Tropical Forest Genetics (Tropical Forestry)

Tropical Forestry Tropical Forestry Volumes Already Published in this Series Tropical Forest Genetics by Finkeldey, R...

Author: Reiner Finkeldey | Hans Heinrich Hattemer

246 downloads 2343 Views 4MB Size Report

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

DOWNLOAD PDF

Tropical Forestry

Tropical Forestry Volumes Already Published in this Series Tropical Forest Genetics by Finkeldey, R., Hattemer, H. 2007 ISBN: 3-540-37396-9 Sampling Methods, Remote Sensing and GIS Multiresource Forest Inventory by Köhl, M. Magnussen, S., Marchetti M. 2006 ISBN: 3-540-32571-9 Tropical Forest Ecology - The Basis for Conservation and Management by Montagnini, F., Jordan C. 2005 ISBN: 3-540-23797-6

Reiner Finkeldey Hans H. Hattemer ●

Tropical Forest Genetics With 44 Figures and 19 Tables

Professor Dr. Reiner Finkeldey Professor Dr. Hans H. Hattemer Institute of Forest Genetics and Forest Tree Breeding Büsgenweg 2 D-37077 Göttingen Germany

ISSN: 1614-9785 ISBN-10 3-540-37396-9 Springer-Verlag Berlin Heidelberg New York ISBN-13 978-3-540-37396-4 Springer-Verlag Berlin Heidelberg New York Library of Congress Control Number: 2006931487 This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. Springer-Verlag is a part of Springer Science+Business Media springer.com © Springer-Verlag Berlin Heidelberg 2007 The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Editor: Dr. Dieter Czeschlik, Heidelberg Desk Editor: Anette Lindqvist, Heidelberg Production: SPi Typesetting: SPi Cover Design: Design & Production, Heidelberg Printed on acid-free paper

3/3152-HM

5 4 3 2 1 0

Preface

This book is intended to provide information on fundamental genetic processes in tropical forests. It is based on lecture notes by the authors prepared for postgraduate students, mainly from tropical countries, of the M.Sc. course “Tropical and International Forestry” at the Faculty of Forest Sciences and Forest Ecology, Georg August University Göttingen, Germany. The intended readership is students, researchers and practitioners interested in the genetic variation of species, in particular forest trees, living in complex forest ecosystems in the tropics. Particular emphasis is placed on the human impact on forest genetic resources in the tropics. Readers should be familiar with basics of classical and molecular genetics such as the structure and function of DNA (double-helix structure, replication), polypeptide synthesis (transcription and translation), and the transmission of genetic information during sexual reproduction (“Mendel’s rules”). This knowledge is easily available from recently published genetics textbooks; we recommend the book by Griffiths et al. (2000). The development of biochemical and molecular marker techniques and their application to species of tropical forests have greatly improved our knowledge of genetic variation patterns of “wild” plants in the tropics during the last two decades. Many examples in this book are based on gene marker techniques. We discuss the results of selected molecular studies and their implications for the development of sustainable management strategies for forest genetic resources. However, we refrain from providing details on laboratory techniques and statistical data analyses. Readers should consult the cited references or the book by Weising et al. (2005) if they require additional information in this regard. No advanced knowledge concerning modern molecular or statistical methods for the analysis of genetic variation patterns is necessary to follow the argumentation of this book. Comprehensive data on DNA sequences are becoming available for more and more plants, including forest trees. The first project aimed at sequencing the full genome of a forest tree (Populus trichocarpa) has recently been completed (Brunner et al. 2004). However, only very few genomic data are currently

VI

Preface

available for tropical forest trees; tropical forest genetics is still in a “pregenomic era.” Thus, the focus of this book in on population genetics processes in natural and managed tropical forests rather than on genomics and related, newly emerging fields of research in forest genetics.

Acknowledgements The authors are grateful to Oliver Gailing for critically reading Sect. 12.5.4 to Barbara Vornam and Martin Ziehe for assistance in literature search and in word processing, and to Wolfgang Tambour for preparing Figures 7.1, 7.2, 7.5, and 12.4. July 2006

Reiner Finkeldey Hans H. Hattemer

Contents

1

Introduction – Genetics of Tropical Forests

Part A Genetic Processes in Tropical Forests 2 2.1 2.2 2.3 2.3.1 2.3.1.1 2.3.1.2 2.3.2 2.3.3 2.3.4 2.4 2.4.1

1 3 5 5 5 6 6 7 12 13 14 15 16 16

2.5 2.6 2.7

Population Genetics – an Overview Introduction The Population Variation at Gene Loci The Molecular Basis of Genetic Variation Molecular Markers Biochemical Markers – Isozymes The Gene As a Unit of Heredity The Mode of Inheritance Definition and Classification of Gene Markers Genetic Structures Within Populations Allelic and Genotypic Structures Example 2.1: Genetic Structures at an Isozyme Gene Locus in Dalbergia sissoo Variation at Uniparentally Inherited Markers Example 2.2: Diversity of cpDNA Haplotypes in D. sissoo Evolution and Evolutionary Factors Phenotypic Variation Recommended Literature

3 3.1 3.2 3.3 3.3.1 3.3.1.1 3.3.1.2 3.3.2 3.3.2.1 3.3.2.2

Genetic Variation of Tropical Forest Plants Introduction Genetic Inventories Measurement of Genetic Variation Genetic Variation Within Populations Genetic Multiplicity Genetic Diversity Genetic Differentiation Among Populations Genetic Distances Genetic Differentiation

23 23 23 24 24 25 25 27 27 27

2.4.2

17 18 18 19 21 22

VIII

Contents

3.4 3.5

28 31

3.6

Genetic Variation in Tropical Forest Species – General Trends Case Studies on Genetic Variation Patterns Example 3.1: Genetic Variation at Isozyme Gene Loci in Natural Populations of Acacia auriculiformis Example 3.2: Genetic Differentiation Among Populations of Eugenia dysenterica from the Brazilian Cerrado Example 3.3: Genetic Variation of Swietenia macrophylla Across the Brazilian Amazon Example 3.4: Genetic Variation of Endangered Australian Species of the Genus Fontainea Assessed with RAPD Markers and by DNA Sequences Example 3.5: Genetic Variation of Shorea leprosula and Shorea parvifolia in Indonesia Assessed at AFLP Loci Example 3.6: Genetic Variation of Cedrela odorata at cpDNA Recommended Literature

4 4.1 4.2 4.2.1 4.2.2 4.2.3 4.3 4.3.1 4.3.2 4.4

Sexual and Asexual Reproduction in Tropical Forests Introduction Sexual Reproduction Sexual Types and Sexual Systems Sexual Systems Sexual Structures and Sexual Function Asexual Reproduction Vegetative Reproduction Apomixis Recommended Literature

41 41 41 42 44 47 49 50 51 52

5 5.1 5.2 5.2.1 5.2.2 5.2.2.1 5.2.2.2 5.2.3 5.2.3.1

Gene Flow and Migration Introduction Gene Flow Through Pollen Pollination and Fertilization Pollen Vectors in Tropical Forests Pollination by Wind (Anemogamy) Pollination by Animals (Zoogamy) Pollen Dispersal Assessed by Marker-Based Studies Dispersal of Rare Alleles Example 5.1: Gene Dispersal in a Clonal Seed Orchard of Teak (Tectona grandis) 5.2.3.2 Paternity Analyses Example 5.2: Gene Flow Through Pollen in Neobalanocarpus heimii and Dipterocarpus tempehes (Dipterocarpaceae) Example 5.3: Pollen Dispersal in Three Neotropical Tree Species with Different Spatial Distribution Patterns on Barro Colorado Island

31 33 33

34 35 37 39

53 53 54 54 54 54 55 60 60 60 61 62

63

Contents

5.2.3.3 Model-Based Estimates of Pollen Dispersal Example 5.4: Pollen Dispersal in Dinizia excelsa (Fabaceae) 5.2.4 Efficiency of Pollen Vectors for Gene Flow 5.3 Migration of Seeds 5.3.1 Seed Vectors 5.3.1.1 Abiotic Seed Dispersal 5.3.1.2 Biotic Seed Dispersal 5.3.2 Efficiency of Seed Dispersal 5.4 Long-Distance Gene Flow and Migration in Tropical Forest Species 5.5 Recommended Literature

64 64 65 65 65 65 66 66 67 68

6 6.1 6.2 6.2.1 6.2.2 6.3 6.3.1

69 69 70 70 72 73 74

6.3.2 6.3.3 6.4 6.4.1 6.4.2 6.5 6.5.1 6.5.2 6.6 6.7 7 7.1 7.2

Mating Systems Introduction Random Mating and Panmixis Heterogeneity of Pollen Allele Frequencies Panmixis and Hardy–Weinberg Structures Selfing and Outcrossing Rates Estimates of Selfing Rates Based on Rare Alleles Example 6.1: Estimates of Selfing Rates in Teak (Tectona grandis) Populations Estimates of Selfing Rates Based on a Mixed Mating Model Estimates of Selfing Rates Based on Nonmaternal Alleles Inbreeding and Inbreeding Depression Genetic Consequences of Inbreeding Inbreeding Depression Example 6.2: Inbreeding Depression in Eucalypts Incompatibility and Self-Sterility Incompatibility Self-Sterility Environmental Effects on Mating Recommended Literature

74 76 76 78 78 80 80 82 82 84 84 85 87 87 87

7.4

Adaptation and Coevolution Introduction Physiological and Evolutionary Adaptation Example 7.1: Viability Selection During Early Life Stages of Platypodium elegans Species Interactions and Coevolution Example 7.2: Figs and Their Pollinators Recommended Literature

8 8.1 8.2

Phylogenies and Evolution Above the Species Level Introduction The Evolution of Species Diversity in the Tropics

99 99 99

7.3

89 90 91 97

IX

X

Contents

8.2.1 8.2.2 8.3

8.4

Species Diversity in Tropical Forests Evolution Above the Species Level Molecular Phylogenies Example 8.1: A Molecular Phylogeny of Indonesian Dipterocarpoideae Example 8.2: Evolution of the Genus Inga Recommended Literature

99 101 102 103 104 106

Part B Applications of Genetics to Tropical Forestry

109

9 9.1 9.2

111 111 113

9.3 9.4 10

Fragmentation of Forests Introduction The Genetic Status of Fragmented Tree Populations Example 9.1: Genetic Impact of Fragmentation on a Rain Forest Canopy Tree Example 9.2: Connectivity of Population Fragments of a Medium-Sized Dry-Forest Tree Genetic Preconditions for Restoration and Persistence Recommended Literature

116 119 123 127

10.5 10.6

Genetic Aspects of the Management of Natural Tropical Forests Introduction Selective Logging Effective Population Density After Logging The Question of Dysgenic Selection Natural Regeneration Genetic Aspects of the Manipulation of Dynamic Processes in Natural Forests Genetic Aspects of Sustainability in Natural Tropical Forests Recommended Literature

137 140 146

11 11.1 11.2 11.3 11.4 11.4.1 11.4.2 11.4.3 11.4.3.1 11.4.3.2 11.4.3.3

Provenance Research Introduction Definitions Historical Development Provenance Tests Types of Field Experiments in Provenance Research Traits Studied in Field Trials Design, Conduct, and Analysis of Provenance Trials Collection of Material for Provenance Trials Production of Planting Stock Experimental Design

147 147 148 150 151 153 155 156 156 159 159

10.1 10.2 10.2.1 10.2.2 10.3 10.4

129 129 130 131 135 136

Contents

11.4.3.4 Choice of Experimental Sites 11.4.3.5 Number and Distribution of Locations of an Experimental Series 11.4.3.6 Observation Period of Field Experiments 11.5 Provenance Differentiation and Geographic Variation Pattern 11.6 Choice of Provenances in Tropical Forestry Example 11.1: Provenances of Eucalyptus camaldulensis and Their Growth Performance Example 11.2: International Provenance Trials in Teak (Tectona grandis) 11.7 Recommended Literature 12 12.1 12.2 12.3 12.3.1 12.3.2 12.3.3 12.4 12.4.1 12.4.2 12.5 12.5.1 12.5.2 12.6 12.6.1 12.6.2 12.6.3 12.6.4 12.7

Domestication and Breeding of Tropical Forest Trees Introduction Domestication Genetic Controlledness of Phenotypic Traits Individuals Related by Descent Pair Comparisons in the Field Response to Natural Viability Selection Linear Model of Genetic Effects on a Phenotypic Trait Phenotypic Trait Expressions Genetic Variance Components and Heritability Estimation of Breeding Parameters; Progeny Testing Progenies of Open-Pollinated Trees Progeny Tests After Controlled Pollination Methods of Selection Selection of Plus Trees Selection in Progeny Tests Selection of Clones Marker-Assisted Selection Heterosis Breeding in Forest Trees Example 12.1: Breeding Eucalypts in Aracruz, Brazil 12.8 Propagation of Breeding Products 12.8.1 Clonal Seed Orchards Example 12.2: Clonal Seed Orchards of Teak (Tectona grandis) 12.8.2 Seedling Seed Orchards 12.8.3 Mass Multiplication of Clones 12.9 Multiple Population Breeding 12.10 Breeding Strategy 12.11 Genetic Consequences of Domestication and Breeding 12.11.1 General Considerations 12.11.2 Monitoring of Breeding Projects by Use of Genetic Markers 12.12 Recommended Literature

161 162 163 163 165 169 170 172 173 173 174 176 176 177 179 179 179 182 183 183 184 186 187 188 189 189 192 194 196 196 198 201 204 205 207 209 209 210 211

XI

XII

Contents

13 13.1 13.2

13.3 13.4 13.5 13.5.1 13.5.2 13.5.3 13.6 13.7 13.7.1 13.7.2 13.8 13.9 13.10 13.11 14 14.1 14.2 14.3 14.4 14.4.1 14.4.2 14.4.3 14.5 14.5.1 14.5.2 14.6 14.6.1 14.6.2 14.6.3 14.6.4 14.6.5 14.7

14.8 14.9

Genetic Aspects of Plantation Forestry in the Tropics Introduction Plantations of Exotic Tree Species Example 13.1: The Origin of Early Plantations of Acacia mangium in Sabah Plantations of Indigenous Species Basic and Reproductive Material Production and Collection of Seed Seed Production Areas Provenance Resource Stands Seed Orchards Collection and Storage of Seed Planting Stock Production Seedlings Clonal Multiplication Establishment and Development of Plantations Natural Regeneration of Plantations Use of Breeding Products Recommended Literature

213 213 215

Conservation of Genetic Resources in Tropical Forest Trees Introduction Development of Tree-Conservation Genetics Defining Priorities Conservation Objectives Objective 1: Preservation of the Potential for Particular Trait Expressions Objective 2: Preservation of Maximum Variation Objective 3: Preservation of Adaptability Selection of Genetic Resources Inventory of Genetic Marker Loci Inventory of Adaptive Trait Expressions and Adaptive Markers Conservation Methods Dynamic Conservation in Situ Dynamic Conservation ex Situ in Man-Made Forests Conservation of Seeds in Gene Banks Vegetative Propagation for the Conservation of Forest Genetic Resources Biotechnological Methods of Preservation Regeneration of Forest Genetic Resources Example 14.1: Conservation of the Genetic Resources of Pinus merkusii in Thailand Programs for the Protection of Forest Genetic Resources Recommended Literature

237 237 239 240 241

References Index

217 218 220 222 222 223 224 226 227 227 228 229 230 231 235

241 242 243 243 245 247 250 251 253 255 256 256 257 259 264 265 267 307

Introduction – Genetics of Tropical Forests

. . . the significance of forest genetics lies not only in squeezing extra gain out of marginal investments, but more importantly in understanding the genetics of forests, in appreciating what forest ecosystems are and how they operate, and in better stewards of forest resources (Namkoong 1989).

The importance of forest ecosystems for environmental protection, global biodiversity, and human welfare can hardly be overestimated. Genetic resources are widely recognized as the basis of biodiversity and the main components of ecosystems. Conservation and utilization of genetic resources are crucial aspects of sustainable forest management practices. However, our current understanding of forest genetic resources is poor at best. Genetic variation patterns have been studied for a small fraction of forest trees only. Evolutionary processes resulting in variation within species are difficult to study for long-living organisms such as forest trees, and human impact on forest genetic resources is often subtle. Thus, it is hardly surprising that misconceptions about forest genetic resources and implications of human impact on genetic resources are commonplace. This holds specifically true for tropical forests, which deserve particular attention for the following reasons. First, tropical forests are centers of biological diversity at least at the level of species diversity. Second, human impact on forest ecosystems is particularly severe in the tropics. Alterations of land use, notably deforestation, endanger forest genetic resources in most tropical countries. Common silvicultural practices such as selective cutting systems based on target diameters threaten genetic resources in managed forests. On the other hand, many tropical countries have launched ambitious reforestation and afforestation programs. The choice of reproductive material used for plantation establishment is a matter of vital importance for the success of reforestation programs. Utilization of genetic resources (tree improvement) plays an important role in this context. This book is aimed at familiarizing readers with the fundamental principles related to genetic resources in tropical forests. We will concentrate our discussion on forest trees since trees are by definition the main structural and functional

1

2

C HAPTER 1 Introduction – Genetics of Tropical Forests

components of forest ecosystems. However, other organisms will also be considered. This holds in particular for animals interacting with trees by distributing their genetic information through the dispersal of pollen or seeds. Priority is given to species of considerable economic importance, mainly trees grown in plantations for wood production (e.g., eucalypts) and keystone species serving vital functions in particular ecosystems (e.g., figs in many Asian forests). The analysis of patterns of genetic variation is in the center of our discussion. The dynamics of genetic structures of selected forest taxa without human interference is discussed in Part A of this book. The role of the evolutionary factors for shaping genetic variation patterns within and among populations is described with emphasis on the reproduction of forest trees (sexual systems, gene flow, and mating systems). Part B is devoted to the human impact on forest genetic resources in the tropics. Both deliberate changes of genetic structures owing to tree improvement and genetic implications of other human activities, including forest destruction, fragmentation, and silvicultural practices, are considered. We describe mutual dependencies between environmental conditions and evolutionary change in primary and secondary forests as well as in plantations. Genetic processes are not seen in isolation but are discussed in an ecological context following the fundamental work of Klaus Stern, the founder of the authors’ institute (Stern and Roche 1974). Genetic variation is regarded as a key aspect of biodiversity and as the basis for evolutionary adaptability. The development of management and conservation strategies for natural and man-made forests in the tropics requires the consideration of genetic aspects. The main focus of the book is on mutual dependencies between environmental conditions in the tropics and genetic variation patterns of selected species, in particular forest trees. The species richness of many tropical forests has numerous implications in this regard.

Genetic Processes in Tropical Forests

Part A

The enormous species richness of tropical forests is mirrored by multiple strategies to ensure the survival of species in tropical forests. Survival implies adaptedness to prevailing environmental conditions, but also, at least in the long run, adaptability to environmental change. Genetic variation is the basis for evolutionary adaptive change. In this part our current understanding of the dynamics of genetic variation in time and space is reviewed for species living in tropical forests. Basics of population genetics and the concept of gene marker loci are briefly introduced to readers not familiar with the analysis of genetic variation (Chap. 2). Surprisingly high levels of genetic variation within populations have been reported for the majority of, but not all, tropical forest tree species (Chap. 3). The maintenance of genetic variation during reproduction (Chap. 4) is due to particular features of gene flow and migration mechanisms (Chap. 5), and the mating system (Chap. 6) of tropical forest trees. Changes of genetic structures due to selection are difficult to monitor for long-living organisms growing in complex forest ecosystems, but are crucial for patterns of adaptive variation (Chap. 7). Finally, evolution and genetic differentiation above the species level, i.e., within genera, families, or even higher taxonomic units, are discussed (Chap. 8).

Population Genetics – an Overview

Population geneticists spend most of their time doing one of two things: describing the genetic structure of populations or theorizing on the evolutionary forces acting on populations. On a good day, these two activities mesh and true insights emerge (Gillespie 1998).

2.1 Introduction Population genetics provides the framework for an understanding of the dynamics of genetic structures or evolution. The following considerations are neither specific for nor confined to forest species. Most aspects mentioned in this chapter are as relevant for tropical forest species as for any other higherplant or animal species. However, the population genetics approach taken throughout this book to discuss genetic processes in tropical forests requires a basic understanding of the most important fundamentals concerning genetic structures and their dynamics in time and space, and introduces the concept of genetic markers as well as the most important molecular and biochemical gene markers.

2.2 The Population The genetic information or genotype of an organism does not change throughout its lifetime. It is necessary to observe and to compare genetic information of many plants or animals in order to describe genetic variation and its dynamics in time. The fundamental units for studies of genetic variation are collectives of individuals exchanging their genetic information among each other for the sexual production of the next generation. Such units of (sexual) reproduction are defined as populations. A species rarely consists of one single population. It is usually composed of several more or less isolated populations. Isolation means the absence of mating

2

6

C HAPTER 2 Population Genetics – an Overview

events between plants (or animals) from different populations in this context. However, an assessment of reproductive isolation usually requires a fairly detailed understanding of the reproduction system of a species. Furthermore, the definition of a population is operational only if an occasional (but rare) exchange of genetic information among populations is taken into consideration. Thus, interfertile, spatially clustered plants are often regarded as belonging to the same population, while spatially separated plants are regarded as parts of different populations. The delineation of populations is particularly difficult for immobile organisms with poorly described spatial distribution patterns and largely unknown means of gene dispersal through seed and pollen. In most cases, it is far from obvious which trees in a species-rich tropical forest belong to the same population and how big tree populations are. This has far-reaching consequences for levels of genetic variation of tropical forest trees, as will be discussed in later chapters. Analyses of spatial patterns of genotypes without a priori information on the delineation of populations or partially isolated subpopulations have recently been proposed as an alternative to the “traditional” approach of dividing the overall variation into components within and among demes or populations (Diniz-Filho and Telles 2002; Manel et al. 2003). Spatially explicit analyses of genetic variation patterns might be particularly informative for plant species occurring in low density in tropical forests; however, currently available information is mainly confined to comparatively small areas (Ng et al. 2004).

2.3 Variation at Gene Loci 2.3.1 The Molecular Basis of Genetic Variation

The genetic information of all organisms is stored as a sequence of the four bases or nucleotides adenine, thymine, cytosine, and guanine in their DNA. It is multiplied by the process of replication and translated to a sequence of amino acids by the processes of transcription (from DNA to messenger RNA, mRNA) and translation (from mRNA to polypeptides) (Griffiths et al. 2000). A cistron is a part of the DNA coding for a particular polypeptide, for example, an enzyme. From this point of view, the variation at a particular gene or gene locus can be assessed by observing differences in the DNA sequence within a cistron. The total number of “genes” is unknown for any tropical forest plant; however, it may be assumed to be at least on the order of more than 25,000 estimated for the first fully sequenced flowering plant, the herbaceous

2.3 Variation at Gene Loci

Arabidopsis thaliana (The Arabidopsis Genome Initiative 2000). The number of genes of the first sequenced tree species (Populus trichocarpa) is estimated to be even higher (http://genome.jgi-psf.org/Poptr1/Poptr1.home.html). The observation of variation at the most basic level of DNA sequences is routine for tropical trees just as for other organisms, but it is time-consuming and costly if sequences of hundreds of organisms need to be compared. To date, only a few population studies have been conducted based on the observation of variation of DNA sequences within a single or a few closely related tropical forest species (but see, e.g., Ishiyama et al. 2003). Sequencing is more frequently applied to clarify taxonomic subdivisions and phylogenies (Chap. 8). Numerous alternatives to sequencing are available. The choice of the most appropriate method depends on several factors, including available resources and the purpose of a study. A comprehensive overview of currently available molecular tools to assess patterns of genetic variation in plants has been given by Weising et al. (2005). 2.3.1.1 Molecular Markers

The most important methods to assess patterns of genetic variation within and among species rely on the amplification of short DNA fragments by means of polymerase chain reaction (PCR). PCR allows short fragments of the DNA of a target organism to be selectively amplified (multiplied) for further analyses. The sequence of two short oligonucleotides (primers), each usually between ten and 25 base pairs in length, is decisive for the amplified DNA region. The following marker types, all based on the PCR method, have gained particular importance in the study of genetic variation patterns of tropical forest trees. (Partial) Gene Sequences and Single Nucleotide Polymorphisms

Selective primers can be designed to amplify the DNA coding for a particular gene (cistron). DNA regions translated into a polypeptide (exons) and nontranslated regions (introns) of structural genes as well as regulatory genes can be analyzed. It is possible to sequence the amplified PCR products, often after a cloning step. For example, Kamiya et al. (2005) investigated the phylogeny of the species-rich genus Shorea (Dipterocarpaceae) on the basis of a (partial) sequence of the PgiC gene (Example 8.1). An investigation of genetic variation is particularly rewarding at gene loci with known or putative function and a directly observable effect on phenotypic traits responsible for the adaptation to particular environmental conditions (candidate genes). Populations can be screened for single nucleotide polymorphisms (SNPs) without the need for repeated sequencing (Morin et al. 2004).

7

8


The characterization of expressed sequence tags, studies of molecular variation at the level of expressed genes, and analyses of variation at SNPs are expected to greatly improve our understanding of adaptive genetic variation and its dynamics in space and time. However, these recently developed marker types are not discussed in detail owing to the lack of published reports for tropical forest species (but see Sect. 7.2). Microsatellites (Simple Sequence Repeats)

Microsatellites or simple sequence repeats (SSRs) are short DNA fragments of usually only two or three base pairs in length which are repeated several times in a particular location of the DNA (Fig. 2.1). SSRs can be studied by the development of primers in conserved DNA regions flanking the microsatellite. The development of primers is rather costly and time-consuming (Squirrell et al. 2003). Nuclear microsatellites developed for a particular species can only be transferred to closely related species, usually within the same genus. Nuclear microsatellites are important genetic markers owing to their usually high variability and codominant mode of inheritance (Sect. 2.3.2, Chap. 3). The variation of microsatellite fragment sizes is due to a different number of repeated motifs resulting in slightly different sizes of amplified fragments (Fig. 2.2). The availability of highly variable markers is particularly useful for analyses of the mating system and gene flow (Chaps. 5, 6). Microsatellites have been developed for a number of tropical forest tree species, including Pithecellobium elegans (Chase et al. 1996a), Swietenia humilis (White and Powell 1997), Gliricidia sepium (Dawson et al. 1997), Shorea curtisii (Ujino et al. 1998), Caryocar brasiliense (Collevatti et al. 1999), Neobalanocarpus heimii (Iwata et al. 2000), and Prosopis spp. (Mottura et al. 2005). Microsatellite motifs are not only found in DNA of the nucleus, but also in DNA of chloroplasts (cpDNA; Sect. 2.3.2). Single nucleotide repeats are particularly common in cpDNA. Universal primers are available to investigate variation at cpSSRs in many different species of angiosperms (Weising and Gardner 1999) and gymnosperms (Vendramin et al. 1996).

AAGGATAAGTTAAA ACACACACACACACACAC GTTGCCTCCATTT TATGATGTATGAAT ACACACACACACACACAC CCCACCTGGTTTT TTGACGTCACACAG ACACACACACACACACACAC TCTCTCATCCACA

Fig. 2.1. Three different DNA sequences of Prosopis chilensis containing the microsatellite repeat motif AC in nine (lower and middle) or ten (upper) copies (underlined). A adenine, C cytosine, G guanine, T thymine. (From Mottura, unpublished)

2.3 Variation at Gene Loci 195

200

205

210

215

220

225

230

235

240

245

250

255

260 1000 500

1000 500

600 400 200

2000 1000

1500 1000 500

900 600 300

900 400 200

2000 1500 1000 500

Fig. 2.2. Length variation for eight Prosopis spp. trees at a microsatellite gene locus (Mottura et al. 2005) separated in an automated sequencer (ABI 3100). Each line refers to one tree. Different sizes of amplified fragments are visualized as peaks at different positions in the chromatogram. Trees with two amplified fragments are heterozygous; tree 5 is homozygous with only one amplified fragment. (From Mottura, unpublished)

Random Amplified Polymorphic DNA and Amplified Fragment Length Polymorphism

Random amplified polymorphic DNA (RAPD; Newbury and Ford-Lloyd 1993; Fig. 2.3) and amplified fragment length polymorphism (AFLP; Vos et al. 1995; Fig. 2.4) are genetic fingerprinting techniques resulting in more or less complex DNA patterns. Unlike most other techniques including microsatellites no previous sequence information is necessary for RAPD and AFLP studies. The presence (+ or 1) or absence (− or 0) of a DNA fragment of a particular size is scored for each sample plant investigated. This information is the basis for a matrix of size n × m (n is the number of plants investigated; m is the number of DNA fragments observed) with elements 1 (fragment present) or 0 (fragment absent). RAPDs and AFLPs are usually interpreted as dominant markers (see later). No

9

10


Fig. 2.3. Low variation at a random amplified polymorphic DNA marker in teak (Tectona grandis). (From Finkeldey, unpublished)

information on the function of a particular DNA fragment or its mode of inheritance is available; hence, RAPDs and AFLPs are “anonymous” markers. The RAPD technique is a simple molecular tool to assess genetic variation within and among populations which has been widely used for tropical forest plants such as Gliricidia spp. (Chalmers et al. 1992), Cedrela odorata (Gillies et al. 1997), Caesalpinia echinata (Cardoso et al. 1998), Prunus africana (Dawson and Powell 1999), and many others. The reproducibility of the RAPD technique has been under dispute (Rabouam et al. 1999). The AFLP method is more demanding with regard to laboratory equipment and experience, but allows more DNA fragments to be investigated from a single PCR reaction, and shows higher reproducibility in comparison with RAPDs. The potential of the AFLP technique to assess genetic variation patterns in tropical trees was recently reviewed by Kremer et al. (2005). AFLP studies to assess genetic variation within tropical tree species were conducted, for example, for Moringa oleifera (Muluvi et al. 1999), Calycophyllum spruceanum (Russel et al. 1999), and Acer skutchii (Lara-Gomez et al. 2005). Restriction Fragment Length Polymorphism

A simple method to observe variation among DNA fragments from different plants is to digest the fragments into smaller pieces by restriction enzymes. Restriction enzymes cut DNA at a particular sequence. These recognition sequences are usually between four and six base pairs in length. The restriction fragment length polymorphism (RFLP) technique is based on the use of restriction enzymes to observe genetic variation. For example, nuclear RFLPs

2.3 Variation at Gene Loci 220

222

224

226

228

230

232 300 200 100

200 100

300 200 100 221.05

300 200 100 222.04

400 300 200 100

200 150 100 50

200 150 100 50

Fig. 2.4. Variation at amplified fragment length polymorphisms in seven Shorea parvifolia trees after separation in an automated sequencer (ABI 3100). Only fragments in the range 220–232 bp are shown. The fragment with a size of 220 bp (arrows) is “diagnostic” for the population Berau (Borneo), since it was not observed in six other populations. (From Cao, unpublished)

were developed for Acacia mangium to assess genetic diversity and differentiation among populations (Butcher et al. 1998) and with the objective to incorporate marker-based approaches in breeding programs (Butcher et al. 2000; Butcher 2004). More frequently, the RFLP techniques is combined with a PCR step and is used to assess variation of cpDNA. For example, Tsumura et al. (1996) and Indrioko et al. (2006) (Fig. 2.5) analyzed phylogenetic relationships among Southeast Asian dipterocarps using the PCR-RFLP technique (Example 8.1), and Cavers et al. (2003) observed strong genetic differentiation among five cpDNA haplotypes of Cedrela odorata in Central America (Example 3.6).

