 1976 Berry et al., 1982 Schiel, Kingsford & Choat, 1986

19°N 18°N

120 212

30 31

25 15

Jones, 1968; Hobson, 1974 Randall, 1967

10°N 8°S

233 106

52a 30b

22 28

Hiatt & Strasburg, 1960 Talbot, 1965

aTotal

bTotal

includes 9 scarids not listed as herbivores by Hiatt & Strasburg (1960). includes 15 scarids not listed as herbivores by Talbot (1965).

herbivorous fishes appear to be limited to a range of latitude from about 40°N to about 40°S. An exception may be the stichaeid Xiphister mucosus, a year-round herbivore in central California waters (Horn, Murray & Edwards, 1982) that occurs as far north as southeastern Alaska (Eschmeyer, Herald & Hammann, 1983) at a latitude of at least 55°N. Whether X. mucosus is still a herbivore at this latitude is, however, unknown. A few fishes living beyond the 40-degree belt are known to consume algae on a seasonal or facultative basis. For example, filamentous algae were found in the stomach contents of most fish species in a kelp bed habitat at 55°N in southern Chile (Moreno & Jara, 1984). The authors, however, believed that this plant material was consumed simultaneously with invertebrate prey and, therefore, was not evidence for herbivory. Also, the nototheniid Notothenia neglecta was found (Daniels, 1982) to crop macroalgae and harvest diatom mats during the spring and summer at 61°S in an Antarctic habitat. The fish switches from being an omnivore in the spring and summer to a carnivore through the autumn and winter. Several hypotheses have been offered to explain the rarity of herbivorous fish species in temperate, boreal and arctic latitudes (Gaines & Lubchenco, 1982). These are discussed in turn. (1) Insufficient time has lapsed to allow cold-adaptation and range expansion to temperate and polar seas. Medd (1970) observed that the Perciformes, the largest order of teleostean fishes, dominates the tropical marine fauna, especially that of coral reefs, and that the diversity and abundance of these fishes decline sharply poleward away from the tropics. He speculated that the explosive radiation of the group in the tropics was mainly the result of the development of the ability to eat and digest benthic plant material. Those perciform families in which herbivorous capabilities evolved (e.g. Acan thuridae, Scaridae, Siganidae) appear to be just the forms that have not radiated in any major way into temperate seas. Mead believed the perciform radiation to be a relatively recent phenomenon (i.e. in the Cenozoic) and argued that time favours cold adaptation even if it is not the driving force. His analysis suggested that the highly diverse tropical Indo-Pacific region was the major colonising source for temperate waters off Australia and New Zealand. The relatively close proximity of the Indo-Pacific to Australia and New Zealand through the Cenozoic may help explain the higher herbivore diversity in these regions as compared to that in the eastern North Pacific (see Table VIII). Mead’s analysis, however, is largely untestable and weakened because it was based on the evolution and radiation of the Perciformes, which is now considered to be a polyphyletic assemblage (Lauder & Liem, 1983; Johnson, 1984). Cladistic analyses of apparently related groups that contain both herbivorous and carnivorous species and whose distributions span both tropical and temperate regions should prove to be valuable.

208

MICHAEL H.HORN

(2) The effects of low temperature on fish digestive physiology may largely exclude herbivorous fishes from temperate and higher latitudes. Gaines & Lubchenco (1982) proposed that if digestive efficiency is dramatically reduced relative to energy demands in colder waters, a primarily herbivorous diet would be energetically infeasible. This hypothesis is unstudied and appears to be testable, but it could also be testable if gut transit time rather than digestive efficiency were the focus of the hypothesis as follows. If the rate of movement of food through the gut is reduced at low temperatures, as expected, so that the stimulus to feed is decreased, thereby reducing the consumption rate, then energy intake from an algal diet could be inadequate to meet the fish’s requirements. (3) Regions where suitable plant food is seasonally unavailable may be habitable only by invertebrate herbivores. If herbivorous fishes are more limited than invertebrates in the length of time they can persist without food, then non-migratory, strictly herbivorous fishes might be excluded from such regions. This hypothesis is weakened by at least two observations. First, many temperate shores have large standing stocks of seaweeds throughout the year and may even be more productive during the winter months (Mann, 1973; Chapman & Lindley, 1980). Secondly, even though brown algae, the seaweeds usually comprising the bulk of the winter biomass in these localities, has been considered to be inferior to green and red algae as fish food and seldom consumed by fishes (Montgomery & Gerking, 1980), several species of herbivorous fishes regularly eat brown seaweeds in both north (e.g. Harris et al., 1984) and south (e.g. Russell, 1983) temperate waters. (4) Latitudinal differences in algal toughness, chemical defences or nutritional quality may restrict temperate fish herbivory. Little or no evidence exists to provide convincing support for this hypothesis. Although Bakus (1969) suggested that the toughness of the larger temperate seaweeds might deter feeding by fishes, several relatively tough brown algae make up a large part of the diets of temperate herbivores such as girellids, kyphosids and odacids (Russell, 1983). Moreover, toughness and calcification seem to be important though variable deterrents to fish feeding in a variety of tropical seaweeds (Littler, Taylor & Littler, 1983b; Lewis, 1985; Paul & Hay, 1986). The incidence of defensive chemicals appears to be more common in seaweeds of tropical and subtropical waters than in those of temperate habitats (Hay & Fenical, 1988). Nevertheless, temperate brown algae, especially the fucoids, are rich in phlorotannins that are known to deter feeding by herbivorous invertebrates (Steinberg, 1984, 1985). They may deter some herbivorous fishes because no north temperate fish is known to consume fucoid algae. In Australian and New Zealand waters, however, where phlorotannin concentrations in fucoid algae are two to three times as great as those in north temperate fucoids (Estes & Steinberg, 1988; Steinberg, 1988), kyphosid and odacid fishes consume these algae as major portions of their diets (Russell, 1983). The lack of predation by sea otters or other marine mammals on invertebrate herbivores, especially sea urchins, in the temperate South Pacific as compared to the eastern North Pacific during recent earth history has been proposed (Estes & Steinberg, 1988) as the cause of the phlorotannin-rich brown algal flora in Australia and New Zealand. Sea urchins, unchecked by predators, would have fed intensively on the algae and provided the selective pressure for the algae to evolve increased chemical defences. In turn, the herbivores would have developed greater tolerances to the seaweed defences. Herbivorous fishes may have been part of these tightly coupled plantherbivore interactions, which could explain the apparently high tolerances of kyphosid and odacid fishes to the phlorotannins and other secondary chemicals in the brown algae they regularly consume. Finally, with regard to nutritional quality, no strong evidence is available to suggest that temperate seaweeds are less nutritious than tropical species although the quality and digestibility of brown algae in both temperate and tropical regions deserves TABLE IX

BIOLOGY OF MARINE HERBIVOROUS FISHES

209

Densities of marine herbivorous fishes in 7 temperate and 14 tropical habitats Latitude Habitat

No. per 1000 m2

References

Temperate New Zealand (Pacific) California (Pacific) New Zealand (Pacific) New Zealand (Pacific) North Carolina (Atlantic) California (Pacific) New South Wales (Tasman)

37°S 36°N 36°S 36°S 35°N 34°N 34°S

Rocky subtidal Rocky intertidal Rocky subtidal Rocky subtidal Rock jetty Kelp bed Seagrass bed

0–39 230 100 0–270 7640 9–38 10–430

15 230 100 63 7640 25 170

Choat & Ayling, 1987 Jones, 1981 Meekan, 1986 Russell, 1977 Hay, 1986 Ebeling et al., 1980 Conacher, Lanzing & Larkum, 1976

Tropical Bermuda Islands (Atlantic) Gulf of Aqaba (Red)

32°N 29°N

Coral reef Coral reef

32a 44–234

32a 127

Hawaii (Pacific) Hawaii (Pacific) Virgin Islands (Caribbean) Puerto Rico (Caribbean) Tahiti (Pacific) Belizean Barrier Reef (Caribbean) Dahlak Archipelago (Red)

21°N 19°N 18°N 18°N 18°S 17°N

Coral reef Coral reef Coral reef Coral reef Coral reef Coral reef

1100a 172–848a 172–279a 114–152a 312–663a 88–398

1100a 530a 225a 133a 488a 211

Bardach, 1959 Bouchon-Navaro & Harmelin-Vivien, 1981 Brock, 1979 Hobson, 1974 Randall, 1963 Barlow, 1975 Galzin, 1977 Lewis & Wainwright, 1985

16°N

Coral reef

55a

55a

Great Barrier Reef (Pacific) Guam (Pacific) Enewetak Atoll (Pacific) Aldabra Atoll (Indian)

15°S 13°N 11°N 9°S

Coral reef Coral reef Coral reef Coral-reef

137–366 57–214a 8a 557a

277 130a 8a 557a

Phoenix Islands (Pacific)

4°S

Coral reef

17–6700a 1839a

Locality (ocean/sea) Range Mean

aDensities

Clark, Ben-Tuvia & Steinitz, 1968 Choat & Bellwood, 1985 Jones & Chase, 1975 Bakus, 1967 Robertson, Polunin & Leighton, 1979 Grovhoug & Henderson, 1978

are from Bouchon-Navaro & Harmelin-Vivien (1981).

further investigation. Comparative studies and manipulative experiments are required before the latitudinal differences in the diversity of herbivorous fishes can be better understood. Abundance A prevalent opinion is that herbivorous fishes are not only more diverse but more abundant in tropical waters than in temperate seas (Bakus, 1969; Medd, 1970; Earle, 1972; Ogden & Lobel, 1978; Gaines & Lubchenco, 1982). Examination of the crude densities of herbivorous fishes compiled from the data on a variety of temperate and tropical habitats (Table IX), however, reveals no sharp or obvious differences

210

MICHAEL H.HORN

between abundances in the two regions. Almost a 1000-fold difference in densities of fish herbivores is represented in the 21 localities, but the data are insufficient to allow rigorous comparisons. The large range of variation shown in the densities from both low and high latitude habitats most likely reflects differences in fish and habitat characteristics, in sampling methods, and in the conditions under which the density values were obtained. Conditions more often associated with temperate habitats—reduced visibility, turbulent waters and low temperatures—may have lead to under-estimates of herbivore densities in higher latitudes. On the other hand, the more uniformly warm temperatures in the tropics probably means that herbivorous fishes are more consistently abundant in the habitat on a year-round basis. The continuous presence of herbivorous fishes on a tropical reef may be a major reason for the apparently greater impacts of these fishes on algal community structure than those of their temperate counterparts. Temperate herbivores may achieve densities rivaling those of tropical species primarily during the summer months. For example, the high densities of herbivores recorded (Hay, 1986) for a North Carolina rock jetty (Table IX) were reached in midsummer by a single species, the sparid Diplodus holbrooki. This species and several other seaweed-eating fishes become more omnivorous during other parts of the year and most individuals move offshore during the winter (Hay, 1986). That temperate herbivorous fishes can achieve densities as high as those of tropical species, if only on a seasonal basis, is consistent with the limited evidence presented earlier (p.251) that these higher latitude herbivores exert some influence on the structure of algal communities. RECOMMENDATIONS FOR FUTURE RESEARCH Virtually all aspects of marine fish herbivory would benefit from further investigation. The following topics seem particularly important for expanding knowledge of the biology of herbivorous fishes in the sea. (1) Factors affecting food choice. Diet selection is a complex process in herbivorous fishes. The relative importance of food quality and antiherbivore defences ought to be carefully assessed. Food preference studies should examine the roles of vision and chemosensory organs in the food selection process by fishes. Such an approach might reveal whether fishes can assess food quality directly and how chemically defended seaweeds are evaluated. In other words, the mechanisms and relative importance of acceptance and avoidance could be determined. (2) Physiological effects of seaweed secondary compounds. Although the recent evidence that secondary metabolites deter feeding by herbivorous fishes is convincing, almost nothing is known about the effects of these chemicals on the digestive and other systems of herbivorous fishes. This topic is important for investigation because herbivorous fishes appear to vary greatly in their reaction to, and tolerance of, these compounds. (3) Physiology and biochemistry of digestion. Our understanding of digestive mechanisms in herbivorous fishes is superficial and rudimentary. Much more information is required on the importance of gut pH, endogenous enzymes and microbial symbionts before an understanding can be reached on how the great variety of seaweeds consumed by fishes are digested and their nutrients assimilated. Increased knowledge of digestive physiology should aid in explaining how secondary compounds are tolerated or perhaps neutralised by certain herbivorous fishes. (4) Growth. Almost nothing is known about growth in herbivorous fishes. Growth experiments combined with studies of digestive physiology should be designed to determine whether these fishes feed to satisfy an energy or a protein requirement, how efficiently they use available protein and how they achieve such rapid population growth on a low protein diet.

BIOLOGY OF MARINE HERBIVOROUS FISHES

211

(5) The ecological importance of small herbivorous fishes in warm water habitats. Small plant-eating fishes such as certain blenniids and gobiids, are often highly abundant on tropical and subtropical reefs but their cryptic behaviour and small size make it difficult to assess their impacts on algal communities. (6) The ecological impacts of temperate-zone herbivorous fishes. Recent studies show that fish herbivores in both north temperate and south temperate latitudes reach high abundances, but these investigations have only provided hints of the potential influence of such fishes on algal community structure. Experiments should be designed with the specific objective of assessing the role of these fishes in temperate communities. (7) The impacts of territorial damselfishes on algal production and herbivore consumption. The relative effects of weeding, cropping and fertilising by damselfishes on the productivity of algae in their territories are poorly understood but required for a greater understanding of the influence of these fishes on coral reef communities. Moreover, studies of the impacts of territorial damselfishes (and surgeonfishes as well) on these communities should be expanded to encompass larger spatial scales on the reef than has heretofore been done. (8) Why are herbivorous fishes more diverse in tropical waters than in temperate waters? Several hypotheses have been proposed to explain this latitudinal pattern, but few are testable. The hypothesis that food processing rates are constrained at low temperatures so that energy demands are not met at the low end of the environmental temperature range seems testable and may help explain the rarity of strictly herbivorous fishes in temperate latitudes. (9) Phylogenetic analysis of related fish groups that contain both herbivorous and carnivorous species and span both temperate and tropical latitudes. A hypothesis of relationships using cladistic techniques and a data set comprising of digestive tract characters could be developed. This approach might help elucidate the evolution of herbivory in different monophyltic groups and help explain the distributional patterns of herbivorous fishes. Candidates for this type of analysis include groups of apparently related taxa such as (1) girellids, kyphosids and scorpidids, and (2) labrids, scarids and odacids. ACKNOWLEDGEMENTS Numerous colleagues aided in the preparation of this review by providing theses, articles in press, and published papers. In this regard, I particularly wish to thank T.A.Anderson, J.H.Burk, J.H.Choat, W.H.Fenical, M.E.Hay, M.A.Hixon, M.J.Kingsford, M.G.Meekan, W.L.Montgomery, S.N.Murray, V.J.Paul, D.W.Rimmer, P.D.Steinberg, T.E.Targett, and D.A.Thomson. Thanks are also due to D.N.Waugh who compiled some of the data, N.E.Caudill who obtained many articles on short notice through interlibrary loan and M.B.Fris who delivered an important fish specimen on even shorter notice. I am especially grateful to C.D.Irelan who gave indispensable help in preparing the figures, compiling the tables and assembling the manuscript. M.B.Fris, C.D.Irelan, S.N.Murray, and M.A. Neighbors offered constructive comments on the manuscript in its various stages. My research on herbivorous fishes has been supported by the National Science Foundation (currently by grant OCE-8716368). REFERENCES Adams, S.M., 1976. Trans. Am. Fish. Soc., 105, 514–519. Alevizon, W.S., 1976. Copeia, 1976, 796–798. Al-Hussaini, A.H., 1947. Publ. Mar. Biol. Stn Ghardaga, Red Sea, No. 5, pp. 1–61. Al-Hussaini, A.H., 1949. Q.J.Microsc. Sci., 90, 109–139.