11

12

C HAPTER 2 Population Genetics – an Overview M 1 2 3 4

5 6 7 8 9 10 11 12

Fig. 2.5. Polymerase chain reaction restriction fragment length polymorphisms of Indonesian Dipterocarpaceae after amplification with the primer pair TrnL–TrnF and restriction with TaqI. M marker (size standard), 1 Hopea celebica, 2 Vatica rassak, 3 V. pauciflora, 4 H. bancana, 5 Dipterocarpus oblongifolius, 6 S. javanica, 7 V. granulata, 8 Upuna borneensis, 9 S. mecistopteryx, 10 D. rigidus, 11 S. mecistopteryx, 12 H. grifithii. (From Indrioko, unpublished)

2.3.1.2 Biochemical Markers – Isozymes

The sequence of nucleotides of the DNA is converted to a sequence of amino acids of polypeptides by the processes of transcription and translation. Enzymes are the most important group of polypeptides catalyzing all kinds of biochemical reactions in the metabolism; thus, there is a close relation between enzymes and controlling genetic information according to the “one gene – one polypeptide hypothesis.” Isozymes are enzymes with similar or even identical functions. Electrophoresis of isozymes followed by biochemical staining is a simple way to observe differences with regard to their electric load and/or spatial structure (Rothe 1994; Fig. 2.6). It is often possible to directly relate the variation observed by isozyme electrophoresis to genetic variation (Bergmann and Hattemer 1998). A formal inheritance study (Sect. 2.3.2) is strongly recommended prior to the use of isozymes as genetic markers. Isozyme gene loci became the first widely used markers to assess patterns of genetic variation of tropical forest species at single gene loci. The ease of isozyme


Tree: 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16

A1 A2 A2A2

A2A2

A1A2

A1A2

A2A2

A2A2

A1A2

A1A2 A1A2

A1A2

A1A1

A1A2

A1A2

A2A2

A1A2 A2A2

Fig. 2.6. Zymogram showing variation of 16 Dalbergia sissoo trees from Nepal at the enzyme system alcohol dehydrogenase. The variation is controlled by a gene locus ADH-A. The inferred genotypes of the respective trees are indicated. (From Pandey, unpublished)

inventories and the comparatively low costs have made them useful tools for population genetics studies, although they were often substituted by various types of molecular DNA markers during the last decade. 2.3.2 The Gene As a Unit of Heredity

Genetics as the science of heredity dates back to the experiments of the Austrian monk Gregor Mendel (1866), if not earlier. The definition of a gene as a unit of heredity was introduced in the early twentieth century, long before the role of DNA as the material basis of genetic information was recognized. A gene in this sense is identified by the observation of segregation within progenies of particular parents. The basic idea for the identification of a gene as a unit of heredity has remained unchanged since the experiments Mendel (1866). He investigated segregation ratios in progenies after controlled pollination at simple morphological traits such as the color of petals in peas. If a plant is heterozygous at a controlling gene locus, it will transmit only one of the two different alleles to a particular progeny (the terms heterozygosity and allele are described later). Thus, segregation is expected in progenies if at least one parent was heterozygous at a controlling gene locus. It is possible to compare the observed segregation in a sample of progenies after controlled pollination with an expectation based on the assumption of a simple “Mendelian” inheritance of the trait. The controlling gene or the gene locus is identified, if the

13

14


differences between the observed and expected values are not significant (Hattemer 1991). Suitable statistical tests are, for example, described by Sokal and Rohlf (1998, p. 686 ff.; c2 test or goodness-of-fit test, G test). The formal identification of a genetic marker locus by an observation of segregation is possible only for environmentally stable, genetic traits. It is also advisable for the identification of gene loci based on the observation of biochemical or molecular markers. For example, Moran and Bell (1983) observed segregation at the isozyme system GOT in progenies of Eucalyptus regnans after controlled pollination and found no evidence to reject the hypothesis of the genetic control of the variation by a single gene locus. Forest trees, in particular those of the tropics, are extraordinarily difficult to cross owing to their large size, short viabilities of pollen, unknown reproductive biology, and other obstacles. Thus, the identification of a gene locus by inheritance studies is preferentially conducted on the basis of alternative approaches such as the observation of segregation in the haploid megagametophyte (the “endosperm”) of gymnosperms (Bergmann 1974) or the observation of segregation in the progenies of putatively heterozygous seed trees after open pollination (Gillet and Hattemer 1989). The first method has been used to identify enzyme gene loci in Pinus merkusii from Thailand (Changtragoon and Finkeldey 1995b); the latter approach was used to clarify the inheritance of isozymes in Pterocarpus indicus from the Philippines (Finkeldey et al. 1998). 2.3.3 The Mode of Inheritance

The analysis of variation patterns at environmentally stable, genetic traits (Chap. 3) requires an understanding of the transmission of the trait from parents to progenies and from one generation to the next. Two considerations are of particular importance in this context: the transmission of genetic information from both parents or from a single parent only, and the impact of a single or two alleles at a gene locus on the observed trait expression (the phenotype). Most of the DNA in a plant cell is located in the nucleus. Nuclear DNA (nDNA) is inherited by both parents. For diploid organisms, one set of chromosomes is transmitted by each of the two parents to their common progenies; thus, nDNA is typically biparentally inherited. In addition, DNA of plants is located in two types of plastids: mitochondria (mitochrondrial DNA, mtDNA) and chloroplasts (cpDNA). DNA in plastids, both mitochondria and chloroplasts, is usually transmitted by a single parent to its progenies (uniparental inheritance). mtDNA is mainly inherited from the female or seed parent only


(maternal inheritance). The same holds for cpDNA in angiosperms. However, cpDNA of gymnosperms is typically paternally inherited, i.e., only the genetic information of the pollen parent is transferred to a progeny. cpDNA was investigated in most studies on genetic variation of tropical forest plants based on uniparentally inherited markers (Example 3.6). Maternally inherited mtDNA was studied in several animal species such as Central American frogs (Crawford 2003) and the giant panda in China (Lu et al. 2001). The different types of genetic information at any biparentally inherited gene locus are called alleles. Since two sets of chromosomes are inherited (one from the seed parent, the other from the pollen parent), each nuclear gene is represented in two copies in each progeny of a diploid species. Polyploids are not discussed in this introduction. A tree is homozygous (e.g., A1A1) at a particular locus A, if two identical alleles (both A1) were inherited from its parents. A heterozygous tree (e.g., A1A2) received two different alleles (A1 from the seed parent and A2 from the pollen parent or vice versa) from its parents. An allele is dominant, if its possession in a single copy is sufficient to express a particular phenotype. If A1 is dominant, the phenotypic expressions of trees with the genotypes A1A1, A1A2, A1A3, etc. are identical. The other alleles (A2 and A3 in the example) are recessive. If it is possible to distinguish all heterozygotes (e.g., A1A2) from all homozygotes (A1A1 and A2 A2), the alleles are defined as codominant. A simple observation of genetic structures (Sect. 2.4) is only possible at codominant marker loci. 2.3.4 Definition and Classification of Gene Markers

On the basis of the previous considerations, it is possible to define a gene marker in different contexts: ●

●

A gene marker (synonyms gene locus and marker locus) in a wide sense is an environmentally stable trait; thus, the variation is determined by genetic factors only. A gene marker in a narrow sense is an environmentally stable, biparentally inherited, codominant trait.

Although many different traits may be regarded as gene markers in a wide sense, only the molecular and biochemical markers briefly described in Sect. 2.3.1 have considerable practical importance to assess genetic variation patterns of tropical forest species. A rough classification of important, currently available marker types with regard to their mode of inheritance and their variability is presented in Table 2.1.

15

16


Table 2.1. Classification of genetic markers in a broad sense commonly used in studies on genetic variation of tropical forest species with regard to their mode of inheritance and their variability Variation

Uniparental

Anonymousa

Codominantb

High

Chloroplast simple sequence repeats; sequences of chloroplast DNA Polymerase chain reaction restriction fragment length polymorphisms of chloroplast DNA and mitochondrial DNA

Amplified fragment length polymorphisms Random amplified polymorphic DNAs

Microsatellites (simple sequence repeats) Isozymes

Moderate to low

a

Mode of inheritance usually unknown; interpretation as dominant markers Gene markers in a narrow sense

b

2.4 Genetic Structures Within Populations The final result from a laboratory study to assess the genetic constitution of a plant at a particular biochemical or molecular gene marker system is a specific pattern, which can be used: ●

●

●

To infer a genotype in case of gene markers in a narrow sense (e.g., isozymes, SSRs) To identify a haplotype in the case of uniparentally inherited markers (e.g., PCR-RFLPs of cpDNA) To assign a genetic fingerprint usually consisting of many anonymous marker loci to the plant (e.g., AFLPs)

Genetic structures are frequency distributions of such genetic types in populations. The most important frequency distributions to assess levels of genetic variation within and among populations are allelic and genotypic structures. 2.4.1 Allelic and Genotypic Structures

Genetic variation is assessed by an investigation of all plants or a (random) sample of plants from a population. At a gene marker in a narrow sense, the genotypes of these plants are known, and it is possible to calculate the (relative) frequency of a particular genotype Ai Aj as Pij = Nij /N,

(2.1)

2.4 Genetic Structures Within Populations

where Nij is the (absolute) frequency of genotype AiAj in the population (or sample) and N is the population size (or sample size). The frequency vector of all (relative) frequencies of genotypes is defined as the genotypic structure of the population at the respective marker gene locus. It obviously holds that // Pij = 1. i

j

Each genotype of a diploid species consists of two alleles; thus, the total number of alleles in a population (or a sample) is twice the number of plants investigated, and the frequency of a particular allele Ai can be calculated as p i = Ni / 2N,

(2.2)

where Ni is (absolute) frequency of allele Ai in the population (or sample) and N is the population size (or sample size). The respective allele Ai is counted twice in case of homozygous plants AiAi. The frequency vector of all (relative) frequencies of alleles is defined as the allelic structure of the population at the respective marker gene locus. Again, it holds that / p i = 1. i If the genotypic structure of a population is known, it is easily possible to compute the allelic structure at the respective gene locus. The frequency of allele Ai (pi) can be computed by calculating the sum of the frequency of the corresponding homozygote (Ai Ai ) and half of the frequencies of all heterozygotes where the respective allele occurs (Ai Aj , Ai Ak , . . .). However, the calculation of genotypic structures based on allele frequencies requires additional information or assumptions (e.g., Sect. 6.1.2). Example 2.1: Genetic Structures at an Isozyme Gene Locus in Dalbergia sissoo The observed variation at the isozyme gene locus ADH-A is illustrated for 16 Dalbergia sissoo trees in Fig. 2.6. Three different genotypes were observed. Six trees (nos. 1, 3, 4, 7, 15, and 16) show the homozygote genotype A2A2, tree 10 is the only tree with genotype A1A1, and the remaining nine trees exhibit the heterozygous genotype A1A2. Thus, the genotypic structure of the sample is as follows: P11=1/16=0.0625, P12=9/16=0.5625, P22=6/16=0.3750. Eleven of the 32 observed alleles (each tree has two alleles) are A1, the remaining 21 alleles are of type A2. The allelic structure may also be computed as follows: p1 = P11 + 0.5P12 = 0.0625 + 0.5625/2 = 11/32 = 0.34375, p2 = P22 + 0.5P12 = 0.375 + 0.5625/2 = 21/32 = 0.65625.

17

18


2.4.2 Variation at Uniparentally Inherited Markers

The uniparental inheritance of cpDNA and mtDNA implies the absence of recombination among genes. Since cpDNA and mtDNA are usually transmitted without any changes in maternal or, in the case of cpDNA of gymnosperms, paternal lineages, it is possible to interpret variation patterns observed at different loci together and to assign particular cpDNA or mtDNA haplotypes to different plants. The interpretation of genetic diversity within and among populations (Chap. 3) is conventionally based on these haplotypes rather than on variation at particular loci. Example 2.2: Diversity of cpDNA Haplotypes in Dalbergia sissoo Six to ten seeds from randomly selected, different seed trees were investigated in each of ten populations of D. sissoo in Nepal. The inventory was conducted in five regions, each represented by one natural population and a neighboring plantation. Variation was detected by the PCR-RFLP technique (Sect. 2.3.1) by restriction of the amplified fragment Trn K1/K2 with the enzymes AluI and RsaI, and by amplification of the chloroplast microsatellites (cpSSRs) ccmp6 and ccmp7 (Weising and Gardner 1999). A total of eight haplotypes with two or three haplotypes in each population were observed (Table 2.2). Haplotypes 1–3 were only observed in natural populations, while haplotypes 4–8 occurred mainly in plantations (Pandey et al. 2004). Table 2.2. Variation of chloroplast DNA haplotypes in five natural populations and five plantations of Dalbergia sissoo in Nepal. (Adapted from Pandey et al. 2004) Region

Population

Natural populations A Hetauda N B Shivapur C Hattisar D Godawari E Pipariya Plantations A Hetauda P B Surai C Thakurdwara D Attaria E Shuklaphanta N sample size

N

10 10 6 10 10 10 10 10 10 10

Chloroplast DNA haplotypes 1

2

9 8 5 4

1

3

4

5

6

4

6

7

8

7 3 4 5

2 3

2 1 6 9

1

1 4 6 5

2.5 Evolution and Evolutionary Factors

2.5 Evolution and Evolutionary Factors Today’s biodiversity is the result of the biological evolution which started at least three and a half billion years ago on earth. The number of existing species is still unknown, since a large proportion of the species living on our planet have not yet been described. The majority of “unknown” species are insects living in the canopy of tropical forests (Erwin 1988). Current estimates of the global species diversity vary from approximately two million to 30 or even 50 million species (Erwin 1997). The long history of evolution on earth does not preclude the study of its effects in short periods. Evolution is permanently ongoing and it is possible to observe its effects even in long-living organisms such as forest trees. Furthermore, the majority of evolutionary change is within existing species and does not immediately result in the creation of new taxa. The main principle of evolution was recognized and described by Charles Darwin (1859): Biological evolution is based on heritable variation within species or populations. Darwin recognized the key role of natural selection in this context. Fundamental observations, which led to the development of the selection theory, were made both by Darwin and Alfred Russel Wallace, who independently developed a similar theory, after extended studies in tropical areas (Lefèvre 1984; Sect. 8.2). Our advanced knowledge about the process of inheritance allow us to define evolution more precisely today that during Darwin’s era. Evolution is a change of the genetic (allelic or genotypic) structure of a population at one or several gene loci. Only changes improving the adaptedness of a population to particular environmental conditions were regarded as evolution by Dobzhansky et al. (1977); however, Kimura (1983) proposed including all changes of genetic structures including those at selectively neutral loci into the concept. Evolutionary factors are the causes of changes of genetic structures (Hedrick 2000). The significance of evolutionary factors for the dynamics of genetic structures of tropical forest plants is the main topic in later chapters of this book; thus, only a brief description is given here: ●

Mutation is the prerequisite for any genetic variation. Mutations are random changes of genetic information of an organism which might affect a single nucleotide (a spontaneous change of a single nucleotide resulting in a SNP; Sect. 2.3.1), the chromosomal structure (e.g., fissions, translocations, inversions), or the number of genes (e.g., duplications, polyploidy). Most mutations are detrimental or neutral for an organism. Only few advantageous mutations increase the fitness, i.e., the capacity to produce offspring.

19

20


●

●

●

●

Mutations are rare events; mutation rates are estimated to be in the range between 10−6 and 10−8 for most genes. However, the frequency of mutations is not uniform. For example, considerably higher estimates of mutation rates were reported for microsatellites (Goldstein and Pollock 1997). Gene flow and migration increase the genetic multiplicity of a population, if genes which were previously restricted to certain populations migrate to other populations. Transport of pollen (gene flow in a narrow sense) and seeds (migration) are the main processes involved in the dispersal of genes for most plant species (Chap. 5). By definition, gene flow and migration among populations are rare events. The mating system is decisive for the fusion of male and female gametes. Gene flow through pollen is an obvious prerequisite for mating between two seed plants. Thus, gene flow and the mating system are closely connected and are the two most important aspects of the reproduction system of a plant species. The mating system decides on the combination of alleles to genotypes. The genotypic structure of the progeny generation is influenced by the mating system, which primarily does not cause changes of allelic structures. Inbreeding due to selfing or mating among relatives is a particularly important aspect of the mating system (Chap. 6). Random fluctuations of genetic structures are described as genetic drift. Genetic drift is a consequence of limited population size and becomes stronger in small populations. The strength of genetic drift can be described as a function of the population size, although its final outcome is unpredictable. Alleles, in particular rare alleles, are likely to completely disappear from populations as a result of genetic drift, and previously polymorphic (variable) gene loci may become monomorphic (fixed). A sudden reduction of the population size for one or several generations is often described as a genetic “bottleneck.” Bottlenecks result in reduced genetic variation even after later population expansion owing to the effects of genetic drift while populations are small. Genetic differentiation among partially or completely isolated populations is expected to increase as a result of genetic drift. Selection is a consequence of the differing abilities of organisms to produce offspring (fitness). Selection does not increase multiplicity, but often contributes to the maintenance of genetic diversity by stabilizing polymorphisms, for example, owing to a selective advantage of heterozygous genotypes. In general, selection results in an improved adaptedness (Chap. 7). The basis of selection is the different survival of genotypes (viability selection) and the different contributions of genotypes to the formation of a progeny generation (fertility selection).

Population geneticists are concerned with the effect of these evolutionary factors on patterns of genetic variation within and among populations

2.6 Phenotypic Variation

(Gillespie 1998). Genetic structures result from complicated interactions between the different evolutionary factors in natural and managed populations. The role of different evolutionary factors for the creation and maintenance of the extraordinary species richness of tropical forests has been widely discussed in the past. A high frequency of selfing and restricted gene flow and migration resulting in small populations, which become differentiated from each other mainly owing to the effects of genetic drift, have been proposed as the main forces for the evolution of the enormous species richness in tropical forests by Fedorov (1966). This view was challenged by Ashton (1969), who argued in favor of more efficient means of gene flow through seed and pollen than previously thought. The evolution of species diversity is promoted by selective differentiation in partially isolated subpopulations which eventually develop reproductive barriers. The latter proposition is greatly supported by makerbased assessments on genetic variation patterns in tropical forest species (Chap. 3) and on the reproduction system of tropical forest plants (Chaps. 5, 6).

2.6 Phenotypic Variation The vast majority of directly observable traits of trees are an outcome of the interaction between the genetic constitution of a plant at a large number of gene loci (genotype, G) and the environmental conditions (E). Thus, it holds for the phenotype (P) that P= G×E, where × symbolizes the interaction between environmental and genotypic effects, which often differs from simple additivity of both components. Phenotypic traits deserve particular interest if they are important for the adaptation of plants to their environment (adaptive traits) or for the value of a tree (economic traits). Adaptive traits such as the tolerance against biotic (pests) or abiotic (e.g., draught) stress and economic traits such as volume growth are controlled by both genetic and environmental factors for most plant species. The analysis of phenotypic traits in a breeding context will be discussed in Chap. 12 in more detail. It is also described for forest trees in other textbooks (Wright 1976; Zobel and Talbert 1984; Williams et al. 2002). The majority of the gene markers described in Sect. 2.3.1 are unlikely to have strong, recognizable effects on adaptive traits or traits of economic importance. They are regarded as “neutral” or “nearly neutral” markers with regard to an adaptation to the environment; however, this does not preclude their application to monitor adaptive processes (Ziehe et al. 1999; Sect. 7.2). In line with the definition of genetic structures, the frequency distribution of particular phenotypes is defined as the phenotypic structure of a population. Direct and indirect effects of the environment on the phenotypic structure are illustrated in Fig. 2.7. Apart from the direct impact of the environment

21

22


Migration

Genotypic structure

Mode of gene action

Phenotypic structure

Modification Viability selection

Mating system

Environmental influence on mating

Mutation

Environment

Population size Fertility selection

Gene flow

Allelic structure

Fig. 2.7. The environmental impact on phenotypic structures. (Adapted from Hattemer and Müller-Starck 1990)

on the phenotype (P = G× E), there are numerous impacts of the environment on genotypic structures. The environmental conditions influence mutation rates, are crucial for population sizes and, hence, the importance of genetic drift, cause changes of allelic structures owing to fertility selection and genotypic structures owing to viability selection (Chap. 7), and have an impact on the mating system (Chap. 6). Thus, manifold human alterations of environmental conditions do not only directly effect phenotypes of forest plants, but also change the genotypic structure of populations.

2.7 Recommended Literature The molecular basics of genetics is covered in many recently published textbooks, such as the one written by Griffiths et al. (2000). A useful and detailed description of genetic markers and their application to study plant genetics is given by Weising et al. (2005). Altukhov and Salmenkova (2002) provide a comprehensive survey on currently available markers to study genetic variation of plants. The need for inheritance studies in order to identify gene markers was emphasized by Hattemer (1991). The basics of population genetics are covered by Hedrick (2000). The book of Altukhov (2006) represents another introduction to population genetics putting genetic variation in an ecological context.

Genetic Variation of Tropical Forest Plants

. . . the main difficulty one faces in a subject like population genetics (. . .) is not the mathematics itself, or the biology itself: it is how to fit them together (p. ix in Maynard Smith 1998).

3.1 Introduction A main application of genetic markers is the observation of genetic variation patterns within and among populations. Variation at phenotypic traits is influenced both by genetic and by environmental factors. Thus, it is impossible to conclude on levels of genetic variation at single loci on the basis of the observation of phenotypes at adaptive traits or traits of economic importance, although an assessment of variance components in field trials allows the degree of genetic control of a phenotypic trait to be estimated (p. 49 ff. in Eriksson and Ekberg 2001). The application of gene markers has greatly enhanced our understanding of genetic variation of tropical forest plants.

3.2 Genetic Inventories Genetic inventories aim at describing extant patterns of genetic variation within a single or, more frequently, within and among several populations. These patterns are the result of the impact of the evolutionary factors acting on populations in the past (Sect. 2.5). On the other hand, the evolutionary adaptive potential of populations depends on genetic variation patterns (Sect. 7.2). For practical reasons, it is only possible to observe a small and very often presumably not representative set of marker gene loci for a given species. Variation at most of the tens of thousands of gene loci remains unexplored. Thus, an assessment of evolutionary adaptive potentials of populations is likely to be unreliable if it is based on the small set of loci investigated only. This holds irrespective of the type of marker loci investigated. Still, the limited variation

3

24

C HAPTER 3 Genetic Variation of Tropical Forest Plants

observed at marker loci is often useful to elucidate the evolutionary past of populations and species. This argument is particularly valid if genetic structures have been shaped by evolutionary factors acting on all genes in principally the same way. For example, a “bottleneck” in the evolutionary past of a population is expected to reduce genetic variation due to drift at all loci, and particular types of mating systems, for example, strong inbreeding, result in characteristic genotypic structures (Chap. 6). On the other hand, it is difficult to assess the impact of selection since its effects are restricted to the selected and associated loci (Sect. 7.2). The conclusions drawn on the evolutionary past of populations and species can be used to assess their evolutionary adaptive potential (Finkeldey and Mátyás 1999).

3.3 Measurement of Genetic Variation A detailed description of the various methods to analyze genetic variation patterns is beyond the scope of this book. The book by Nei (1987) should be consulted for a comprehensive account. Berg and Hamrick (1997) and Weising et al. (2005, p. 207ff.) present brief overviews. Only simple measures and methods frequently used to analyze patterns of genetic variation of tropical forest species are briefly introduced in this section. Statistical problems related to the estimation of genetic variation based on samples rather than on full inventories of all plants are not discussed in detail. The focus is on the population genetics interpretation of allelic and genotypic structures at gene markers in a narrow sense, i.e., nuclear, codominant markers such as microsatellites or isozymes. The most complete and comprehensive information on genetic variation patterns at a particular marker locus is the genotypic and allelic structures at that locus (Sect. 2.4); however, it is difficult to compare variation levels among loci or populations on the basis of these genetic structures. Thus, the genotypic and allelic frequency distributions are “condensed” to measures characterizing particular aspects of variation within or among populations. This is particularly important if the aim is to investigate and to compare genetic variation at several gene loci. The gene pool is the average or mean of genetic variation estimates from several single loci. 3.3.1 Genetic Variation Within Populations

The genetic variation within populations is characterized by measures of genetic multiplicity, which only take into account the occurrence of different genetic types, but not their frequencies, and measures of genetic diversity, which also consider the frequency of different genetic types (alleles or genotypes).

3.3 Measurement of Genetic Variation

3.3.1.1 Genetic Multiplicity

Proportion of Polymorphic Loci The simplest estimate of genetic multiplicity is the proportion of polymorphic (variable) loci (PPL). It is calculated by dividing the number of polymorphic gene loci (PL) by the number of all loci investigated, which also includes loci showing no variation (monomorphic loci, ML): PPL = PL / (PL + ML).

(3.1)

The PPL is often reported as a percentage after multiplication by 100. Some authors regard a locus as polymorphic only if the frequency of the most frequent allele is below 95% (or 99%). Other authors regard a locus as polymorphic if at least two types (irrespective of their frequencies) occur. Number of Alleles per Locus and Allelic Richness

Another simple measure of genetic multiplicity is the average number of alleles per gene locus. It is calculated by counting the number of alleles (ni) at each of L gene loci, summing over all loci investigated, and dividing this total number of alleles by the number of loci L. L

/n Number of alleles per locus = i = L1

i

(3.2)

Since the probability of observing a (rare) allele in a population strongly depends on sample size, comparisons of the average number of alleles among populations possibly lead to false conclusions if the number of sampled plants differs among populations. Effects of different sample sizes are taken into consideration by the computation of the allelic richness (Petit et al. 1998; Comps et al. 2001). 3.3.1.2 Genetic Diversity

The genetic diversity of a population does not only depend on the number of genetic types (alleles or genotypes), but also on their (relative) frequencies (pi , pij). Effective Number of Alleles

The “effective” number of alleles at a gene locus (Crow and Kimura 1970) has also been defined as allelic diversity v (Gregorius 1978) of a population at that locus. If n is the total number of alleles at a locus and pi are the relative frequencies of the alleles, the effective number of alleles v is calculated as follows:

25

26


1#v=

1 # n. / p i2

(3.3)

i

The effective number of alleles v reaches its maximum value for a given n if the frequencies of all alleles are identical (1/n). In this case, v = n. The allelic diversity v decreases with increasing frequency of one allele and consequently decreasing frequencies of the remaining alleles. The effective number of alleles of the gene pool is the harmonic mean of the v at single loci. The effective number of alleles from Example 2.1 is v=

1

/p i

2

=

i

1 = 1 . 1.82. p 12 + p 22 0.34375 2 + 0.65625 2

Expected Heterozygosity

The “expected heterozygosity” He is the most frequently used measure of genetic diversity within populations. The “expectation” is based on the assumption that alleles are randomly combined to genotypes, i.e., that Hardy–Weinberg structures are realized (Sect. 6.2.2). In this case, it is possible to calculate the frequency Pii of a particular homozygote with genotype Ai Ai as pi2 (pi is the frequency of allele Ai). The frequency of the expected homozygosity is calculated by summing over all alleles (Σpi2), the remainder being the expected heterozygosity. Accordingly, the expected heterozygosity is calculated as H e = 1 - / p i2 .

(3.4)

i

Gene pool values of the expected heterozygosity are the arithmetic mean of single locus values. The expected heterozygosity from Example 2.1 is H e = 1 - (p 12 + p 22 ) = 1 - (0.34375 2 + 0.65625 2 ) . 0.451. Observed Heterozygosity

A simple measure of genetic diversity is the proportion of all heterozygotes from all plants (homozygotes and heterozygotes) investigated in a population: n

n

n

Ho = / / Pij = 1 - / Pii . i = 1j = 1 i ! j

(3.5)

i = 1

This observed heterozygosity equals the expected heterozygosity if the population exhibits Hardy–Weinberg structures. The gene pool value of the observed heterozygosity is calculated as the arithmetic mean of Ho from single loci. The observed heterozygosity of the Dalbergia sissoo population in Example 2.1 is Ho= P12 = 1 – (P11 + P 22) = 0.5625.

3.3 Measurement of Genetic Variation

It is also possible to calculate the degree of individual heterozygosity of a single plant by dividing the number of heterozygous gene loci by the number of observed loci. 3.3.2 Genetic Differentiation Among Populations

Genetic differences between two populations are quantified by measures of genetic distance; the differences between an arbitrary number of populations are measured as genetic differentiation. 3.3.2.1 Genetic Distances

Out of a large choice of commonly used genetic distances only the minimum genetic distance according to Nei (1973) is mentioned here. For two populations k and l it is defined as 2 d min (k , l ) = 12 / 7 p i (k) - p i (l )A , (3.6) i where pi(k) and pi(l ) are the frequencies of the ith allele in populations k and l, respectively. This distance reaches its minimum value if the allelic structures of both populations are identical [pi(k)=pi(l) for all alleles i], and has a maximum value of 1 if the populations are fixed on different alleles. More commonly used distance measures such as Nei’s standard genetic distance are described by Nei (1987). If more than two populations are investigated, it is possible to compute a matrix with all pairwise genetic distances, and to use this information as a basis for the illustration of genetic differentiation patterns by means of cluster analyses (e.g., Fig. 3.4). Populations are grouped on the basis of their genetic distances by different algorithms. A simple and commonly used method for grouping in population genetics is the unweighted pair-group method with arithmetic means (UPGMA) method (Sneath and Sokal 1973). Similar methods are used to illustrate the relatedness of higher taxa in molecular phylogenies (Sect. 8.3). 3.3.2.2 Genetic Differentiation

The standard method to assess genetic differentiation among an arbitrary number of populations is the calculation of GST (or FST) (Nei 1973). The total gene diversity in a population composed of several subpopulations is calculated as HT. HT is the expected heterozygosity (Eq. 3.4) with the frequencies of

27

28


alleles (pi) calculated on the basis of their occurrence in the total population. The average gene diversity within populations (HS) is calculated as mean gene diversity (He) within the subpopulations. It can be shown that the difference HT – HS equals the average minimum genetic distance of Nei (see before) between all population pairs (DST = HT – HS) (Finkeldey 1994). GST (or FST) is the relative contribution to the total gene diversity which is due to differences among populations: GST = FST = DST /HT = (HT – HS)/HT.

3.4 Genetic Variation in Tropical Forest Species – General Trends Levels and patterns of genetic variation at single loci depend on numerous factors. Apart from species-specific or population-specific variability, the type of the marker(s) investigated is of crucial importance. An assessment of genetic variation levels is meaningful only if results are related to those of comparable studies. Most available data on patterns of genetic variation of tropical forest plants were based on isozyme inventories, although the number of studies using molecular markers is constantly increasing. In view of the multitude of tropical forest plants it is hardly surprising that levels and patterns of genetic variation considerably vary among species (Table 3.1). In tropical forests, plant species with low, intermediate, and high levels of genetic diversity within populations are encountered. Mean levels of genetic variation of woody plants, mainly forest trees, from different climatic regions were compared by Hamrick et al. (1992) (Table 3.2). At isozyme gene loci, the measures of genetic multiplicity (PPL; number of alleles per locus) are on average slightly lower for woody plants from the tropical region as compared with temperate and boreal species. However, this does not hold for the average expected heterozygosity (He) as the most commonly used measure of genetic variation. Thus, there is no evidence for fundamental differences with regard to levels of genetic variation of tropical and temperate trees, even though the majority of tropical trees occur in low density, while many temperate and boreal species form large, pure stands. The impact of density on levels of genetic variation was studied in more detail by Hamrick and Murawski (1991). Genetic variation of 16 uncommon tropical forest tree species (density less than 0.5 trees per hectare) occurring on Barro Colorado Island, an artificial island in the Panama Channel, was assessed at isozyme gene loci, and compared with variation levels of 16 “common” species (density more than four trees per hectare). Variation levels within uncommon species varied from low (PPL = 14.3%, He = 0.026 for Tricanthera gigantea) to high (e.g., PPL = 76%, He = 0.257 for Myrospermum fructescens). The average variation of the uncommon species is lower than the variation of more common species (Table 3.3); thus, there is evidence for losses of genetic

3.4 Genetic Variation in Tropical Forest Species – General Trends

Table 3.1. Examples of tropical tree species showing low and high levels of genetic variation within populations, respectively Species Low variation Acacia mangium Acer skutchii Moringa oleifera Pentaclethra macroloba Pinus merkusii High variation Acacia albida (=Faidherbia albida) Calycophyllum spruceanum Gliricidia sepium Pinus caribaea Pterocarpus macrocarpus Shorea leprosula Swietenia macrophylla

Country/region

Marker type He

Reference

Australia, Papua New Guinea Mexico India, East Africa Costa Rica

Isozyme

0.017

Moran et al. (1989a)

SSR AFLP Isozyme

0.112 0.067 0.074

Lara-Gomez et al. (2005) Muluvi et al. (1999) Hall et al. (1994a)

Thailand

Isozyme

0.058

Changtragoon and Finkeldey (1995a)

West Africa

Isozyme

0.442

Joly et al. (1992)

Peru

AFLP

Russel et al. (1999)

Central America Caribbean Islands Thailand

Isozyme Isozyme Isozyme

0.249– 0.349 0.238 0.26 0.246

Chamberlain et al. (1996) Zheng and Ennos (1999) Liengsiri et al. (1995)

Malaysia Brazil

Isozyme SSR

0.406 0.781

Lee et al. (2000) Lemes et al. (2003)

He average expected heterozygosity within populations, SSR simple sequence repeat, AFLP amplified fragment length polymorphism

Table 3.2. Average genetic variation at isozyme gene loci in long-lived woody species from different climatic regions. (Adapted from Hamrick et al. 1992) Climatic zone

No. of species

Average no. of populations

PPL (%)

A/L

He

Boreal – temperate Temperate Temperate – tropical Tropical

26 122 5 38

8.2 12.2 16.6 21.3

82.5 63.5 62.2 57.9

2.58 2.27 1.89 1.87

0.206 0.166 0.169 0.191

PPL proportion of polymorphic gene loci, A/L average number of alleles per locus, He expected heterozygosity

variation due to small reproduction effective population sizes resulting in genetic drift for very uncommon tree species in tropical forests. However, the average variation of uncommon tree species on Barro Colorado Island is higher than the average variation of 486 plant species investigated at isozyme gene loci and is roughly comparable to the average variation found for 115 long-lived woody perennial species, mainly trees (Table 3.3). Most of the genetic variation within a species is usually found within populations, with only a minor fraction of the overall diversity found among

29

30


Table 3.3. Mean genetic variation at isozyme gene loci for 16 uncommon (rare; density less than 0.5 trees per hectare) and 16 common (density more than four trees per hectare) tree species on Barro Colorado Island in comparison with average variation of 486 plants and 115 long-lived woody species. (Adapted from Hamrick and Murawski 1991) Group

No. of species

Average no. of gene loci

PPL (%)

He

Uncommon species Common species All plant species Long-lived woody plants

16 16 486 115

25.8 27.2 16.5 17.0

41.8 60.9 34.2 50.0

0.142 0.211 0.113 0.149

He expected heterozygosity

populations. The mean population differentiation (GST) for 37 tropical woody plant species included in a survey by Loveless (1992) was 0.109; thus, on average, less than 11% of the total genetic variation is due to differences among populations. This result is not confined to the majority of tropical forest plants, but has also been found in plants, in particular forest trees, of other climatic regions (Austerlitz et al. 2000). Differentiation among populations at isozyme gene loci is even lower for temperate (GST = 0.092) and boreal (GST = 0.038) species as compared with 26 tropical species (GST = 0.119) included in the review by Hamrick et al. (1992). Low differentiation at single gene loci does not preclude considerable differentiation with regard to genetic components of the variation observed at phenotypic traits of adaptive or economic significance (Kremer et al. 1997; Chap. 11). In summary, the following conclusions can be drawn from the multitude of marker-based studies on genetic variation of tropical forest trees: ●

●

●

●

Forest trees are among the genetically most variable of all organisms. This result was first obtained for a few species of the temperate and boreal zone and later confirmed for tropical tree species. The genetic variation within populations at isozyme gene loci is similar for temperate and tropical tree species (Table 3.2). Several important tropical tree species are characterized by low levels of genetic variation within populations, while other species are highly diverse (Table 3.1). It is not possible to reliably predict genetic variation levels for a particular species on the basis of its life-history characteristics. Endemic species are, on average, less variable within populations than species with a wide distribution range (Loveless 1992). Uncommon species, i.e., those occurring in very low population densities in tropical forests, are, on average, less variable than species occurring in

3.5 Case Studies on Genetic Variation Patterns

●

higher density (Table 3.3). However, the expected heterozygosity is high for many tree species even if the population density is low. The degree of genetic differentiation among populations (GST) is slightly higher for tropical forest tree species than for tree species of temperate forests. Tropical tree species with abiotic means of seed dispersal (barochorous and anemochorous species) show, on average, a much higher differentiation among populations (GST = 0.138) than zoochorous species (GST = 0.050) (Sect. 5.3). Seed dispersal by animals is usually very efficient and results in small genetic differentiation among populations of zoochorous tree species (Loveless 1992).

3.5 Case Studies on Genetic Variation Patterns Since predictions of genetic variation patterns are very unreliable for a particular species, experimental investigations on the distribution of genetic variation within and among populations are mandatory for a detailed understanding of variation patterns of a target species. Six case studies were selected as examples to illustrate genetic variation patterns in tropical forest species out of a wide variety of published reports. The case studies were chosen since they represent different: ●

● ●

Marker types: isozymes, microsatellites, random amplified polymorphic DNA (RAPDs), amplified fragment length polymorphisms (AFLPs), variation of chloroplast DNA (cpDNA) Regions of the tropics (continents and climatic conditions) Species with regard to succession stage (pioneer–climax species), dispersal mechanism of seeds, abundance (rare or common), and other life-history traits

The focus is on genetic variation patterns and variation levels in forest tree populations without or with only minor human impact. Genetic effects of human disturbance of tropical forest ecosystems will be discussed in later chapters. Example 3.1: Genetic Variation at Isozyme Gene Loci in Natural Populations of Acacia auriculiformis Acacia auriculiformis A. Cunn. Ex Benth. is one of the most important Acacia species for plantation establishment worldwide. The natural distribution of the species is confined to Australia, Papua New Guinea, and a few Indonesian

31

32


islands (Fig. 3.1). A genetic inventory was conducted at 22 isozyme gene loci coding for 17 enzyme systems in 18 natural populations of the species from Papua New Guinea and Australia (Wickneswari and Norwati 1993). Clear differences with regard to the level of genetic variation measured as expected heterozygosity were observed (Fig. 3.1). Populations from Papua New Guinea revealed the highest level of genetic variation (average He = 0.133 or 13.3%), which is comparable to the average for long-lived woody plant species (Table 3.3). Intermediate levels of genetic diversity were observed in populations from Queensland, and low levels in populations from the Northern Territory of Australia. Similar results were also reported for other measures of genetic variation (PPL; alleles per locus; data not shown). Thus, Papua New Guinea is a center of genetic diversity of the species. Wickneswari and Norwati (1993) explained the observed patterns of genetic variation by the evolutionary history of the species. They suspect that the species survived less favorable climatic conditions in small, scattered refuges in Papua New Guinea, and expanded only recently via land bridges connecting Papua New Guinea and Australia until the end of the last glaciation. According to this hypothesis, only a few trees colonized Queensland and, later, the Northern Territory. Thus, the lower levels of genetic variation in Australia as compared with Papua New Guinea are explained by bottleneck effects (genetic drift) during recent recolonization (From Wickneswari and Norwati 1993).

Papua New Guinea

Natural distribution

Bensbach

Investigated populations

Northern Territory

Manton River

He 5.0 %

Reynolds River

0.7 %

Douglas River

7.5 %

He 14.3 %

Morehead R. Rouku 10.3 % North Mibini

18.0 %

Mibini Swamp Old Tonda Village

13.0 % 11.9 %

Mai Kussa River

12.0 %

mean (Papua New Guinea) 13.3 %

Gerowie Creek

3.0 %

South Alligator River

0.2 %

East Alligator River

2.7 %

Wenlock River

Goomadeer River

8.7 %

South Coen

9.1 %

Morehead River

9.1 %

Mount Molloy Kings Plain

5.5 % 9.4 %

mean (Northern Territory) 4.0 %

Queensland

mean (Queensland)

He 6.5 %

7.9 %

Fig. 3.1. Natural distribution and genetic variation of Acacia auriculiformis. He expected heterozygosity. (Adapted from Wickneswari and Norwati 1993)


Example 3.2: Genetic Differentiation Among Populations of Eugenia dysenterica from the Brazilian Cerrado Eugenia dysenterica DC. is a fruit tree species growing in the Cerrado region of Brazil. Ten populations were investigated for variation at eight isozyme gene loci (Telles et al. 2003). The species proved to be highly polymorphic with an expected heterozygosity between 0.223 and 0.377 in the populations and an average number of alleles per locus ranging from 2.4 to 3.6. However, levels of observed heterozygosity (Ho) were considerably lower than the respective values for He in all populations. Accordingly, positive fixation indices (F; inbreeding coefficients; Sect. 6.4.1) were observed at all marker loci. The overall genetic differentiation among populations was estimated as 0.154, indicating considerable genetic differences between populations. A highly significant correlation between the genetic distance (Nei 1972) and the geographical distance between populations was observed (Fig. 3.2). Two spatially isolated populations are clearly distinct from the remaining eight populations. The two genetically distinct groups should be treated as independent units for conservation (From Telles et al. 2003). Example 3.3: Genetic Variation of Swietenia macrophylla Across the Brazilian Amazon Mahogany (Swietenia macrophylla King) is one of the most important tropical timber species worldwide. Timber is harvested in plantations and logged in natural populations. Seven natural populations from the southern arc of the Brazilian Amazon basin were investigated at eight microsatellite (simple sequence repeat, SSR) gene loci. Variation was very high, with a mean number

Genetic distance (Nei)

0.25 r − 0.725 (P= 0.003)

0.20 0.15 0.10 0.05 0.00

0

50

100 150 200 Geographic distance (km)

250

Fig. 3.2. Comparison of genetic distances (Nei 1972) and geographic distances between ten populations of Eugenia dysenterica. (From Telles et al. 2003)

33

34


of 18.4 alleles per locus observed in all populations and a mean expected heterozygosity of 0.781 (Table 3.4). The inbreeding coefficient was positive and significantly different from 0 in four populations, indicating a slight excess of homozygotes relative to Hardy–Weinberg proportions (Sect. 6.4.1). The total genetic differentiation among populations was estimated as 0.097. This is only slightly lower than the average reported for tropical forest trees (GST = 0.109; Sect. 3.4) and may be regarded as a “moderate” level of genetic differentiation among populations. High levels of genetic variation were observed since microsatellites (SSRs) were chosen as marker types. SSRs are hypervariable markers showing higher levels of genetic variation in comparison with other codominant markers such as isozymes (Sect. 2.3.1). However, genetic variation in the Amazon basin was higher than in a directly comparable study on mahogany in Central America (Novick et al. 2003). On the other hand, the differentiation among populations was slightly higher among Central American populations. The significant genetic differentiation among populations in both regions stresses the importance of conserving as many populations as possible rather than of concentrating on one or a few populations in a region to be conserved in situ (Chap. 14) (From Lemes et al. 2003). Example 3.4: Genetic Variation of Endangered Australian Species of the Genus Fontainea Assessed with RAPD Markers and by DNA Sequences Fontainea oraria Jessup and Guymer is an extremely rare and endangered, dioecious, small tree species known to exist on a single site, a littoral rain forest in New South Wales (Australia) (Rossetto et al. 2000). Only ten adult plants and 52 seedlings grow on a 600-m-long strip. Sequencing of a cpDNA fragment and of the nuclear ITS2 region (Sect. 8.3) revealed a close relation to Table 3.4. Genetic variation at eight microsatellite gene loci in seven populations of Swietenia macrophylla from the Brazilian Amazon. (Adapted from Lemes et al. 2003) Population

A/L

He

Ho

F

A. Azul Cach. A Maraj P. Lacerda C. Mendes Cach. E P. Bueno All populations

8.4 10.7 9.2 9.7 10.6 10.2 7.6 9.5

0.761 0.785 0.793 0.812 0.754 0.810 0.754 0.781

0.753 0.781 0.740 0.812 0.709 0.780 0.680 0.750

0.012 0.005 0.068 −0.004 0.060 0.042 0.100 0.038

Deviationa

**

*** * ** ***

He expected heterozygosity, Ho observed heterozygosity, F inbreeding coefficient a Significant deviation from Hardy–Weinberg proportions at *P < 0.05, **P < 0.01, ***P < 0.001


other species of the genus Fontainea and in particular to F. australis, another endangered species of the genus restricted in distribution to four sites. It was possible to distinguish all F. oraria plants (adults and progenies) by an investigation based on seven RAPD primers which amplified a total of 57 RAPD bands. The level of genetic variation at RAPD markers was similar for F. oraria and F. australis; however, variation was lower in the progenies of F. oraria as compared with the parental population. An uneven contribution of adult plants to the production of the seedlings (uneven fitness) is likely. There is evidence that a single tree dominates the production of progenies while other adults do not participate in reproduction. The example shows that even an extremely low population size does not preclude the observation of genetic variation in a species (From Rossetto et al. 2000). Example 3.5: Genetic Variation of Shorea leprosula and Shorea parvifolia in Indonesia Assessed at AFLP Loci Shorea leprosula Miq. and Shorea parvifolia Dyer are common and widespread emergents of lowland dipterocarp forest in Indonesia and other Southeast Asian countries. They are very important timber trees and are the main sources of light red meranti wood. Genetic variation was assessed by AFLP markers in 12 natural populations (six for each species) from Sumatra and Kalimantan (Borneo), and for S. leprosula in addition in one plantation from Java (Fig. 3.3) (Cao et al. 2006). Levels of genetic variation (He; Table 3.5) were estimated based on the assumption of Hardy–Weinberg structures in both populations. Variation is higher for S. leprosula in comparison with S. parvifolia (Table 3.5). Low variation was observed in both S. parvifolia populations sampled in Borneo. The population Asialog from Sumatra clearly showed the highest variation for both species. Variation levels measured as He decreased in the same order for the five locations where both species were sampled (Asialog > Pasir Mayang > Bukit Tiga Puluh National Park > Sari Bumi Kusuma > Nanjak Makmur). Considerable differentiation was not only observed among species, but also among populations. The proportion of the total genetic variation within species due to differentiation among populations is high for S. leprosula (GST = 0.20) and even higher for S. parvifolia (GST = 0.31). Populations of S. leprosula from Borneo are clearly separated from populations of the same species from Sumatra (Fig. 3.4). The cluster diagram (Fig. 3.4) suggests the origin of the planted population (Haurbentes, Java) as Sumatra. Conservation of genetic resources of dipterocarps needs to take into account the strong genetic differentiation among populations at presumably neutral AFLP loci. The presence of numerous “diagnostic” characters occurring in high frequency in only one or a few populations suggests that conservation in as

35

36


Sumatra

Kalimantan (Borneo)

SB

(S. lepr. S. parv.)

BB

(S. parv.)

PS

TS

(S. lepr. S. parv.)

(S. lepr. S. parv.)

NS

(S. lepr. S. parv.)

AS

TB

(S. lepr. S. parv.)

(S. lepr.)

HA

(S. lepr.)

Java Fig. 3.3. Location of sampled populations of Shorea leprosula and S. parvifolia in Indonesia. Abbreviations as in Table 3.5. (From Cao et al. 2006) Table 3.5. Genetic variation at AFLP loci in two dipterocarp species (Shorea leprosula and Shorea parvifolia) in Indonesia. (Adapted from Cao et al. 2006) Species

Population

Abbreviation

Location

PPL (%)

He

S. leprosula

Haurbentes Tering Sari Bumi Kusuma Asialog Pasir Mayang Bukit Tiga Puluh National Park Nanjak Makmur Mean Total Batu Ampar Sari Bumi Kusuma Asialog Pasir Mayang Bukit Tiga Puluh National Park Nanjak Makmur Mean Total

HA TB SB AS PS TS

Java Borneo Borneo Sumatra Sumatra Sumatra

60.71 64.29 60.71 82.14 62.50 60.71

0.186 0.146 0.174 0.258 0.178 0.177

NS

Sumatra

BB SB AS PS TS

Borneo Borneo Sumatra Sumatra Sumatra

50.00 61.93 92.86 33.93 46.43 73.21 60.71 46.43

0.151 0.177 0.222 0.115 0.124 0.221 0.151 0.127

NS

Sumatra

50.00 54.57 82.14

0.119 0.150 0.206

S. parvifolia

He expected heterozygosity

3.5 Case Studies on Genetic Variation Patterns S. lepr._AS S. lepr._HA S. lepr._NS S. lepr._PS S. lepr._TS S. lepr._TB S. lepr._SB S. parv._SB S. parv._BB S. parv._TS S. parv._AS S. parv._PS S. parv._NS

Fig. 3.4. Unweighted pair-group method with arithmetic means tree based on Nei’s standard genetic distance illustrating genetic differentiation among populations of two dipterocarps from Indonesia. Abbreviations as in Table 3.5. (From Cao et al. 2006)

many populations as possible is important. The identification of centers of genetic diversity of dipterocarps seems feasible since ordering of populations with regard to their variation (He) revealed the same trend for both species, with a single location (Asialog) clearly showing the highest diversity for S. leprosula and S. parvifolia (From Cao et al. 2006). Example 3.6: Genetic Variation of Cedrela odorata at cpDNA Cedrela odorata is an important timber species of the family Meliaceae with a wide natural distribution from Mexico to northern Argentina. A study of variation of cpDNA in Central American populations revealed five different haplotypes (Cavers et al. 2003). Populations are extremely differentiated with only three from 29 populations investigated showing polymorphism (Table 3.6); 96% of the total variation is distributed among populations (GST = 0.96). Maternal inheritance of cpDNA frequently results in strong differentiation of cpDNA haplotypes among populations. Repeated migration from South America to Central America is a likely cause of the strong differentiation. The ancestors of trees with the “northern” haplotypes 1 and 2 found in Guatemala and Mexico might have migrated to Central America early, possibly even before the land bridge connecting South America and North America was formed. The differentiation between populations with the “central” haplotype 3 and the “southern” haplotypes 4 and 5 coincides with considerable differences of climatic conditions and morphological and physiological traits. Likewise, investigations of anonymous RAPD markers revealed strong differentiation among these ecotypes. Thus, repeated colonization of Central America during eras with different climatic conditions is a likely cause for the observed differentiation.

37

38


Table 3.6. Variation of chloroplast DNA haplotypes in Central American populations of Cedrela odorata. (Adapted from Cavers et al. 2003) Population

Haplotype 1

Panama Gualaca Las Lajas San Francisco Costa Rica Horizontes Puriscal Canas Palo Verde Hojancha Jimenez Talamanca Upala Pacifico Sur Perez Zeledon Nicaragua Ometepe Masatepe Wabule La Trinidad Honduras Meambar Taulabe Comayagua La Paz Guatemala Los Esclavos El Idolo Tikal San Jose Mexico Calakmul Bacalar Zona maya Escarcega Total

2

3

4

5 5 5 1

4

20 20 20 20 20 20 20 19 20 20

1

20 20 20 20 20 20 20 20 4 5 5 5 20 20 20 20 99

1

1

259

106

9

The strong differentiation among populations from different regions offers good opportunities to apply molecular markers to control the origin of reproductive material (seeds or seedlings) (From Cavers et al. 2003).

3.6 Recommended Literature

3.6 Recommended Literature A comprehensive discussion of methods to analyze genetic variation within and among populations is given by Nei (1987). Short introductions to the analysis of genetic variation based on isozymes and several nuclear DNA markers, respectively, are provided by Berg and Hamrick (1997) and Nybom (2004). Weising et al. (2005, p. 207 ff) describe the analysis of genetic variation at marker loci. Older literature on isozyme variation of tropical forest trees has been reviewed by Loveless (1992). Useful reviews on genetic variation within and among plant populations at isozyme loci were also presented by Hamrick and Godt (Hamrick and Godt 1989, 1996, 1997; Hamrick et al. 1992).

39

Sexual and Asexual Reproduction in Tropical Forests

Flowering plants have evolved one of the most complex and “sexiest” life cycles on earth (Armstrong 1996).

4.1 Introduction Reproduction is an obvious prerequisite for the survival of species and populations. Reproduction in tropical forests has attracted much interest from ecologists, botanists, and conservation biologists since an understanding of the processes taking place during reproduction is fundamental for an analysis of the evolution and maintenance of the high (species) diversity in these forests (Sect. 7.2). A population is defined as a unit of reproduction (Sect. 2.2), and the recombination of genes during (sexual) reproduction implies that changes of genetic structures are particularly likely during the reproduction phase. Thus, an understanding of genetic processes during reproduction is crucial for the analysis of the temporal dynamics of genetic structures within populations, in particular for long-living organisms such as forest trees. The basic genetic principles connected to sexual and asexual reproduction in tropical forests are reviewed in this chapter. The distribution of genetic information during reproduction is discussed in more detail in Chap. 5, and the new combination of genes to genotypes by the mating system is the topic of Chap. 6. Chapters 4–6 are mainly concerned with the reproduction process in natural forests. The implications of human impact on the reproduction system of tropical forests plants and the production of forest reproductive material for enrichment planting and plantation establishment are discussed in Part B of this book.

4.2 Sexual Reproduction Sexual reproduction is characterized by the fusion of two gametes of opposing sex to a zygote. The formation of gametes is preceded by meiosis, i.e., a (sexual) cell division resulting in gametes with only half of the chromosome number as

4

42

C HAPTER 4 Sexual and Asexual Reproduction in Tropical Forests

compared with the chromosome number in somatic cells of the parental plant. The consequences of this process for genetic structures of a progeny generation derived from a parental stand (t) and the possible causes of deviations of the genetic structures in the progeny generation (t +1) from the population of adults are illustrated in Fig. 4.1. Trees of generation t produce female and male gametes. Differences in the ability to produce gametes result in fertility differences among trees (fertility selection). Male gametes (pollen) are distributed during the flowering period and external pollen from other populations is carried in the population (pollen influx). Mating preferences or incompatibilities possibly result in differences between allelic structures of the “pollen cloud” reaching a stigma and the fertilizing or effective pollen, for example, owing to differences in pollen tube growth. Viability selection starts early after fertilization and results in differences of genetic structures between fertilized ovules and fully developed seeds. Viability selection continues during seed germination, seedling growth, and throughout stand development. For practical reasons, it is not easily possible to observe the genetic structures of female and male gametes (pollen and ovules) or zygotes. Thus, changes of genetic structures are often monitored by comparing the adult population (generation t) with the mature seeds (generation t +1; comparison 1 in Fig. 4.1), by comparing the seeds with the population of seedlings, saplings, or adults (comparison 2 in Fig. 4.1; Example 7.1), or by comparing two generations of the same or, in case of long-living organisms like forest trees more frequently, of different age (comparison 3 in Fig. 4.1).

4.2.1 Sexual Types and Sexual Systems

Only gametes of opposite sex can fuse and jointly produce the zygote which will further develop by mitotic cell divisions into an embryo, a seedling, and eventually into an adult plant capable of producing gametes itself. The production of these two different types of gametes (female and male) is a requirement for successful sexual reproduction in a population. The sexual type of an organism depends on the types of gametes it produces: It is either cosexual (female and male gametes are produced) or unisexual female or male. Plants not capable of producing any functioning gametes are sterile. The determination of the sexual type of an organism is usually under strict genetic control although an environmental influence on the sexual type has been observed for some animals. The sexual type of most animals and some plants is determined by particular chromosomes (gonosomes, i.e., sexual chromosomes). Females often have two identical gonosomes: they are homogametic. Males of such species have two different gonosomes: they are heterogametic. An example of sex determination by gonosomes are the X and

4.2 Sexual Reproduction

production of and gametes

gametes

fertility selection

distribution of flowering over time pollen transport pollen influx

trees of generation t pollination

mating preferences incompatibility

comparison 1 fertilization

comparison 3

viability selection (self-sterility etc.)

mature seeds of generation t + 1

seed production

viability selection (germination, mortality) comparison 2

stand

n development

ele

ys

lit bi

io ct

via

trees of generation t + 1

Fig. 4.1. Causes for changes of genetic structures during and after reproduction in consecutive generations. (From Hattemer 1999)

Y chromosomes of humans. Females have two X chromosomes; males one X and a smaller Y chromosome. The sex of some insects mostly belonging to the Hymenoptera (e.g., bees and wasps) is determined by the ploidy level. Unfertilized, haploid eggs develop into males. Diploid, fertilized eggs develop into females (Cook 1993). Flowering plants are principally cosexual; however, the separation of sexes (dioecy) repeatedly evolved in plants (Sect. 4.2.2). Heteromorphic sex

43

44


chromosomes have been identified for a few dioecious plants, for example, in the herbaceous genera Silene and Rumex (Dellaporta and Calderon-Urrea 1993). A particular region on the LG1 chromosome has been identified as carrying sex-determining genes in papaya (Carica papaya) (Charlesworth 2004). Sex expression is regulated by hormones in many dioecious plants (Dellaporta and Calderon-Urrea 1993), but the mechanism of sex determination is still unknown for most woody plant species and for virtually all tropical forest species. A possible determination of the sex by age (or size) of trees has been described for the neotropical tree Iryanthera macrophylla. The plants of this species produce only one type of flower (female or male) in a particular reproduction period; however, Ackerly et al. (1990) found evidence that small, young trees of this species flower male, but later produce female flowers. Thus, the sex expression may not be determined by genetic factors, but by the age or size of plants, and the species may not be dioecious, but “diphasic” (Ackerly et al. 1990). 4.2.2 Sexual Systems

Since the formation of a progeny generation depends on the production of gametes of either sex, only the following combinations of the sexual types of individual plants to the sexual system of a species are possible: ● ●

●

●

●

Cosexuality: all plants produce both types of gamete (female and male) Dioecy: plants produce only one type of gamete – they are female (female) or male Trioecy: some plants produce both types of gamete (female and male), others are female or male Gynodioecy: some plants produce both types of gamete (female and male), others are female (no males) Androdioecy: some plants produce both types of gamete (female and male), others are male (no females)

Cosexuality is the archetypal and most common sexual system for plants. Two forms of cosexuality are distinguished: Hermaphroditic plants have complete flowers producing both types of gametes in a single flower. Monoecious plants produce both types of gamete in different flowers. Thus, single flowers are either male or female in monoecious species, but these flowers are produced by the same plant (Table 4.1). Different types of pollination can be distinguished for hermaphroditic species depending on the origin of the fertilizing pollen (Fig. 4.2). Autogamy is the


Table 4.1. Sexual systems of tropical plants with examples Sexual system

Sexual types

Examples

Cosexuality: hermaphroditism Cosexuality: monoecy

( ) in single flowers ( ) in different flowers on single plants ( ), ( )

Majority of tropical tree species, e.g., Dipterocarpaceae, Tectona, Eucalyptus Pinus, Castanopsis, and other Fagaceae

Dioecy

( ), ( ), ( ) ( ), ( ) ( ), ( )

Trioecy Gynodioecy Androdioecy

Calamus spp., Diospyros spp., most Podocarpaceae Pachycereus pringlei, Fraxinus excelsior a Rhizophora mangle, Ocotea tenera, Ficus spp.b Castilla elastica, Bauhinia pauletia, Bauhinia ungulata

a

Temperate tree species Gynodioecious figs are functionally dioecious (Example 7.2)

b

Allogamy Xenogamy Geitonogamy

Autogamy

Single flower

Selfing

Outcrossing

Fig. 4.2. The origin of fertilizing pollen for a flower for a hermaphroditic tree species. (Adapted from Hattemer et al. 1993)

fertilization of an ovule by pollen from the same flower. It is just like geitonogamy, the fertilization with pollen from a different flower of the same tree, a form of allogamy resulting in selfing (self-pollination; Sect. 6.3). Xenogamy is pollination from a different plant and results in outcrossing. Autogamy is not possible for monoecious species, and no form of allogamy, neither autogamy nor geitonogamy, is possible for dioecious species, which are, in consequence, obligatory outcrossed.