212

MICHAEL H.HORN

Anderson, T.A., 1986. J.Morphol., 190,109–119. Anderson, T.A., 1987. J.Fish Biol., 31, 221–228. Anderson, T.A., 1988. J.Fish Biol., 32, 911–921. Atkinson, J.L., Hilton, J.W. & Slinger, S.J., 1984. Can. J.Fish. Aquat. Sci., 41, 1384–1386. Avila, E.A., 1986. Zool. Anz., Jena, 217, 178–191. Ayling, T., & Cox, G.J., 1982. Collins Guide to the Sea Fishes of New Zealand. Collins, Auckland, 343 pp. Bakus, G.J., 1964. Allan Hancock Found. Occas. Pap., 27, 29 pp. Bakus, G.J., 1967. Micronesica, 3, 135–149. Bakus, G.J., 1969. Int. Rev. Gen. Exp. Zool., 4, 275–369. Bardach, J.E., 1959. Limnol. Oceanogr., 4, 77–85. Bardach, J.E, 1961. Science, 133, 98–99. Barlow, G.W., 1974a. J. Zool., 174, 333–340. Barlow, G.W., 1974b. Am. Zool., 14, 9–34. Barlow, G.W., 1975. Mar. Biol., 33, 281–293. Barton, M.G., 1982. Calif. Fish Game, 68, 174–182. Bell, J.D., Burchmore, J.J. & Pollard, D.A., 1980. Aust. Zool., 20, 391–05. Bernstein, B.B. & Jung, N., 1979. Ecol. Monogr., 49, 335–355. Berry, P.F., Van der Elst, R.P., Hanekom, P., Joubert, C.S.W. & Smale, M.J., 1982. Mar Ecol. Prog. Ser., 10, 49–55. Birkeland, C., 1977. Proc. 3rd Int. Coral Reef Symp., Miami, 1, 15–21. Bjorndal, K.A., 1985. Bull. Mar. Sci., 36, 224–230. Bjorndal, K.A., 1987. Copeia, 1987, 714–720. Blaxter, J.H.S. & Hunter, J.R., 1982. Adv. Mar. Biol., 20, 1–223. Borowitzka, M.A., 1981. Endeavour, 5, 99–106. Bouchon-Navaro, Y. & Harmelin-Vivien, M.L., 1981. Mar. Biol., 63, 79–86. Bowen. S.H., 1976. Nature, (London), 260, 137–138. Bowen, S.H., 1979. Ecol. Monogr., 49, 17–31. Bowen, S.H., 1982. In, The Biology and Culture ofTilapias, edited by R.S.V.Pullin & R.H.Lowe-McConnell, ICLARM Conf. Proc.7, International Center for Living Aquatic Resources Management, Manila, Philippines, pp. 141–156. Bowen, S.H., 1984a. In, Evolutionary Ecology of Neotropical Freshwater Fishes, edited by T.Zaret, Junk Publishers, The Hague, The Netherlands, pp. 59–66. Bowen, S.H., 1984b. Acta Hydrochim. Hydrobiol, 12, 55–59. Bowen, S.H., 1987. Can. J.Fish. Aquat. Sci., 44, 1995–2001. Bradbury, R.H. & Goeden, G., 1974. Proc. 2nd Int. Coral Reef Symp., Brisbane, 1, 167–178. Brawley, S.H. & Adey, W.H., 1977. Environ. Biol. Fish., 2, 45–51. Bray, R.N. & Ebeling, A.W., 1975. Fish. Bull, 73, 815–829. Brett, J.R. & Groves, T.D.D., 1979. In, Fish Physiology, Vol. VIII, Bioenergetics and Growth, edited by W.S.Hoar et al, Academic Press, New York, pp. 279– 352. Brock, R.E., 1979. Mar. Biol., 51, 381–388. Bryan, P.G., 1975. Pacif. Sci., 29, 269–277. Buckman, N.S. & Ogden, J.C., 1973. Ecology, 54, 1377–1382. Buddington, R.K., 1980. Trans. Am. Fish. Soc., 109, 653–656. Buddington, R.K., 1987. J. Comp. Physiol. B, 157, 677–688. Buddington, R.K., Chen, J.W. & Diamond, J., 1987. J. Physiol, 393, 261–281. Burchmore, J.J., Pollard, D.A., Bell, J.D., Middleton, M.J., Pease, B.C. & Matthews, J., 1985. Bull. Mar. Sci., 37, 70–85. Butman, C.A., 1987. Oceanogr. Mar. Biol. Annu. Rev., 25, 113–165. Cancino, J.M. & Castilla, J.C., 1988. J.Nat. Hist., 22, 249–261. Carpenter, R.C., 1981. J. Mar. Res., 39, 749–765.

BIOLOGY OF MARINE HERBIVOROUS FISHES

213

Carpenter, R.C., 1983. In, The Ecology of Deep and Shallow Coral Reefs, edited by M.L.Reaka, Symposium Series for Undersea Research, Vol.1, Office of Undersea Research, NOAA, Rockville, Maryland, pp. 113–118. Carpenter, R.C., 1986. Ecol. Monogr., 56, 345–363. Carpenter, R.C., 1988. Proc. Nat. Acad. Sci., 85, 511–514. Castilla, J.C. & Paine, R.T., 1987. Rev. Chilena Hist. Nat., 60, 131–151. Chapman, A.R.O., 1986. Adv. Mar. Biol, 23, 1–161. Chapman, A.R.O. & Lindley, J.E., 1980. Mar. Biol, 57, 1–5. Charlock, M.A., 1983. Biotropica, 15, 117–121. Choat, J.H., 1982. Annu. Rev. Ecol. Syst., 13, 423–449. Choat, J.H. & Ayling, A.M., 1987. J. Exp. Mar. Biol. Ecol, 110, 257–284. Choat, J.H. & Bellwood, D.R., 1985. Mar. Biol, 89, 221–234. Christensen, M.S., 1978. Fish. Bull, 76, 389–401. Clark, E., Ben-Tuvia, A. & Steinitz, H., 1968. Bull Sea Fish. Res. Stn, Istael, 49, 15–31. Clarke, T.A., 1970. Ecol. Monogr., 40, 189–212. Clements, K.D., 1985. MSc thesis, University of Auckland, Auckland, New Zealand, 211pp. Clements, K.D. & Bellwood, D.R., 1988. Aust. J.Mar. Freshwater Res., 39, 87– 107. Cloud Jr, P.E., 1959. U.S.Geol. Surv. Prof. Paper, No. 280-K, pp. 361–45. Coetzee, D.J., 1981. S.Afr. J. Zool, 16, 14–20. Coleman, N., 1980. Australian Sea Fishes South of 30° South, Doubleday & Company, Sydney, 302 pp. Collette, B.B., McGowen, G.E., Parin, N.V. & Mito, S., 1984. Am. Soc. Ichthyol. Herpetol. Spec. Publ., No. 1, pp. 335–354. Collins, M.R., 1981. J. Fish Biol., 19, 307–315. Conacher, M.J., Lanzing, W.J.R. & Larkum, A.W.D., 1979. Aust. J.Mar. Freshwater Res., 30, 387–100. Connell, J.H., 1978. Science, 199, 1302–1310. Connell, J.H. & Slatyer, R.O., 1977. Am. Nat., 111, 1119–1144. Conover, R.J., 1966. Limnol. Oceanogr., 11, 338–345. Cowey, C.B. & Sargent, J.R., 1979. In, Fish Physiology, Vol. VIII, Bioenergetics and Growth, edited by W.S.Hoar et al., Academic Press, New York, pp. 1–69. Craigie, J.S., 1974. In, Algal Physiology and Biochemistry, edited by W.D.P.Stewart, University of California Press, Berkeley, pp. 206–235. Crawley, M.J., 1983. Herbivory, the Dynamics of Animal-Plant Interactions. University of California Press, Berkeley, 437 pp. Dabrowski, K., 1982. Acta Hydrochim. Hydrobiol., 10, 265–271. Dahl, A.L., 1973. Bot. Mar., 16, 171–175. Daniels, R.A., 1982. Fish. Bull, 80, 575–588. Day, R.W., 1977. Mar. Biol, 44, 1–5. Demment, M.W. & Van Soest, P.J., 1985. Am. Nat., 125, 641–672. De Silva, S.S. & Perera, M.K, 1983. J. Fish Biol, 23, 675–684. Doak, W., 1978. Fishes of the New Zealand Region, revised edition. Hodder & Stoughton, Auckland, New Zealand, 135 pp. Doherty, P.J., 1983. Ecology, 64, 176–190. Eakin, C.M., 1987. Symbiosis, 4, 275–288. Earle, S.A., 1972. In, Results of the Tektite Program: Ecology of Coral Reef Fishes, edited by B.B.Collette & S.A.Earle, Nat. Hist. Mus. Los Angeles Cty. Sci. Bull, No. 14, pp 17–44. Ebeling, A.W., 1957. Copeia, 1957, 173–185. Ebeling, A.W. & Bray, R.N., 1976. Fish. Bull., 74, 703–717. Ebeling, A.W., Larson, R.J., Alevizon, W.S. & Bray, R.N., 1980. Fish. Bull., 78, 361–377. Edwards, T.W. & Horn, M.H., 1982. Mar. Biol., 67, 247–253.

214

MICHAEL H.HORN

Eisenberg, J.F., 1978. In, The Ecology of Arboreal Folivores, edited by G.G. Montgomery, Smithsonian Institution Press, Washington, D.C., pp. 135–152. Emery, A.R., 1973. Bull. Mar. Sci., 23, 649–770. Emery, A.R. & Thresher, R.E., 1980. Bull. Mar. Sci., 30, 145–146. Eschmeyer, W.N., Herald, E.S. & Hammann, H., 1983. A Field Guide to Pacific Coast Fishes of North America from the Gulf of Alaska to Baja California. Houghton Mifflin Company, Boston, 336 pp. Estes, J.A. & Steinberg, P.D., 1988. Paleobiology, 14, 19–36. Fänge, R. & Grove, D., 1979. In, Fish Physiology, Vol. VIII, Bioenergetics and Growth, edited by W.S.Hoar et al, Academic Press, New York, pp. 161–260. Ferraris, R.P. & Ahearn, G.A., 1983. J. Comp. Physiol, 152, 79–90. Fish, G.R., 1960. Hydrobiologia, 15, 161–178. Fishelson, L., Montgomery, W.L. & Myrberg Jr, A.H., 1985. Science, 229, 49–51. Fishelson, L., Montgomery, W.L. & Myrberg Jr, A.H., 1987. Mar. Ecol. Prog. Ser., 39, 37–47. Foster, M.S., 1972. In, Proc. 7th Int. Seaweed Symp., edited by K.Nisizawa, University of Tokyo Press, Tokyo, Japan, pp. 55–60. Foster, M.S., 1975. Mar. Biol., 32, 331–342. Foster, M.S. & Schiel, D.R., 1985. The Ecology of Giant Kelp Forests in California: A Community Profile. U.S.Fish Wildl. Serv. Biol. Rep., 85 (7.2), 152 pp. Foster, S.A., 1985a. Oecologia (Berlin), 67, 499–505. Foster, S.A., 1985b. Anim. Behav., 33, 782–792. Foster, S.A., 1987. J.Exp. Mar. Biol. Ecol, 105, 1–20. Frydl, & Stearn, C.W., 1978. J. Sediment. Petrol., 48, 1149–1157. Gaines, S.D. & Lubchenco, J., 1982. Amu. Rev. Ecol. Syst., 13, 111–138. Gaines, S.D. & Roughgarden, J., 1987. Science, 235, 479–481. Galzin, R., 1977. Thèse 3ème cycle, Université Sciences et Techniques du Languedoc, France, 109 pp. Geevarghese, C., 1983. J. Nat. Hist., 17, 731–741. Gerking, S.D., 1984. Trans. Am. Fish. Soc., 113, 378–387. Gibson, R.N., 1968. Vie Milieu, Ser. A., 19, 447–456. Gibson, R.N., 1972. J. Anim. Ecol., 41, 189–207. Glynn, P.W., 1973. In, Biology and Geology of Coral Reefs, Vol. II: Biology 1, edited by O.A.Jones & R.Endean, Academic Press, New York, pp. 271–324. Gohar, H.A.F. & Latif, A.F.A., 1959. Publ. Mar. Biol. Stn Ghardaga, Red Sea, No. 10, pp. 145–189. Gohar, H.A.F. & Latif, A.F.A., 1961a. Publ. Mar. Biol. Stn Ghardaga, Red Sea, No. 11, pp. 95–126. Gohar, H.A.F. & Latif, A.F.A., 1961b. Publ. Mar. Biol. Stn Ghardaga, Red Sea, No. 11, pp. 127–146. Gohar, H.A.F. & Latif, A.F.A., 1961c. Publ. Mar. Biol. Stn Ghardaga, Red Sea, No. 11, pp. 215–234. Gohar, H.A.F. & Latif, A.F.A., 1963. Publ. Mar. Biol. Stn Ghardaga, Red Sea, No. 12, pp. 3–42. Goldschmid, A., Kotrschal, K. & Wirtz, P., 1984. Zool. Anz., Jena, 213, 145–150. Gomon, M.F. & Paxton, J.R., 1985. Indo-Pacific Fishes, No. 8, 57 pp. Gosline, W.A., 1965. Ecology, 46, 823–831. Grant, J.J., Wilson, K.C., Grover, A. & Togstad, H.A., 1982. Mar. Fish. Rev., 44, 53–60. Gray, R.D., 1987. In, Foraging Behavior, edited by A.C.Kamil et al., Plenum Press, New York, pp. 69–140. Grossman, G.D., 1986. J. Zool. B, 1, 317–355. Grove, D.J. & Crawford, C., 1980. J. Fish Biol., 16, 235–247. Grovhoug, J.G. & Henderson, R.S., 1978. Atoll Res. Bull., No. 221, pp. 99–158. Gygi, R.A., 1969. Bermuda Biol. Stn, Spec. Publ., No. 2, pp. 137–143. Gygi, R.A., 1975. Ecologae Geol. Helv., 68, 327–359. Hallacher, L.E. & Roberts, D.A., 1985. Environ. Biol. Fish., 12, 91–110. Harris, L.G., Ebeling, A.W., Laur, D.R. & Rowley, R.J., 1984. Science, 224, 1336– 1338. Hatcher, B.G., 1981. Proc. 4th Int. Coral Reef Symp., Manila, 2, 515–524.