45

46


Hermaphroditism is the most common sexual system in tropical forests, followed by dioecy. Monoecious species are widespread in temperate and boreal forests, but are rarely encountered in tropical forests (Table 4.2). The domination of monoecious species in temperate forests reported for North America in Table 4.2 is also observed in central Europe: the dominating species of the families Fagaceae (Fagus, Quercus) and Pinaceae (Pinus, Picea, Abies) are monoecious. All other sexual systems (trioecy, gynodioecy, androdioecy) are extremely rare, both in tropical and in temperate forests. The sexual system of figs (Ficus spp.), which are either monoecious or gynodioecious, is explained in detail in Example 7.2. Bauhinia ungulata is pollinated by bats. It is morphologically a hermaphrodite, but seems to be functionally androdioecious. Trees produce flowers with either short or long pistils, but seeds do not develop on trees with short pistils (Ramirez et al. 1984). The floristic composition of a forest has an influence on the frequency of sexual systems since particular sexual systems are common in certain families. For example, many species of the Salicaceae [willows (Salix spp.) and poplars (Populus spp.)] are dioecious. The proportion of this family has a strong impact on the frequency of dioecy in a community of woody plants in a temperate climate. Tropical tree families with many dioecious species are, for example, the Meliaceae and the Ebenaceae. The high proportion of dioecious species in a South Nigerian rain forest (Table 4.2) is explained by the high frequency of these two families in the area investigated. The high frequency of dioecious species in species-rich forests of the central tropics is surprising in view of the low densities of many populations, since dioecious species are obligatory outcrossing. Thus, pollen vectors need to transport pollen from male to female plants for the production of sexual progenies. The advantages of outcrossing as compared with selfing (Sect. 6.4.2) are Table 4.2. Frequency (%) of common sexual systems in tropical and temperate forests. (Adapted from Bawa and Opler 1975) Location Tropics Costa Rica South Nigeria Sarawak Florida Temperate zone Ohio Ohio Arkansas Kentucky New Hampshire

Cosexual: hermaphroditic

Cosexual: monoecious

Dioecious

68 47 60 61

10 13 14 12

22 40 26 27

6 27 0 15 13

83 60 83 70 81

11 13 17 15 6


assumed to be the main reason for the evolution of dioecy for many originally self-compatible species of tropical forests (Bawa and Opler 1975). The frequency of sexual types in a forest is related to main pollinators (Sect. 5.2.2). Monoecious species are mainly wind-pollinated. The complete flowers of hermaphroditic species are predominantly pollinated by animals. 4.2.3 Sexual Structures and Sexual Function

The sexual structure of a population is the frequency distribution of its different sexual types. In the case of a cosexual (hermaphroditic or monoecious) species, it simply holds that P = 1, where P is the relative frequency of cosexual type from all f lowering plants. For a dioecious species, an even frequency of sexual types (P = P = 0.5) is often assumed, where P (P ) is the relative frequency of female (male) plants. The frequencies of male and female flowering plants were calculated for various dioecious species in Costa Rica (Table 4.3) (Opler and Bawa 1978). The 23 species investigated belong to 13 different families. This supports the hypothesis that dioecy evolved independently during the evolution of angiosperms in Table 4.3. Sexual structures of tree species in a tropical forest in Costa Rica. Test for significant deviation from an even sex ratio (male-to-female ratio of 1; com. X2) and test for significant heterogeneity of the sex ratio among populations (int. X2). (Adapted from Opler and Bawa 1978) Family

Species

No. of Population

No. of males

Anacardiaceae Boranginaceae

Spondias nigrescens Cordia collococca Cordia panamensis Bursera simarouba Bursera tomentosa Erythroxylon rotundifolium Bernardia nicaraguensis Trichilia anisopleura Trichilia cuneata Cecropia peltata Chlorophora tinctoria Pisonia macranthocarpa

5 6 6 9 5 1

57 127 110 114 74 30

58 85 52 95 60 14

0.98 1.50 2.11 1.20 1.23 2.14

2

23

22

1.04

NS

–

5 5 9 7 5

58 115 104 79 61

42 93 100 64 45

1.38 1.24 1.04 1.23 1.36

NS NS NS NS NS

NS NS NS NS NS

Burseraceae Erythroxylaceae Euphorbiaceae Meliaceae Moraceae Nyctaginaceae

No. of Male-to- Com. Int. X2 females female X2 ratio NS

NS

**

**

**

NS NS NS –

NS NS **

(Continued)

47

48


Table 4.3. Sexual structures of tree species in a tropical forest in Costa Rica. Test for significant deviation from an even sex ratio (male-to-female ratio of 1; com. X2) and test for significant heterogeneity of the sex ratio among populations (int. X2). (Adapted from Opler and Bawa 1978)––Cont’d Family

Species

No. of Population

No. of males

No. of Male-to- Com. Int. X2 females female X2 ratio

Polygonaceae

Coccoloba caracasana Coccoloba floribunda Ruprechtia costata Triplaris americana Allophyllus occidentalis Simarouba glauca Albertia edulis Genipa caruto Randia spinosa Randia subcordata Zanthoxylum setulosum

6 1 1 8 4

55 23 21 157 74

136 27 30 220 50

0.41 0.85 0.70 0.71 1.48

4 4 6 9 7 4

111 83 76 154 157 115

102 49 77 57 77 60

1.08 1.69 0.99 2.70 2.04 1.92

Sapindaceae Simaroubaceae Rubiaceae

Zanthoxylaceae

**

**

NS NS

– – NS

** *

**

NS NS

NS NS NS

**

**

**

NS NS

**

**

NS not significant * P < 0.05; **P < 0.01

various taxa. A significant deviation from an expected even ratio of male and female plants was observed for ten out of the 23 species (com. X2 in Table 4.3). Significantly more males than females (male-to-female ratio greater than 1 and significant com. X2) were observed for eight species. A significant excess of females (male-to-female ratio less than 1 and significant com. X2) was observed for only two species. Significant heterogeneity of the sex ratio among populations (int. X2) was observed for four species with P < 0.01. Deviations from a balanced sex ratio were also observed in populations of functionally dioecious figs (Ficus spp.) and rattan (Calamus spp.). In most cases an excess of male plants has been reported (Table 4.3). Possible reasons for an unbalanced sex ratio are: ●

●

●

Mode of inheritance of sex. Sex determination is unknown for most tropical forest plants (Sect. 4.2.1). Gametic selection favoring one sex. Pollen carrying the information for a male (or female) progeny is significantly more successful in fertilization that pollen carrying the information for the opposite sex of the progeny. Viability selection. Female (or male) plants show significantly lower survival as compared with survival rates of the opposite sex. Different resource allocation with regard to vegetative growth and investment in reproduction is

4.3 Asexual Reproduction

●

●

likely for many dioecious plants. Female plants need to supply energy and nutrients for the production of female flowers and seeds which are often embedded in large fruits. These resources are not available for vegetative growth. This possibly results in reduced viability as compared with that of male plants producing “only” pollen in male flowers. Earlier maturity of one sex, usually males, as compared with the opposite sex, or higher frequency of flowering of one sex. For example, annual flowering for males but a biannual cycle for females results in an excess of males in each flowering period. Preferential asexual reproduction of one sex. Strict genetic control of the sex results in identical sex expression of all plants from vegetative reproduction of a particular (female or male) tree.

The sexual function of a plant is the proportion of successful female gametes from all successful gametes (female and male) produced by a plant. A gamete is regarded as “successful” or effective if it has contributed its genetic information to a zygote. There are two sexual functions for dioecious species. Female plants have a sexual function of 1, since they only produce female gametes. The sexual function of males is 0, since they never produce female gametes. If the female and male gametes of a cosexual plant are successful in even proportions, its sexual function is 0.5. Observation of flowering intensities in populations strongly suggests deviations from a sexual function of 0.5 for many monoecious species. Some trees flower predominantly or, at least in a particular flowering season, exclusively male, while others flower predominantly female. Most of these investigations were conducted for monoecious temperate forest trees such as Pinus sylvestris, Fagus sylvatica, or the trioecious Fraxinus excelsior (p. 161 ff. in Hattemer et al. 1993). Clones in a Pinus sylvestris seed orchard strongly differed with regard to their male and female reproductive success (MüllerStarck and Ziehe 1984). Hermaphroditism precludes the estimation of sexual functions based on an observation of flowering, since male and female flowering cannot be easily distinguished. Thus, the variation of sexual functions in populations of tropical trees is largely unknown owing to the predominance of hermaphroditism as a sexual system in the tropical forests.

4.3 Asexual Reproduction Many plant species and some animal species have the potential for asexual reproduction. In this case, offspring is produced from tissue of a single organism by mitotic cell divisions only; meiosis and recombination do not occur.

49

50


Thus, the genotypes of asexually produced offspring are identical copies of the genotype of the original plant. Obviously, asexually produced progenies of a single plant are also genetically identical among each other. A collective of genetically identical organisms is called a clone.

4.3.1 Vegetative Reproduction

Vegetative reproduction is not only an important tool for the production of forest reproductive material such as cuttings (Hartmann and Kester 1983), but is also frequently encountered in natural and secondary forests of the tropics. The importance of vegetative propagation for tree improvement, plantation forestry, and conservation of forest genetic resources will be discussed in later chapters. The following remarks mainly refer to vegetative reproduction under natural conditions. Not all plants which are commonly vegetatively propagated by humans are capable of vegetative reproduction in natural forests. Root suckers, i.e., the development of sprouts from roots, are the most common form of vegetative reproduction in natural forests. Coppicing may be regarded as a form of vegetative growth or of reproduction; thus, coppice shoots exemplify the transition from vegetative growth to vegetative multiplication. Many tropical forest plants use vegetative reproduction as a complementary method in addition to sexual reproduction via seeds. Root suckers are an important means of reproduction for Cordia alliodora. They are particularly common in dry and semidry forests of the tropics. For example, Prosopis juliflora and other legumes in dry forests frequently reproduce by root suckers. Bamboos are another important species group which predominantly reproduces vegetatively. Flowering of bamboos is a rare event, resulting in the death of a flowering plant after seed production for many species (Janzen 1976). Bambusa vulgaris is a commonly cultivated species throughout the tropics. It very rarely flowers and the production of viable seeds has never been observed for this species (Banik 1995); thus, B. vulgaris is one of the very few taxa which exclusively multiplies vegetatively. Vegetative reproduction is not only common for tropical trees, but also for other plants growing in forests of the lower latitudes. For example, vegetative reproduction was expected in clusters of the neotropical epiphytic orchid Laelia rubescens Lindley. Clonal reproduction within clusters was confirmed by marker-based inventories, but the degree of sexual reproduction and, hence, genetic diversity within clusters was surprisingly high (Trapnell et al. 2004).

4.3 Asexual Reproduction

4.3.2 Apomixis

A sexual embryo develops in the seed of flowering plants after fertilization of the ovule and fusion of the genetic information of two gametes (female and male) to a zygote. The development of an embryo without the processes of fertilization and fusion of gametes is called agamospermy or, more commonly, apomixis in a narrow sense. Typically, the embryo develops from a somatic cell of the seed parent (apospory). In this case, it is genetically identical to the seed parent. All seeds containing an embryo from apopory from the same seed parent are clones. The development of an embryo from a macrospore mother cell after a meiotic cell division is called diplospory. The resulting embryos from the same plant are not necessarily genetically identical; hence, diplospory is not a form of asexual reproduction in a strict sense. Apospory is a method to propagate and distribute genetically identical organisms (clones) through seeds. This is particularly advantageous if pollen transport from other trees (xenogamy) is either absent or very limited, for example, owing to a large spatial distance between trees growing in low density. Seeds developing in the absence of “foreign” pollen after selfing usually show strong inbreeding depression (Sect. 6.4.2). Inbreeding is avoided and individual variation (heterozygosity) maintained if apomictic seeds are produced. The potential for apomixis is particularly advantageous for a plant if it is combined with the ability to produce sexual progenies if foreign pollen is available (facultative apomixis). For example, the walnut (Juglans regia) produces sexual offspring if pollen is available, but is capable of apomixis in the case of a lack of pollination (p. 47 in Stern and Roche 1974). Trees are usually sterile if meiotic cell divisions are disturbed owing to an unbalanced chromosome number (e.g., triploidy) since no functioning gametes are produced; however, such trees may produce viable seeds by means of apomixis. Thus, many apomictic species are polyploid; the somatic cells of those plants contain more than two full sets of chromosomes. For example, triploid Casuarina nana plants reproduce through apomixis (Barlow 1958). The West African species Pachira oleaginea (Bombacaceae) is a hermaphrodite pollinated by bats. It is a polyploid species with 72 chromosomes in somatic cells. Seeds are produced only after pollination and fertilization, but self-fertilization is possible; thus, Pachira oleaginea seems to be self-fertile although other species of the genus Pachira have an incompatibility system (Sect. 6.5.1). A sexual proembryo develops in the ovule after fertilization, but up to 15 adventive embryos start to develop almost simultaneously in the same ovule from tissue of the nucellus. The mature seeds contain only two to five

51

52


embryos. Only one or two of these develop into seedlings. In most cases, the sexual embryo dies during seed development since adventive embryos grow faster. Thus, seed production requires regular fertilization, but embryos of most mature seeds have an apomictic origin (Baker 1960). This form of seed development (“pseudogamy”) requiring pollination but resulting in apomictic embryos has also been described for fruit species of the genus Citrus. At least facultative apomixis has been described for several dipterocarps, including Stemonoporus oblongifolius (Murawski and Bawa 1994), Hopea odorata (Wickneswari et al. 1995), and the autotetraploid Shorea ovalis ssp. sericea (Ng et al. 2004). Apomixis is suggested to be an evolutionary strategy allowing the maintenance of high genetic diversity in populations of Shorea ovalis ssp. sericea (Ng et al. 2004). Apomixis in H. odorata is possibly related to the occurrence of triploid trees in this species (Jong 1976). Kaur et al. (1978) suggested that apomixis may be common not only for dipterocarps, but also for other tropical forest plants. The observation of abundant seed crops of spatially isolated trees of dioecious or self-incompatible species suggests apomixis as a likely explanation. However, marker-based studies rarely confirmed apomixis in tropical forests. In many cases, long-distance transport of effective pollen seems to be responsible for the production of outcrossed seeds even for trees’ “isolated” by distances of several hundred meters from other flowering trees of the same species serving as potential pollen donors (Sect. 5.2.3).

4.4 Recommended Literature Early studies of Bawa and Opler (Bawa 1974, 1980; Bawa and Opler 1975; Opler and Bawa 1978) on the sexual system of neotropical trees revealed the high frequency of dioecy and self-incompatibility. Kaur et al. (1978) suspected that apomixis may be widespread among tropical forest plants. This idea stipulated in-depth studies which, however, only rarely supported their hypothesis.

Gene Flow and Migration

The level of gene exchange between populations is determined by a multitude of factors, including the size, density, and shape of the donor and recipient populations, plant height and breeding system, characters of the surrounding vegetation, terminal velocity of pollen and seeds, pollen and seed prodction, and the foraging behavior of pollen and seed vectors, as well as the distance between populations. To further confound matters, many of these factors vary in both time and space (Levin and Kerster 1974).

5.1 Introduction Genetic information is transported within and among populations of tropical forest species. The main focus of the following discussion is on plants, but animals are also considered since they are important vectors for the dispersal of pollen and seeds, and since gene flow is crucial for the spatial distribution of genetic information both in plants and in animals. Higher plants and their female gametophytes, which are part of the gynoecium, are immobile; thus, the principal methods for the transportation of genetic information in plants are the movement of the male gamete (pollen), and of seeds. For some species, vegetative reproduction results in shortdistance (e.g., root suckers) or long-distance (e.g., bulbils) transport of genetic information. Seedlings are dispersed in rare cases such as in viviparous mangrove species. The terms “gene flow” and “migration” are in common use for the transportation of genetic information in plants. In a broad sense, both terms refer to any transportation of genetic information, either by pollen or by seed movement. We will use the term gene flow in its narrow sense only for the movement of pollen, and will describe the transportation of seeds as migration.

5

54

C HAPTER 5 Gene Flow and Migration

5.2 Gene Flow Through Pollen 5.2.1 Pollination and Fertilization

Sexual reproduction is based on the fusion of a male and a female gamete. The transportation of the male gametophyte to the female plant is required in seed plants. This requires a series of events in angiosperms: First, pollen is transported to the stigma of the same (autogamy) or a different (geitonogamy or xenogamy; Fig. 4.2) flower. The pollination is followed by the growth of a pollen tube through the style to the ovule. Then, after entering the ovule through the micropyle, fertilization of the embryo sac takes place (p. 32 ff. in Proctor et al. 1996). The genotypic structures of the progeny generation are influenced both by the dispersal of pollen and by genetic differences of the pollen with regard to germination, pollen tube growth, and fertilization success. Pollen dispersal from anthers to stigmas or, in the case of gymnosperms, to the micropyle is discussed in this chapter. Aspects related to the processes of the germination of pollen, pollen tube growth, and fertilization are covered in Sect. 6.5.1. 5.2.2 Pollen Vectors in Tropical Forests

A detailed description of pollination and pollen transport by biotic and abiotic vectors was given by Proctor et al. (1996). Prance (1985) reviewed the pollination and pollinators of Amazonian plants. Only the most important considerations related to pollen transportation in tropical forests are briefly reviewed here. Pollen is dispersed by abiotic or biotic means. Wind is the principal abiotic factor facilitating pollen dispersal; different animals are pollen dispersal agents of most tropical forest plants. 5.2.2.1 Pollination by Wind (Anemogamy)

Wind is the most important pollen vector of trees in the temperate and boreal zone. For example, the principal stand-forming species in central Europe belonging to the families Fagaceae (Fagus sylvatica, Quercus spp.) and Pinaceae (Pinus sylvestris, Picea abies) are wind-pollinated (anemogamous). Wind pollination is rare in the humid central tropics (Regal 1982); only 1–2% of plants in the central

5.2 Gene Flow Through Pollen

tropics disperse their pollen by the wind (Bawa 1990). However, this proportion increases in semidry and dry tropical forests. Tropical pines (Pinus spp.) and many other gymnosperms such as Podocarpus species are wind-pollinated (Regal 1982). Pioneers and species commonly found in secondary forests such as some Cecropia spp. are also anemogamous. Wind has been suggested as a pollen vector for some eucalypts, but these statements are not supported by experimental evidence (p. 190 in Eldridge et al. 1994). The success of pollen parents in pollinating a particular seed tree depends on numerous factors, including the spatial structure and the distance between trees, environmental factors, including the preferential wind direction, velocity, and competing vegetation acting as a “pollen filter,” species-specific criteria such as pollen morphology and pollen size, and individual characteristics, including the number of pollen produced during a flowering period. For example, there is evidence for limited pollen dispersal and preferential matings among near neighbors for the wind-pollinated Araucaria angustifolia (Sousa and Hattemer 2003; Sect. 11.4.3). A large number of studies conducted on wind-pollinated tree species in temperate forests revealed both preferential pollen transport from directly neighboring trees (e.g., for Pinus sylvestris; Müller 1977) and a large proportion of successful long-distance pollination (e.g., for Quercus macrocarpa; Dow and Ashley 1998). Observations in clonal seed orchards (Sect. 12.8.1) of wind-pollinated conifers in temperate and boreal climates proved a high frequency of contamination by external pollen (Wheeler and Jech 1992); thus, spatial isolation of clonal seed orchards from conspecific forest stands does frequently not result in reproductive isolation owing to efficient long-distance dispersal of pollen by the wind. A high frequency of matings after long-distance pollen dispersal has been suggested by Lindgren et al. (1995) for Scots pine (Pinus sylvestris). The authors assume that viable pine pollen can be transported over hundreds of kilometers before becoming effective in the boreal climate of Scandinavia. The high species diversity and, hence, the low population density is likely to be the main reason for the rarity of wind-pollination in evergreen moist tropical forests, since “wind-pollination is usually not as effective as animal pollination where inter-individual distances are large” (Regal 1982). 5.2.2.2 Pollination by Animals (Zoogamy)

Zoogamy is dominant in tropical forests, and numerous animals are involved in pollen dispersal. This is in sharp contrast to temperate and boreal forests, where fewer species are pollinated by animals, and where only certain insects play a significant role in pollination.

55

56


Vertebrates

Only small flying vertebrates (birds and bats) are important pollinators for tropical plants. Nonflying mammals such as opossums and monkeys have been reported to act as pollinators for a few plant species with robust flowers (Janson et al. 1981), but most flowers are severely damaged by visiting nonflying mammals. Bats

Many Bombacaceae such as the baobab (Adansonia digitata) and Ceiba pentranda (Lobo et al. 2005), but also some Parkia spp. (Mimosaceae) and Bauhinia spp. (Caesalpinaceae; Heithaus et al. 1974) are predominantly pollinated by bats. Flowers visited by bats are usually very big and white or light yellow. They open at dawn or during the night, smell strong, and produce large quantities of nectar (p. 36 in Whitmore 1975). Inflorescences either stand upright over the crown or hang from long twigs. Sometimes they are produced by modified twigs originating directly from the stem (cauliflory, e.g., the “cannonball tree” Couroupita guianensis; Fig. 5.1). The production of flowers suitable for bat pollination is energy-demanding for a tree. On the other hand, most bats fly long distances. Accordingly, they are efficient vectors for the transport of pollen over long distances. The efficiency of bats as pollinators critically depends on the behavior of each species as well as preferred resting areas (Heithaus et al. 1974; Stuart and Marshall 1976).

Fig. 5.1. Flowers of the cauliflorous cannonball tree Couroupita guianensis (Lecthidaceae). (Photo: R. Finkeldey)


Birds

Only comparatively few small bird species act as pollinators for tropical forest plants. Hummingbirds are well-known pollinators of forest plants in tropical America. Their function is fulfilled by various other species belonging to the Zosteropidae, Dicaeidae, Nectariniidae, and Meliphagidae in other parts of the tropics. Many flowers visited by birds are bright red and big. They produce large quantities of nectar, which is the main energy source for the pollinators, but the flowers rarely have strong smells. Many inflorescences are brushlike with long filaments. The conspicuous flowers of Erythrina spp. and Bromeliaceae are pollinated by birds. Heliconia spp. (Musaceae) are frequently planted ornamentals originating from America. Their beautiful flowers are adapted to pollination by birds. Several taxa, such as species of the genus Calliandra, are pollinated by birds, bats, and possibly also insects. The production of “bird-flowers” is as energyconsuming as the production of “bat-flowers.” The size and the behavior of birds suggests that they are efficient vehicles for long-distance transport of pollen; however, the actual distances between two trees which can be bridged by birds critically depends on the sizes of their territories. Pollination by Insects

Insects are the most important pollen vectors in the tropics. Many “generalist” species visit flowers of a multitude of plants. Other pollinating insects coevolved with particular plant species. The obligatory symbiosis between figs (Ficus spp.) and their pollinators (fig wasps; Agaonidae) is a particular wellknown example of a close coevolution between tropical plants and insects (Janzen 1979a; Example 7.2). Hymenoptera

The Hymenoptera (bees, wasps, bumble bees, etc.) are an extraordinary species-rich group of pollinators. Although most bees forage on flowers of numerous plant species, they often prefer a particular species at a given time. The flight behavior of bees and hence their efficiency as pollinators depends on numerous factors, including their size, their sex (male or female; Ackerman et al. 1982), and their social or solitary life cycle. There are few uniform features of flowers which are exclusively or predominantly pollinated by Hymenoptera. The flower morphology greatly varies even within a single tree family such as the neotropical Lecythidaceae, which is predominantly pollinated by bees (Prance 1985). Some plant species attract numerous bees and wasps. For example, more than 70 different bee species were counted visiting the flowers of a single Andira inermis tree (Bawa 1990).

57

58


Bees prefer to fly short distances between visited flowers. Accordingly, most visits are within the crown of a single tree, and geitonogamy is expected to be common (Sect. 4.2.2). Strategies to minimize geitonogamy and hence selffertilization include the evolution of incompatibility mechanisms (Sect. 6.5.1) and a particular flowering phenology (see Example 6.1 for teak). Butterflies and Moths

Flowers mainly pollinated by butterflies typically combine bright pink or red colors with a sweet scent (Proctor et al. 1996). Nectar is usually accessible through a slender corolla. Night-active moths are attracted by a heavy sweet smell; flowers pollinated by moths are often pale. A close coevolution has been described for Heliconius spp. butterflies and vines belonging to the genera Anguria and Gurania in South America (Prance 1985). The butterflies learn the routes between plants widely scattered in the forest and repeatedly visit male flowering plants on their way. The dioecious plants depend on the butterflies as pollinators for reproduction, and the pollen is an important source of protein for the butterflies (p. 82 in Proctor et al. 1996). Moths are less conspicuous than butterflies, but they are important pollinators of many trees of the Rubiaceae, Meliaceae, Mimosaceae, and Apocynaceae and other families. Flies and Beetles

Flies (Diptera) and beetles (Coleoptera) are ubiquitous, extraordinary speciesrich groups of insects, but they play an important role in pollination of woody plants only in the tropics. Flowers predominantly pollinated by flies are often sweet-scented. Occasionally, they have an unattractive, excremental scent. Some flowers adapted to pollination by flies are adapted to “trap pollination” and “sapromyiophily,” i.e., pollination associated with decaying organic material (p. 294 ff. in Proctor et al. 1996). Long-term “imprisonment” of flies in flowers is common in several genera such as Aristolochia (Aristolochiaceae) and Ceropegia (Asclepiadaceae). The huge flowers of Rafflesia arnoldii, a flowering plant parasitic on climbers in tropical Southeast Asia, are pollinated by flies. Many angiosperms belonging to the Annonaceae and other “primitive” families are pollinated by beetles. Thus, this type of pollination seems to be archetypal for angiosperm plants. Beetle pollination has also been reported for species of the gymnosperm family Cycadaceae, and for younger families such as many palms (Palmae) (Eguiarte et al. 1993; Listabarth 1996). Flowers pollinated by beetles often attract other insects as well. Various beetles destroy flowers or inflorescences at least partially by feeding on them.


Other Insects

Other, frequently inconspicuous groups of insects serve important functions for forest ecosystems since they are the main pollinators of keystone plant species. Tiny thrips, cicadas, and bugs belong to this group. Dipterocarps from the central tropics show the phenomenon of simultaneous mass flowering in irregular intervals. Many of them, in particular Shorea and Hopea spp. with small flowers, are mainly pollinated by tiny thrips (Thysanoptera) (Ashton 1988). Another example of the enormous diversity of pollinating insects is the climber Uvaria elmeri (Annonaceae), which is pollinated by cockroaches (Nagamitsu and Inoue 1997). Pollen Vectors: General Patterns and Trends

The species diversity of tropical forests is mirrored by a great diversity of pollination mechanisms. Some plants, for example, most figs (Ficus spp.), depend on a single animal species and evolved an obligatory symbiosis to its pollinator (Example 7.2). Other forest plants are pollinated by a great variety of animals. Pollination by birds, bats, opossums, and large bees has been described for Mabea fistulifera (Euphorbiaceae) in Brazil (Viera and de Carvalho-Okano 1996). The relative importance of different main groups of pollinators has been estimated for evergreen rain forests by Bawa (1990) (Table 5.1). Distinct patterns appear for species of the canopy and of the undergrowth. For example, birds and beetles are important pollinators chiefly for species of the undergrowth. A significant role of bees as main pollinators is evident both for canopy trees and for the undergrowth.

Table 5.1. Relative frequency of different pollen vectors in the canopy and the undergrowth of an evergreen tropical rain forest. (Adapted from Bawa 1990) Pollen vector

Canopy

Undergrowth

No. of species Bats Birds Large bees Small bees Beetles Butterflies Moths Wasps Various small insects Wind

52 3.8% 1.9% 44.2% 7.7% – 1.9% 13.5% 3.8% 23.1% –

220 3.6% 17.7% 21.8% 16.8% 15.5% 4.5% 7.3% 1.8% 7.7% 3.2%

59

60


5.2.3 Pollen Dispersal Assessed by Marker-Based Studies

The physical distribution of pollen can be studied by observation of the behavior of pollinators. Pollinators can be repeatedly trapped and marked, for example, with color stains, to reveal their mobility (Ackerman et al. 1982). The pollen itself can be marked by fluorescent dyes, and its distribution can be studied (Webb and Bawa 1983). Minimum pollination distances can be estimated for dioecious species as the distance between a seed-bearing female plant and its nearest male neighbor (Midgley 1989). The observation of physical pollen transport can be used to estimate the potential gene flow by pollen since transportation of pollen is an obvious requirement for successful pollination and fertilization; however, the actual gene flow by pollen also depends on the success of male gametes. Only gametes incorporated into zygotes and eventually seeds and seedlings contribute to the movement of genetic information. The measurement of actual gene flow through pollen among plants gives estimates of distances between plants in mating contact; hence, it provides crucial information for population delineation. Modern methods to estimate actual gene flow through pollen are based on the observation of genetic markers, often in two consecutive generations, i.e., adults and their progenies. 5.2.3.1 Dispersal of Rare Alleles

A straightforward method to directly assess the gene dispersal of a male flowering tree which is the only carrier of a unique allele is the observation of the distribution of this allele in progenies of surrounding trees (Example 5.1). The estimation is feasible for unique alleles at biparentally inherited, codominant loci and rests on the assumption of regular segregation in the case of a heterozygous potential pollen parent. If the potential pollen tree is homozygous for the rare allele, its success as a pollen parent can be reliably assessed by the observation of the allele in the heterozygous state in the progeny of a seed tree in the vicinity. In the case of heterozygosity of the potential pollen parent for the unique allele, only half of its progenies carry the marker allele. Example 5.1: Gene Dispersal in a Clonal Seed Orchard of Teak (Tectona grandis) A tree in a clonal seed orchard of teak (Tectona grandis) comprising 16 clones was found to be the unique carrier of a rare allele A2 at the isozyme gene locus PGM (genotype A1A2). Since no other tree in the investigated compartment of the seed orchard exhibited this allele, it is unlikely that the tree was a ramet of

Relative success of marker tree as pollen parent

5.2 Gene Flow Through Pollen 1997

40%

1996 1996 and 1997

30%

20%

10%

20

40 60 80 100 Distance from marker tree (m)

Fig. 5.2. Success of a teak (Tectona grandis) tree as a pollen parent for progenies of seed trees growing at different distances in a clonal seed orchard in Thailand. (Adapted from Finkeldey 2006)

a selected clone. The distribution of the allele in progenies of surrounding seed trees was observed in two consecutive years (1996 and 1997) and the mating success of the “marker tree” was estimated for each seed tree. It was found that the intensely flowering tree was successful as a pollen parent in its immediate neighborhood, but that the success of pollination was low or absent in seed trees growing a distance of more than 20 m from the marker tree (Fig. 5.2). The estimate of the selfing rate of the marker tree in two consecutive years (distance assumed to be 0 m in Fig. 5.2) is discussed in Example 6.1 (From Finkeldey 2006). 5.2.3.2 Paternity Analyses

The observation of unique alleles allows the paternity of progenies harvested from a tree to be assigned to a particular pollen parent; thus, it is a particularly simple case of paternity assignment, applicable only if rare, ideally unique, alleles are observed in reproducing populations. A more general approach is the exclusion of paternity on the basis of the simultaneous observation of genotypes of seed trees, their progenies, and a set of potential pollen parents at multiple gene loci. Ideally, all but one of the potential pollen parents can be excluded for a particular progeny owing to an impossible combination of its genotype, the genotype of the seed parent, and the genotypes of the potential pollen parents. A potential pollen tree is excluded as an actual pollen parent of

61

62


a particular progeny if the tree does not carry an allele which was paternally transmitted to the progeny at any gene locus. Obviously, the reliability of paternity exclusion procedures increases with an increasing number of observed gene loci and with increasing variability at the observed loci. Hypervariable, codominant “microsatellites” (simple sequence repeats; Sect. 2.3.1) are particular suitable tools for paternity analyses (Example 5.2). The calculation of paternity likelihoods by special software such as FAMOZ (Gerber et al. 2003) is often preferable to simple exclusion procedures. Example 5.2: Gene Flow Through Pollen in Neobalanocarpus heimii and Dipterocarpus tempehes (Dipterocarpaceae) Thirty potentially flowering trees of the dipterocarp Neobalanocarpus heimii and 248 of their progenies growing in the Pasoh Forest Reserve, Malaysia, were investigated at four microsatellite gene loci (Konuma et al. 2000). Seedlings and saplings were collected within a radius of 10 m around five putative seed parents. The putative seed parent was excluded as an actual seed parent for 16% of the progenies; thus, seeds were also dispersed by an unknown vector. Gene flow by pollen was observed on the basis of paternity exclusion procedures. The estimated minimum average distance between the five seed trees and the successful pollen parents was 524 m. All pollen parents were excluded for 71 out of the 248 progenies investigated (28.6%); thus, these progenies originated from matings with pollen parents out of the plot studied. N. heimii is mainly pollinated by stingless bees (Trigona spp.) and honey bees (Apis spp.). The results of the paternity analyses point towards the efficiency of bees to bridge large distances of, on average, more than 500 m between pollen parents and seed trees (Konuma et al. 2000). Pollen flow of Dipterocarpus tempehes (Dipterocarpaceae) was studied in two years in Lambir Hills National Park, Sarawak, Malaysia (Kenta et al. 2004). Mass flowering was observed in 1996; 1998 was a year of less-intensive flowering. A total of 277 trees in four spatially distinct subpopulations in a 70-ha area as well as 147 (from seed year 1996) and 188 (from seed year 1998) progenies, respectively, were investigated at six simple sequence repeat loci. Average selfing rates were low in both years (7% and 4% for 1996 and 1998, respectively), and mean distances of pollen dispersal were similar and lower than the estimates for N. heimii (222 m in 1996 and 192 m in 1998). Pollen flow among subpopulations was slightly higher in 1996 (44%) than in 1998 (33%). The main pollinators in 1996 were giant honey bees (Apis dorsata), while several moths were suspected of being chiefly responsible for pollen transport in 1998. These differences in the main pollinators did not result in fundamental differences of the pollen dispersal patterns and outcrossing rates; however, the mean relatedness of progenies from the same seed parent was slightly lower in 1996 than in 1998 (From Konuma et al. 2000; Kenta et al. 2004).


Example 5.3: Pollen Dispersal in Three Neotropical Tree Species with Different Spatial Distribution Patterns on Barro Colorado Island The impact of population density and population structure on pollen dispersal was studied for three insect-pollinated species growing on Barro Colorado Island, an artificial island in the Panama Canal (Stacy et al. 1996). Paternity exclusion was conducted for each species on the basis of the observation of variation at five to nine polymorphic isozyme gene loci. All adult or subadult trees in an area of 50 or 84 ha were sampled, and seeds were collected under reproducing trees. Minimum estimates of pollen dispersal were calculated assuming that the nearest not-excluded potential pollen parent according to the paternity analysis was the “true” pollen parent. No selfing was observed for any of the populations investigated, and considerable distances between pollen parents and seed trees were bridged by the pollinators; however, the pollination patterns differed among the species studied and depended on the spatial distribution of trees. The distribution of the 25 adult trees of Calophyllum longifolium (Clusiaceae)was rather uniform throughout the area investigated (84 ha). The nearest neighbor was rarely identified as a pollen parent in the case of spatially isolated seed trees, and the successful pollen parent was separated from the seed tree by a distance of at least 210 m for 62% of the progenies investigated. If trees were widely separated, pollen flow ranged up to several hundred meters. However, cross-pollination seemed to be frequent within the only “cluster” of two directly neighboring trees, which could not be excluded as a pollen parent for 72 and 73% of the progenies investigated from the other tree in this cluster, respectively. Fourteen adults of Spondias mombin (Anacardiaceae) were counted in the same area of 84 ha. Pollination was mainly within clusters, and the nearest neighbor could not be excluded as a pollen parent for roughly 90% of the progenies investigated for this species. On the other hand, a low level of gene flow through pollen over distances of at least 300 m was also confirmed, and pollen flow from outside the area studied was confirmed by the occurrence of rare alleles in progenies which were not observed in any adult in the plot. Only five of the 30 potentially reproducing trees of Turpinia occidentalis (Staphyleaceae) in the 50-ha plot studied produced abundant seeds in the season studied (1993). These five trees occurred in two small clusters of two and three trees with a diameter of 35 and 57 m, respectively. The clusters were separated by a distance of approximately 235 m. Pollen flow was mainly within clusters, and pollen flow over a distance of more than 130 m was confirmed for only one out of 115 progenies investigated. These results point towards a considerable importance of the spatial population structure on pollen dispersal by small insects in tropical forests. Predominance of

63

64


pollination among near neighbors was observed mainly within clusters and small groups of trees. Pollen dispersal is more far-reaching and not restricted to nearest neighbous in the case of a more random distribution of reproducing adults in the forest (From Stacy et al. 1996). 5.2.3.3 Model-Based Estimates of Pollen Dispersal

The investigation of pollen dispersal on the basis of paternity analyses requires the availability of highly polymorphic marker systems and a full inventory of all potential pollen parents in the selected area. Alternative marker-based approaches such as the neighborhood model (Adams 1992) and the TWOGENER method (Smouse et al. 2001; Austerlitz et al. 2004) are less demanding with regard to the degree of polymorphism and the sampling design. However, estimates of pollen dispersal based on these methods rest on model assumptions which are difficult to test and often unlikely to be fulfilled. The investigation of a limited number of seed trees and of a sample of their separately harvested and investigated progenies allows the assessment of the allelic structures in successful pollen contributions. The heterogeneity among pollen allele frequencies of single seed trees is not only a suitable measure of deviation from random mating (Sect. 6.2.1), but is also used by the TWOGENER method to estimate pollen dispersal parameters. Example 5.4: Pollen Dispersal in Dinizia excelsa (Fabaceae) Pollen dispersal was estimated on the basis of an analysis of five microsatellite loci and the TWOGENER method for the Amazonian endemic tree Dinizia excelsa (Fabaceae) in Brazil (Dick et al. 2003). Single-tree progenies were harvested in a highly disturbed, fragmented landscape dominated by pastures, and in an undisturbed continuous forest. Selfing rates were slightly higher in disturbed (s =14%) than in undisturbed (s =10%) conditions. Average pollination distances were estimated to be 212 m in undisturbed forests, where Dinizia excelsa occurs in very low effective densities (approximately 0.1 trees per hectare). The estimate for average pollination distances was considerably higher in the fragmented landscape (1,264 m). Native stingless bees and beetles are the main pollinators in undisturbed forests. Introduced African honey bees (Apis mellifera scutellata) are the principal pollinators in the disturbed landscape. The different types of pollinators result in considerably modified pollination and mating patterns. African honey bees are efficient pollen vectors linking remnant trees in open areas with trees growing in forest fragments (From Dick et al. 2003).

5.3 Migration of Seeds

5.2.4 Efficiency of Pollen Vectors for Gene Flow

The examples of marker-based studies on pollen transport and pollination distances point towards the efficiency of pollinators to bridge distances of at least several hundred meters between reproducing trees in species-rich tropical forests. Similar results were observed in many other comparable studies. For example, Nason et al. (1998) reported large estimates of pollination distances for central American figs (Ficus spp.; Example 7.2). In summary it appears that pollination by animals (zoogamy) is an efficient mechanism to facilitate contacts among widely dispersed, spatially, but not reproductively isolated trees. Thus, population sizes may be large even though the population density is low for the majority of plant species occurring in tropical forests.

5.3 Migration of Seeds Regeneration and colonization of new habitats requires the movement of diaspores. These processes also involve transportation of genetic information usually by means of seed migration both for gymnosperms and for angiosperms. Diaspores are either the seeds themselves or complete fruits or parts of fruits containing seeds. The seedling develops from one or several embryos within a seed. 5.3.1 Seed Vectors

Seeds are dispersed by biotic and abiotic means. Water and wind are the principal abiotic dispersal agents. Biotic seed dispersal is facilitated by animals. Humans play an enormous role in today’s distribution of plants and patterns of genetic variation within species. 5.3.1.1 Abiotic Seed Dispersal

The diaspores of many seed plants contain structures promoting the dispersal of seeds by the wind. The name of the tree family Dipterocarpaceae refers to two wings attached to the fruits of many dipterocarps. However, some dipterocarps have more than two wings, and the wings of several species are only

65

66


rudimentarily developed. The hairs attached to the light seeds of many Bombacaceae promote their dispersal by the wind. Another important abiotic means for the dispersal of seeds is water. Hydrochory requires special adaptations to ensure the survival of seeds in water. Seeds of trees growing along the coastline, along rivers, and in swamp or peat forests are potentially dispersed by water. The weight of diaspores is of crucial importance for their ability to be transported by the wind. Trees with light diaspores such as Eucalyptus spp. and Casuarina spp. have a wide “seed shadow,” i.e., they disperse their seeds in a large area. Barochorous trees produce heavy diaspores and have no efficient means of abiotic or biotic dispersal; their seed shadow is narrow since diaspores are “dispersed” by gravity alone. Typically, pioneer species have a larger seed shadow than species of later successional stages. Most barochorous species belong to climax communities. 5.3.1.2 Biotic Seed Dispersal

Vertebrates, in particular birds and mammals, are by far the most important biotic seed dispersers of tropical forest plants (Loiselle et al. 1996). Endozoochorous seed dispersal requires a passage of seeds through the gastrointestinal tract of an animal. Otherwise, dispersal of seeds by animals is exozoochorous. Seed dispersal of tree species with fleshy, juicy fruits is typically endozoochorous. Exozoochorous dispersal is typical for nuts. Most bird species and both flying (bats, flying foxes) and nonflying (rodents, pigs, primates, and others) mammals disperse seeds. The seeds of the small mountain tree Dunalia arborescens (Solanaceae) are dispersed by at least 16 different bird species in Jamaica (Cruz 1981). 5.3.2 Efficiency of Seed Dispersal

The type of seed vector and, in the case of biotic seed dispersers, their behavior are crucial factors for the seed shadow of forest plants. A comparison between seed dispersal agents in temperate and tropical American forests revealed a dominance of endozoochorous seed dispersal only in the tropics. Endozoochorous, exozoochorous, and wind dispersal are of similar importance in temperate forests of North America (Howe and Smallwood 1982). In general, endozoochorous seed dispersal results in a wider dispersal of seeds as compared with that from other mechanisms since the fruits containing seeds are digested before the seeds are defecated. Endozoochorous seed dispersal may be wide but clustered, for example, under the rest trees of seed-dispersing birds

5.4 Long-Distance Gene Flow and Migration in Tropical Forest Species

and bats, and may result in nonrandom dispersal patterns (Loiselle et al. 1996). The comparatively low differentiation among populations of the rare European wild service tree (Sorbus torminalis; Rosaceae) at maternally inherited chloroplast DNA (cpDNA) has been attributed to the endozoochorous dispersal of seeds by birds resulting in frequent migration among spatially separated stands (Oddou-Muratorio et al. 2001). Limited dispersal of seeds results in complex relations between spatial distribution patterns and kinship relations. In general, neighboring trees are expected to be more closely related to each other and hence to be genetically more similar than spatially separated trees of the same population. The existence of “family structures” can be studied if spatially explicit genetic data are available. Ng et al. (2004) found a significant spatial aggregation in small- and medium-diameter classes for two dipterocarp species (Shorea leprosula and Shorea ovalis ssp. sericea) in a 50-ha plot in the Pasoh Forest Reserve, Malaysia. They explained the result by limited dispersal of both seed and pollen for Shorea leprosula, and an apomictic mode of reproduction in the tetraploid Shorea ovalis ssp. sericea. Strong genetic differentiation among populations separated by less than 10 km from each other has been found for the Amazonian tree Corythophora alta at cpDNA haplotypes (Hamilton 1999). Seed dispersal is very restricted in Corythophora alta, resulting in areas as large as 10 ha founded by a single maternal lineage.

5.4 Long-Distance Gene Flow and Migration in Tropical Forest Species Genetic structures of populations are the outcome of evolutionary processes acting over long time periods. Global changes of environmental conditions in the past did not only affect the distribution of species, but also patterns of genetic variation within species. The most fundamental global events with considerable implications for the distribution of species and their genetic variation patterns were the Quaternary glaciations. The ice ages did not only have a farreaching impact on plants in areas directly affected by growing ice sheets and glaciers, but changed the climate and the sea water levels worldwide (Sect. 8.2). The impact of glacial refugia and postglacial recolonization on genetic variation patterns has been studied for numerous temperate species (Hewitt 2000), including forest trees such as oaks (Quercus spp.; Petit et al. 2002). While it is likely that the Quaternary climatic variability had a considerable impact on the distribution of the regional flora and plant species also in the tropics (Spichiger et al. 1995; Sect. 8.2), much less is known about migration patterns of single species and the resulting genetic variation patterns in the central tropics.

67

68


Maternally inherited markers are particularly useful to analyze migration patterns of plants. The distribution of maternally inherited cpDNA haplotypes among Central American populations of Cedrela odorata is related to its migration history (Example 7.6). Strong differentiation of cpDNA haplotypes among populations has also been observed in brazilwood (Caesalpinia echinata Lam.; Lira et al. 2003). The species occurs in three spatially isolated regions, all characterized by particular cpDNA haplotypes. This pattern points towards the existence of separate glacial refugia as an explanation for the spatial isolation of distribution areas even prior to the beginning of anthropogenic forest fragmentation (Lira et al. 2003). Very restricted migration between orangutan (Pongo pygmaeus) orang-utan populations on the two different sides of the Kinabatangan river in Sabah, Malaysia, was estimated on the basis of the observation of variation at nuclear microsatellites (Goossens et al. 2005). Thus, the river is an efficient natural barrier for the gene flow for the endangered great ape. However, genetic structures point towards frequent migration between distant locations on the same side of the river, indicating relatively free and long movement of orang-utans.

5.5 Recommended Literature Gene flow in seed plants has been thoroughly reviewed by Levin and Kerster (1974), Ellstrand (1992) and, with a focus on forest trees, Adams (1992). The book by Proctor et al. (1996) describes many aspects of the reproductive biology of plants with regard to pollination. The availability of uniparentally inherited markers and hypervariable gene loci has greatly promoted experimental studies in this field as pointed out by Cruzan (1998), by McCauley (1995) for cpDNA polymorphisms, and by Chase et al. (1996a) for biparentally inherited microsatellites with emphasis on tropical forest trees. The impact of palaeoclimatic conditions on migration and patterns of genetic variation has been reviewed by Hewitt (2001).

Mating Systems

The importance of the generation and use of scientifically based information for genetic conservation is well illustrated by changes in views of mating patterns in tropical trees (Boshier 2000).

6.1 Introduction Mating is possible only between individuals producing gametes of opposite sex. The sexual system (Sect. 4.2.1) describes the possible mating events. Mating among conspecific plants requires the movement of the male gamete to the female gamete and, hence, gene flow (Sect. 5.2). Both the sexual system and gene flow are closely related to the analysis of the actual mating events in a population. The genotypic structure of a progeny generation is crucially affected by the successful mating events in a population. Random mating is a useful reference, although it is rarely achieved for plant species occurring in tropical forests (Sect. 6.2). Selfing (Sect. 6.3) and mating among relatives result in inbreeding; inbred forest trees frequently suffer from strong inbreeding depression (Sect. 6.4). The statement at the beginning of this chapter by Boshier (2000) refers to the accumulation of scientific evidence that inbreeding and in particular selfing is not the rule for tropical forest trees, as it was assumed prior to the conduct of numerous case studies on mating patterns of tropical forest trees. Most forest trees are either dioecious or self-incompatible (Sect. 6.5), and even the majority of self-compatible trees in tropical forests are predominantly outcrossed. The mating system of a species is not uniform, but can vary among different populations and among years. The impact of the population density and other environmental factors on the mating systems of tropical forest plants is mentioned in Sect. 6.6.

6

70

C HAPTER 6 Mating Systems

6.2 Random Mating and Panmixis Ideally, a population consists of a group of individuals producing progenies at random, i.e., without any preferences for mating events between particular plants or groups of plants. Random mating is a useful reference even though in actual plant populations mating preferences for neighboring plants or plants with overlapping flowering periods are likely to exist. A prerequisite for random mating is the even distribution and success of male gametes in the population. Thus, all seed parents receive homogeneous pollen contributions in a population with a random mating system. The degree of the deviations from random mating can be assessed by measuring the heterogeneity among pollen contributions to single seed plants (pollen allele frequencies). 6.2.1 Heterogeneity of Pollen Allele Frequencies

The constitution of the successful paternal contributions to the progenies of a particular seed plant is its “genetically effective pollen cloud.” The term “pollen cloud” is illustrative for wind-pollinated species, but may also be used for zoogamous species. Genetic differentiation among pollen allele frequencies results from an unbalanced success of trees as pollen parents for the progenies of particular seed trees as illustrated by a fictitious example in Fig. 6.1. The pollen allele frequency of a seed tree is assessed by estimating the allelic structure of successful pollen after an investigation of a sample of its progenies (seeds). If the haploid megagametophyte and the corresponding embryo in the seed of a gymnosperm are simultaneously analyzed, the pollen contribution can be reliably assessed for each progeny (Müller 1976). Maximum-likelihood estimates of pollen allele frequencies can be obtained if progenies from angiosperm plants are analyzed (Gillet 1997). The separate investigation of progenies from several seed trees of a single population allows the differentiation among them to be assessed with methods used for an analysis of population differentiation (Sect. 3.3.2). The differentiation among pollen allele frequencies is the most important parameter used to estimate pollen dispersal with the TWOGENER method (Sect. 5.2.3). This illustrates the close affinities between gene flow and mating system analyses. Considerable differences with regard to the heterogeneity of pollen allele frequencies at isozyme gene loci appear between populations occurring in pure stands, populations occurring in intermediate density in tropical forests, and rare species distributed in low density in tropical forests (Finkeldey 2002; Fig. 6.2). Extraordinary low differentiation among pollen clouds was observed

6.2 Random Mating and Panmixis

I

II

Fig. 6.1. Mating of two trees in a population. Only conspecific trees in the forest are shown. Tree I is fertilized by only two other trees contributing effective pollen in somewhat unequal proportions; therefore, the effective number of its male mating partners is close to unity and the majority of its offspring are full-sibs. Neither can the effective pollen cloud of tree II be considered representative of that of the virtual pollen cloud of the population. The two trees share only one common male mating partner. Their effective pollen clouds are expected to be markedly differentiated

in three pure Norway spruce (Picea abies) plantations in Germany (Finkeldey 1995). Even this low level of differentiation was statistically significant at most loci in one or several populations, but no strong deviations from random mating were observed. Differentiation was slightly higher in three populations of teak (Tectona grandis) in Thailand, all characterized by comparatively high density (Finkeldey 2006). Most studies on genetic differentiation among pollen allele frequencies were conducted on Barro Colorado Island (Murawski and Hamrick 1991). Significant, but moderate heterogeneity was observed for most tropical forest species occurring in intermediate density (approximately one to ten trees per hectare). The highest differentiation among pollen allele frequencies and, thus, the strongest deviations from random mating were observed for species occurring in low density (less than one flowering tree per hectare) in tropical forests (Murawski and Hamrick 1991).

71

C HAPTER 6 Mating Systems GST 0.30 0.25 0.20 0.15 0.10

>50 Ind/ha

1-10 Ind/ha

Tachigali versicolor

Platypodium elegans II

Platypodium elegans I

Ceiba pentandra

Cavanilessia platanifolia III

Cavanilessia platanifolia II

Cavanilessia platanifolia I

Brosimum alicastrum

Trichilia tuberculata

Pterocarpus indicus

Sorocea affinis II

Sorocea affinis I

Quararibea asterolepis

Beilschmedia pendula

Tectona grandis Plant.

Tectona grandis Nat. Pop.

Picea abies III

0.00

Picea abies II

0.05

Picea abies I

72

pi2 for all alleles i at several gene loci) is expected to result from inbreeding (Sect. 6.4). Slightly more heterozygotes than expected under Hardy–Weinberg conditions were observed at isozyme gene loci in adult trees of a subdivided population of the pioneer species Cecropia obtusifolia in Mexico (Alvarez-Buylla et al. 1996); however, homozygotes exceeded the expectation in the seeds collected after distribution, and in seeds collected from the soil. The proportion of homozygotes decreased from seedlings to juveniles and adults. Mixing of partially differentiated subpopulations owing to seed dispersal explains the homozygote excess in seeds after dispersal better than inbreeding for this dioecious species. Selection favoring heterozygotes possibly accounts for the increase of heterozygosity with age (Alvarez-Buylla et al. 1996). A similar temporal dynamics of genotypic structures with a significant excess of homozygotes in particular in seeds and young seedlings has also been reported for several other tropical tree species such as Ardisia escallonioides (Pascarella 1997), Symphonia globulifera (Aldrich et al. 1998), and Platypodium elegans (Hufford and Hamrick 2003; Example 7.1). The frequency of deviations from Hardy–Weinberg structures in populations of tropical forest trees is another indication that random mating is a useful reference, but not the rule in tropical forests.

6.3 Selfing and Outcrossing Rates Cosexual (hermaphroditic and monoecious) species produce gametes of both sexes on a single plant; thus, the pollination of a female gamete by the male gamete of the same plant is possible (Sect. 4.2.2; Fig. 4.2). The relative frequency of these selfings among all successful mating events of a single plant or a population is the selfing rate s, and its complement is the outcrossing rate t = 1 – s. The selfing rate is a fundamental aspect of the mating system of a population. Many studies were conducted to estimate selfing rates in populations of tropical forest species (Table 6.2). Most experimental studies aimed at the estimation of selfing rates were based on the observation of biochemical or molecular gene markers. Occasionally, other marker types have been used to estimate selfing rates of trees. For example, the mangrove species Rhizophora mangle is a predominantly selfing species according to a study based on the observation of chlorophyll deficiency as a color trait marker (Lowenfeld and Klekowski 1992).

73

74


6.3.1 Estimates of Selfing Rates Based on Rare Alleles

Selfing can be reliably assessed if a tree is homozygous at a biparentally inherited, codominant gene locus for a unique allele Ax, i.e., if no other tree in the population carries this allele. In this case, all progenies from selfing are homozygous Ax Ax, and all progenies from outcrossing are heterozygous Ax Ay (with y = i, j, k, . . .). However, the homozygous occurrence of a rare allele in a single tree is an extremely unlikely and hence rare event. Usually, trees are heterozygous for unique variants x (genotype Ax Ay). In this case, only a fraction of the progenies from selfing can be reliably identified: All progenies homozygous for the rare allele (genotype Ax Ax) arise from selfing. If regular segregation and random fusion of gametes are assumed, the expected segregation within the selfings Ax Ay × Ax Ay is 0.25Ax Ax , 0.5Ax Ay , 0.25Ay Ay. Hence, the proportion of genotypes Ax Ax (Pxx) among a sample of progenies investigated equals one quarter of all selfings, and the selfing rate of the tree can be estimated as s = 4 × Pxx. Example 6.1: Estimates of Selfing Rates in Teak (Tectona grandis) Populations A tree heterozygous for a unique allele A2 at the isozyme gene locus PGM-A was identified in a clonal seed orchard of teak (Example 5.1). Genotypic structures of progenies harvested from this tree were observed in two consecutive years (Table 6.1). On the basis of these observations, the selfing rate was estimated as s = 4 × P22 = 4 × 0.060 = 0.240 in 1996 and s = 4 × 0.94 = 0.376 in 1997. The method was also applied to estimate selfing rates of 24 teak trees carrying a rare allele at a marker locus in natural and planted populations. Estimates of the selfing rate range from s = 0 to s = 0.62 for single trees. Population estimates based on a mixed mating model were between s = 0.07 and s = 0.17 (Finkeldey 2006). Teak flowers are complete (Fig. 6.3a). A single flower is receptive and releases its pollen during a few hours of a single day; however, flowering of an inflorescence (Fig. 6.3b) usually extends over several weeks since only a few flowers are Table 6.1. Genotypic structures at the gene locus PGM-A in progenies of a teak tree (genotype A 1A2) in a clonal seed orchard in Thailand in 1996 and 1997. (Adapted from Finkeldey 2006)

1996 1997 N sample size

N

P11

P12

P22

383 245

0.400 0.429

0.540 0.477

0.060 0.094

6.3 Selfing and Outcrossing Rates

a b Fig. 6.3. Single flower (a) and inflorescence (b) of teak (Tectona grandis). (Photos: R. Finkeldey)

receptive each day. The extended flowering of teak inflorescences reduces the probability of selfing due to geitonogamy since pollinators find only a limited number of receptive flowers available within single inflorescences and hence are forced to move among inflorescences, and possibly trees. Observations of the mating system in natural and breeding populations confirmed that teak is a predominantly outcrossing species. The existence of a system of self-incompatibility has been suggested for teak (Sect. 6.5.1), but considerable levels of selfing were observed in all populations studied in Thailand. High variation of the selfing rate of individual trees was observed. Some trees are (almost) completely outcrossing, while surprisingly high estimates of selfing rates were estimated for other trees. Significantly different selfing rates were estimated for the same tree in different seed years, suggesting an impact of the environmental conditions during flowering (temperature, activity of pollinators, etc.) on selfing rates in teak (Finkeldey 2006).

75

76


6.3.2 Estimates of Selfing Rates Based on a Mixed Mating Model

Selfing is associated with characteristic changes of genotypic structures in the progenies, since it is the strongest form of inbreeding (Sect. 6.4). Thus, it is possible to estimate selfing rates on the basis of an investigation of the progenies from a sample of seed parents in a population. The estimation is based on the assumption that each progeny results either from selfing or from random outcrossing. Random outcrossing is rarely expected to be realized in particular for tropical trees occurring in low density (Sect. 6.2). Accordingly, assumptions of the mixed mating model are often violated, and results are possibly biased. This applies in particular to estimates based on single marker loci and individual seed trees. Multilocus estimates for populations are expected to be more robust against violations of model assumptions. Estimates of the multilocus outcrossing rate tm are most frequently reported (Ritland and Jain 1981). Most estimates of outcrossing rates of tropical trees were obtained from studies based on moderately variable isozyme gene markers using the mixed mating model. Population estimates of tm are shown in Table 6.2 for selected tropical trees species. If more than a single population was investigated in a particular study, the minimum and maximum estimates of tm are reported. Outcrossing rates are close to 1 (or 100%) for the majority of the tropical tree species investigated. Thus, selfing rates (s = 1 – tm) are usually low, and even self-compatible species with a mixed mating system are mainly outcrossing. Exceptions are the self-compatible Cavanillesia platanifolia (Sect. 6.6), some populations of Ardisia escallonioides (Pascarella 1997), and Pinus merkusii in Thailand. Extraordinary low estimates of outcrossing rates for Pinus merkusii were estimated in a few populations only. Overaging, degradation, and lack of flowering synchrony are likely causes for a presumably low availability of foreign pollen in these populations (Changtragoon and Finkeldey 1995a). High outcrossing rates were estimated for Pinus merkusii in a natural stand and seed orchards in Indonesia (Siregar and Hattemer 2001). 6.3.3 Estimates of Selfing Rates Based on Nonmaternal Alleles

The observation of a nonmaternal allele Ak (k ≠ i, j) in a progeny of a seed parent with genotype Ai Aj at any biparentally inherited gene locus is sufficient to exclude selfing; thus, it is possible to calculate a minimum estimate of the outcrossing rate by computing the proportion of progenies with a

6.3 Selfing and Outcrossing Rates

Table 6.2. Estimates of multilocus population outcrossing rates tm based on a mixed mating model for tropical forest tree species Species

tm

Reference

Acacia auriculiformis Ardisia escallonioides Astrocaryum mexicanum Beilschmelia pandula Bertholletia excelsa Brosimum alicastrum Calophyllum longifolium Carapa guianensis Carapa procera Cavanillesia platanifolia Cedrela odorata Ceiba pentandra Cordia alliodora Dipterocarpus cf. condorensis Eucalyptus citriodora E. delegatensis E. grandis E. kitsoniana E. obliqua E. pauciflora E. regnans E. rhodantha E. saligna E. stellulata E. stoatei Jacaranda copaia Pterocarpus indicus Pterocarpus macrocarpus Pinus merkusii

0.673–0.951 0.390–1.051 0.933–1.050 0.918 0.849 0.875 1.030–1.031 0.967–0.986 0.78 0.213–0.661 0.969 0.689 0.966 0.68–0.81 0.86 0.79 0.84 0.77 0.76 0.63 0.69 0.58–0.60 0.77 0.77 0.82 0.982 0.908 0.719– 0.959 0.017–0.843

Platypodium elegans Quararibea asterolepis Shorea congestiflora Shorea megistophylla Shorea trapezifolia Sorocea affinis Spondias mombin Stemmadenia donnel-smithii Stemonoporus oblongifolius Tachigali versicolor Trichilia tuberculata Turpinia occidentalis

0.898–0.924 1.008 0.874 0.71–0.87 0.519–0.602 0.969–1.089 0.989–1.304 0.896 0.84 0.937 1.077 1.006–1.071

Wickneswari and Norwati (1995) Pascarella (1997) Eguiarte et al. (1992) Murawski and Hamrick (1991) O’Malley et al. (1988) Murawski and Hamrick (1991) Stacy et al. (1996) Hall et al. (1994b) Doligez and Joly (1997) Murawski and Hamrick (1992) James et al. (1998) Murawski and Hamrick (1991) Boshier et al. (1995) Luu (2005) Moran and Bell (1983) Moran and Bell (1983) Moran and Bell (1983) Moran and Bell (1983) Moran and Bell (1983) Phillips and Brown (1977) Moran and Bell (1983) Sampson et al. (1989) Moran and Bell (1983) Moran and Bell (1983) Moran and Bell (1983) James et al. (1998) Finkeldey et al. (1999) Liengsiri et al. (1998) Changtragoon and Finkeldey (1995a) Murawski and Hamrick (1991) Murawski and Hamrick (1991) Murawski et al. (1994a) Murawski et al. (1994b) Murawski et al. (1994a) Murawski and Hamrick (1991) Stacy et al. (1996) James et al. (1998) Murawski and Bawa (1994) Murawski and Hamrick (1991) Murawski and Hamrick (1991) Stacy et al. (1996)

77

78


nonmaternal allele at any gene locus investigated. The estimation is a special case of paternity exclusion (Sect. 5.2.3) with the seed parent regarded as the only potential pollen parent. The reliability of the estimate increases with the number of gene loci investigated and with the variability at these loci. Thus, the method is preferably based on an investigation of several “hypervariable” loci; nuclear microsatellites (simple sequence repeats) are particularly useful to identify nonmaternal alleles owing to their codominance and high variability (Example 7.1). High outcrossing rates were estimated for a Dipterocarpus tempehes population (t = 0.93 and 0.96 in two different seed years) on the basis of paternity exclusion and taking possible genotyping errors into account (Kenta et al. 2004; Example 5.2). Comparable results were obtained for the outcrossing rate in an undisturbed natural forest of another dipterocarp, Shorea curtisii (t = 0.96); however, estimates of individual outcrossing rates were considerably lower for five Shorea curtisii tress in a selectively logged forest (t = 0.52; Obayashi et al. 2002).