BIOLOGY OF MARINE HERBIVOROUS FISHES

215

Hatcher, B.G. & Larkum, A.W.D., 1983. J. Exp. Mar. Biol. Ecol., 69, 61–84. Hay, M.E., 1981a. Aquat. Bot., 11, 97–109. Hay, M.E., 1981b. Am. Nat., 118, 520–540. Hay, M.E., 1984a. Oecologia (Berlin), 64, 396–407. Hay , M.E., 1984b. Ecology, 65, 446–454. Hay, M.E., 1985. Proc. 5th Int. Coral Reef Congr., Tahiti, 4, 29–34. Hay, M.E., 1986. Am. Nat., 128, 617–641. Hay, M.E., Colburn, T. & Downing, D., 1983. Oecologia (Berlin), 58, 299–308. Hay, M.E., Duffy, J.E., Pfister, C.A. & Fenical, W., 1987a. Ecology, 68, 1567–1580. Hay, M.E. & Fenical, W., 1988. Annu. Rev. Syst. Ecol., 19, 111–145. Hay, M.E., Fenical, W. & Gustafson, K., 1987b. Ecology, 68, 1581–1591. Hay, M.E. & Goertemiller, T., 1983. In, The Ecology of Deep and Shallow Coral Reefs, edited by M.L.Reaka, Symposium Series for Undersea Research, Vol. 1, Office of Undersea Research, NOAA, Rockville, Maryland, pp. 97–102. Hay, M.E., Paul, V.J., Lewis, S.M., Gustafson, K., Tucker, J. & Trindell, R.N., 1988a. Oecologia (Berlin), 75, 233–245. Hay, M.E., Renaud, P.E. & Fenical, W., 1988b. Oecologia (Berlin), 75, 246–252. Hay, M.E. & Taylor, P.R., 1985. Oecologia (Berlin), 65, 591–598. Hiatt, R.W. & Strasburg, D.W., 1960. Ecol. Monogr., 30, 65–127. Hinds, P.A. & Ballantine, D.L., 1987. Aquat. Bot., 27, 299–308. Hixon, M.A., 1983. In, The Ecology of Deep and Shallow Coral Reefs, edited by M.L.Reaka, Symposium Series for Undersea Research, Vol. 1, Office of Undersea Research, NOAA, Rockville, Maryland, pp. 79–87. Hixon, M.A., 1985. In, The Ecology of Coral Reefs, edited by M.L.Reaka, Symposium Series for Undersea Research, Vol. 3, Office of Undersea Research, NOAA, Rockville, Maryland, pp. 27–32. Hixon, M.A., 1986. In, Contemporary Studies on Fish Feeding, edited by C.A. Simenstad & G.M.Cailliet, Dr W.Junk Publishers, Dordrecht, The Netherlands, pp. 235–257. Hixon, M.A. & Brostoff, W.N., 1981. Proc. 4th Int. Coral ReefSymp., Manila, 2, 507–514. Hixon, M.A. & Brostoff, W.N., 1982. In, Gutshop 81, Fish Food Habits Studies, Proceedings of the Third Pacific Workshop, edited by G.M.Cailliet & C.A. Simenstad, Washington Sea Grant Publ., University of Washington, Seattle, pp. 249–257. Hixon, M.A. & Brostoff, W.N., 1983. Science, 220, 511–513. Hixon, M.A. & Brostoff, W.N., 1985. Bull. Mar. Sci., 37, 200–213. Hobson, E.S., 1972. Fish. Bull., 70, 715–740. Hobson, E.S., 1974. Fish. Bull., 72, 915–1031. Hobson, E.S. & Chess, J.R., 1978. Fish. Bull., 76, 133–153. Hofer, R. & Schiemer, F., 1981. Oecologia (Berlin), 48, 342–345. Horn, M.H., 1983. Oecologia (Berlin), 58, 345–350. Horn, M.H., Murray, S.N. & Edwards, T.W., 1982. Mar. Biol., 67, 237–246. Horn, M.H., Murray, S.N. & Seapy, R.R., 1983. Bull. South. Calif. Acad. Sci., 82, 79–94. Horn, M.H. & Neighbors, M.A., 1984. Trans. Am. Fish. Soc., 113, 388–396. Horn, M.H., Neighbors, M.A. & Murray, S.N., 1986. J. Exp. Mar. Biol. Ecol., 103, 217–234. Horn, M.H., Neighbors, M.A., Rosenberg, M.J. & Murray, S.N., 1985. J. Exp. Mar. Biol. Ecol. , 86, 241–253. Howe, H.F. & Westley, L.C., 1988. Ecological Relationships of Plants and Animals. Oxford University Press, New York, 273 pp. Hughes, D.E., Wimpenny, J.W.T. & Lloyd, D., 1971. In, Methods in Microbiology, Vol. 5B, edited by J.R.Norris & D.W.Ribbons, Academic Press, London, pp. 1–54. Hughes, R.N., 1980. Oceanogr. Mar. Biol. Annu. Rev., 18, 423–181. Hughes, T.P., Reed, D.C. & Boyle, M.-J., 1987. J. Exp. Mar. Biol. Ecol., 113, 39– 59. Hutchings, P.A., 1986. Coral Reefs, 4, 239–252.

216

MICHAEL H.HORN

Hyatt, K.D., 1979. In, Fish Physiology, Vol. VIII, Bioenergetics and Growth, edited by W.S.Hoar, et al, Academic Press, New York, pp. 71–119. Irvine, G.V., 1982. In, Gutshop 81, Fish Food Habits Studies, Proceedings of the Third Pacific Workshop, edited by G.M.Cailliet & C.A.Simenstad, Washington Sea Grant Publ., University of Washington, Seattle, pp. 240–248. Ishida, J., 1935. Annot. Zool. Jpn, 15, 182–189. Ishida, J., 1936. Annot. Zool. Jpn, 15, 263–284. Jobling, M., 1980. J. Fish Biol., 17, 187–196. Jobling, M., 1987. J. Fish Biol., 30, 299–314. John, D.M. & Pople, W., 1973. J. Exp. Mar. Biol. Ecol., 11, 81–90. Johnson, G.D., 1984. Am. Soc. Ichthyol. Herpetol. Spec. Publ., No. 1, pp. 464–98. Jones, J.A., 1981. MA thesis, California State University, Fullerton, Fullerton, California, 62 pp. Jones, R.S., 1968. Micronesica, 4, 309–361. Jones, R.S. & Chase, J.A., 1975. Micronesica, 11, 127–148. Kapoor, B.G., Smit, H. & Verighina, I.A., 1975. Adv. Mar. Biol, 13, 109–239. Karasov, W.H., Buddington, R.K. & Diamond, J.M., 1985. In, Transport Processes, Iono-, and Osmoregulation, edited by R.Gilles & M.G.Gilles-Baillien, Springer-Verlag, Berlin, pp. 227–239. Kaufman, L.S. & Liem, K.F., 1982. Breviora, No. 472, 19 pp. Kennelly, S.J., 1983. J. Exp. Mar. Biol. Ecol, 68, 257–276. Kingsford, M.J., Schiel, D.R. & Battershill, C.N., 1988. Polar Biol., 9, 179–186. Klumpp, D.W., McKinnon, D. & Daniel, P., 1987. Mar. Ecol. Prog. Ser., 40, 41– 51. Klumpp, D.W. & Nichols, P.D., 1983. Mar. Ecol. Prog. Ser., 12, 207–216. Knights, B., 1985. In, Fish Energetics: New Perspectives, edited by P.Tytler & P. Calow, The Johns Hopkins University Press, Baltimore, Maryland, pp. 309–340. Kotrschal, K. & Thomson, D.A., 1986. Oecologia (Berlin), 70, 367–378. Lassuy, D.R., 1980. Bull. Mar. Sci., 30, 304–312. Lassuy, D.R., 1984. Environ. Biol. Fish., 10, 183–193. Lauder, G.V. & Liem, K.F., 1983. Bull. Mus. Comp. Zool. Harv. Univ., 150, 95– 197. Lee, D.J. & Putnam, G.B., 1973. J. Nutr., 103, 916–922. Leighton, D.L., 1971. Nova Hedwigia, 32, 421–453. Lessios, H.A., Robertson, D.R. & Cubit, J.D., 1984. Science, 226, 335–337. Levitan, D.R., 1988. J. Exp. Mar. Biol. Ecol., 119, 167–178. Lewis, S.M., 1985. Oecologia (Berlin), 65, 370–375. Lewis, S.M., 1986. Ecol. Monogr., 56, 183–200. Lewis, S.M., Norris, J.N. & Searles, R.B., 1987. Ecology, 68, 636–641. Lewis, S.M. & Wainwright, P.C., 1985. J. Exp. Mar. Biol. Ecol, 87, 215–228. Lewis, V.P. & Peters, D.S., 1984. J. Exp. Mar. Biol Ecol, 84, 95–100. Liem, K.F. & Greenwood, P.H., 1981. Am. Zool, 21, 83–101. Limbaugh, C., 1955. Fish Life in the Kelp Beds and the Effects of Kelp Harvesting. Univ. Calif. Inst. Mar. Res. IMR Ref. 55–9, 158 pp. Lindsay, G.J.H. & Harris, J.E., 1980. J. Fish Biol, 16, 219–233. Littler, M.M., 1976. Micronesica, 12, 27–41. Littler, M.M. & Littler, D.S., 1980. Am Nat., 116, 25–44. Littler, M.M., Littler, D.S. & Taylor, P.R., 1983a. J.Phycol, 19, 229–237. Littler, M.M., Taylor, P.R. & Littler, D.S., 1983b. Coral Reefs, 2, 111–118. Littler, M.M., Taylor, P.R. & Littler, D.S., 1986. Coral Reefs, 5, 63–71. Livingston, R.J., 1982. Mar. Ecol Prog. Ser., 7, 1–12. Lobel, P.S., 1980. Bull. Mar. Sci., 30, 273–289. Lobel, P.S., 1981. J. Fish Biol, 19, 365–397. Lobel, P.S. & Ogden, J.C., 1981. Mar. Biol, 64, 173–183.

BIOLOGY OF MARINE HERBIVOROUS FISHES

217

Lubchenco, J., 1978. Am. Nat., 112, 23–39. Lubchenco, J. & Gaines, S.D., 1981. Annu. Rev. Ecol Syst., 12, 405–437. Lundberg, B. & Lipkin, Y., 1979. Bot. Mar., 22, 173–181. Mackie, W. & Preston, R.D., 1974. In, Algal Physiology and Biochemistry, edited by W.D.P.Stewart, University of California Press, Berkeley, pp. 40–85. Mann, K.H., 1973. Science, 182, 975–981. Marais, J.F.K., 1980. J. Exp. Mar. Biol. Ecol., 44, 193–209. Marliave, J.B., 1977. J. Exp. Mar. Biol Ecol., 27, 47–60. Mattson Jr, W.J., 1980. Annu. Rev. Ecol. Syst., 11, 119–161. Medd, G.W., 1970. In, Scientific Explorations of the South Pacific, edited by W.S. Wooster, National Academy of Sciences, Washington, D.C., pp. 236–251. Meekan, M.G., 1986. MSc thesis, University of Auckland, Auckland, New Zealand, 79pp. Menge, B.A. & Lubchenco, J., 1981. Ecol. Monogr., 51, 429–450. Menzel, D.W., 1959. J. Cons., Cons. Int. Explor. Mer, 29, 308–313. Miller, A.C., 1982. Pacif. Sci., 36, 467–482. Miller, K.A. & Marshall, W.H., 1987. Calif. Fish Game, 73, 37–44. Millikin, M.R., 1982. Fish. Bull., 80, 655–686. Milton, K., 1981. Am. Nat., 117, 496–505. Montgomery, W.L., 1977. Copeia, 1977, 178–182. Montgomery, W.L., 1980a. Bull. Mar. Sci., 30, 290–303. Montgomery, W.L., 1980b. J. Exp. Mar. Biol. Ecol, 47, 9–24. Montgomery, W.L. & Gerking, S.D., 1980. Environ. Biol. Fish., 5, 143–153. Montgomery, W.L., Gerrodette, T. & Marshall, L.D., 1980. Bull. Mar. Sci., 30, 901–908. Montgomery, W.L. & Pollak, P.E., 1988. Mar. Ecol. Prog. Ser., 44, 7–13. Moran, M.J. & Sale, P.F, 1977. Mar. Biol., 39, 121–128. Moreno, C.A. & Jara, H.F., 1984. Mar. Ecol. Prog. Ser., 15, 99–107. Moriarty, D.J.W., 1973. J. Zool., 171, 25–39. Moriarty, D.J.W., 1976. J. Exp. Mar. Biol. Ecol., 22, 131–143. Moriarty, D.J.W., Pollard, P.C., Algoni, D.M., Wilkinson, C.R. & Gray, J.S., 1985. Proc. 5th Int. Coral Reef Congr., Tahiti, 3, 457–469. Morton, E.S., 1978. In, The Ecology of Arboreal Folivores, edited by G.G.Montgomery, Smithsonian Institution Press, Washington, D.C., pp. 123–130. Murray, S.N. & Horn, M.H., 1989. Bot. Mar. (in press). Nagy, K.A. & Shoemaker, V.H., 1984. Physiol. Zool., 57, 281–290. Nelson, J.S., 1984. Fishes of the World, John Wiley & Sons, New York, 2nd edition, 523 pp. Nelson, S.G. & Wilkins, S.DeC., 1988. J. Fish Biol., 32, 817–824. Niederholzer, R. & Hofer, R., 1979. J.Fish Biol, 15, 411–16. Norris, J.N. & Fenical, W.H., 1985. In, Handbook of Phycological Methods. Ecological Field Methods: Macroalgae, edited by M.M.Littler & D.S.Littler, Cambridge University Press, Cambridge, England, pp. 121–145. Norris, J.S., Norris, D.O. & Windell, J.T., 1973. J. Fish. Res. Board Can., 30, 201– 204. Norris, K.S. & Prescott, J.H., 1959. Copeia, 1959, 275–283. North, W.J., 1972. Annual Report, Kelp Habitat Improvement Project, 1971–1972. California Institute of Technology, Pasadena, pp. 57–99. North, W.J., 1973. Annual Report, Kelp Habitat Improvement Project, 1972–1973. California Institute of Technology, Pasadena, pp. 41–56. Nursall, J.R, 1977. J. Zool., 182, 205–223. Nursall, J.R., 1981. Zool. J. Linn. Soc., 72, 69–92. Odum, H.T. & Odum, E.P., 1955. Ecol. Monogr., 25, 291–320. Odum, W.E., .1968. Limnol. Oceanogr., 13, 92–98.