6.4 Inbreeding and Inbreeding Depression Plants (and animals) are genetically related, if they have at least one common ancestor. If two related plants mate, they possibly transmit a copy of the same allele, which both of them may have inherited from their common ancestor, to their progeny. In result, mating among relatives leads to a certain probability that two alleles in the progeny are identical by descent (p. 179 ff. in Hedrick 2000); thus, mating among relatives results in an increased probability that the two alleles transmitted to a progeny are identical, i.e., that the progeny is homozygous at the respective gene locus. The same applies in the case of a progeny resulting from gametes produced by the same parent after selfing. Inbreeding is defined as mating among relatives, including selfing, which is regarded as the strongest form of inbreeding. It results in increased homozygosity as compared with matings among nonrelated plants. The increase in average homozygosity due to inbreeding depends on the degree of relatedness of the parents. 6.4.1 Genetic Consequences of Inbreeding

The average level of inbreeding in a population is frequently quantified by a comparison of the heterozygosity He expected under Hardy–Weinberg conditions

6.4 Inbreeding and Inbreeding Depression

(Sect. 6.2.2) and the observed heterozygosity Ho (see Sect. 3.3.1 for a computation of He and Ho). It holds for the “inbreeding coefficient” F F = 1 − Ho/He.

(6.3)

In case of Hardy–Weinberg structures, Ho equals He, resulting in F = 0. If the observed heterozygosity exceeds the expected heterozygosity, F becomes negative. Negative values of F cannot be explained by inbreeding but have other causes, for example, selection favoring heterozygotes at the gene locus investigated. If fewer heterozygotes were observed than expected for corresponding Hardy–Weinberg structures, F becomes positive. If positive inbreeding coefficients F are observed at several gene loci, and if the values for F are homogeneous across loci, inbreeding is a plausible explanation for the observed genotypic structures. Examples of inbreeding coefficients F observed in populations of tropical forest plants are shown in Table 6.3. Weak or no impact of inbreeding on genotypic structures is suggested by low inbreeding coefficients for Pterocarpus indicus, D. condorensis, Dipterocarpus cf. condorensis, and Swietenia macrophylla. The inbreeding coefficient was low, but significantly different from zero in four out of seven populations of Swietenia macrophylla studied (Lemes et al. 2003; Sect. 3.5.3; Table 3.4). The genotypic structures of Ancistrocladus korupensis (Foster and Sork 1997) and the five Pinanga species studied by Shapcott (1999) are strongly influenced by inbreeding. Very high selfing rates and preferential mating among neighboring, closely related trees are likely to account for the high inbreeding coefficients observed for both species.

Table 6.3. Average inbreeding coefficients F = 1 – Ho/He observed in populations of tropical forest plants Species

Plant group

Marker type

F

Reference

Ancistrocladus korupensis

Liana

Isozyme

0.455

Dipterocarpus cf. condorensis Pinanga spp. (5 species) Pterocarpus indicus

Timber tree

Isozyme

0.051

Foster and Sork (1997) Luu (2005)

Palm Timber tree

Isozyme Isozyme

0.640–0.856 0.015

Pterocarpus macrocarpus

Timber tree

Isozyme

0.099

Swietenia macrophylla

Timber tree

Microsatellite

0.038

Shapcott (1999) Finkeldey et al. (1999) Liengsiri et al. (1995) Lemes et al. (2003)

79

80


6.4.2 Inbreeding Depression

The loss of heterozygosity as a consequence of inbreeding is often reflected at phenotypic traits of adaptive or economic significance. Many complex traits, including survival, growth rate, and quality traits such as stem straightness, are negatively influenced by inbreeding of predominantly outcrossing species. The negative impact of inbreeding on phenotypes and the fitness is defined as inbreeding depression. Predominantly or exclusively selfing species, including many agricultural crops such as rice and soy bean, do not suffer from inbreeding depression since inbreeding does not further reduce the level of heterozygosity for these species. Inbreeding depression can be explained by a slight superiority of heterozygotes in comparison with homozygotes at many gene loci (“overdominance hypothesis”; Ziehe and Roberds 1989) or by a strong inferiority of particular homozygotes for deleterious alleles at a few adaptive gene loci (“dominance hypothesis”; p. 220 ff. in Allard 1960). Deleterious recessive alleles seem to be a more important cause of inbreeding depression in plants than overdominance (Ritland 1996a). Most tropical forest tree species are highly variable (Sect. 3.4; Tables 3.2, 3.3) and predominantly outcrossing (Sect. 6.3; Table 6.2); thus, the observed heterozygosity is usually high, and strong inbreeding depression is expected for most trees regardless of the exact cause of the negative impact of low levels of (average) heterozygosity on fitness and phenotypic traits (Ledig 1986b). Inbreeding depression acts on numerous traits affecting fitness throughout the life cycle of plants. Inbreeding effects have been studied particularly well for eucalypts (Example 6.2) and a few other species grown in plantations. It is difficult to investigate inbreeding depression in natural and naturally regenerated forests (but see Example 7.1), but a negative effect of inbreeding on fitness is likely to be a main cause for the evolution of efficient pollen dispersal mechanisms (Sect. 5.2) and of incompatibility systems (Sect. 6.5) for many tropical forest plants. Thus, the disruption of mating contact among nonrelated, widely separated plants poses a main threat to populations since it results in increased inbreeding (Finkeldey 2002). Example 6.2: Inbreeding Depression in Eucalypts Eucalypts (Eucalyptus spp.) are hermaphroditic, self-compatible trees. A mixed mating system is typical for eucalypts. Outcrossing usually dominates, but selfing rates of 20–40% are frequently encountered (Table 6.2). Inbreeding depression of eucalypts has attracted considerable interest since it directly affects the performance of eucalypt plantations.

6.4 Inbreeding and Inbreeding Depression

Inbreeding depression during an early life stage was studied for E. delegatensis (Moran and Brown 1980). Fruits of this species remain on seed trees for several years, allowing the investigation of seed cohorts of different age. The outcrossing rate tm was estimated in 1-, 2, and 3-year-old seed cohorts on the basis of isozyme studies and a mixed mating model with the following result: tm = 0.66 (1-year-old seeds), tm = 0.78 (2-year-old seeds), tm = 0.85 (3-year-old seeds). Thus, outcrossing seems to increase and, hence, selfing to decrease with increasing age. The most likely reason for the effect is early mortality of inbred seeds. Preferential survival of outcrossed progenies is a plausible explanation for the apparent increase of the outcrossing rate. The growth of progenies from controlled selfing, open pollination, and controlled outcrossing was compared for seed parents of E. regnans in a field trial in Australia. Survival and height growth were only slightly affected by selfing during the early development of the trial, but strong and highly significant differences were observed 12 years after planting. Only 31.2% of the selfed trees survived; their average diameter was 20.4 cm. This compares with 71.9% survival with an average diameter of 24.2 cm for the open-pollinated trees. The completely outcrossed family was even superior to the trees from open pollination (Griffin and Cotterill 1988). Many studies proved strong inbreeding depression in E. grandis. Inbreeding results in inferior growth of young plants and has a negative impact on quality traits like stem form (Hodgson 1976a). Inbreeding depression in E. grandis is not restricted to selfing, but has also been described for matings among fulland half-sibs (van Wyk 1981). Its strength depends on the relatedness of the parent trees. The growth performance of E. grandis progenies from single seed trees after open pollination is correlated to the outcrossing rate tm of the seed parents estimated by a mixed mating model on the basis of isozyme investigations (Burgess et al. 1996). The mean family heights of 20 open-pollinated families with high estimates of outcrossing rates (tm) were quite uniform for 2.5-yearold plants (Fig. 6.4). Some families with low estimates of outcrossing rates reached comparable mean heights, but the mean height growth of most families with low tm was considerably below the mean for the reference populations with high tm. Thus, the mean growth rate of a family is negatively affected by a low outcrossing rate, presumably owing to inferior growth of selfed progenies. The strong inbreeding depression of eucalypts has considerable practical importance. Eucalypt plantations are established worldwide on a large scale (Eldridge et al. 1994), and the origin of reproductive material is often unknown. Most eucalypts flower and fruit prolifically and regularly even if they grow as isolated individuals or in small stands. Uncontrolled harvest of seeds from isolated trees or small stands poses a risk of producing inbred reproductive material which is expected to suffer from strong inbreeding depression.

81

C HAPTER 6 Mating Systems 12 Families with low outcrossing rates

Families with high outcrossing rates

11 Mean family height (m)

82

10

9

8

7

6 0

0.2

0.4

0.6 0.8 1 Family outcrossing rate

1.2

1.4

1.6

Fig. 6.4. Family mean height of 2.5-year-old Eucalyptus grandis seedlings in relation to an estimate of the family outcrossing rate tm for 40 families with contrasting estimates of tm. (From Burgess et al. 1996)

Negative effects from inbreeding depression are particularly likely to occur if seeds are harvested from well-performing, but isolated trees or small stands close to villages for the establishment of small-scale plantations in rural areas.

6.5 Incompatibility and Self-Sterility Since most tropical forest trees are hermaphroditic, selfing by means of autogamy and geitonogamy is possible. However, selfing is the strongest form of inbreeding and is associated with strong inbreeding depression. In consequence, mechanisms preventing selfing offer considerable evolutionary advantage. The most important mechanisms acting prior to fertilization are incompatibility systems. Cosexual species without incompatibility reactions often evolve systems of self-sterility which result in early abortion of embryos after fertilization. 6.5.1 Incompatibility

The genotypes of two plants which are able to produce a common progeny are compatible. If plants belong to different biological species, they are incompatible

6.5 Incompatibility and Self-Sterility

since they cannot produce common progenies. On the other hand, incompatibility also often occurs between gametes produced by a single cosexual plant. For example, controlled pollination of Gmelina arborea (Verbenaceae) is not successful if pollen from the seed parent is used (Bolstad and Bawa 1982). This suggests that G. arborea is completely self-incompatible. Observations of the flower development after pollination suggest that teak (T. grandis Tectona grandis), another important timber species of the Verbenaceae, is also at least partially self-incompatible (Tangmitcharoen and Owens 1997); however, controlled self-pollination in teak results in low success rates (Bryndum and Hedegart 1969), and marker-based analyses of the mating system after open pollination revealed that teak is a mixed mating species (Example 6.1). Self-incompatibility is controlled by a single locus or only a few incompatibility loci (S-loci). The incompatibility usually results from the inability of pollen to germinate on the pistil of particular plants, or from inhibition of the growth of the pollen tube. The incompatibility system does not only exclude self-pollen from successful fertilization since the genetic information at the S-locus is crucial for the reaction. Two main types of incompatibility systems are distinguished (Boshier 2000). In the case of gametophytic self-incompatibility, the genetic constitution of the male gametophyte at the S-locus is decisive for the incompatibility reaction. Pollen is excluded, if it carries any of the two alleles of the seed parent. Thus, self-pollination is impossible just like pollination between two related or unrelated trees with identical genotypes at the S-locus. If two trees share only one allele at the S-locus, they are semicompatible since half of their pollen is compatible but the other half is excluded from pollination success. Obviously, all plants are heterozygous at the S-locus, and populations need at least three alleles at this locus in order to be able to produce progenies. Usually, the allelic diversity at the S-locus is very high. The genotypic information of the pollen parent at the S-locus, and not only the allelic constitution of haploid pollen, is decisive for the incompatibility reaction in the case of sporophytic systems. Two plants with identical genotypes at the S-locus are incompatible; thus, selfing is obviously impossible. Sporophytic systems are often heteromorphic with regard to flower morphology. For example, two types of flowers are observed in case of distyly. Flowers have either a long style and short stamina, or they are short-styled with long stamina. Pollen from short stamina is excluded in long-styled flowered and vice versa. Since plants produce only one of the two flower morphs, they are selfincompatible. Many other mating combinations are also excluded. Distyly is controlled by a sporophytic incompatibility system with only two alleles and complete dominance of one of the alleles. An incompatibility system is regarded as “cryptic” if selfing and other forms of “incompatible” matings are disadvantageous, but not completely excluded. For

83

84


example, cryptic incompatibility has been reported for Brazil nut (Bertholetia excelsa) (Moritz and Lüdders 1993). Cryptic incompatibility results in a strong dominance of outcrossing, if “foreign,” fully compatible pollen from other trees is available, but does not preclude the production of seeds by sexual processes in the case of the absence of foreign pollen. Incompatibility systems are very common among hermaphroditic plants in tropical forests, but the genetic control has rarely been investigated. Gametophytic incompatibility has been reported for Inga spp. (Koptur 1984) and Acacia retinodes (Kenrick and Knox 1985). Cordia pringlei and Cordia dentata are distylous (Opler et al. 1975). Other species of the genus Cordia (Boraginaceae) are dioecious (C. inermis, C. collococca, and C. panamensis), and Cordia alliodora is self-compatible, but predominantly outcrossing (Boshier et al. 1995; Table 6.2). The system of self-incompatibility is particularly complicated in cacao (Theobroma cacao). Several loci are involved; the genetic control is both sporophytic and gametophytic. Most trees in the South American region of origin of cacao are self-incompatible, but there are also self-compatible varieties (Warren et al. 1995). Bawa (1974) estimated a frequency of 54% of self-incompatible hermaphrodites among 130 species investigated in a tropical lowland rain forest in Costa Rica. Since 22% of the species in this community are dioecious, 76% are likely to be obligatory outcrossing. The remaining trees are either self-compatible hermaphrodites (14%) or monoecious (10%). 6.5.2 Self-Sterility

Postzygotic effects of selfing have been studied in detail for many conifers (p. 67 f. in Stern and Roche 1974; Savolainen et al. 1992). The flower morphology of gymnosperms does not permit the selective discrimination of particular pollen types. The early abortion of zygotes from selfing is not controlled by a single locus or a few gene loci, but is the effect of many interacting genes. Thus, self-sterility may be regarded as a very early form of inbreeding depression acting on young zygotes and embryos and resulting in strong viability selection against inbred genotypes. The dominance of angiosperms and the frequent observation of incompatibility systems in tropical forests suggest that systems of self-sterility are of lesser importance.

6.6 Environmental Effects on Mating The mating system is not fixed or exclusively determined by genetic factors within a species, but it is also influenced by environmental conditions; thus,


human impact on environmental conditions, for example, by means of forest fragmentation and silvicultural management, has considerable implications for the realized mating systems (Finkeldey 2002). Environmental variation is common both in managed and in unmanaged forests and hence the mating system is likely to vary both in time and in space. An effect of population density on the mating system is evident from comparisons of the heterogeneity of pollen allele frequencies among species occurring in different density in tropical forests (Sect. 6.2.1). In general, low population density results in strong deviation from random mating and strong differentiation among pollen allele frequencies of single seed trees owing to a restricted number of effective pollen parents contributing to the production of progenies from single seed trees. The “reproduction effective” population density has an influence on selfing rates of the self-compatible Cavanillesia platanifolia (Murawski and Hamrick 1992). The outcrossing rate of a population studied on Barro Colorado Island in Panama was very low in a year with only a few flowering trees (32% of censused trees that flowered: tm= 0.213; estimates based on isozyme variation and a mixed mating model, see Sect. 6.3.1), but increased when more trees participated in reproduction in other seed years (49% of censused trees that flowered: tm= 0.347; 74% of censused trees that flowered: tm= 0.569). The estimate of the outcrossing rate was tm= 0.661 in a reference population with higher density. No obvious impact of population density and stand structure on outcrossing rates was observed in five populations of the endemic Dipterocarpus cf. condorensis in Vietnam, but the populations varied with regard to the estimates for the effective number of pollen parents fertilizing single seed trees (Luu 2005). An impact of different spectra of pollinators on gene flow and the mating system may be assumed for species with a wide spatial distribution and species occurring in different ecosystems. However, very little is known about the consequences of different natural “pollination environments” on the mating system of tropical forest trees.

6.7 Recommended Literature The contribution of Boshier (2000) to the book edited by A. Young et al. (2000) gives a concise introduction to many aspects of plant mating systems, including inbreeding and incompatibility. The genetic consequences of inbreeding in plants were discussed by Ritland (1996a). Inbreeding depression has been reviewed with emphasis on conifers by Williams and Savolainen (1996) and with a focus on tropical forest trees by Griffin (1990). Incompatibility is discussed in detail in the book of de Nettancourt (2001).

85

Adaptation and Coevolution

The results of natural selection, the evolutionary force most responsible for adaptation to the environment, are evident everywhere, yet it is remarkably difficult to observe the time course of changes brought about by selection (p. 49 in Gillespie 1998).

7.1 Introduction Plants are able to survive and to reproduce in different environments. The longevity and immobility of trees implies that they need to withstand different environmental conditions during their lifetime. Thus, adaptation of trees to their environment is ubiquitous and necessary for survival. Adaptedness implies a high likelihood of survival and reproduction in a given environment. Adaptation is the process leading to a state of adaptedness in response to environmental variability. The adaptive potential or adaptability describes the ability to achieve a state of adaptedness to a range of different environmental conditions (p. 59 in Eriksson and Ekberg 2001).

7.2 Physiological and Evolutionary Adaptation Since adaptation is a general aspect in the life cycle of plants it is important to distinguish different strategies and levels for adaptive processes. Each plant reacts to changing environmental conditions by adjustment of its metabolism. These physiological adaptations or “phenotypic plasticity” (Schlichting 2002) of plants is eventually manifested in morphological features such as sun- and shadow-leaves. Many plants can be cloned and copies of the same genotype can be grown in different environments. This allows the phenotypic response of genotypes to different environments (their “norms of reaction”) to be assessed and compared. Clonal tests revealed large differences of physiological adaptability among genotypes for many tropical trees (Foster and Bertolucci 1994).

7

88

C HAPTER 7 Adaptation and Coevolution

Different norms of reactions are not only observed among clones, but are also the rule among genetically heterogeneous demes such as families or provenances. The use of genotypes in plantation forestry requires their adaptedness to the respective site conditions. Thus, traditional tree improvement (Chap. 12) is based on differences of the physiological adaptedness among demes compared in field trials although the “domestic fitness” rather than the natural fitness of genotypes and demes matters in applied tree breeding (p. 94 ff. in Eriksson and Ekberg 2001). Drastic environmental change results in the loss of adaptedness of particular genotypes. Preferential survival or higher reproductive capacities of “adapted” genotypes result in changes of genetic structures owing to viability selection or fertility selection. Evolutionary adaptations are caused by selective changes of genetic structures; hence, the population, and not the single plant, is the basic entity to investigate evolutionary adaptations, which can take place only in genetically diverse populations. There is no doubt that evolutionary changes of genetic structures due to viability and fertility selection contributed to the adaptation of trees to their respective environments in boreal, temperate, and tropical forests. Most tropical tree species are highly variable, and most of the genetic variation at single loci resides within populations (Chap. 3). Thus, evolutionary adaptations were not only involved in the creation of the high species diversity in tropical forests, but they continue to play an important role for the reactions of forest ecosystems to environmental change. The urgent need to conserve genetic resources of tropical forest trees (Chap. 14) is motivated by the importance to maintain evolutionary adaptive potentials in tropical forests. Selection primarily acts on single gene loci; thus, it is much more difficult to investigate in natural ecosystems than evolutionary factors affecting all gene loci in principally the same way. For example, drift results in losses of genetic variation at all loci, and inbreeding reduces heterozygosity at all biparentally inherited, nuclear genes. The dynamics of genetic structures at putatively “neutral” gene loci can be used to monitor selective processes due to associations between selected loci and marker loci (Finkeldey and Ziehe 2004). Comparisons of genetic structures at isozyme gene loci between “tolerant” and “sensitive” adult beech (Fagus sylvatica) trees in environmentally stressed forests in Germany suggest a selective response of trees to drastic environmental change (Müller-Starck 1985; Ziehe et al. 1999). The potential for drastic changes of genetic structures due to viability selection without encountering critically low population sizes is highest during early life stages of trees since most trees produce a large number of progenies. Changes of genetic structures were observed during the development of beech seedlings in greenhouses and under field conditions (Ziehe et al. 1999). Viability selection was also observed during early life stages of the tropical tree

7.2 Physiological and Evolutionary Adaptation

Platypodium elegans (Fabaceae; Example 7.1); however, the selection is mainly against inbred progenies (Sect. 6.4.2) and, thus the observed changes of genetic structures essentially result from inbreeding depression (Hufford and Hamrick 2003). These investigations suggest that it is in principle feasible to monitor adaptive processes at putatively neutral markers even in long-living organisms like forest trees. Attempts to identify single genes of adaptive significance (Remington and Purugganan 2003) and to relate genetic variation at these genes to phenotypic trait expressions in field trials and to different environmental conditions at the growing sites of populations are hampered by a large number of “candidate genes” and a poor understanding of their physiological significance for most forest trees. The knowledge of the complete genome sequence of poplar (Populus spp.) (Brunner et al. 2004) and a rapidly growing number of expressed gene sequences in other trees (e.g., Pinus spp., Cryptomeria japonica; Strabala 2004) is an important basis to significantly improve our understanding of adaptive processes in forest trees. A draft genomic sequence of Eucalyptus camaldulensis might become available in 2007 (Poke et al. 2005). The first results on sequence variation in candidate genes of putative adaptive significance have been published (Krutovsky and Neale 2005). Simple markers such as single nucleotide polymorphisms will be developed characterizing the variation in genes with putatively adaptive function in model species and related taxa (Morin et al. 2004; Sect. 2.3.1). Unfortunately, no comparable data are currently available with regard to adaptive changes of genetic structures at single gene loci for most tropical forest trees. Genomic resources will only be available for few tropical tree taxa like eucalypts (Poke et al. 2005) in the foreseeable future. Our current understanding of evolutionary adaptation in tropical forest trees mainly relies on the observation of phenotypic variation in field trials; thus, it is restricted to a rather low number of plantation species, and refers to the phenotypic consequences rather than to the genetic basics of adaptation. Many trials have proved strong differentiation among provenances, families, and clones with regard to important adaptive traits, including survival (Chap. 12). Example 7.1: Viability Selection During Early Life Stages of Platypodium elegans Genetic structures at four hypervariable microsatellite loci were observed at three different early life stages (aborted fruits, mature seeds, and seedlings from the same cohort) of progenies from 12 P. elegans (Fabaceae) trees and a population of adults. The trees grow on Barro Colorado Island, an artificial island in the Panama channel. A slight increase in heterozygosity was observed from mature seeds to seedlings at two loci. The average observed heterozygosity increased from Ho = 0.750 in aborted fruits to Ho = 0.761 in mature seeds,

89

90


Ho = 0.768 in seedlings, and Ho = 0.797 in adult trees. This suggests a slight selective advantage of heterozygosity. The inbreeding coefficient F (Sect. 6.4.1) decreased with increasing observed heterozygosity. The outcrossing rate t was estimated for each cohort on the basis of the observation of nonmaternal alleles (Sect. 6.3.1). It increased from t = 0.75 for aborted fruits to t = 0.78 for mature seeds and t = 0.87 for young seedlings. Comparisons of genetic structures during early life stages of P. elegans revealed selection against homozygotes. However, the observed increase of heterozygosity with increasing age is likely due to the selective disadvantage of inbred progenies and not related to a direct heterozygote advantage at the observed microsatellite loci. Thus, inbreeding depression results in changes of genetic structures due to selection in particular during the phase from mature seeds to the establishment of seedlings. The observed genotypic structures in adults suggest that the selective disadvantage of inbred progenies continues and leaves an essentially completely outcrossed adult population with Hardy–Weinberg structures (Sect. 6.2.2). (From Hufford and Hamrick 2003)

7.3 Species Interactions and Coevolution Plants do not only need to be adapted to their chemical and physical environment, but also to interacting organisms. The biological environment is particularly complex for plant species in species-rich tropical forests. It consists of other higher plants, animals, fungi, and microorganisms which interact either positively (symbioses) or negatively (pests and diseases) with plants. Interactions with animal pollinators are particularly important from a genetic point of view (Chap. 5). The coevolution between tropical figs (Ficus spp.) and fig wasps (Agaonidae) is an interesting and well-studied example of a coevolution between plants and their animal pollinators (Example 7.2). The impact of pests and diseases on the distribution of forest trees and their regeneration has been suggested as one of the main explanations for the high species diversity in tropical forests (Baker 1970; Janzen 1970; Hill and Hill 2001). Breeding for disease resistance is an important goal in tree improvement (Namkoong 1991). It is based on genetic variation with regard to the susceptibility to particular diseases, which is commonly observed in forest trees. The preservation of disease-resistant genotypes is an important objective in conservation of forest genetic resources (Byrne 2000). Symbiotic associations between trees and mycorrhiza constitute an important aspect of biotic interactions in tropical forests and possibly explain the dominance of particular species in some low-diversity tropical forests (Connel and Lowman 1989). Little is known about mutual preferences between genotypes of

7.3 Species Interactions and Coevolution

plants and their mycorrhiza in tropical forests. A comparison of two orchid species revealed different preferences for particular mycorrhizal fungi. Tolumnia variegata is a generalist in its association with mycorrhiza; Ionopsis utricularoides is more specialized with regard to its fungal symbionts (Otero et al. 2004). Example 7.2: Figs and Their Pollinators The pantropical genus Ficus (Moraceae) is one of the most species-rich genera of woody tropical plants, with more than 800 species. The distribution center is in Asia, with more than 500 species. Figs are Keystone Species

Tropical figs produce fruits throughout the year and are an important diet of many animal species. Figs are an indispensable part of the nutrition of many birds and mammals such as bats and primates particularly during periods of low availability of fruits from other plants (Janzen 1979a); thus, figs are regarded as keystone species since their disappearance would have a profound impact on the structure and function of many ecosystems (Terborgh 1986). Case studies proved the crucial importance of figs for ecosystems in particular in Asia (Leighton and Leighton 1983) but not in African forests (Gautier-Hion and Michaloud 1989). Morphology of Figs and Their Pollinators

Figs evolved many different growth habits: free-standing trees, shrubs, climbers, epiphytes, and hemiepiphytes (“strangling figs”). The root systems of strangling figs germinating on a particular host tree tend to fuse; thus, many adult strangling fig trees are composed of several distinct genotypes – they are chimeras (Thomson et al. 1991). Population sizes of strangling figs are often underestimated since a single “tree” may be a small population (Hamrick and Murawski 1991). The inflorescence of figs (syconium) is fleshy and folded inside to form a cave (Fig. 7.1). Many simple, male or female flowers are inserted on the inner wall inside the cave, which is closed by apical bracts leaving a small entrance for pollinating insects, the ostiole (Verkerke 1989). Figs produce two types of female flowers: short-styled “gall” flowers and long-styled flowers. With only a few exceptions (Molbo et al. 2004) each fig species lives in obligatory symbiosis with a single wasp species, which is the sole pollinator (Fig. 7.2); thus, the number of species of fig wasps (Agaonidae) roughly equals the number of fig species. Molecular phylogenies based on chloroplast and mitochondrial genes revealed similarities between the evolutionary relationships among figs and among their pollinators, suggesting a true coevolution and a high stability of the symbiosis (Herre et al. 1996). Similar phylogenies were also found for pollinating fig wasps and associated, nonpollinating wasps (Machado et al. 1996).

91

92


Fig. 7.1. Inflorescence (A) and single male flowers (B1), short-styled gall flowers (B2), and long-styled female flowers (B3) of a monoecious fig (Ficus spp.). (Adapted from Hill 1967a)

Fig. 7.2. Female and male Blastophaga quadraticeps, a fig wasp, the pollinator of Ficus spp. (Adapted from Barth 1982)


The Reproduction Cycle of Figs and the Life Cycle of Fig Wasps

The reproduction cycle of figs and the life cycle of the pollinators are closely connected (Janzen 1979a; Wiebes 1979; Bronstein and McKey 1989; Fig. 7.3). Female wasps are attracted by pheromones and enter the syconia of a “receptive,” female flowering syconium, through the ostiole. They fertilize the long-styled female flowers with pollen brought from another fig of the same species and lay eggs through the style of short-styled gall flowers. Afterwards they usually die within the syconium; only a few wasps manage to leave a syconium they entered (Giebernau et al. 1996). The emerging larvae feed on the developing seeds of their gall flowers, thereby destroying it. The heterostyly within syconia allows the simultaneous development of seeds and larvae of wasps in different flowers of the same inflorescence. A new generation of wasps develops within a few weeks. Male insects appear first, fertilize the females, and die shortly thereafter inside the syconium. The male fig flowers shed their pollen only after the female fig wasps have emerged. Thus, figs are distinctly protogynous, i.e., female flowering is earlier than male flowering. The fertilized female fig wasps take up pollen actively or are passively loaded with pollen. They leave the syconium and start searching for another syconium in the “receptive” phase. The seeds continue to develop after the female wasps have left the syconium. The shape and the taste of a syconium change when the seeds are ripe. The seeds are endozoochorously dispersed by mammals and birds. Fig flower

flower receptive pollination, fertilization

Wasp gall flower developed

wasp

wasp

enters syconium

wasp lays eggs wasp eggs, larvae develop imago

imago

fertilization flowering

pollen uptake seeds

leaves syconium, flight

Fig. 7.3. The reproduction cycle of monoecious figs, and the life cycle of fig wasps

93

94


Flower Phenology

The flowering cycle of a single fig plant is usually strictly synchronized; all syconia of a plant are in the same stage of floral development. Only a few fig species growing in a seasonal climate show a certain degree of individual asynchrony (Cook and Power 1996). Thus, emerging wasps do not find receptive syconia on the same plant but have to search for other trees of the same species. The wasps survive only for a few days outside of syconia. In consequence, asynchronous flowering of the fig plants of a population is a necessary requirement for the preservation of the pollinators (Fig. 7.4). Asynchronous flowering results in the production of fruits throughout the year. The importance of figs as a more or less continuously available energy source for vertebrates is explained by the need to produce flowers as breeding grounds for their pollinators, and hence also fruits, throughout the year. Gynodioecious Figs

The previous description of the life cycle of figs refers to monoecious species. The eatable fig F. carica and many Asian fig species are gynodioecious (Berg 1989). In these species, the cycle for the preservation of the pollinating wasps is confined to male and female flowering plants. All female flowers in these plants are short-styled gall flowers; hence, the pollinators are able to lay eggs in all female flowers, and no mature seeds develop in this type of syconia. In consequence, gynodioecious figs are functionally dioecious.