218

MICHAEL H.HORN

Odum, W.E., 1970. In, Marine Food Chains, edited by J.H.Steele, University of California Press, Berkeley, pp. 222–240. Ogburn, M.V., 1984. MS thesis, University of North Carolina, at Wilmington, Wilmington, North Carolina, 74 pp. Ogden, J.C., 1976. Aquat. Bot., 2, 103–116. Ogden, J.C., 1977. In, Reefs and Related Carbonates—Ecology and Sedimentology, edited by S.H.Frost et al., Am. Assoc. Petrol. Geol. Stud. Geol. No. 4, pp. 281– 288. Ogden, J.C., Brown, R. & Salesky, N., 1973. Science, 182, 715–717. Ogden, J.C. & Buckman, N.S., 1973. Ecology, 54, 589–596. Ogden, J.C. & Lobel, P.S., 1978. Environ. Biol. Fish., 3, 49–63. Paine, R.T., 1966. Am. Nat., 100, 65–75. Paine, R.T. & Vadas, R.L., 1969. Mar. Biol., 4, 79–86. Pandian, T.J. & Vivekanandan, E., 1985. In, Fish Energetics: New Perspectives, edited by P.Tytler & P.Calow, The Johns Hopkins University Press, Baltimore, Maryland, pp. 99–124. Pantastico, J.B., Baldia, J.P. & Reyes Jr, D.M., 1986. Aquaculture, 56, 169–178. Paul, V.J., 1987. Bull. Mar. Sci., 41, 514–522. Paul, V.J. & Fenical, W., 1983. Science, 221, 747–749. Paul, V.J. & Hay, M.E., 1986. Mar. Ecol. Prog. Ser., 33, 255–264. Paul, V.J., Hay, M.E., Duffy, J.E., Fenical, W. & Gustafson, K., 1987. J. Exp. Mar. Biol. Ecol. , 114, 249–260. Paul, V.J., Sun, H.H. & Fenical, W., 1982. Phytochemistry, 21, 468–69. Paul, V.J. & Van Alstyne, K.L., 1988. Coral Reefs, 6, 263–269. Payne, A.L, 1978. J. Fish Biol, 13, 627–629. Penry, D.L. & Jumars, P.A., 1986. Bioscience, 36, 310–315. Penry, D.L. & Jumars, P.A. 1987. Am. Nat., 129, 69–96. Pierce, G.J. & Ollason, J.G., 1987. Oikos, 49, 111–118. Pillay, T.V.R., 1953. Proc. Natl. Inst. Sci. India, 19, 777–827. Polunin, N.V.C., 1988. J. Exp. Mar. Biol. Ecol., 123, 61–76. Polunin, N.V.C. & Koike, L, 1987. J. Exp. Mar. Biol. Ecol., 111, 285–296. Potts, D.C., 1977. J. Exp. Mar. Biol. Ecol., 28, 207–216. Pough, F.H., 1973. Ecology, 54, 837–844. Power, M.E., 1984. Ecology, 65, 523–528. Power, M.E. & Matthews, W.J., 1983. Oecologia (Berlin), 60, 328–332. Prejs, A. & Blaszczyk, M., 1977. J. Fish Biol, 11, 447–452. Pyke, G.H., 1984. Annu. Rev. Ecol. Syst., 15, 523–575. Pyke, G.H., Pulliam, H.R. & Charnov, E.L., 1977. Quart. Rev. Biol, 52, 137–154. Quasi, J.C., 1968. In, Utilization of Kelp-Bed Resources in Southern California, edited by W.J.North & C.L.Hubbs, Calif. Dept. Fish Game Fish Bull. No. 139, pp. 109–142. Ragan, M.A. & Glombitza, K.-W., 1986. Prog. Phycol Res., 4, 130–241. Ralston, S.L. & Horn, M.H., 1986. J. Exp. Mar. Biol. Ecol, 98, 35–50. Rand, A.S., 1978. In, The Ecology of Arboreal Folivores, edited by G.G.Montgomery, Smithsonian Institution Press, Washington, D.C., pp. 115–122. Randall, J.E., 1961a. Pacif. Sci., 15, 215–272. Randall, J.E., 1961b. Ecology, 42, 812 only. Randall, J.E., 1963. Carib. J. Sci., 3, 31–47. Randall, J.E., 1965. Ecology, 46, 255–260. Randall, J.E., 1967. Stud. Trop. Oceanogr., 5, 665–847. Randall, J.E., 1968. Caribbean Reef Fishes. T.F.H. Publications, Jersey City, New Jersey, 318 pp. Randall, J.E., 1974. Proc. 2nd Int. Coral Reef Symp., Brisbane, 1, 159–166. Randall, J.E., 1986. In, Smith’s Sea Fishes, edited by M.M.Smith & P.C.Heemstra, Springer-Verlag, Berlin, pp. 706–714.

BIOLOGY OF MARINE HERBIVOROUS FISHES

219

Reese, E.S., 1973. Copeia, 1973, 145–149. Richards, W.J. & Leis, J.M., 1984. Am. Soc. Ichthyol. Herpetol Spec. Publ, No. 1, pp. 542–547. Riegle, K.C., 1976. MA thesis, California State University, Fullerton, Fullerton, California, 72 pp. Rimmer, D.W., 1986. Mar. Biol, 92, 443–148. Rimmer, D.W. & Wiebe, W.J., 1987. J. Fish Biol, 31, 229–236. Risk, M.J. & Sammarco, P.W., 1982. Oecologia (Berlin), 52, 376–380. Roberts, C.M., 1987. J. Exp. Mar. Biol Ecol, 111, 61–75. Robertson, A.I. & Klumpp, D.W., 1983. Mar. Ecol. Prog. Ser., 10, 197–201. Robertson, D.R. & Gaines, S.D., 1986. Ecology, 67, 1372–1383. Robertson, D.R., Hoffman, S.G. & Sheldon, J.M., 1981. Ecology, 62, 1162– 1169. Robertson, D.R. & Polunin, N.V.C., 1981. Mar. Biol, 62, 185–195. Robertson, D.R., Polunin, N.V.C. & Leighton, K., 1979. Environ. Biol Fish., 4, 125–170. Robertson, D.R., Sweatman, H.P.A., Fletcher, E.A. & Cleland, M.G., 1976. Ecology, 57, 1208–1220. Robins, C.R., Ray, G.C. & Douglass, J., 1986. A Field Guide to Atlantic Coast Fishes of North America. Houghton Mifflin Company, Boston, 354 pp. Robison, B.H., 1984. Mar. Biol, 84, 119–123. Rosenblatt, R.H. & Hobson, E.S., 1969. Copeia, 1969, 434–453. Rozin, P. & Mayer, J., 1964. Am. J.Physiol, 206, 1430–1436. Russ, G., 1984a. Mar. Ecol. Prog. Ser., 20, 23–34. Russ, G., 1984b. Mar. Ecol Prog. Ser., 20, 35–44. Russ, G.R., 1987. J. Exp. Mar. Biol. Ecol., 110, 1–17. Russell, B.C., 1977. N.Z.J.Mar. Freshwater Res., 11, 23–36. Russell, B.C., 1983. N.Z.J.Mar. Freshwater Res., 17, 121–145. Russell-Hunter, W.D., 1970. Aquatic Productivity: An Introduction to Some Basic Aspects of Biological Oceanography and Limnology. Collier-MacMillan, London, 306 pp. Ruyter van Steveninck, E.D. de, 1984. Aquat. Bot., 20, 11–19. Ruyter van Steveninck, E.D.de, Kamermans, P. & Breeman, A.M., 1988. Mar. Ecol. Prog. Ser., 44, 77–84. Sale, P.F., 1977. Am. Nat., 111, 337–359. Sale, P.F., 1980. Oceanogr. Mar. Biol. Annu. Rev., 18, 367–421. Sammarco, P.W., 1982a. J. Exp. Mar. Biol. Ecol., 61, 31–55. Sammarco, P.W., 1982b. J. Exp. Mar. Biol. Ecol, 65, 83–105. Sammarco, P.W., 1983. Mar. Ecol. Prog. Ser., 13, 1–14. Sammarco, P.W. & Carleton, J.H., 1981. Proc. 4th Int. Coral Reef Symp., Manila, 2, 525–535. Sammarco, P.W., Carleton, J.H. & Risk, M.J., 1986. J. Exp. Mar. Biol. Ecol., 98, 1–19. Sammarco, P.W., Levinton, J.S. & Ogden, J.C., 1974. J. Mar. Res., 32, 47– 53. Sammarco, P.W., Risk, M.J. & Rose, C., 1987. J. Exp. Mar. Biol. Ecol., 112, 185– 199. Sammarco, P.W. & Williams, A.H., 1982. Mar. Ecol. Prog. Ser., 8, 53–59. Santelices, B., 1987. Rev. Chilena Nat. Hist., 60, 153–158. Schiel, D.R., Kingsford, M.J. & Choat, J.H., 1986. N.Z.J.Mar. Freshwater Res., 20, 521–535. Schluter, D., 1981. Am. Nat., 118, 139–147. Schoener, T.W., 1971. Annu. Rev. Ecol. Syst., 11, 369–404. Schoener, T.W., 1987. In, Foraging Behavior, edited by A.C.Kamil et al, Plenum Press, New York, pp. 5–67. Scoffin, T.P., Stearn, C.W., Boucher, D., Frydl, P., Hawkins, C.M., Hunter, I.G. & MacGeachy, J.K., 1980. Bull. Mar. Sci., 30, 475–508. Sell, J.L., Hasiak, R.J. & Owings, W.J., 1985. Poult. Sci., 64, 1527–1535. Sibly, R.M. & Calow, P., 1986. Physiological Ecology of Animals, an Evolutionary Approach. Blackwell Scientific Publications, Oxford, England, 179 pp. Smit, H., 1967. Comp. Biochem, Physiol., 21, 125–132.

220

MICHAEL H.HORN

Smit, H., 1968. In, Handbook of Physiology, Vol. 5, edited by C.F.Code, American Physiological Society, Washington, D.C., pp. 2791–2805. Smith, D.C. & Douglas, A.E., 1987. The Biology of Symbiosis. Edward Arnold, London, 302 pp. Smith, J.L.B. & Smith, M.M., 1986. In, Smiths’ Sea Fishes, edited by M.M.Smith & P.C.Heemstra, Springer-Verlag, Berlin, pp. 580–594. Smith, R.L. & Paulson, A.C., 1974. Copeia, 1974, 796–799. Stearns, S.C. & Schmid-Hempel, P., 1987. Oikos, 49, 118–125. Steinberg, P.D., 1984. Science, 223, 405–07. Steinberg, P.D., 1985. Ecol Monogr., 55, 333–349. Steinberg, P.D., 1986. Oecologia (Berlin), 69, 628–630. Steinberg, P.D., 1988. J. Exp. Mar. Biol. Ecol., 120, 221–237. Steneck, R.S., 1983. Paleobiology, 9, 44–61. Steneck, R.S. & Watling, L., 1982. Mar. Biol., 68, 299–319. Stenseth, N.C. & Hansson, L., 1979. Am. Nat., 113, 373–389. Stephens, D.W. & Krebs, J.R., 1986. Foraging Theory. Princeton University Press, Princeton, New Jersey, 247 pp. Stephens Jr, J.S. & Zerba, K.E., 1981. Environ. Biol. Fish., 6, 111–121. Stephenson, W. & Searles, R.B., 1960. Aust. J.Mar. Freshwater Res., 2, 241–267. Stickney, R.R., 1975. Proc. 29th Ann. Conf. Southeastern Assoc. Game Fish Commissioners, pp. 282–287. Stickney, R.R. & Shumway, S.E., 1974. J. Fish Biol, 6, 779–790. Stoner, A.W., 1980. Fish. Bull., 78, 337–352. Stoner, A.W. & Livingston, R.J., 1984. Copeia, 1984, 174–187. Sun, H.H., Paul, V.J. & Fenical, W., 1983. Phytochemistry, 22, 743–745. Suyehiro, Y., 1942. Jpn J. Zool., 10, 1–303. Taborsky, M. & Limberger, D., 1980. P.S.Z.N.I. Mar. Ecol., 1, 143–153. Tacon, A.G.J. & Cowey, C.B., 1985. In, Fish Energetics: New Perspectives, edited by P.Tytler & P.Calow, The Johns Hopkins University Press, Baltimore, Maryland, pp. 155–183. Taghon, G.L., 1981. Am. Nat., 118, 202–214. Talbot, C., 1985. In, Fish Energetics: New Perspectives, edited by P.Tytler & P. Calow, The Johns Hopkins University Press, Baltimore, Maryland, pp. 125–154. Talbot, F.H., 1965. Proc. Zool. Soc. Land., 145, 431–470. Targett, N.M., Targett, T.E., Vrolijk, N.H. & Ogden, J.C., 1986. Mar. Biol., 92, 141–148. Targett, T.E., 1981. Mar. Ecol. Prog. Ser., 4, 243–263. Thayer, G.W., Bjorndal, K.A., Ogden, J.C., Williams, S.L. & Zieman, J.C., 1984. Estuaries, 7, 351–376. Thomson, D.A., Findley, L.T. & Kerstitch, A.N., 1979. Reef Fishes of the Sea of Cortez, the Rocky-Shore Fishes of the Gulf of California, John Wiley & Sons, New York, 302 pp. Thomson, D.A. & Lehner, C.E., 1976. J. Exp. Mar. Biol. Ecol., 22, 1–29. Thomson, J.M., 1954. Aust. J.Mar. Freshwater Res., 5, 469–485. Thomson, J.M., 1959. Aust. J.Mar. Freshwater Res., 10, 365–374. Thomson, J.M., 1966. Oceanogr. Mar. Biol. Annu. Rev., 4, 301–335. Thresher, R.E. & Colin, P.L., 1986. Bull. Mar. ScL, 38, 253–272. Titcomb, M., 1972. Native Use of Fish in Hawaii. University of Hawaii Press, Honolulu, 2nd edition, 175 pp. Townsend, C.R. & Hughes, R.N., 1981. In, Physiological Ecology: An Evolutionary Approach to Resource Use, edited by C.R.Townsend & P.Calow, Sinauer Associates, Sunderland, Massachusetts, pp. 86–108. Troyer, K., 1984. Physiol. Zool., 57, 1–8. Tsuda, R.T. & Bryan, P.G., 1973. Copeia, 1973, 604–606. Urquhart, K.A.F., 1984. MA thesis, California State University Fullerton, Fullerton, California, 93 pp. Van den Hoek, C., Breeman, A.M., Bak, R.P.M. & Van Buurt, G., 1978. Aquat. Bot., 5, 1–46. Vaughan, F.A., 1978. Bull. Mar. Sci., 28, 527–536. Vine, P.J., 1974. Mar. Biol., 24, 131–136.

BIOLOGY OF MARINE HERBIVOROUS FISHES

Walker, D.L, 1984. DPhil thesis, University of York, York, England, 188 pp. Wanders, J.B.W., 1977. Aquat. Bot., 3, 357–390. Weinstein, M.P., Heck Jr, K.L., Giebel, P.E. & Gates, J.E., 1982. Bull. Mar. Sci. 32, 791–795. Wellington, G.M., 1982. Ecol. Monogr., 52, 223–241. Wellington, G.M. & Victor, B.C, 1985. Oecologia (Berlin), 68, 15–19. Westernhagen, H.von, 1974. Aquaculture, 3, 109–117. Westoby, M., 1974. Am. Nat., 108, 290–304. Westoby, M., 1978. Am. Nat., 112, 627–631. Wheeler, A., 1980. In, The Shore Environment, Vol. 2: Ecosystems, edited by J.H. Price, D.E.G.Irvine & W.F.Farnham, Academic Press, London, pp. 677–698. White, T.C.R., 1985. Oecologia (Berlin), 67, 596–597. Wiegert, R.F. & Owen, D.F., 1971. J.Theor. Biol., 69, 69–91. Wilkins, M.B., 1984. Advanced Plant Physiology. Pitman, London, 514 pp. Wilkinson, C.R. & Sammarco, P.W., 1983. Mar. Ecol. Prog. Ser., 13, 15–19. Williams, A.H., 1980. Am. Nat., 116, 138–142. Williams, A.H., 1981. Ecology, 62, 1107–1120. Williams, G.C. & Williams, D.C., 1955. Calif. Fish Game, 41, 203–208. Wilson, R.P. & Halver, J.E., 1986. Annu. Rev. Nutr., 6, 225–244. Wolf, N.G., 1985. Proc. 5th Int. Coral Reef Congr., Tahiti, 4, 23–28. Wylie, C.R. & Paul, V.J., 1988. Mar. Ecol. Prog. Ser., 45, 23–32. Yamamoto, N. & Lopez, G., 1985. J. Exp. Mar. Biol. Ecol., 90, 209–220. Zeller, D.C., 1988. Mar. Ecol. Prog. Ser., 44, 85–93.

221

Oceanogr. Mar. Biol. Annu. Rev., 1989, 27, 273–335 Margaret Barnes, Ed. Aberdeen University Press

THE BENGUELA ECOSYSTEM PART VI SEABIRDS A.BERRUTI* Sea Fisheries Research Institute, Private Bag X2, Rogge Bay 8012, Cape Town, South Africa N.J.ADAMS and S.JACKSON Percy FitzPatrick Institute of African Ornithology, University of Cape Town, Rondebosch 7700, Cape Town, South Africa

ABSTRACT The ecology of seabirds in the Benguela upwelling system off western southern Africa is reviewed. The marine avifauna comprises a distinctive assemblage of 12 breeding seabirds (including nine endemic taxa) and 36 species which are regular non-breeding visitors. Research has concentrated on the resident seabirds, particularly on the interaction between the three most abundant species, the Cape gannet oM rus ac nep sis, aj ckass penguin nehpS i scus ed rem sus, and Cape cormorant ahP al rc ocorax ac ep nsis, and their commercially-exploited fish prey, notably the pilchard aS rdinop s oce al t us and anchovy nE rg aul is aj op nicus ac nep sis . Changes in the population sizes and th e diets of these three species are consistent with changes in the catches of the purse-seine fisheries. The pelagic ecology of the non-breeding seabirds and the responses of seabirds to small-scale and mesoscale variability are poorly known. Bottomtrawling activities appear, however, to be an important determinant of the distribution of the larger non-breeding species. At present population levels, seabirds probably play a relatively minor role in energy and nutrient cycling in the Benguela ecosystem. Although the processes regulating primary production in the Benguela ecosystem appear to be intact, man has brought about major transformations. Seabird numbers have been reduced in the past by the collection of eggs, chicks, adults, and guano and by the over-exploitation of epipelagic fish prey. aM n has also reduced the numbers of predators thought to make prey available to seabirds in mixed-species feeding assemblages.

INTRODUCTION The Benguela ecosystem is one of the world’s four major eastern-boundary current regions (Wooster & Reid, 1963), where primary production is enhanced by wind-driven upwelling (Cushing, 1969) and which are dominated by similar fish assemblages (Parrish, Bakun, Husby & Nelson, 1983; Crawford, 1987).

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

223

Various aspects of the Benguela ecosystem have been reviewed: physical features and processes (Shannon, 1985), chemical processes (Chapman & Shannon, 1985), plankton (Shannon & Pillar, 1986), important fish invertebrate resources (Crawford, Shannon & Pollock, 1987), and coastal zone ecology (Branch & Griffiths, 1988). This paper reviews the literature on the ecology of seabirds in the Benguela ecosystem and attempts to identify key ecological processes which affect seabird populations. A full bibliography of all ornithological literature referring to the seabirds of southern Africa up to 1980 is contained in Cooper & Brooke (1981) and is partly updated by Brown (1985). Aspects of the interaction between the Cape gannet Morus capensis, jackass penguin Spheniscus demersus, and Cape cormorant Phalacrocorax capensis and their commercially valuable fish prey are discussed by Crawford et al. (1987). SEABIRD RESEARCH IN THE BENGUELA ECOSYSTEM Seabirds are conspicuous top predators in marine ecosystems, often competing directly with man for food (Furness, 1982). The interaction between large populations of resident seabirds and their commercially important fish prey has provided the impetus for much of the seabird research in the Benguela ecosystem (Frost, 1981; Hockey, Cooper & Duffy, 1983). This topic was the subject of intensive research in the 1950s (e.g. Davies, 1955; Rand, 1959, 1960a,b, 1963a,b; Matthews, 1961) and again from the late 1970s (e.g. Crawford & Shelton, 1978, 1981). Changes in the diet and population dynamics of the abundant resident seabirds relative to changes in fish resources over a period of 30 years (e.g. Crawford & Shelton, 1978, 1981) are probably better known in the Benguela system than for any other large marine ecosystem over a similar length of time. Since the 1970s, the scope of marine ornithological research has broadened and the biology of most species breeding in southern Africa has been studied at the nest (Hockey et al. 1983). The pelagic ecology of only the jackass penguin and Cape gannet are, however, well known (Wilson, 1985a; Berruti, 1987; Heath & Randall, in press; Wilson, Wilson & Duffy, 1988). Comparatively little is known of the ecology of the non-breeding seabirds which seasonally visit the Benguela ecosystem (Cooper, 1981a). This imbalance is reflected in the unequal coverage given to these two groups in this review. Seabirds are defined as birds which obtain all or much of their food from the open ocean. Birds which feed in the exposed intertidal region or on the shoreline, or which seldom occur at sea, are excluded. THE OCEANOGRAPHY OF THE BENGUELA ECOSYSTEM AND AGULHAS BANK Seabird assemblages are influenced at different spatial and temporal scales by dynamic oceanographic processes which affect the availability, distribution and abundance of their marine prey (Schneider & Piatt, 1986; Hunt & Schneider, 1987). The reviews of these processes and their effects (Chapman & Shannon, 1985; Shannon, 1985; Shannon & Pillar, 1986) are used to describe briefly salient features of the Benguela ecosystem which may affect seabirds. We define the northern boundary of the Benguela System as the Angola- Benguela front, situated between 15c and 17°S (Shannon, Agenbag & Buys, 1987). The 1000-m depth contour is taken as the offshore boundary, and includes the entire continental shelf and trawling grounds for bottom-fish (Fig 1) (Shannon, 1985). Cape Point does not constitute a biological boundary (Duffy, Siegfried & Jackson, 1987b) and extensive migrations of epipelagic fish occur between the west coast and Agulhas Bank (Crawford et

*Present address: Durban Natural History Museum, PO Box 4085, Durban 4000, South Africa

224

A.BERRUTI, N.J.ADAMS AND S.JACKSON

Fig 1.—The Benguela Ecosystem, Agulhas Bank and the continental shelf of southern Africa as defined by the 1000 mdepth contour. Place names mentioned in the text are given.

al. 1987). In the area east of Cape Agulhas, several resident seabirds of the Benguela ecosystem breed in greatly reduced numbers, or not at all. We consider the western Agulhas Bank as part of the Benguela ecosystem. The latitude of the mouth of the Orange River is arbitrarily used to divide the Benguela ecosystem into northern and southern sectors. Differing marine water masses, which may constitute distinct habitats for seabirds, are temporally and spatially mobile (Shannon, 1985). The Benguela ecosystem is characterised by upwelling and equatorward surface flow (Fig 1). The area influenced by upwelling is 150–200 km wide. The frontal region between this zone and warmer oceanic waters is dynamic and is characterised by plumes, filaments, eddies, and the advection of filaments of warm Agulhas Current water into the southern Benguela ecosystem (Lutjeharms & Stockton, 1987). Plumes of upwelled waters are separated from oceanic waters by sharp thermosaline gradients and are most marked during active upwelling. Upwelling occurs at different temporal and spatial scales. Eight upwelling cells exist between 15c and 35°S, including the Cape Peninsula and Cape Columbine (Lutjeharms & Meeuwis, 1987). In the southern Benguela off the Cape Peninsula, upwelling is pulsed with a periodicity of 3–6 days but is less variable and slower north of Cape Columbine. In the northern Benguela north of Lüderitz, upwelling is most intense in winter and spring, but is perennial at Lüderitz. In the southern Benguela, upwelling is most intense in summer, but advection of warmer offshore waters may restrict the surface area of upwelling, and the thermal front may lie close inshore when active upwelling relaxes (Shelton, Boyd & Armstrong, 1985). In winter, the water column is well mixed and the zone of cool water along the coast is broad. During the process of upwelling and subsequent warming of waters transported away from the coast, the nutrient levels become depleted as phytoplankton biomasses increase, lagged by zooplankton blooms. Secondary plankton blooms may occur. There are interannual differences in the intensity of upwelling. The spatial distribution of phytoplankton is highly variable in the southern Benguela, but dense concentrations of phytoplankton are often located offshore in aged, upwelled waters. Nutrient levels and plankton production are consistently high in the southern Benguela north of Cape Columbine. Zooplankton levels are highest but most variable inshore, but are higher in St Helena Bay than on the Agulhas Bank. The extensive upwelling area at Lüderitz (Lutjeharms & Meeuwis, 1987) apparently acts as a barrier to epipelagic, neritic fish, separating stocks of the same fish species (Crawford et al., 1987; Agenbag & Shannon, 1988). The Agulhas Bank is a wide bulge of the continental shelf off the southern tip of Africa east of the Benguela system (see Fig 1), which deflects the Agulhas Current farther south. Upwelling occurs along the eastern edge of the Agulhas Bank. In summer and late autumn, waters of the Agulhas Bank are characterised by thermoclines and sub-surface phytoplankton maxima, but in winter, waters are well-mixed (Boyd, Tromp & Horstman, 1985; Carter, McMurray & Largier, 1987). Nutrient upwelling is limited. Moderate levels of nutrients occur only in spring; levels in summer are lower because of nutrient depletion in strongly stratified waters (Brown & Hutchings, 1985). The structural features of the southern Benguela system proposed by Bang (1971) are variable spatially (both longshore and offshore) and temporally, and this variability should be recognised in investigations of the structure and role of seabird assemblages, communities or habitats (cf Abrams, 1985a).

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

225

THE SEABIRD ASSEMBLAGE OF THE BENGUELA ECOSYSTEM Seabirds are the most mobile and conspicuous of marine organisms and a large proportion of the seabird species recorded in southern African waters are vagrants or very rare (Brooke & Sinclair, 1978; Clancey, Brooke, Irwin & Markus, 1980; Clancey, Brooke, Crowe & Mendelsohn, 1987; Ryan & Rose, 1989). An assessment of the regularity of occurrence and abundance of seabird species recorded in the Benguela ecosystem is therefore necessary. Breeding seabirds are defined as those species which breed in the Benguela ecosystem; non-breeding species are usually seasonal visitors. THE BREEDING SEABIRDS Of the 14 species of birds usually regarded as the breeding seabirds of southern Africa (Brooke, 1981a; Cooper, Williams & Britton, 1984; Berruti, 1989), 12 breed within the Benguela ecosystem, the greyheaded gull Lams cirrocephalus and roseate tern Sterna dougallii being excluded (Table I). Although the greyheaded gull occurs at the Cunene estuary, along the central Namibian coast and very sparsely along the coastline of the southern Benguela (Whitelaw, Underbill, Cooper & Clinning, 1978; Cooper, Robertson & Shaughnessy, 1980; Ryan, Cooper, Stutterheim & Loutit, 1984b; Cooper, Hockey & Ryan, in prep.), it mainly inhabits freshwater and estuarine habitats and scavenges along the shoreline (Whitfield, 1977; De Kock & Randall, 1984). Only one reference to feeding at sea in southern Africa was located—Miller (1951) stated that the greyheaded gull occasionally feeds on pilchards Sardinops ocellatus off Natal during the “sardine run”. This species should not be regarded as a seabird until it is shown that it regularly feeds at sea. At present, the roseate tern breeds only in Algoa Bay in southern Africa (Randall & TABLE I The breeding seabird species of the Benguela ecosystem Species Sphenisciformes jackass penguin, Spheniscus demersus Pelecaniformes white pelican, Pelecanus onocrotalus Cape gannet, Morus capensis Cape cormorant, Phalacrocorax capensis bank cormorant, P.neglectus crowned cormorant, P.coronatus whitebreasted cormorant, P.carbo lucidus Charadriiformes kelp gull, Larus dominicanus Hartlaub’s gull, L.hartlaubii Damara tern, Sterna balaenarum swift tern, S.bergii bergii Caspian tern, Hydroprogne caspia

Status endemic widespread endemic endemic endemic endemic widespread endemic subspecies endemic endemic endemic subspecies widespread

Randall, 1980). Although it has previously bred in very small numbers at Dyer Island in the Benguela ecosystem, the bulk of the small southern African population has always bred at localities in the Algoa Bay

226

A.BERRUTI, N.J.ADAMS AND S.JACKSON

region (Randall & Randall, 1980). Roseate terns occur rarely in the Benguela ecosystem (e.g. Cooper et al., in prep.) and this species is therefore regarded as a rare vagrant. Two breeding species with a tenuous claim as seabirds are the white pelican Pelecanus onocrotalus and Caspian tern Hydroprogne caspia. White pelicans rarely feed at sea in the southern Benguela (Guillet & Crowe, 1981; Brooke, 198la) but in the northern Benguela they occur regularly in the Walvis Bay lagoon, in Sandwich Harbour (40 km south of Walvis Bay), and along the central Namibian coast (Berry & Berry, 1975; Whitelaw et al., 1978; Cooper et al., 1980), where they presumably feed on marine prey. Although the Caspian tern mainly occurs in freshwater and estuarine habitats, it feeds regularly in calm marine bays. The degree to which the two species rely, however, on marine prey requires further study. NON-BREEDING SEABIRDS A total of 77 non-breeding seabird species has been recorded in southern Africa (Ryan & Rose, 1989) (Table II). We exclude six of these species, the lesser blackbacked gull Larus fuscus, herring gull LOTUS argentatus, Franklin’s gull Larus pipixcan, blackheaded gull Larus ridibundus, gullbilled tern Gelochelidon nilotica, and whitewinged tern Chlidonias leucoptera, because they are vagrants or only occasionally marine in the Benguela system. TABLE II Seabird species which visit but do not breed in the Benguela ecosystem, divided into species which occur regularly and species which are vagrant, irruptive or very rare Regular species Podicipediformes blacknecked grebe Procellariiformes wandering albatross shy albatross blackbrowed albatross greyheaded albatross yellownosed albatross northern giant petrel southern giant petrel Antarctic fulmar pintado petrel greatwinged petrel softplumaged petrel broadbilled prion whitechinned petrel Cory’s shearwater great shearwater sooty shearwater Manx shearwater