...

...

flowering phase development flowering phase

wasp flight crop abortion wasp extinction

Fig. 7.4. A hypothetical example on the phenology of figs and its impact on pollinating wasps. (Adapted from Bronstein et al. 1990)


If a wasp enters the syconium of a solely female flowering plant, the flowers are fertilized by the pollen carried by the insect. All female flowers in these syconia are long-styled; thus, the wasp cannot deposit its eggs and seeds develop from all fertilized flowers. Since the wasp is irreversibly trapped in the syconium, it will not be reproductively successful if it enters a solely female flowering inflorescence without gall flowers (Fig. 7.5). Genetic Consequences of the Symbiosis

Self-pollination of monoecious figs is usually excluded owing to the strong protogyny and the synchronized flowering of single plants. Asynchronous development of a few syconia is observed in some species (Cook and Power 1996). It helps to preserve the pollinators, but is likely to result in selfed offspring of figs. Gynodioecious figs are functionally dioecious and hence obligatory outcrossing.

Fig. 7.5. Reproduction of gynodioecious figs. (Adapted from Hill 1967b). Further explanation in the text

95

96


Mating preferences among figs are likely to be strongly influenced by their complicated flowering phenology, and to a lesser degree by spatial proximity as long as the distance can be bridged by pollinators. Investigations of seven monoecious strangler fig species in Central America revealed that their seeds were fertilized by a surprisingly high number of pollen parents (Nason et al. 1996, 1998). Owing to the low density of figs, the average movement of pollen was estimated to be between 5.8 and 14.2 km. Breeding units are as large as 106–632 km2, which is a considerably larger area than for any other plant species investigated (Herre 1996). The surprisingly efficient long-distance transport of fig pollen is possibly facilitated by passive wind-mediated dispersal of fig wasps flying high above the canopy (Compton et al. 2000). The wasps are likely to be attracted by volatile chemicals if they come close to a receptive fig tree. These results indicate that fig populations can be maintained in fragmented landscapes as long as the pollinators are capable of bridging the fragments; thus, even widely isolated remnants of forests may be important for the maintenance of figs and their pollinators. The potential for local adaptation of figs to specific sites is limited owing to the effective gene flow and migration by pollen and seeds. The genetic structures of consecutive-age cohorts of single seed trees are likely to be strongly differentiated from each other owing to changing pollen parents. The flowers of a single syconium are frequently fertilized by a single wasp only (Janzen 1979b). The progenies produced in such a syconium are full-sibs (Nason et al. 1998). The sex of fig wasps is, like the sex of most Hymenoptera, determined by a system of haplodiploidy. Male insects develop from unfertilized eggs and are consequently haploid (Godfray 1988). Female insects are able to control the sex ratio by facultative admittance of sperms to fertilize their egg cells (Frank 1985). Female fig wasps are fertilized before they leave the syconium; thus, the number of females entering and laying eggs in a syconium is crucial for the degree of genetic relatedness of the offspring produced. Matings between siblings are frequent since syconia are often entered by only a single wasp and rarely by more than four wasps. Inbreeding is frequent in wasp populations (Molbo et al. 2004). The very important conservation of tropical figs is only possible if their pollinators are also preserved. Minimal population sizes for the preservation of fig populations should be based on models which allow the population dynamics of the pollinators as the most important criterion to be assessed (Bronstein et al. 1990). Actual population sizes critically depend on phenology and the migration behavior of female fig wasps. Experimental results suggest


conserving even widely isolated fig trees which possibly belong to viable populations owing to the high efficiency of fig wasps as pollinators.

7.4 Recommended Literature The concept of physiological adaptation or phenotypic plasticity of plants has been reviewed by Schlichting (1986, 2002). Evolutionary adaptations and the role of genetic variation for adaptability have been analyzed by Dobzhansky (1968) and Dobzhansky et al. (1977). Plant–pollinator interactions in tropical forests were reviewed by Bawa (1990). The reproduction of figs was concisely described by Janzen (1979a).

97

Phylogenies and Evolution Above the Species Level

The wealth of species in humid tropical and subtropical forests has always fascinated botanists – and not only taxonomists but also ecologists, geneticists, and students of evolution. Much research remains to be done in these forests,. . . (p. 210 in Stern and Roche 1974).

8.1 Introduction One of the most obvious and fundamental differences between tropical and temperate forests refers to their floral diversity and in particular the diversity of tree species (Whitmore 1975). The number of tree species in a single hectare in a tropical lowland forest is often considerably higher than the number of trees native to central Europe. Public and scientific discussions are often focused on the conservation of this diversity rather than its evolution (Lugo 1988; Raven 1988; Sayer and Whitmore 1990). Conservation of biological diversity continues to deserve the highest priority in view of the undiminished speed of forest destruction in many parts of the tropics (FAO 2002; http://www.fao.org/forestry/site/fra/en). However, improved knowledge of the evolution of the high species diversity in tropical forests and its maintenance in undisturbed forest ecosystems will help us to develop sound strategies for its conservation.

8.2 The Evolution of Species Diversity in the Tropics 8.2.1 Species Diversity in Tropical Forests

The enormous floral diversity was noticed during early expeditions to all tropical regions and quickly became a field of intensive research (e.g., von Humboldt 1808). The rich biological diversity in the tropics stipulated evolutionary

8

100

C HAPTER 8 Phylogenies and Evolution Above the Species Level

thinking of researchers during the nineteenth century (Lefèvre 1984). Alfred Wallace and Charles Darwin independently developed the theory of evolution by natural selection after personal observations of the diversity of life in the tropics. Floristic diversity is not evenly distributed among continents and regions. The overall number of flowering plants is considerably higher in tropical America (approximately 90,000 species) than in Asia (approximately 40,000 species) and Africa (approximately 35,000 species; estimates from Thomas 1999). Regional centers of floral diversity are presumably related to the climatic variability throughout the Quaternary (Cook et al. 1990) and possibly represent refugial areas during eras with a less favorable climate (see later). The identification of regional centers of species diversity is of obvious importance for the development of conservation strategies (Lugo 1988; Wilson 1988; Gentry 1992). Variation of species diversity has also been described at the local level by detailed floristic inventories. Extraordinary high numbers of tree species per hectare were counted in local inventories in the Neotropics (Gentry 1986) and in tropical lowland forests in Southeast Asia (p. 5 ff. in Whitmore 197). Simulation models predict a positive correlation between levels of genetic diversity within species (Sect. 3.4) and species diversity (Vellend 2005). However, no empirical studies have been conducted in tropical forests, and it is unknown at which spatial scale correlations possibly exist. Even the type of the expected correlation is a matter of speculation: Refugial areas might be hotspots both of species diversity and of diversity within species. On the other hand, a high diversity of tree species implies low population densities for most species, and genetic variation may be lost owing to drift, if low population density results in low population sizes. Not all natural tropical forests are characterized by a high species diversity. The number of species declines with decreasing temperature in mountainous areas and extended dry seasons in marginal regions of the tropics (Richards 1981). Azonal edaphic conditions also frequently result in decreasing species diversity (Whitmore 1975). Forests with a low diversity of tree species occur even in the humid central tropics (Hart et al. 1989; Hart 1990). Biotic interactions possibly play an important role for their existence and stability (Connel and Lowman 1989). The floral diversity represents only a small part of the overall biodiversity in tropical forests. Estimates concerning the number of existing species on earth vary greatly since only a small proportion of the most species-rich group, the insects, have been described (Adis 1990). The diversity of insects in the canopy of tropical forests constitutes the biggest share of global species diversity (Erwin 1988).

8.2 The Evolution of Species Diversity in the Tropics

8.2.2 Evolution Above the Species Level

The floral diversity of tropical forests is fairly well described; however, its evolution and the mechanisms ensuring its maintenance are only poorly understood. The most important theories trying to explain the species diversity of tropical forests were reviewed by Baker (1970) and, more recently, by Hill and Hill (2001) and Montagnini and Jordan (2005, p. 27 ff.). Most of these theories are chiefly concerned with the ecological requirements for the coexistence of many species in a given area (“ecological niches”, environmental heterogeneity, disturbance) and species interactions, including competition and predation. Two hypotheses with a focus on the factors ruling the emergence of the high species diversity in tropical forests are concerned with the duration of periods of an “undisturbed” ecosystem development. A long period of “cumulative” evolution undisturbed by annual periods of unfavorable environmental conditions such as low temperatures and long dry seasons is regarded as an important prerequisite for the diversification of species. A more or less constant climate for extended periods of millions of years is likely to be even more important (“museum model”). The evolution of angiosperms in tropical forests lasted at least since the beginning of the Tertiary for approximately 65 million years.“It is likely that in some at least of the regions where Rain forest is now the climatic climax, similar vegetation has existed uninterruptedly since a very remote geological period.” (p. 14 in Richards 1981). Thus, even a slight excess of newly evolving species in comparison with presumably low species extinction rates resulted in the accumulation of many species during this long era. According to Richards (1981, p. 16), the floristic richness of tropical forests is largely due to its great antiquity. Multiple evidence suggests a strong impact of the climatic variability during the Pleistocene not only on the flora in temperate regions, but also on tropical rain forest vegetation (Haffer and Prance 2001). “Allopatric” speciation in isolated refugia during dry periods possibly increased species diversity since populations isolated from each other in separate refugia eventually developed reproductive barriers (Hill and Hill 2001); however, the size and the location of pleistocene refugia are under dispute. The fossil pollen record suggests stability and continuity of the forest cover over the Amazonian lowlands (Colinvaux and De Oliveira 2001). Sequence variation patterns at the mitochondrial cytochrome b gene in different mammals from North America and Amazonia suggest rapid population growth for the North American species after the end of the last glaciation, but no or only minor demographic expansion for small mammals in Amazonia (Lessa et al. 2003). This result also suggests only a limited impact of climatic variability on the population sizes of small mammals in

101

102


the Neotropics and hence contradicts the refugia hypothesis. However, a molecular phylogeny of the species-rich genus Inga suggests a recent speciation, possibly as a consequence of the climatic variability during the pleistocene (Richardson et al. 2001; Example 8.2). Ecological research on the dynamics of tropical forests, including their regeneration, greatly profited from the establishment and repeated analysis of intensively studied plots on Barro Colorado Island (Panama), Pasoh Forest Reserve (Malaysia; Lee 1995), and in other tropical regions (Hubbell and Foster 1992). Numerous genetics studies were conducted in these plots; several of them are described in this book (Table 3.3, Fig. 6.2; Examples 5.2, 5.3). However, the focus of these investigations was on the maintenance of species diversity, the ecological and genetic consequences of a high species diversity, and the role of species interactions in this context. The limited time for observations does not permit the evolution of species diversity for long-living organisms like forest trees to be studied. Taxonomic studies are the most important basis for the description of the biodiversity in tropical forests. The relationships among existing species as revealed by the differentiation of morphological traits of taxonomic significance offer opportunities not only to classify taxa, but also to understand their origin. During the past few decades, traditional taxonomic surveys were complemented by molecular approaches to reveal phylogenies. Early caryological studies used chromosome polymorphisms and in particular polymorphisms of chromosome numbers to infer relationships and the evolution of plants in the tropics (Mehra and Bawa 1969). More recently, molecular genetics markers based on variation of DNA became commonly applied tools not only to reveal genetic variation within species (Sect. 3.4), but also to clarify phylogenetic relations among more or less closely related taxa occurring in tropical forests (Sect. 8.3).

8.3 Molecular Phylogenies The evolutionary history or phylogeny of a group of organisms can be analyzed by investigating similarities at particular characters, which possibly reflect relatedness or a common origin. Thus, if two taxa A and B are more similar to each other than both are to taxon C, a more recent common origin and, hence, a higher degree of relatedness of taxa A and B may be assumed in comparison with taxon C. Similarities need to be measured at particular traits or characters, and the observation of different traits often results in contradicting conclusions. Various reasons can account for similarity apart from a common origin. Convergent evolution in very distantly related organisms eventually results in similarities at morphological traits in response to particular environmental conditions. Frequently cited examples for plants are a herbaceous growth habit,

8.3 Molecular Phylogenies

which repeatedly evolved from woody ancestors, dioecy (Sect. 4.2.3), or leafless plants in adaptation to a dry environment in Neotropical Cactaceae and Euphorbiaceae of the Old World. Thus, similar trait expression may not only be due to a common origin (homology), but characters may also be similar although they are not derived from a common ancestor (homoplasy). The observation of variation at molecular traits offers obvious advantages in comparison with morphological characters. This applies in particular to variation of DNA, which is not influenced by environmental conditions. Thus, some of the methods briefly described in Sect. 2.3.1 are not only commonly used to investigate genetic variation within species, but are also used to clarify differentiation among species and to elucidate evolutionary relationships by phylogenetic analyses. The observation of variation at nonrecombining chloroplast DNA (cpDNA) is of particular importance for plants. Low mutation rates of cpDNA are responsible for, in general, low variation within species, but ample and easily interpretable differentiation among species at various levels. Sequences of cpDNA were used to investigate phylogenetic relationships among plants at virtually all taxonomic levels. Variation of cpDNA is informative to study splits during the earliest evolution of land plants (Goremykin and Hellwig 2005), and also to reveal phylogenetic relations within single genera (Gernandt et al. 2005). Phylogenies are not only based on cpDNA, but also on nuclear DNA (nDNA) and anonymous markers. Commonly used nDNA sequences for phylogenetic analyses are ribosomal DNA internal transcribed spacer (ITS) regions. For example, molecular phylogenies of the pioneer tree genus Macaranga (Euphorbiaceae) were based on ITS sequence data (Blattner et al. 2001) and amplified fragment length polymorphisms (Bänfer et al. 2004). The evolution and phylogeny of flowering plants has recently been reviewed largely based on molecular data (The Angiosperm Phylogeny Group 2003). Two approaches are commonly used to construct phylogenetic trees based on DNA variation (Hall 2001). The most important algorithmic methods are the unweighted pair-group method with arithmetic means (UPGMA) and the neighbor joining (NJ) method. Both are distance methods based on the computation of a matrix of all pairwise differences between samples (Sect. 3.3.2). Parsimony is the most commonly used tree-searching method. Many trees are constructed and compared. The most likely (maximum parsimonious) tree is the one requiring the lowest number of evolutionary steps, including homoplasies, to explain the data. The reliability of phylogenetic trees is often statistically tested by bootstrapping (p. 58 ff. in Hall 2001). Example 8.1: A Molecular Phylogeny of Indonesian Dipterocarpoideae The family Dipterocarpaceae consists of three subfamilies; the Asian Dipterocarpoideae is by far the most important and most diverse subfamily,

103

104


consisting of about 470 species in 13 genera. Nine genera with more than 250 species are native to Indonesia. The Dipterocarpoideae are commonly divided into two tribes (Ashton 1982). Tribe Dipterocarpae, with a basic chromosome number of x = 11, consists of the genera Anisoptera, Cotylelobium, Dipterocarpus, Upuna, and Vatica. Tribe Shoreae (x = 7) consists of Dryobalanops, Hopea, Parashorea, and the species-rich genus Shorea. Genus Dryobalanops has a basic chromosome number of x = 7, but shows intermediate characters between the tribes concerning flower morphology (sepal aestivation) and wood anatomical features (vessels, resin canals). Variation of cpDNA was studied by means of polymerase chain reaction restriction fragment length polymorphisms and chloroplast simple sequence repeats (Sect. 2.3.1). A total of 129 samples from 58 species were investigated. The African dipterocarp Monotes kerstingii (subfamily Monotoideae) was used as an outgroup to root the phylogenetic tree (Indrioko et al. 2006). No variation was observed within species. The differentiation among species at cpDNA is in good agreement with the classification based on morphological characters (Fig. 8.1). One major clade consists of Upuna, Cotylelobium, Anisoptera, Vatica, Dipterocarpus (tribe Dipterocarpae), and Dryobalanops (tribe Shoreae) in a basal position. A second clade consists of Hopea, Parashorea, and Shorea (tribe Shoreae). The genera are, in general, well resolved and strongly supported. The intermediate position of genus Dryobalanops is confirmed by the analyses. Parashorea groups with Shorea in a single big, poorly resolved clade. Both genera have close botanical affinities and the phylogenetic relevance of their distinction is questionable. Similar results were obtained by previous studies based on cpDNA variation of dipterocarps (Tsumura et al. 1996; Kajita et al. 1998; Kamiya et al. 1998; Gamage et al. 2003). Phylogenetic analyses based on cpDNA failed to consistently group the species-rich genus Shorea into well-supported clades. However, the variation at the nuclear PgiC gene locus is consistent with the distinction of different timber groups (balau, red meranti, white meranti, yellow meranti) and sections according to Ashton (1982) within genus Shorea (Kamiya et al. 2005). (From Indrioko et al. 2006) Example 8.2: Evolution of the Genus Inga Molecular phylogenies can be used to test the alternative hypotheses of speciation according to the “museum model” or recent speciation in response to environmental variation during the past few million years (Sect. 8.2.2). The slow evolution of species diversity as a result of a long period of favorable climatic conditions (museum model) is expected to result in a well-resolved phylogeny, each species distinguished from others by numerous mutations (nucleotide substitutions at the level of DNA). Recent allopatric speciation in

8.3 Molecular Phylogenies

100

67 95 63

Cotylelobium lanceolatum Upuna borneensis Anisoptera costata Anisoptera marginata 99 Anisoptera reticulata 68 Vatica bantamensis Vatica bella Vatica granulata 99 83 Vatica pauciflora Vatica rassak 94 Vatica venulosa Dipterocarpus grandiflorus Dipterocarpus oblongifolius 71 Dipterocarpus retusus 100 Dipterocarpus rigidus Dipterocarpus tempehes Dryobalanops aromatica 99 Dryobalanops lanceolata Hopea bancana Hopea celebica Hopea odorata 85 Hopea sangal Hopea dryobalanoides Hopea nigra Hopea griffithii Hopea mengarawan 66 Parashorea lucida Parashorea globosa Shorea blumutensis Shorea guiso Shorea montigena Shorea scaberrima Shorea seminis Shorea acuminata Shorea andulensis Shorea javanica Shorea johorensis Shorea leprosula Shorea macroptera Shorea mecistopteryx 69 Shorea ovalis Shorea palembanica Shorea parvifolia Shorea platyclados Shorea xanthophylla Shorea balangeran 59 Shorea selanica Shorea splendida Shorea macrophylla 55 Shorea pinanga Shorea stenoptera Shorea acuminatissima Shorea dasyphylla Shorea faguetiana 77 Shorea multiflora Shorea fallax Shorea materialis 98 60 Shorea virescens Monotes kerstingii

x=11; valvate

Dipterocarpeae

sub valvate

x=7; imbricate Shoreae

valvate

x=7

Fig. 8.1. Strict consensus tree of Dipterocarpoideae (outgroup Monotes kerstingii) Numbers at nodes indicate bootstrap values. Typical characters (valvate, subvalvate or imbricate sepal aestivation; basic chromosome number x) and tribes are indicated. (From Indrioko et al. 2006)

refugia should be reflected in few nucleotide substitutions differentiating species, and a poorly resolved phylogeny (Fig. 8.2). The genus Inga (Fabaceae; Mimosoideae) contains about 300 species growing in neotropical rain forests. Sequence variation was observed for 32 species at the ITS of nuclear ribosomal DNA and for 31 species at a cpDNA fragment. A total of 45 species representing most sections of the genus were analyzed (Richardson et al. 2001). Only a few substitutions were observed and the phylogenies were unresolved, conforming Fig. 8.2b. The result is consistent

105

106


Fig. 8.2. a A well-resolved phylogeny due to the accumulation of mutations over long periods and b a poorly resolved phylogeny due to a low number of mutations differentiating species (a–h) indicating recent speciation. (Adapted from Richardson et al. 2001)

a b c d e f g

a

h a b c d e f g

b

h

with a recent speciation within the genus Inga. Rough estimates of the divergence time suggest that speciation occurred within the last ten million years, and approximately 30% of the species diverged from their closest relative during the Pleistocene (within the last two million years). Shorea (Dipterocarpaceae) contains only slightly fewer species than Inga. The molecular phylogeny within the genus Shorea is also poorly resolved (Example 8.1), likewise suggesting comparatively recent speciation within this genus, at least in Southeast Asia. However, the diversity of tropical forests does not only result from a few species-rich genera like Inga, Shorea, Ficus, and others, but is also reflected in a large number of genera, families, and higher taxonomic units present in most tropical forests (Bermingham and Dick 2001). Thus, the species diversity of tropical forests is presumably a consequence of both their great antiquity (museum model) and recent speciation for certain taxa like Inga. (From Richardson et al. 2001)

8.4 Recommended Literature The causes of the species richness of many tropical forests were reviewed by Baker (1970), Hill and Hill (2001), and Montagnini and Jordan (2005, p. 27 ff.).


Hart (1990) pointed out that not all tropical forests are characterized by a high species diversity. Phylogenetic methods are described among other topics related to phylogeography in the book by Avise (2000). Their application to plants was reviewed by Schaal et al. (1998). An update of the classification of the orders and families of angiosperms largely based on molecular phylogenetic methods has recently been given by The Angiosperm Phylogeny Group (2003). The rationale behind molecular phylogenies and the commonly used software PAUP for phylogenetic analyses are described in the book by Hall (2001).

107

Applications of Genetics to Tropical Forestry

Part B

The majority of tropical tree species possess high genetic diversity (Sect. 3.4). The elements of their genetic systems that are responsible for the evolution of this genetic variation and its temporal and spatial dynamics under natural conditions were the topics of Part A. Part B is oriented towards application. It concentrates on both genetic implications of human activities and on planned utilization of genetic variation. Tropical forests are being cleared, burned, fragmented, logged, and overhunted at rates that lack historical precedent (Laurance and Bierregaard 1997a). The consequences of the large-scale conversion of forests to other types of land use and the fragmentation of contiguous forests require analysis and the development of guidelines for appropriate treatment of remnant forests (Chap. 9). Forestry involves logging and the silvicultural management of diverse types of forests. Silvicultural management has far-reaching consequences for the genetic system of tree species and their genetic structures (Chap. 10). Tree breeding including provenance selection (Chap. 11) aims at changing genetic structures of populations in order to increase the plantation yield. Tree improvement and planting activities in tropical forestry commenced a long time before population genetic research on the genetic system of tropical forest trees was taken up (Chaps. 12, 13). Last but not least, in view of the genetic impact of human interference, projects on the conservation of tree genetic resources are an imperative application of genetics. Genetic information is not a renewable resource. These projects require concomitant research in order to be successful (Chap. 14). In most tropical countries forest genetics plays only a minor role in forestry education. In most curricula emphasis is put on tree improvement, if subjects related to genetics are discussed at all. In courses concentrating on breeding, students become acquainted with working methods of quantitative genetic research and relevant statistical methodology. Basics of population genetics, an analysis of the genetic systems of tropical forest trees, the role of genetics in silviculture, and the conservation of forest genetic resources have conventionally been regarded as topics of minor importance as compared with tree improvement. Not too much at variance from the conditions in the temperate zones,

110

PART B Applications of Genetics to Tropical Forestry

forest genetics research is occasionally used as a synonym for the formulation of breeding goals and the conduct of breeding projects. Owing to this misconception, the fundamental importance of an analysis of the genetic system of tropical forest trees by use of population genetics methodology is overlooked. Nonetheless, tree improvement is but an application of forest genetics and must be accompanied by the conservation of forest genetic resources.

Fragmentation of Forests

9.1 Introduction Following the advance of civilization, large forests are completely eliminated and converted to other types of land use on all continents. However, small pieces of forest are left on areas that are of little use owing to their adverse topography or drainage, or other reasons. They are embedded in a largely agricultural matrix (James et al. 1998). They are also looked at as islands surrounded by a sea of modified habitats (Laurance and Bierregaard 1997a). Darby (1956) described the history of the reduction of forests in Europe and documented their increasing fragmentation on a continent-wide scale. When discussing the human impact on forests, Ledig (1992) and Ledig and Kitzmiller (1992) listed fragmentation of forests as next in line after complete destruction (see also Clarke and Young 2000). Habitat fragmentation of trees and other living organisms as an almost inevitable consequence of large-scale forest loss has emerged as a global concern. The size distribution of the fragments, their spatial pattern in landscapes, and the proportion of the area covered by forest affect both the total number of persisting species and the relation between them. Along with the forests, the distribution ranges of tree species and their populations are fragmented. Depending on the distribution pattern of species, fragmentation of forests may disrupt the range of one species but not so of others. This condition is encountered not only in the tropics but also in other climatic zones; however, in species-rich forests of the tropics it is of particular concern to ecologists and forest geneticists. Owing to the species richness particularly of tropical rain forests with a high proportion of endemics (Viana et al. 1997), the number of species being strongly affected or becoming extinct is larger than in other biomes. Owing to local rarity or pronounced clumping of species occurrence, many species may be missing from any particular fragment simply because they never occurred there (Laurance et al. 2002). The same applies to genotypes within species. Also, many species have disappeared even from larger fragments because they are sensitive to disturbance.

9

112

C HAPTER 9 Fragmentation of Forests

Another condition often goes along with fragmentation. Owing to raised pressure of the local human population on those remnants of forest, the trees of the same species occur with reduced overall density. They also have a patchy distribution, because the original density of trees varied with the species and with microsite variation in a heterogeneous environment. Locally, the density may be dramatically reduced. The landscape may not be completely denuded of forest but remaining trees may grow in rows along roads and watercourses or as solitary trees on pastures or abandoned fields. The question arises, to what extent these trees form possibly persistent populations in the genetic sense (Sect. 2.2). The implications of low population density and restricted population size for the reproduction system of trees were discussed in Chaps. 5 and 6. In the present chapter, the very small remnant populations of reduced density are addressed also from the management point of view. The conversion of large areas of contiguous tropical forests to other land uses leaves forest fragments of various sizes. For instance, Sanchez-Azofeifa et al. (1998) reported on deforestation in lowlands of Costa Rica, where more than 80% of the primary forests had been eliminated since 1984. During a 10-year period the annual loss rate was more than 2% and the number of forest fragments doubled, while their average size decreased from 95 to 25 ha. The total area of fragments of less than 100 km2 size or less than 1 km away from an edge in lowland rain forest in Brazilian Amazonia increased from 20.8 × 106 ha in 1978 to 50.8 × 106 ha less than 20 years thereafter (Whitmore 1998). These exemplary data indicate that fragmentation in a largely deforested country is still in progress on many spatial scales and will presumably lead to fewer and smaller fragments and to the extirpation of many plant and animal species. The genetic impact of this process will become ever more severe. While considering the genetic effects of isolation, it must not be overlooked that the amount of available habitat is greatly reduced. The smaller the fragments, the more important become edge effects that penetrate deep into the forest remnants (Laurance et al. 2002). In a case study of the decline of deciduous and semideciduous tropical forest in Bolivia, the increase in the number and in the area of both large and small fragments is documented over a 20-year period. The microclimatic edge effects such as increases in light, temperature, and wind speed affect larger part of the remnant forest (Steininger et al. 2001). The mortality among trees increases (Laurance et al. 2002). In an early paper on forest fragmentation, Curtis (1956) documented the clearing of forest in a certain part of Wisconsin, USA, during a 120-year period. This author pointed out the decrease of forest area, the initially increasing and subsequently decreasing number of fragments, their average size, their distance from one another, and their total periphery. He also described the various ecological correlates of this man-induced process. Loyn and MacAlpine (2001) analyzed different types of forest fragmentation and found their ecological impacts somewhat different.

9.2 The Genetic Status of Fragmented Tree Populations

9.2 The Genetic Status of Fragmented Tree Populations With some time-lag, the extinction of at least some tree populations or species in fragmented tropical forest is likely in view of their reproduction system. Most tropical trees are characterized by predominant outcrossing (Sect. 6.3), strong inbreeding depression (Sect. 6.4.2), and reproductive strategies relying on the presence of other tree and animal species (Ashton 1981). As already indicated, if a forest has been fragmented by human action, its tree populations are suddenly put into an environment to which they are not necessarily fully adapted; therefore, it is the formerly widespread and common plant species that suffer most from (complete) fragmentation of their populations (den Njis and Oostermeijer 1997). The remnant forests possess less biodiversity. They represent disturbed ecosystems comprising merely part of the tree species belonging to the respective forest community. The moderate numbers of individuals of the former keystone species form populations of reduced reproduction-effective size. Even if tree species were intimately mixed, a certain minimum area is required for all species to occur. Consequently, many species are locally lost if the respective fragment is small. The same is true with genotypes belonging to one species. The relict populations are in danger of lacking formerly existing genetic variants due to genetic drift and those losses are no longer compensated by the influx of external effective pollen. Isolation is likely to result in increased genetic differentiation among relict populations owing to nonadaptive changes of genetic structures. It must be emphasized that reduced genetic variation in small populations is presumably a symptom of endangerment rather than its cause (Namkoong et al. 1996; Holsinger 2000). With progressing disturbance of forest ecosystems coadapted other species of plants and of animals are eventually no longer present. Pollen vectors may not survive or may no longer visit remote forest fragments (but see Example 8.2). However, some dispersers of pollen and seed respond positively to forest clearings and to the edge effects implied by fragmentation (Laurance et al. 2002). Pollen vectors support a critical portion of the life cycle of tropical trees and must have a permanent food basis. It is interrupted, if the flower rewards are not offered all-year round. If also the seed vectors have become rare or absent, the occupation of patches of the former habitat becomes increasingly difficult for tree species. First of all, rare widely dispersed tree species will suffer from altered pollinator–plant assemblages (James et al. 1998); therefore, one has to anticipate complex, cryptic responses to fragmentation (Aldrich and Hamrick 1998). The trees in those relict forests are less adapted to the new environment, since they have evolved in obligate associations with groups of pollen and seed dispersers that either are now missing or have greatly changed.