Podiceps nigricollis Diomedea exulans D. cauta D. melanophris D. chrysostoma D. chlororhynchos Macronectes halli M. giganteus Fulmarus glacialoides Daption capense Pterodroma macroptera P. mollis Pachyptila vittata Procellaria aequinoctialis Calonectris diomedea Puffinus gravis P. griseus P. puffinus

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

little shearwater European storm petrel Leach’s storm petrel Wilson’s storm petrel blackbellied storm petrel Phalacrocoracidae reed cormorant Charadriiformes grey phalarope Arctic skua longtailed skua pomarine skua Subantarctic skua Sabine’s gull Sandwich tern common tern Arctic tern Antarctic tern little tern black tern Vagrant or very rare species Sphenisciformes king penguin macaroni penguin rockhopper penguin Procellariiformes royal albatross sooty albatross lightmantled sooty albatross Antarctic petrel Bulwer’s petrel whiteheaded petrel Atlantic petrel Kerguelen petrel blue petrel slenderbilled prion fairy prion grey petrel fleshfooted shearwater whitebellied storm petrel Pelecaniformes

P. assimilis Hydrobates pelagicus Oceanodroma leucorhoa Oceanites oceanicus Fregetta tropica Phalacrocorax africanus Phalaropus fulicarius Stercorarius parasiticus S. longicaudus S. pomarinus Catharacta antarctica Larus sabini Sterna sandvicensis S. hirundo S. paradisaea S. vittata S. albifrons Chlidonias niger

Aptenodytes patagonicus Eudyptes chrysolophus E. chrysocome Diomedea epomophora Phoebetria fusca P. palpebrata Thalassoica antarctica Bulweria bulwerii Pterodroma lessonii P. incerta P. brevirostris Halobaena caerulea Pachyptila belcheri P. turtur Procellaria cinerea Puffinus carneipes Fregetta grallaria

227

228

A.BERRUTI, N.J.ADAMS AND S.JACKSON

greater frigatebird redbilled tropic bird whitetailed tropic bird redtailed tropic bird Australian gannet Charadriiformes south polar skua blacklegged kittiwake roseate tern royal tern sooty tern bridled tern common noddy

Fregata minor Phaethon aethereus P. lepturus P. rubricauda Morus serrator Catharacta maccormicki Rissa tridactyla Sterna dougallii S. maxima S. fuscata S. anaethetus Anous stolidus

Another seven species are not known to have occurred in the Benguela system or Agulhas Bank: Laysan albatross Diomedea immutabilis, Audubon’s shearwater Puffinus lherminieri, wedgetailed shearwater P.pacificus, brown booby Sula leucogaster, blacknaped tern Sterna anaethetus, whitecheeked tern S.repressa, and lesser noddy Anous tenuirostris. The reed cormorant Phalacrocorax africanus is added to the list because it has been recorded as a non-breeding bird on the northern Namibian coast south of the Cunene river and in Algoa Bay (Crawford, Shelton, Brooke & Cooper, 1982b; Every & Spearpoint, 1984; Ryan, Cooper & Stutterheim, 1984a; Ryan et al, 1984b). The non-breeding seabird fauna of the Benguela system consists of 65 species: 36 occur regularly while 29 are regarded as vagrants, very rare or irruptive (Table II) and are not considered as regular members of the marine avifauna of the area. They are ignored in the following discussion. BIOGEOGRAPHY The breeding avifauna of the Benguela System, Agulhas Bank, and southern Agulhas Current (Algoa Bay) is highly distinctive, and is bounded in the north on both the east and west coasts by tropical marine avifaunas (Harrison, 1983; Cooper et al., 1984). No seabirds have been known to breed in Angola (Brooke, 1981b), or in southern Moçambique in this century (Brooke & Cooper, 1982). Of the 12 seabird species of the Benguela ecosystem, nine taxa (seven species and two subspecies) are restricted to southern Africa as breed ing species (see Table I). The specific status of the Cape gannet, Hartlaub’s gull Larus hartlaubii, and crowned cormorant Phalacrocorax coronatus has been debated, but they are at present regarded as full species (Clancey et al., 1980; Crawford et al., 1982b; Clancey et al, 1987). Subspecific status has been accorded to the kelp gull Larus dominicanus vetula and swift tern Sterna bergii bergii which breed in southern Africa (Brooke & Cooper, 1979a; Clancey et al., 1980; Clancey et al, 1987). Freshwater and marine populations of the whitebreasted cormorant Phalacrocorax carbo lucidus in southern Africa are assigned to the same subspecies although the marine population is apparently discrete (Brooke, Cooper, Shelton & Crawford, 1982; Jarvis, 1970). The breeding distribution of seabirds is constrained by the availability of breeding sites free of terrestrial mammalian predators. Offshore islands and guano platforms provide the safest and most important breeding sites in southern Africa (Cooper & Berruti, in press). There are four important groups of islands (including artificial sites) off southern Africa; the guano platforms off central and northern Namibia at Walvis Bay,

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

229

Swakopmund and Cape Cross, the group of islands off southern Namibia (both northern Benguela System), the islands off the southwestern Cape between Dyer Island and Lambert’s Bay (southern Benguela System) and the Algoa Bay islands (Agulhas Current) (see Fig 1). Four species, the whitebreasted cormorant, Damara tern, kelp gull, and Cape cormorant, breed north of Swakopmund in the northern Benguela, and their breeding distributions extend to Algoa Bay on the east coast (Clinning, 1978a; Randall, Randall, Batchelor & Ross, 1981a; Brooke et al., 1982; Cooper, Brooke, Shelton & Crawford, 1982; Crawford, Cooper & Shelton, 1982a). The relatively small number of breeding species off northern Namibia may be a result of the lack of natural breeding sites in this area. The only marine population of the white pelican in southern Africa breeds on a guano platform at Walvis Bay, central Namibia, having bred previously at Sandwich harbour (Crawford et al., 1981). Three species, the bank Phalacrocorax neglectus and crowned cormorants and Hartlaub’s gull breed from central Namibia to the western Agulhas area (Cooper, 1981b; Crawford et al. 1982a; Williams, 1988; Berruti, 1989). Another four species (jackass penguin, swift and Caspian terns, and Cape gannet) breed from central or southern Namibia to Algoa Bay (Clinning, 1978b; Randall et al., 1981a; Crawford et al., 1983b; Shelton, Crawford, Cooper & Brooke, 1984; Williams, 1988). There are relatively large breeding populations of only two (jackass penguin and Cape gannet) of the eight species that occur east of Cape Agulhas (Crawford et al., 1983b; Shelton et al., 1984). The breeding seabird assemblage of Algoa Bay is an extension of the Benguela avifauna. The breeding seabirds of the Benguela ecosystem are mainly derived from tropical orders: the Pelecaniformes and Charadriiformes (Brooke, 1981a). The jackass penguin is the only representative of Sphenisciformes, an order with its origins in the southern cold regions, although at least four fossil penguins are known from southern Africa (Brooke, 19811a). The lack of breeding Procellariiformes is surprising as they breed in other major eastern-boundary upwelling zones. The Benguela System had more breeding seabirds in the geological past. In the late Miocene and early Pliocene, the breeding fauna included penguins and petrels, as well as gannets and cormorants (Rich, 1980). Olson (1985a,b) suggested that the climate and waters of the southeastern Atlantic Ocean were colder at that time. The Benguela upwelling system is thought, however, to have evolved as an eastern-boundary upwelling system in the late Miocene (6–10 Myr ago), while persistent upwelling is thought to have become established in the late Pliocene (about 2 Myr ago) (Shannon, 1985). It is unlikely that the lack of breeding procellariiforms is a result of warmer conditions as suggested by Olson (1983, 1985a,b), because the Benguela ecosystem supports at least 22 procellariiform species as regular visitors and is no warmer than other eastern-boundary currents which support breeding procellariiforms. Changes in the availability of suitable nest sites and competition with abundant breeding seabirds and Cape fur seals Arctocephalus pusillus pusillus may, however, have caused the displacement of procellariiforms from the few flat islands remaining when water levels changed. The present assemblage of breeding seabirds has existed for at least 100 000 years, although the relative proportions of the species appear to have altered (Siegfried, Cooper & Avery, 1982). On a global scale, three major marine faunal zones have been defined: northern cold, tropical, and southern cold (Briggs, 1974). These may be subdivided into the Arctic and boreal zones; southern and northern subtropical and tropical zones; and subantarctic and Antarctic (Ashmole, 1971; Brooke, 19811a). Geographically, southern Africa falls within the southern subtropical zone (Ashmole, 1971; Brooke, 1981a). Of the 36 regular non-breeding seabirds, 22 are procellariiforms, and 18 of these breed in the subantarctic and Antarctic regions (Table III). There are 17 charadriiform visitors which, with two exceptions, breed in the Arctic and boreal regions. Only the Laridae and Phalacrocoracidae have both breeding and non-breed-

230

A.BERRUTI, N.J.ADAMS AND S.JACKSON

TABLE III The taxonomic grouping of breeding seabirds and taxonomic grouping and breeding grounds of seabirds which are regular, non-breeding visitors to the Benguela ecosystem Breeding Breeding species

Non-breeding species

All species

Taxon

Total

Arctic and boreal

N. and S. Subtropics and Tropics

Antarctic and Subantarctic

Total number Resident % Non-resident %

Podicipedida e Pelecanidae Spheniscidae Diomedeidae Procellaridae Hydrobatida e Sulidae Phalacrocora cidae Charadriidae Laridae Stercoraridae Total

0

0

1

0

1

0

100

1 1 0 0 0

0 0 0 1 2

0 0 0 1 0

0 0 5 11 2

1 1 5 13 4

100 100 0 0 0

0 0 100 100 100

1 4

0 0

0 1

0 0

1 5

100 80

0 20

0 5 0 12

1 5 3 12

0 1 0 4

0 1 1 20

1 12 4 48

0 41 0 25

100 59 100 75

ing representatives. The non-breeding seabirds are dominated by the Procellaridae (13 species) and the Laridae (12 species). The zoogeography of southern African seabirds has been discussed by various workers (Liversidge, 1959; Winterbottom, 1972, 1974; Brooke, 1981a). Liversidge (1959) and Winterbottom (1972, 1974) divided the marine area of southern Africa into eastern and western regions based on the distribution of breeding seabirds and in accordance with major differences in the sea surface temperatures of the warm Agulhas and cold Benguela systems. Winterbottom (1974) regarded Cape Agulhas as a convenient dividing point. Neither author commented on the high degree of endemism in southern Africa, partly because taxonomic changes have resulted in the recognition of subsequent endemic taxa. Both authors regarded the oceanic birds (defined by Winterbottom, 1974, as farther than 5 km offshore) as being part of the Southern Oceanic avifauna, although their analyses were based on inadequate lists of non-breeding seabirds. Brooke (19811a) considered both non-breeding and breeding seabirds in dividing the southern African marine area into three zoogeographical divisions: west (Cape Cross to Cape Agulhas), south (a transitional zone from Cape Agulhas to Kei mouth), and east (Kei mouth to Maputo) and suggested that the Moçamedes province of Angola be included as part of the Benguela seabird assemblage. The seabird assemblages show changes from north to south in the Benguela System. Six seabirds reach their northernmost breeding limits at Walvis Bay or between Walvis Bay and Lüderitz. Most of the marine populations in the Benguela ecosystem of the blacknecked grebe Podiceps nigricollis, white pelican, Damara and black Chlidonias niger terns and pomarine Stercorarius pomarinus and longtailed S. longicaudus skuas occur in the northern Benguela (Jensen & Berry, 1972; Clinning, 1978a; Lambert, 1980; Ryan, 1980; Ryan & Rose, 1989). The number of procellariform species and their abundance decreases from Walvis Bay northwards while several

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

231

southern species e.g. Antarctic fulmar Fulmarus glacialoides, greyheaded albatross Diomedea chrysostoma, and Antarctic tern Sterna vittata are very rare north of the Orange River (Cooper, 1976; Ryan & Rose, 1989). Even if the short-term and seasonal variations in abundance of seabirds are taken into account, strict adherence to geographical boundaries in the context of highly mobile seabirds and their habitats, may be biologically meaningless because such boundaries are unlikely to define accurately seabird habitats or even zoogeographic regions. SPECIES RICHNESS The Benguela ecosystem has fewer breeding species than three other eastern-boundary upwelling systems, the California, Canary, and Humboldt Currents (Table IV). The Canary Current is lacking in seabirds which pursue prey underwater, notably cormorants (Le Grand, Emmerson & Martin, 1984). The California Current, with most breeding species, supports eight alcids (Jehl, 1984), which do not breed in any of the other systems. The seabird assemblages of both the Humboldt and Benguela systems have an abundant sulid and a flock-feeding cormorant. Both have a pelican and a penguin. The brown pelican Pelecanus thagus and white pelican are, however, not ecological equivalents. The brown pelican is an abundant plunge-diving marine species, TABLE IV The number of seabird species in each family and order which breed in the Benguela, California, Canary, and Humboldt systems, based on this paper; Duffy, Hays & Plenge (1984b); Jehl (1984); Le Grand, Emmerson & Martin (1984) and Schlatter (1984) Taxon

Current systems

Order Family

Benguela

California

Canary

Humboldt

1

0

0

2

0 0 0

2 5 0

5 3 0

0 4 1

0 1 1 4 0

0 1 0 3 0

1 0 1 0 1

0 1 2 3 0

5 0 0 12

7 1 8 27

4 0 0 15

5 0 0 18

Sphenisciformes Spheniscidae Procellariiformes Procellaridae Hydrobatidae Pelecanoididae Pelecaniformes Phaethontidae Pelecanidae Sulidae Phalacrocoracidae Fregatidae Charadriiformes Laridae Rhynchopidae Alcidae Total

232

A.BERRUTI, N.J.ADAMS AND S.JACKSON

whereas the white pelican is essentially a freshwater species which has colonised calm marine waters where it feeds from the surface. Despite the low number of breeding seabirds, the assemblage of seabird species in the Benguela system is rich. The 48 species, which occur regularly, comprise 15% and 17% of the total number of seabird species listed by Harrison (1983) and Croxall, Evans & Schreiber (1984), respectively. The abundance and species richness of seabirds is dependent on the scale of time and distance over which these parameters are measured (Schneider & Duffy, 1985), particularly in the case of non-breeding seabirds which are not constrained by the distribution of breeding sites. In the southern Benguela, the species richness of seabirds offshore of the 200 m depth contour increases because of the presence of many nonbreeding species (Duffy, Siegfried & Jackson, 1987b), The fishing activities of bottom-trawlers in waters deeper than 200 m affect the distribution of some seabird species (Jackson, 1988; Ryan & Moloney, 1988) and may locally enhance seabird abundance and species richness (Abrams, 1983, 1985a). Fourteen species of Northern Hemisphere seabirds occur in the Benguela ecosystem in the summer with the Arctic tern Sterna paradisaea common as a passage migrant only (Ryan & Rose, 1989). With the exception of the winter-breeding greatwinged petrel Pterodroma macroptera, another 19 species of non-breeding seabirds from the Southern Hemisphere occur in the Benguela ecosystem mainly in winter (Cooper & Dowle, 1976; Ryan & Rose, 1989), resulting in an increase in species richness (Summerhayes, Hofmeyr & Rioux, 1974; Abrams & Griffiths, 1981; Abrams, 1983). The great shearwater Puffinus gravis and blackbellied storm petrel Fregetta tropica are common on passage only, being abundant in September-October and April-May (Ryan & Rose, 1989). Immature and non-breeding individuals of many of the abundant procellariiform species are present in summer in small numbers when the bulk of the populations are present on the breeding grounds (Ryan & Rose, 1989). Changes in the structure of seabird assemblages associated with evolving and decaying upwelling events have not been described in the Benguela System. Detailed surveys of seabird distribution off California (Briggs & Chu, 1987) have demonstrated high densities of planktivores around the edges of upwelling plumes with piscivorous seabirds concentrated downstream of major upwellings in less turbulent waters. The persistence and seasonal predictability of upwelling in the Benguela System may play an important role in the timing and success of seabird breeding (Duffy, Berruti, Randall & Cooper, 1984a). FEEDING ECOLOGY FORAGING GUILDS Foraging guilds provide a useful framework to summarise the known feeding ecology of seabirds (Tables V and VI). The breeding seabirds are classified by prey type, feeding zone, and foraging technique (Table V). The least well known breeding seabirds are the Damara tern, white pelican, Caspian tern, and whitebreasted cormorant. Because the diets of the non-breeding seabirds are so poorly known and because many species now rely on trawler offal as a primary food source, evolutionarily selected foraging differences may be difficult to detect, and the classification of non-breeding species into guilds based on diet and foraging technique (Abrams & Griffiths, 1981; Abrams, 1983, 1985a) is less useful. Duffy et al. (1987b) categorised seabirds into three broad guilds (inshore and benthic, epipelagic, trawler offal) according to major prey types in order to estimate seabird consumption of different food types in the Benguela ecosystem. We have categorised non-breeding seabirds by foraging zone, feeding techniques, geographical area, and seasonal occurrence (Table VI).

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

233

DIET Breeding seabirds The diet of most breeding Benguela seabirds, with the exception of the Caspian tern and white pelican, has been investigated in varying detail (Tables VII–XIV, see Figs 2 and 3). The diet compositions are presented as percentage numerical abundance, mass, and frequency of occurrence as defined by Hyslop (1980). The only reported diet items of the marine population of white pelicans TABLE V Synopsis of feeding ecology of resident seabird species in the Beuguela ecosystem. Brackets indicate secondary (Cape gannet) or unknown importance (Kelp and Hartlaub’s gulls) of particular foraging technique. Foraging techniques are defined according to Harper, Croxall & Cooper (1985). Further references are cited in feeding ecology section (pages 285–303). Scientific names are given in Table I Predominant food type

Where prey caught

Species

Predominant prey

Main feeding techniques

Feeding zone

Major references

Invertebrates

surface

Epipelagic fish

surface

Damara tern

small surface fish, blennies

surface seize, dipping, piracy Siegfried (1977); Shaugnessy (1980); Brooke & Cooper (1979b) surface/ shallow plunging

Walter (1984); Ryan (1987a)

bivalves, crustaceans

small invertebrates nearshore and shoreline

nearshore and shoreline

(kelp gull)

Hartlaub’s gull surface seize, piracy, dipping

sheltered bays mainly N.Benguela

Clinning (1978a)

Caspian tern

fish

sheltered bays

swift tern

anchovy, pelagic goby

(Hartlaub’s gull) white pelican

pelagic goby

surface/ shallow plunging surface/ shallow plunging, dipping surface seize, dipping surface seizing

neritic waters

Rand (1959); Crawford & Shelton (1981); Davies (1955, 1956, 1958); Matthews (1961); Matthews &

fish

near surface

Cape gannet

inner neritic waters

nearshore in N.Benguela sheltered bays in N.Benguela pilchard, anchovy, saury

Walter (1984); Walter, Cooper & Suter (1987) Walter (1984) Berry & Berry (1975) deep plunging

234

A.BERRUTI, N.J.ADAMS AND S.JACKSON

Berruti (1983); Cooper (1984); Batchelor & Ross (1984); Crawford et al. (1985); Berruti (1987) sub-surface

Cape cormorant

pilchard, anchovy, pelagic goby, Cape horse mackerel

pursuit diving

inner neritic waters

jackass penguin

anchovy, pilchard

pursuit diving

inner neritic waters

Sub-surface fish & Crustaceans benthic

sub-surface/ benthic

whitebreasted cormorant

fish (Sparidae)

Rand (1960a); Davies (1955, 1956, 1958); Matthews (1961); Crawford & Shelton (1981); Cooper (1984); Crawford et al. (1985); Wilson (1985a); Randall & Randall (1986) pursuit diving shallow nearshore

bank cormorant

clinids, blennies, crustaceans pursuit diving

pursuit diving

crowned cormorant

clinids

nearshore shallows, rocky substrate

nearshore coastal kelp beds Rand (1960b); Williams & Cooper (1983)

Davies (1955, 1956, 1958); Matthews (1961); Rand (1960a); Crawford & Shelton (1981); Matthews & Berruti (1983); Cooper (1984, 1985a); Crawford et al. (1985); Duffy et al. (1987c)

Rand (1960b); Cooper (1985a)

Rand (1960b)

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

bank cormorant

pelagic goby

pursuit diving

Offal

surface

(kelp gull)

(Cape gannet)

hake

deep plunging

open waters in N. Benguela hake

trawling grounds

235

Crawford et al. (1985) surface seize, piracy, dipping

nearshore, shoreline, trawling grounds

Brooke & Cooper (1979b)

Rand (1959); Crawford & Shelton (1981); Davies (1955, 1956, 1958); Matthews (1961); Matthews & Berruti (1983); Cooper (1984); Batchelor & Ross (1984); Berruti (1987); Crawford et al. (1985)

TABLE VI Synopsis of ecological features relating to dietary differences within feeding guilds of regularly occurring non-resident seabird species occurring in southern African waters. Foraging techniques are defined according to Harper, Croxall & Cooper (1985). Unless otherwise indicated, the major reference for all species is Ryan & Rose 1989 Species

Where prey caught

Feeding technique

Feeding zone

Seasonality

Major reference

blacknecked grebe reed cormorant

benthos/surface

pursuit diving pursuit diving

more numerous in winter present all year

Ryan (1980); Robertson (1981)

sub-surface

sooty shearwater

sub-surface

pursuit diving, surface diving, surface seize pursuit diving, surface diving, surface seize pursuit diving, surface diving, surface seize pursuit diving, surface diving, surface seize

coastal bays in N.Benguela nearshore, extreme N.Benguela mainly inner neritic waters

more numerous in winter

Jackson (1988)

mainly inner neritic waters

September– March

shelf-break, oceanic waters

mainly winter

shelf-break, oceanic waters

SeptemberOctober, AprilMay

Manx shearwater

little shearwater

great shearwater

236

A.BERRUTI, N.J.ADAMS AND S.JACKSON

Species

Where prey caught

Feeding technique

Feeding zone

Seasonality

European stormpetrel Leach’s stormpetrel Wilson’s stormpetrel blackbellied storm-petrel grey phalarope

surface

pattering

shelf-break, trawling grounds shelf-break, oceanic waters shelf-break, trawling grounds oceanic waters

November– March October–January

broadbilled prion

Sabine’s gull

pattering pattering pattering surface seize

hydroplaning, surface seizing, pattering

neritic and oceanic waters, shelf-break mainly neritic waters

more abundant in winter September– November, May summer

May–September

neritic waters

September–May

sandwich tern

pattering, surface plunging surface plunging

coastal waters

common tern

dipping

coastal waters

Arctic tern

little tern

dipping, surface plunging dipping, surface plunging dipping

black tern

dipping

pintado petrel

surface seize

softplumaged petrel

surface seize

greatwinged petrel

surface seize

Cory’s shearwater

surface seize, surface plunging surface seize, surface diving

shelf-break, oceanic waters southern inner neritic waters nearshore coastal waters coastal N. Benguela shelf-break, trawling grounds shelf-break, oceanic water, mainly S. Benguela shelf-break, oceanic waters neritic waters

more abundant in summer more abundant in summer mainly September– November April-November

Antarctic tern

whitechinned petrel

neritic, oceanic waters, trawling grounds

Major reference

Furness (1983)

summer summer May-October May-October

mainly November– March October–April more numerous in winter

Jackson (1988)

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

Antarctic fulmar

surface seize

blackbrowed albatross

surface seize

greyheaded albatross

surface seize

yellownosed albatross shy albatross

surface seize

wandering albatross

surface seize

southern giant petrel

surface seize

northern giant petrel

surface seize

Subantarctic skua

piracy, dipping, surface seize

surface seize

longtailed skua

aerial/surface

piracy, dipping

Arctic skua pomarine skua

aerial

piracy, dipping piracy, dipping

TABLE VII

trawling grounds in S.Benguela shelf-break, oceanic waters, southern trawling grounds shelf-break, trawling grounds in S.Benguela shelf-break, oceanic waters neritic waters, shelf-break, fronts, trawler grounds shelf-break, oceanic waters, mainly S. Benguela neritic and oceanic waters, trawling grounds neritic and oceanic waters, off seal colonies in N. Benguela mainly shelfbreak, trawling grounds, neritic waters neritic waters and trawling grounds in N. Benguela, oceanic in S. Benguela neritic waters mainly neritic waters in N. Benguela

237

June–October particularly common in winter

June–September

more abundant in winter more numerous in winter more numerous in winter

more numerous in winter more numerous in winter

more abundant in winter

Sinclair (1980)

September–May

Lambert (1980)

mainly summer mainly summer

238

A.BERRUTI, N.J.ADAMS AND S.JACKSON

Diet of Hartlaub’s gull Larus hartlaubii, swift tern Sterna bergii, and Damara tern S.balaenarum in the Benguela ecosystem. %N=percentage of diet by numerical abundance, %F=percent frequency of occurrence in diet Species

Hartlaub’s gull

swift tern

swift tern

Damara tern

Time period

Feb 1982

Feb 1982

1977–1986

1975–1977

Area

Possession Is., N.Benguela Possession Is., N.Benguela Saldanha Bay Central region, S.Benguela N.Benguela

Sampling method

otoliths, pellets

otoliths, pellets

regurgitated, stomach contents

otoliths, pellets

Source

Walter (1984)

Walter (1984)

Walter, Cooper & Suter (1987)

Clinning (1978a)

No. of samples

8

35

?

?

No. of items

85

72

1311

22

Diet measured

%N

%F

%N

%F

%N

%N

Sufflogobius bibarbatus Lampanyctodes hectoris Merluccius spp. Engraulis japonicus capensis Hepsetia breviceps Etrumeus whiteheadii Liza richardsonii Trachurus capensis Sardinops ocellatus Unidentified blenny Other fish Pterygosquilla armata Cephalopoda Other invertebrates

39

63

97

91

1

0

61

38

0

0

0

0

0 0

0 0

3 0

10 0

4 51

0 9

0 0

0 0

0 0

0 0

6 6

0 0

0 0

0 0

0 0

0 0

5 3

14 0

0

0

0

0

3

0

0

0

0

0

0

68

0 0

0 0

0 0

0 0

9 8

0 0

0 0

0 0

0 0

0 0

5 1

9 0

Sample size

are the chicks of Cape cormorants, eaten at the guano platform where both species nest (Berry, 1976). Elsewhere, the white pelican eats fish weighing up to 500 g (Cramp & Simmons, 1977) and it is probable that this species feeds on marine fish in central Namibia. The gulls are generalist feeders, and much of their terrestrial food is scavenged from man. The marine diet of the Hartlaub’s gull includes amphipods, isopods and fish (Table VIII; Furness, 1983; Walter, 1984;

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

239

Ryan, 1987a). At Possession Island in the northern Benguela, an analysis of otoliths in regurgitated pellets showed the diet of this gull to consist entirely of pelagic goby Sufflogobius bibarbatus and lanternfish Lampanyctodes hectoris (Walter, 1984). In addition, Hartlaub’s gulls kleptoparasitise other birds (Morant, 1987; Ryan, 1987a). Before the arrival of European man, the Hartlaub’s gull may have relied mainly on swarming crustaceans caught in the nearshore zone and invertebrates caught along the shoreline (Ryan, 1987a). Important marine food items of kelp gulls are bivalves and offal from bottom trawlers and other human sources (Siegfried, 1977; Sinclair, 1978; Brooke & Cooper, 1979b; Ryan & Moloney, 1988). Predation on fish has been recorded (Shelton, De Villiers & Crawford, 1978). Kelp gulls are predators of seabirds, their eggs and young (Cooper, 1974, 1977a) and they kleptoparasitise other birds (Hockey, 1980; Avery, 1983; Furness, 1983). Natural scavenging is frequently recorded (Shaughnessy, 1980). Before European colonisation of southern Africa, kelp gulls may have fed along the shoreline, eating bivalves and scavenging dead animals, as well as catching surface-living fishes and crustaceans at sea (Brooke & Cooper, 1979b). The diet of the Damara tern is not well known, and comprises mainly unidentified juvenile blennies. It also includes the southern mullet Liza richardsonii, anchovy Engraulis japonicus capensis and squid in its diet (Table VII; Clinning, 1978a). The swift tern eats mainly pelagic goby in the northern Benguela and anchovy in the southern Benguela (Walter, 1984; Walter, Cooper & Suter, 1987). The diet of the Caspian tern in the Benguela ecosystem has not been described, but elsewhere it feeds on fish up to 250 mm in length (Cramp & Simmons, 1985). The diets of marine populations of the whitebreasted cormorant are poorly known (Table VIII). Rand (1960b) showed that sparids were important in their diet in the southern Benguela. Cape horse mackerel Trachurus capensis, musselcracker Sparodon durbanensis, crustaceans, and molluscs were also recorded as prey in the Benguela ecosystem. De Kock & Randall (1984) noted freshwater, estuarine, and marine fishes in the diet of whitebreasted cormorants in Algoa Bay. In two Natal estuaries, Mugilidae were the most important prey (Whitefield, 1977; Jackson, 1984). Whitebreasted cormorants were reported to feed on the southern mullet in St Helena Bay (Davies, 1956). In a southern Cape estuary, Whitefield (1986) found the Cape silverside Hepsetia breviceps, a small species, to be the most important prey. In the southern Benguela, bank and crowned cormorants feed mainly on clinids and other slow-moving benthic prey (Table VIII; Rand, 1960b; Williams & Burger, 1978; Williams & Cooper, 1983; Cooper, 1985a). At Dassen Island, the diet of the bank cormorant also includes flat fish (Pleuronectiformes) and platanna-klipfishes (Xenopoclininae) (Cooper, 1985a). At Mercury and Ichaboe islands, the diet consists almost entirely of pelagic gobies (Cooper, 1984; Crawford, Cruickshank, Shelton & Kruger, 1985). A TABLE VIII

240

A.BERRUTI, N.J.ADAMS AND S.JACKSON

Diet of the whitebreasted cormorant Phalacrocorax carbo, bank cormorant P.neglectus, and crowned cormorant P.coronatus in the Benguela ecosystem. N and F as in Table VII Species

whitebreasted cormorant

bank cormorant

bank crowned cormorant cormoran t

crowned crowned cormorant cormorant

Time period

1954–1956

1954–1956

1979– 1980

May 1980 Sept-Oct 1980

Area

S.W.Cape

S.W.Cape

Ichaboe S.W.Cape Is., N.Bengue la

Marcus Marcus Is., Is., S.W.Cape S.W.Cape

Sampling method

shot, stomach contents

shot, stomach contents

regurgitat shot, stomach contents ed stomach contents

regurgitat ed stomach contents

Source

Rand (1960b)

Rand (1960b)

Crawford Rand (1960b) et al. (1985)

Williams Williams & Cooper & Cooper (1983) (1983)

No. of samples

6

68

41

10

7

?

No. of items

9a

470

?

88

143

13

Diet measure

%N

%F

%N

%F

%N

%N

%F

%N

%N

Sparidae Clinidae Blennidae Sufflogob ius bibarbatu s Syngnath us Ammodyt es Triglidae Heteromy cteris Chorisoc hismus Merlucci us spp. Other fish

89 0 0 0

100 0 0 0

0 23 10 0

0 40 16 0

0 0 0 95

0 10 0 0

0 60 0 0

0 75 0 0

0 46 46 0

0

0

0

2

0

15

20

22

0

0

0

8

7

0

0

0

0

0

0 0

0 0

3 2

6 4

0 0

0 28

0 10

0 0

0 0

0

0

3

4

0

1

10

0

0

0

0

0

0

5

0

0

0

0

11

17

7

–

0

4

10

0

0

1952–1956

regurgitat ed stomach contents

Sample size

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

241

Species

whitebreasted cormorant

bank cormorant

bank crowned cormorant cormoran t

crowned crowned cormorant cormorant

Time period

1954–1956

1954–1956

1979– 1980

May 1980 Sept-Oct 1980

Area

S.W.Cape

S.W.Cape

Ichaboe S.W.Cape Is., N.Bengue la

Marcus Marcus Is., Is., S.W.Cape S.W.Cape

Sampling method

shot, stomach contents

shot, stomach contents

regurgitat shot, stomach ed contents stomach contents

regurgitat ed stomach contents

Source

Rand (1960b)

Rand (1960b)

Crawford Rand (1960b) et al. (1985)

Williams Williams & Cooper & Cooper (1983) (1983)

No. of samples

6

68

41

10

7

?

No. of items

9a

470

?

88

143

13

Diet measure

%N

%F

%N

%F

%N

%N

%F

%N

%N

Crustacea ns Cephalop oda Other invertebra tes

0

0

22

53

1

43

30

1

0

0

0

10

41

0

0

0

1

0

0

0

12

–

1

0

0

1

8

1952–1956

regurgitat ed stomach contents

Sample size

a

Cormorants and molluscs not included.

small but important part of the diet of bank cormorant is the rock lobster Jasus lalandii (Rand, 1960b; Avery, 1983; Cooper, 1985a). The diets of Cape gannets, jackass penguins, and Cape cormorants are detailed in Tables IX–XIV. Differential rates of digestion of different food types may result in a bias towards the more durable prey items (Duffy & Laurenson, 1983; Furness, Laugksch & Duffy, 1984; Wilson, La Cock, Wilson & Mollagee, 1985; Jackson & Ryan, 1986). Diet studies are based on stomach contents obtained by shooting (e.g. Davies, 1955; Rand, 1959; Matthews, 1961), regurgitation (Cooper, 1984; Berruti, 1987) or stomach pumping (Wilson, 1985b). The compositions of the diets of the Cape gannet, jackass penguin and Cape cormorant in the 1950s (Davies, 1955, 1956; Rand, 1959, 1960a,b; Matthews, 1961) were recalculated from original data presented in these studies (Tables IX–XIV). Empty stomachs were excluded from the calculation of percentage frequency of occurrence. Percentage mass was recalculated using stomach contents at the time of collection and not reconstituted mass. Items which were noted as present but were not

242

A.BERRUTI, N.J.ADAMS AND S.JACKSON

enumerated, were not included in the recalculation of percent numerical abundance. Diet data presented by Davies (1955, 1956) were confusing. There are unaccountably large differences between the recalculated diet compositions and those presented in figures in Davies (1955, 1956). Although the percentage numerical abundance over-emphasises the contribution of the relatively small and numerous anchovy compared to the larger pilchard (Duffy & Jackson, 1986), it has been the most frequently used measure of diet composition in the Benguela ecosystem. All measures of diet composition, however, demonstrate the large differences in diet between northern and southern Benguela and within the same region over the 30-year period between the 1950s and 1970s and 1980s (see Figs 2 and 3). In Figures 2 and 3, the numerical abundances of prey in the diet are averaged over years at one locality or over several localities in a region. The diets of the Cape gannet, jackass penguin and Cape cormorant tend to be dominated by the same prey species (Tables IX–XIV; Davies, 1955, 1956, 1958; Rand, 1959, 1960a,b; Matthews, 1961; Berry, 1976; Crawford & Shelton, 1981; Matthews & Berruti, 1983; Cooper, 1984; Crawford et al., 1985; Wilson, 1985b; Duffy, Wilson & Berruti, 1985; Berruti, 1987; Duffy, Wilson & Wilson, 1987c). These prey species are pilchard, Cape horse mackerel and anchovy (eaten by all three seabird species), saury Scomberesox saurus (eaten by Cape gannet) and pelagic goby (eaten by jackass penguin and Cape cormorant). Dominant prey, with the exception of trawler offal taken by Cape gannets and cephalopods eaten by jackass penguins (Randall, Randall & Klingelhoeffer, 1981b; Crawford et al., 1985), are epipelagic shoaling fish (Crawford, Shannon & Pollock, 1987; Berruti, 1988). Changes in the species composition of the diets of the Cape gannet, jackass penguin and Cape cormorant from the 1950s to the early 1970s, the late 1970s and early 1980s are consistent with changes in fishery catches of each fish species (Crawford & Shelton, 1981; Cooper, 1984; Randall & Randall, 1986; Berruti, 1987). There are distinct regional similarities in the diets of all three species (Tables IX–XIV), so that dietary changes are shown in relation to the composition of the purse-seine catch off Namibia and the western Cape respectively (Figs 2 and 3). In the 1950s, pilchard dominated the diets of the Cape gannet, jackass TABLE IX

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

243

Diet of the Cape gannet Morus capensis in the northern Benguela. All localities except Walvis Bay (central Namibia) are in southern Namibia. N and F as in Table VII. %M=percentage of diet by mass Area

Walvis Bay

Walvis Bay

Mercury, Mercury Is. Ichaboe Is. Possession Ichaboe Is. and Possession Is.

Time period 1957–1958

1958–1959

1978–1979 Nov 1978– Nov 1978– Dec 1978– Feb 1979 Feb 1979 Feb 1982

Sampling method

shot, stomach contents shot, stomach contents

Source

Matthews (1961)

Sample size 155

regurgitate regurgitate regurgitate regurgitate d, stomach d, stomach d, stomach d, stomach contents contents contents contents

Matthews & Berruti (1983) Crawforda & Shelton (1981)

Crawforda et al. (1985)

Crawforda et al. (1985)

Crawforda et al. (1985)

240

256

116

967

345

Diet measure

%M

%N

%F

%M

%N

%F

%N

%N

%N

%N

Sardinops ocellatus Engraulis japonicus capensis Trachurus capensis Sufflogobi us bibarbatus Merluccius spp. Scomberes ox saurus Other

93

85

81

93

99

76

1

1

1

1

0

0

0

0

0

0

85

69

36

77

3

10

8

1

1

2

0

4

1

1

0

0

0

0

0

0

5

10

28

1

0

0

0

0

0

0

0

5

9

3

0

0

0

0

0

0

6

4

19

7

4

5

22

5

–

21

2

7

6

11

a

Crawford & Shelton (1981) is a subset of Crawford et al. (1985).

244

A.BERRUTI, N.J.ADAMS AND S.JACKSON

TABLE X Diet of the jackass penguin Spheniscus demersus in the northern Benguela. All localities except Walvis Bay (central Namibia) are in southern Namibia. N and F as in Table VII, M as in Table IX Area

Walvis Bay

Mercury and Ichaboe Is.

Halifax and Possession Is.

Mercur Ichabo y Is. e Is.

Halifax Possess Mercur Is. ion Is. y, Ichabo e, Halifax and Possess ion Is.

Time period

1957–1958

1980

1980

Feb 1980

Feb 1980

Jan 1980

Jan 1980

1980

Samplin shot, stomach contents stomach g pumping method

stomach pumping

stomac h pumpin g

regurgi tated, stomac h content s

stomac h pumpin g

stomac h pumpin g

stomac h pumpin g

Source

Matthews (1961)

Crawford & Shelton (1981)

Crawforda & Shelton (1981)

Crawfo rda et al. (1985)

Crawfo rda et al. (1985)

Crawfo rda et al. (1985)

Crawfo rda et al. (1985)

Crawfo rda et al. (1985)

Sample size

19

121

83

50

43

21

45

114

Diet %M measure Sardin ops ocellat us Engrau lis japonic us capensi s Trachu rus capensi s Sufflog obius bibarb atus Merluc cius spp.

%N

%F

%M

%N

%M

%N

%N

%N

%N

%N

%N

94

79

68

0

0

0

0

0

0

0

0

0

0

0

0

1

1

7

21

0

1

0

1

1

1

2

4

0

0

0

0

0

0

0

0

0

0

0

0

71

92

53

17

73

37

0

3

56

0

0

0

24

6

0

0

1

19

0

0

4

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

245

Area

Walvis Bay

Mercury and Ichaboe Is.

Halifax and Possession Is.

Mercur Ichabo y Is. e Is.

Halifax Possess Mercur Is. ion Is. y, Ichabo e, Halifax and Possess ion Is.

Time period

1957–1958

1980

1980

Feb 1980

Feb 1980

Jan 1980

Jan 1980

1980

Samplin shot, stomach contents stomach g pumping method

stomach pumping

stomac h pumpin g

regurgi tated, stomac h content s

stomac h pumpin g

stomac h pumpin g

stomac h pumpin g

Source

Matthews (1961)

Crawford & Shelton (1981)

Crawforda & Shelton (1981)

Crawfo rda et al. (1985)

Crawfo rda et al. (1985)

Crawfo rda et al. (1985)

Crawfo rda et al. (1985)

Crawfo rda et al. (1985)

Sample size

19

121

83

50

43

21

45

114

Diet %M measure

%N

%F

%M

%N

%M

%N

%N

%N

%N

%N

%N

Cephal opoda Other

1

19

12

5

3

30

11

26

44

99

93

40

5

–

–

0

0

10

50

0

1

1

3

1

a

Crawford & Shelton (1981) overlaps with Crawford et al. (1985).

TABLE XI

246

A.BERRUTI, N.J.ADAMS AND S.JACKSON

Diet of the Cape cormorant Phalacrocorax capensis in the Northern Benguela. All localities except Walvis Bay and Swakopmund (both in central Namibia) are in southern Namibia. N and F as in Table VII, M as in Table IX Area

Walvis Bay

Walvis Bay

Swakopmun Merc Merc Ichab Ichaboe Is. d ury, ury oe Is. Posse Is. ssion and Ichab oe Is.

Ichab Merc oe Is. ury Is.

Time 1957–1958 period

1958–1959

1973–1974

1978 – 1979

Nov 1978 – Feb 1980

Nov Nov 1978 1978 – Feb 1980

1982 Feb – 1982 1983

Sampl shot, stomach ing contents metho d

shot, stomach contents

shot, stomach contents

regur gitate d stom ach conte nts

regur gitate d stom ach conte nts

regur regurgitated gitate stomach contents d stom ach conte nts

otolit hs, pellet s

otolit hs, pellet s

Sourc Matthews (1961) e

Matthews & Berruti Berry (1976) Craw Craw (1983) forda forda & et al. Shelt (1985 on ) (1981 )

Craw Coopera forda (1981 5b) et al. (198 5)

Duff ya et al. (198 7c)

Duff ya et al. (1987 c)

Sampl 210 e size

250

101

10

93

176

71

225

19

%M Diet measu re

%N

%F

%M

%N

%F

%M

%N

%M

%N

%N

%N

%F

%Nb

%Nb

Sardi nops ocell atus Engr aulis japon icus cape nsis Trac hurus cape nsis Suffl ogobi us

87

67

87

90

76

71

84

80

0

0

1

0

0

0

0

1

4

1

1

2

1

15

12

28

0

23

1

5

0

0

9

16

21

7

14

13

1

1

1

0

1

0

0

0

0

0

0

0

0

0

0

1

2

65

96

72

100

95

94

100

THE BENGUELA ECOSYSTEM PART VI SEABIRDS

Area

247

Walvis Bay

Walvis Bay

Swakopmun Merc Merc Ichab Ichaboe Is. d ury, ury oe Is. Posse Is. ssion and Ichab oe Is.

Ichab Merc oe Is. ury Is.

Time 1957–1958 period

1958–1959

1973–1974

1978 – 1979

Nov 1978 – Feb 1980

Nov Nov 1978 1978 – Feb 1980

1982 Feb – 1982 1983

Sampl shot, stomach ing contents metho d

shot, stomach contents

shot, stomach contents

regur gitate d stom ach conte nts

regur gitate d stom ach conte nts

regur regurgitated gitate stomach contents d stom ach conte nts

otolit hs, pellet s

otolit hs, pellet s

Sourc Matthews (1961) e

Matthews & Berruti Berry (1976) Craw Craw (1983) forda forda & et al. Shelt (1985 on ) (1981 )

Craw Coopera forda (1981 5b) et al. (198 5)

Duff ya et al. (198 7c)

Duff ya et al. (1987 c)

Sampl 210 e size

250

101

10

%M Diet measu re bibar batus Merl 0 ucciu s spp. Other 4 a b

93

71

225

19

%N

%F

%M

%N

%F

%M

%N

%M

%N

%N

%N

%F

%Nb

%Nb