113

114


The decline of dispersers is followed by the decline of the respective tree species. Changes in the guild of pollinators are not just responsible for pollen limitation and lower seed set. As an environmental factor influencing the mating system (Sect. 6.6), they also affect genetic structures of the next generation of trees. Adaptive changes of genetic structures are induced but they may require generations until the former level of adaptedness is reached; hence, population fragments are genetically vulnerable. In one example from New Zealand, a shift in the guild of pollinators was not accompanied by consistent changes in the mating systems of the partially self-fertilized tree Metrosideros excelsa (Schmidt-Adam et al. 2000). Other birds with similar behavior have presumably taken over the role of two other species of pollinators. As Whelan et al. (2000) found in several species of the protaceous genus Grevillea in fragmented Australian populations, there may be strong interactions between pollinator guilds such as vertebrates and insects depending on the natural or “preferred” mating system of trees. In the completely self-fertile species G. macleayana the proportion of self-fertilization ranged between 20 and more than 90%. In this type of species, changes due to fragmentation are expected to be less intensive. However, the majority of tropical tree species are all but fully self-fertile. Introduced honey bees eventually change the realized mating systems and the concomitant mating success under the now given conditions. These insects have largely displaced honeyeaters as pollinators from some of their food plants, thus contributing to their decline and leading to pollen-limitation in Grevillea spp. and other species (Paton 2000). The disruption of mutualisms between trees and their native pollinators is sometimes compensated by the genetic rescue of remnant trees through introduced African honey bees. They outnumbered native pollinators of the canopy-emergent tree Dinizia excelsa (Fabaceae) in disturbed Brazilian habitats and were the only pollinator of a lone pasture tree of this species, transferring pollen from large distances (Dick 2001). Almost 80% of the canopy-forming trees in Samoa have been found to depend on flying foxes as dispersers of pollen or seed. It is still unknown whether the role of these declining animal species with their vast foraging area can be filled by introduced organisms (Cox and Elmqvist 2000). Studying insect-pollinated perennials in a South African shrub community, Donaldson et al. (2002) reported complex reactions to fragmentation. Ecological characteristics of fragments rather than their size and their distance to a contiguous shrub community had a stronger overall impact on pollinator diversity and plant reproductive success; however, the conditions varied greatly between plant species possessing different reproduction systems. The remaining pollen vectors are able to transfer only reduced amounts of pollen among the trees. Eventually, the distances between trees approach the range of the feeding flights of the pollinators, so the mating contact among the trees is impeded. It may even be interrupted, if tree distances exceed the


activity radius of pollinators. Fragments have then lost their connectivity. Historically old population isolation in Caesalpinia echinata Lam. in coastal Brazilian forest has led to reduced variation in local populations and their strong differentiation at maternally inherited markers (Lira et al. 2003; Sect. 5.4). Depending on the composition and the behavior of the pollinator guild, the amount of remote pollen transferred inevitably, though not continuously, decreases with increasing distance. Consequently, also the remnant groups of trees within the forest fragments are at least partially isolated. They can no longer be considered coherent populations of the type that once prevailed on the respective sites. Reduced mating contact also within population fragments implies an increase in self-pollination. Unless there is no absolute barrier against self-fertilization, some of the arising offspring show inbreeding depression, a decline of fitness in terms of retarded or reduced fertility, reduced growth, and reduced longevity. If also the seed vectors have become less abundant, the trees in a neighborhood that are potential mating partners tend to be related. This biparental inbreeding is to be expected wherever the dispersal of pollen and seed is not fully effective. The genetic effects described are more or less severe depending on the reproduction system of the respective species. In low-density populations or low-density patches the seed shadows of individuals possibly do not overlap, so neighboring seedlings are more closely related, because they have at least their seed parent in common. This family structure is less likely to arise if seed dispersal is very efficient. An initial spatial structure eventually fades with progressing life stages owing to mortality and selection. In an ecological sense, a tree is solitary if it has no neighboring trees as competitors; however, it is not necessarily isolated in the genetic sense. Contrary to human perception, spatial distance to the nearest neighbors does not necessarily reflect their actual gene flow pattern. If a solitary tree produces seed by cross-fertilization or contributes pollen to the progeny of other trees, it is part of their population. A solitary tree that does not reproduce in either way might still possess a vital ecological function, since its flowers support pollinator movement between trees. Tall free-standing intervening trees may also serve as roosts for fruit dispersers and thus direct seeds to nearby patches (Nason et al. 1997; Galindo-González et al. 2000; Dick 2001). From the second generation following fragmentation, another source of inbreeding is likely to occur depending on the reproduction-effective sizes of population fragments. These inbreeding effects accumulate if the reproductive isolation of small fragments persists for several generations (Ledig 1992). Since not all members of a population contribute balanced numbers of gametes to the produced progeny, the effective number of individuals is reduced to less than their census number. The size of the parental pool may even drastically differ from the census number. With help of highly variable genetic markers at

115

116


least the effective size of the parental pool relative to the census number may be estimated. The diminished size is due to imbalances in the production of effective gametes (review by Hattemer in 2006). Stern and Gregorius (1972) have shown that in monoecious populations a further reduction is due, since the individuals may produce gametes of either sex, which represents a further source of kinship. Depending on the degree of the imbalance of gamete contributions and the sexual function of the individual trees, the effective number is more or less strongly reduced. The reproduction-effective population size has two aspects, as Kjær (1996) discussed with reference to tree populations. For instance, a population fragment of 50 trees with an inbreeding-effective size of half the census number is expected to produce the same number of inbred progeny as a panmictically reproducing population of size 25. Furthermore, a collective of 50 trees with a variance-effective number of half the census number is expected to produce progeny with the same amount of nonadaptive gene frequency change as a panmictically reproducing population of size 25. Although the assumption of panmictic reproduction raises the impression of being highly unrealistic, the panmictic population still represents a useful reference (Sect. 6.2.2). Depending on the number of progeny produced, the variation in the effective number of female and male gametes, and the proportion of self-fertilization, the two types of effective population size may differ but they are mostly less than the census number. Kjær (1996) indicated also methods of estimation. The theory of the two parameters introduced was explained and reviewed under a range of circumstances by Caballero (1994). These parameters possess relevance also for breeding and are addressed in Chap. 13. A situation where presumably both parameters are drastically reduced is described in Example 9.1. Example 9.1: Genetic Impact of Fragmentation on a Rain Forest Canopy Tree Aldrich and Hamrick (1998) and Aldrich et al. (1998) studied the Neotropical Symphonia globulifera (Clusiaceae), a late-successional, bird-pollinated canopy tree of Costa Rican rain forests. On the study site, a population had been fragmented 10–30 years before. The conditions in remnant forest patches and on pasture were compared with those in continuous forest at a control site. Seed set was low owing to low pollination, high abortion rates, and rapid removal by frugivores. Seedlings did not recruit on pasture but had higher density in remnant than in contiguous forest. The proportion of seedlings arising from self-fertilization increased from about 1% in continuous forest to 11% in remnant forest and to even 25% in pasture trees. The few pasture trees had produced more of the younger than the older progeny. This indicates a sequential reduction in the number of reproductively effective trees. A variance-effective population size of Symphonia globulifera of 17 was estimated both in continuous and in remnant forest; however, accounting for migration led to estimates of 9.4 in continuous


forest and 2.8 in fragmented forest owing to the lower seed migration rate in the latter. These estimates are extremely low. The reproductive imbalance becomes even clearer if also the respective area sizes are accounted for: the effective population size was much smaller on 38 ha of remnant forest than on only 1 ha of continuous forest. Parentage analysis by means of microsatellite markers revealed that in remnant forest by far most of the progeny were produced by a few large trees left on nearby pastures. The large reproductive success of the pasture trees was positively correlated with their increased crown size owing to the release from competition following fragmentation. This was interpreted by a shift in the foraging pattern of the humming-bird pollinators from traplining towards more pronounced territoriality, i.e., from moving among tree crowns towards moving within the same large tree crown for the sake of saving energy. The size difference between saplings and seedlings reflected their difference in age, which helped in classifying progeny individuals as seedlings or saplings. The adult trees did not show any homozygote excess (F ≈ 0; see Sect. 6.4.1 for an explanation of the inbreeding coefficient F ). Saplings in the remnant forest had but a slight excess of homozygotes. The moderate inbreeding coefficient of the saplings which was presumably due either to mating events prior to fragmentation or to viability selection since then was greatly exceeded by that of the seedlings (F = 0.21). Bats dispersed fruits of the pasture trees over a considerable distance into the remnant forest, thus adding notably to the reproductive imbalance but little, if anything at all, to the variance-effective population size and the genetic variation in the remnant forest. It must be added that the seedlings were established in a suboptimal habitat, and that they belonged to a small number of families, i.e., they were closely related or even inbred. This is consistent with statements made by Ledig (1992) on the decrease of heterozygosity to be expected in future generations assuming restricted population size. (From Aldrich and Hamrick 1998; Aldrich et al. 1998) Hall et al. (1996) have pointed out that the effective population size at a given time may be diminished by variation in flowering phenology and may thus become critical in small fragments. This expectation was supported by results of Fuchs et al. (2003) for bat-pollinated Pachira quinata (Bombacaceae). These authors found no reduction in reproductive output, because the lower percentage of flowers developing into fruit was overcompensated by the larger number of flowers produced by trees in fragments. However, they also estimated lower outcrossing rates and a lower number of sires, leading to a higher degree of relatedness of the seed produced in fragmented forest, particularly by asynchronously flowering trees. The degree of isolation cannot necessarily be judged on a geographic scale alone (Nason and Hamrick 1997). Rather, dispersal delimits the patches. It should have become clear by now that population fragments are reproductively

117

118


isolated only if they are separated by distances that greatly impede the transfer of effective pollen or that cannot even be bridged by pollen or seed vectors at all (Chap. 5). Their behavior and abundance ultimately maintains links among forest fragments. To some degree, low tree density expands pollinator foraging ranges and pollen dispersal distances. It has been shown that tree pollen can travel far even in fragmented terrain (Chase et al. 1996b; Dawson et al. 1997; Nason and Hamrick 1997; Aldrich and Hamrick 1998; Aldrich et al. 1998; James et al. 1998; White and Boshier 2000; Dick 2001; Example 7.2). Nonetheless, one has to assume absolute limits beyond which no effective pollen can be transferred. However, as was shown both in wind-pollinated (Starke 1996; Wang 2001) and in insect-pollinated tree species of the northern temperate zone, fragmentation may also facilitate pollen dispersal. Ballal et al. (1994) reported up to 17% influx of pollen into small population fragments of Acer saccharum close to contiguous forest. This species has a dual system of pollen transfer, because in the absence of pollinators wind-transferred pollen is sufficient for normal seed set (Gabriel and Garrett 1984). Surveying seedlings in samples from a natural park and in patches (comprising 90–646 individuals), Young et al. (1993) found no evidence of reduced genetic variation, reduced gene flow, or increased selfing in the latter, although fragmentation had taken place more than a century before. These authors explain superior diversity of population fragments by increased gene flow. Also Young et al. (1996) came to the conclusion that increased gene flow may counteract if not even compensate or overcompensate the effect of reduced population size. Nason et al. (1997) discussed the results of numerous studies of pollen dispersal in tropical trees and concluded that the absolute reach of transfer distances in species adapted to low population density in species-rich communities is several hundred meters to some kilometers (Sect. 5.2). As Wright (1931) derived from his island model, few migrants or effective pollen per generation are sufficient to counteract loss of genetic diversity and drift-related genetic differentiation. The probability of mating between trees at a certain distance in adjoining fragments is higher than expected on the basis of that distance in a contiguous population. The reason is the absence of pollen-producing trees growing in-between. Bierregaard et al. (1992) reported that in Amazonian forest breaks as little as 80 m wide may diminish interfragment movement of some insects and birds. There must clearly be a certain limit to the gapcrossing ability of any pollinators; however, this gap may be rather wide depending on several circumstances, as shown for figs in Example 7.2 and Swietenia humilis in Example 9.2. This study also shows that fragmentation-induced changes in the foraging pattern of dispersal agents and the concomitant genetic processes in a tree population can differ among fragmented habitats. Apsit et al. (2001) described another situation of substantial pollen immigration into a population fragment of the insect-pollinated Enterolobium cyclocarpum (Jacq.) Griseb. (Fabaceae). A small number of trees received pollen from many outside


pollen donors growing hundreds of meters away. Similar findings were reported by Cascante et al. (2002) studying isolated trees of Samanea saman (Mimosaceae). Isolated trees received pollen from far-distant trees; however, the styles of these trees received lower quantities of pollen and a greater proportion of the pollen tubes were arrested at the bottom of the style, presumably owing to genetic incompatibility. In spite of high reproductive capacity of the isolated trees, they received pollen from fewer pollen donors, which led to an increase in the degree of the relatedness among their progeny. Example 9.2: Connectivity of Population Fragments of a Medium-Sized Dry-Forest Tree In a Honduran study region in seasonally dry forest the population of the largely self-incompatible species Swietenia humilis (Meliaceae) is highly fragmented and comprises small forest remnants and scattered trees in pastures and along the borders of agricultural fields (Fig. 9.1). A large proportion of the trees was present prior to fragmentation; therefore, the present pattern largely reflects the previous more continuous forest. Pollination is by small generalist insects; the seeds are wind-dispersed. The species flowers for about 1 month during the dry season with some individual variation.

Fig. 9.1. Locations of Swietenia humilis trees sampled. Each tree is denoted by a dot; the 17 trees selected for progeny analysis are circled (five in Las Tablas, three in Jiote, six in Butus/Jicarito, and the isolated tree 501). No trees were sampled in the spotted enclosed area adjoining the Las Tablas site marked by U. (From White and Boshier 2000)

119

120


The total number of trees was 98 in the largest fragment, Las Tablas, and was smaller in Butus/Jicarito (44), Jiote (22), and Tablas Plains (7). Owing to immediate losses of rare alleles of highly polymorphic simple sequence repeats, their numbers of low-frequency alleles encountered in the trees amounted to 54, 29, 19, and 8, respectively. The simple sequence repeats provided for complete paternity exclusion except for one, so paternal assignment by exclusion was 100%. Observed maximum pollen transport distances were at least 0.6 km for Las Tablas, 0.9 km for Tablas Plains, 1.5 km for Jiote, 3.6 km for Butus/Jicarito, and 4.5 km for tree 501. The true maximum distances may have been underestimated. As impressive as reports on maximum reach of animal-mediated pollen transfer may ever , the totality of the distribution of the quantities of transferred effective pollen is crucial. Pollen flow had a contrasting pattern. Besides the predominance of near-neighbor matings within 300-m distance, a large proportion of the pollen donors grew outside each fragment in the study area. In Las Tablas, the largest fragment, 64% of the matings were between near neighbors. Within-fragment matings dominated also in Jiote (62%) and Butus/Jicarito (53%). In the smallest fragment, Tablas Plains, only 31% of the effective pollen was produced within that fragment. The lone tree 501 received less than 30% of its effective pollen from distances between 1.2 and 2.4 km but over 70% of its pollen from unidentified male partners at least 4.5 km away It turned out that restricted mating between near neighbors was due to their asynchronous flowering. The flowers of the Meliaceae are as a rule functionally monoecious, i.e., either the male or the female organs are rudimentary (Styles 1972). The male flowering period of a given tree may be much shorter, so one of two synchronously flowering trees may fertilize the other but not vice versa. This reduces the number of near potential mates over and above asynchronous flowering but opens the way to some far-distance pollen becoming effective. Both the identity and the number of pollen donors varied greatly among seed trees. The number of individuals producing seed increased with increasing disturbance in this species. In general, a large proportion of the effective pollen came from the largest fragment rather than the nearest one. Long-distance pollen flow increased with a decrease in population size of the fragments, leading to restoration of alleles lost in the original fragmentation, if not even to an increase in the level of genetic variation within the local subpopulations. This was facilitated by the ability of pollinators to move between the spatially isolated stands. The proportion of pollen flow was governed by the size of the pollen source (large crowns or many trees, respectively) rather than its distance to the seed tree. The latter was also true for a remnant tree in pasture. The levels of heterozygosity among the seeds produced in the various fragments were only little influenced by the sizes of the respective fragments. This was interpreted by


an interbreeding network of trees with sufficient gene exchange to compensate for genetic drift in the parent generation. The widely spaced distribution of Swietenia humilis did not serve as a barrier to gene flow. Consequently, the seed was also very little differentiated between the fragments. Pairwise genetic distances were not correlated with geographic distances. The authors stress the need for measuring gene flow between fragments in order to see whether reduction of the size of fragments is really followed by “theorized genetic depletion.” In view of the rather efficient pollen dispersal, moderate distances of seed dispersal rather than persistent biparental inbreeding must be responsible for the existing spatial genetic structure of the old trees. Autocorrelation analysis revealed a significant excess of relatives (presumably half-sibs) in the near neighborhood (up to 50 m) of trees. The production of seeds increased with increased degree of disturbance. Since fragmentation does not necessarily have severe consequences for the connectivity of fragments, also remnant stands have long-term conservation value. Lone trees in pastures act as focal points or stepping-stones for pollinator movement. The results for their progeny contrasted the prediction made by Murawski and Hamrick (1991) that isolated trees receive pollen from fewer donors and are more likely to deviate from random mating. The authors conclude that a system of nonspecialist insect pollinators is less susceptible to habitat disturbance but stress the hypothesis that tropical tree species possess the capacity of adaptation to fragmentation of their populations (From White et al. 1999; White and Boshier 2000). Increased atmospheric turbulence over a treeless area supports both the transfer of airborne pollen and pollinator movement. It is, therefore, dangerous to extrapolate from data on pollen transfer in undisturbed populations to draw conclusions about the impacts under fragmentation (White and Boshier 2000). The large size of breeding units of figs (Example 7.2) suggests that forest fragments isolated by several kilometers can be part of a large viable population at least of these species. As Law and Lean (1999) have shown, bats as pollinators in Australian rain forest trees were highly mobile in a fragmented landscape. Their estimated home ranges reached 1,800 ha and might cover several forest fragments. Since hand-pollinations using pollen from different fragments yielded more seed than those within the same fragment, it may be inferred that the pollen transferred by bats between fragments is more effective, since the mating partners are unrelated. The transfer of genes from other subpopulations thus contributes to the “genetic rescue” of a subpopulation as defined by Richards (2000) by increasing the probability of its survival. Unfortunately, the future density of bats in the fragmented landscape is uncertain because of increased predation by owls. In Panama, Nason et al. (1997) reported reduced reproductive success of Spondias mombin, a canopy

121

122


tree, from small islands. Seed viability was up to more than 30% less than in contiguous forest. The use of genetic markers is indispensable in measuring tree population size as determined by the activity range of insects, birds, and bats (Nason et al. 1997; Sect. 5.2.3). In the case of only recent fragmentation, the search for private alleles as indicators of reproductive isolation makes little sense. The fragments are expected to share most genetic variants except the rarest ones, unless they are very small and short-distance gene flow prevailed before human interference. In general, genetic divergence may be due to pre-existing spatial genetic structure (Aldrich et al. 1998). Forest fragmentation does not only affect tree populations. For instance, the decline in the number of animal species with decreasing size of forest fragments has been studied in four Malagasy forest birds. However, owing to the relatively large size of the fragments surveyed and the high migratory potential of the birds studied, human interference was not reflected in genetic differentiation of populations (Andrianarimisa et al. 2000). Autopolyploidy is expected to partially offset the risk of genetic erosion and the decay of heterozygosity. The incidence of polyploidy is high in some tropical tree families such as the Bombacaceae (Bawa 1973). It is, however, not known how many of these species are autopolyploids. Several authors presented empirical data on the inverse relationship between the size of isolated conspecific natural tree populations and the amount of their genetic variation. Nason et al. (1997) focused our attention on the differential short-term and long-term genetic consequences of bottlenecks. Under the assumption of equal initial variation in the past, smaller genetic multiplicity due to the loss of rare genes is to be expected in smaller populations. After several generations, overall genetic diversity of smaller populations is affected more strongly (Nei et al. 1975). It has to be awaited whether empirical studies will provide the evidence for these genetic changes in completely isolated fragments. These authors also analyzed the quality of pre-existing spatial genetic structure and of the spatial scale of fragmentation. Data presented by Moran and Hopper (1983), McClenaghan and Beauchamp (1986), and Billington (1991) largely confirmed the negative relationship between population size and genetic variation. However, the establishment of a strict correlation takes many generations of continuously small size of the reproduction-efficient population until the results of multiple bottlenecks become manifest in diversity, because in every generation with finite size a renewed size-dependent risk exists that genetic variants will become lost. This risk also depends on the initial frequency structure of those variants. One may derive from this that the sizes of man-made fragments of a spatially structured population are moderately appropriate predictors of their genetic variation, as was found by Young et al. (1996), Sork et al. (1998), and Aldrich et al. (1998). The findings of

9.3 Genetic Preconditions for Restoration and Persistence

Young et al. (1999) at allozyme gene loci in a perennial herb are consistent with those in trees. Further to tropical trees, an inverse relationship was also found by Hall et al. (1996) on the basis of their study of Pithecellobium elegans and by White and Boshier (2000) studying highly variable microsatellites of Swietenia humilis in recent forest fragments (Example 9.2). Little impact of fragmentation on the genetic system of another tree of the Meliaceae family, Swietenia macrophylla, was reported by Céspedes et al. (2003). Also Prober and Brown (1994) found moderately strict relationships between diversity and population size; however, after discussing incomplete isolation and other reasons, they considered the relationships surprisingly strong. Be the relationship called strong or only moderate, the tendency involved in the results is beyond any doubt. Among natural populations with more or less unknown past history, the estimated degree of determination of genetic variation by population size cannot possibly be expected to be as high as in artificial populations of model insects in the laboratory. When summarizing the results of the genetic impact of forest fragmentation, it has to be stated that there exists, of course, no empirical information on the genetic change during a full tree generation. Nonetheless, according to the estimates of various population genetics parameters, the reproduction system of some tree populations is subject to major change with regard to mating systems, gene flow, and migration. Spatial genetic structures are affected by fragmentation both on a landscape scale and on a regional scale.

9.3 Genetic Preconditions for Restoration and Persistence The attempt to precisely restore a forested area to its presumed original vegetation would certainly be all but promising. This applies even more to the genetic structure of tree populations. However, taking account of some basic preconditions for an adequate population size of tree species is helpful in making decisions. Since fragments of tropical tree populations are in a critical situation, managers have to consider the size of a population that is worthwhile for preservation and possible later enlargement. The persistence of tree populations in tropical forests over generations has certain preconditions, such as a minimum population size, since small populations are prone to extinction sooner or later. Besides the reduction of evolutionary adaptability due to the loss of genetic variation (Sect. 7.2) caused by drift, the accumulation of detrimental inbreeding effects is another genetic reason for increased vulnerability of small isolated populations. Information on the size and structure of neighboring fragments and their locations is important for making decisions on the restorability of

123

124


a fragment or a group of fragments (Loyn and MacAlpine 2001). Two approaches towards conceiving population sizes adequate for persistence of population fragments shall be briefly presented. If mating is assumed to be random and selection to be absent, heterozygosity is expected to decrease at a rate that depends exclusively on population size (Crow and Kimura 1970). If the reproduction-effective size of the population is 50, the decay of heterozygosity is expected to amount to just 1% per generation; hence, 50 is the minimum population size that keeps the speed of this decay equal to or below 1% at least over short periods of time. Considering the number 50 as operationally adequate is simplistic, since it does not account for other factors such as genetic load and selection that influence the dynamics of heterozygosity (Sherwin and Moritz 2000). Particularly in trees, the density and the spatial distribution of the population members play a role in setting the scene for inbreeding, as was described in Sect. 9.2. This number is still referred to as the basic rule of conservation (short-term rule). The tenfold population size of 500 individuals is not only expected to keep the decay of heterozygosity below 0.1%. It is also expected to provide for a balance between the loss of genetic variants and their origination through mutation. A population size of 500 individuals (long-term rule) is said to be required to avoid immediate loss of diversity and to ensure the long-term preservation of genetic variation and thus of evolutionary adaptability (Franklin 1980; Frankel and Soulé 1981). Experimental investigations in caged Drosophila populations showed that the loss of genetic variation in populations of 500 reproducing insects due to genetic drift is compensated by mutations; however, this finding does not take account of the identity of the genetic variants and their adaptational relevance in nature. Furthermore, the time factor has to be accounted for, since the generation period is much longer in trees than in insects. The much larger number of 500 individuals can also be seen as a gross safety measure. The 50/500 rule is weakly supported by experimental studies on feral animal species and even more so on plant species but is still frequently reported in the relevant literature (Frankel and Soulé 1981). Last but not least, the reproduction system of plants is much more complex than that of animals. It has to be stressed that besides population size, imbalances in the mating success of individual trees have to be accounted for, as was outlined in Sect. 9.2. Consequently, reproduction-effective population sizes have to be considered when applying the basic rules. Fragments of species-rich tropical forests of only a few hectares in size are likely to be insufficient for the preservation of numerous tree species, if the 50/500 rule is adopted for forest tree populations. The long-term persistence of species in tropical forests is only possible if certain additional preconditions according to a different concept are fulfilled. The estimation of minimum viable population (MVP) sizes as first conceived by Shaffer (1981) have attracted considerable interest among conservation biologists (Soulé 1987). The MVP

9.3 Genetic Preconditions for Restoration and Persistence

concept refers to two aspects of populations. One is the totality of critical factors of the genetic structure of populations above a level below which inbreeding depression and the loss of adaptive capacity would become a problem for their continued survival. The other one is the size of an isolated population of a given species in a given habitat that keeps the risk of extinction low. The MVP is to ensure at some acceptable level of probability that the isolated population persists in a viable condition for a given interval of time. Some authors have specified these parameters as 95% probability and 1,000 years, respectively. This period is rather extended at first sight; however, it is equivalent to a small number of generations of forest trees. Apart from genetic considerations, other factors such as demographic stochasticy, i.e., random fluctuations of the sizes and perturbations of the age structures of populations without associated changes in their genetic structure, may also pave the way towards extinction (Lande 1988). The long-term growth rate of a population is determined by the geometric mean of the growth rates over time. The geometric mean is less than the arithmetic mean. Consequently, the long-term growth of a population may be negative, although it presently shows positive growth. Several nongenetic factors share in the determination of the expected persistence time of small populations. Shrinking populations require constant support and management. Also deterministic effects such as the incidence of herbivores, pests and disease, competition, and the colonization by invasive nonindigenous species represent threats to persistence. The long-term conservation of forest fragments is more promising, the better the habitat quality of pollinators and frugivores as seed dispersers (Nason et al. 1997). Reduced density of fruit-eating animals in small forest fragments leads to lower recruitment of large-seeded animal-dispersed trees, although the trees bear fruit (Cordeiro and Howe 2001). If the animal disperser of a tree species is absent owing to subsistence hunting in fragmented forest, the seeds are only gravity-dispersed. This leads to a closer relationship of the seedlings under parental canopies. Therefore, defaunation also has an impact on the dynamics of spatial genetic structures in trees which will become crucial at reproduction (Pacheco and Simonetti 2000). Fragmentation of tropical forests creates conditions that are similar to a metapopulation structure. In unfragmented forests, the habitats of extinct local subpopulations can be recolonized and mating contact between temporarily isolated local subpopulations can sometimes be regained (Céspedes et al. 2003). In permanently or temporarily isolated subpopulations an increase of inbreeding is expected, and genetic drift causes local losses of genetic variation and evolutionary adaptability. Céspedes et al. (2003) concluded that metapopulation models could be used to describe the genetic structure of plants in highly fragmented landscapes. However, as Gilpin (1991) pointed out, not only the sizes of the subpopulations but also the size of the total metapopulation matter greatly. The question of MVP is then extended to the question of the

125

126


minimum number of interconnected local populations necessary for the longterm persistence of the total metapopulation. Both of these quantities are likely to be strongly reduced by man-induced fragmentation. Genetic variation decays faster in the smaller subpopulations and these are hardly the source of genetic variants that were lost in adjacent larger subpopulations (Aldrich et al. 1998). Therefore, both recolonization and gene flow originating from the larger donor subpopulations are crucial for the persistence of patterns of partially isolated population fragments (Hall et al. 1996). Habitat corridors may provide avenues for gene dispersal between patches and first of all large trees may contribute to the local genetic dynamics of the adjoining patches (Nason et al. 1997). Only little pollen influx from genetically differentiated adjoining fragments through corridors is sufficient to counteract drift-related genetic change. Strip lines such as revegetated margins of streams and watercourses, live fences, and regrowth on abandoned fields or pastures may support pollinator movement (Law and Lean 1999). Restoration efforts aimed at increasing the number of keystone species should therefore include the establishment of those links between fragments (Nason et al. 1997; Galindo-González et al. 2000; White and Boshier 2000). If the effectiveness of corridors for connectivity of local populations of various tree species can be proven to be valid in general, many tree species may then persist as metapopulations, which often makes forest fragments worthy of conservation. This is particularly true in regions that have been heavily degraded. Few scattered big trees left in pastures serve as traps for genetic information, as pollen from distant trees may sire their seeds and enrich the local variation. Therefore, the favorable conditions offered on abandoned pastures for survival and growth of seedlings and saplings of some species are augmented by favorable genetic perspectives, although opposite effects were described for Symphonia globulifera in Example 9.1. Remote-sensing data may help to select priority areas and to design forest reserves that will have a good chance of conserving biodiversity (Saatchi et al. 2001). Some techniques of restoration in various tropical regions were covered by Lamb et al. (1997). As already expressed in this section, our methods of defining minimum population sizes are admittedly rather restricted. Nevertheless, the considerations of the type presented may help in making decisions. Since the 50/500 rule does obviously not account for various nongenetic menaces to population fragments of a given species in a given environment, minimum sizes much larger than 50 should be maintained. If long-term preservation of forest with a large spectrum of tree species is concerned, population sizes at least of all of the keystone species must be much above 500. If a given population fragment falls short of this lower bound, it may still be eligible for short-term maintenance. Decisions of this type are supported by the presence of tree species that are important in terms of ecology and/or economy. It has to be kept in mind that numbers such as 50 or 500 must refer to the reproduction-efficient population


size introduced in Sect. 9.2 rather than the census number of reproducing trees. In contiguous populations of anemophilous species this parameter can be estimated from the variation in flower counts (Kjær 1996). More reliable in tropical forest fragments is counting effective gametes on the basis of the reconstruction of descent of seeds or even seedlings by means of highly variable genetic markers such as microsatellites (Aldrich et al. 1998). This requires much effort, since relevant genetic surveys have to be repeated in order to account for variation in time (Murawski and Hamrick 1992). Eventually it is sufficient to sample progeny of a restricted number of adult trees and to estimate the size of the reproduction-effective population relative to the census size. This ratio may serve as an appropriate reduction factor. In view of the multitude of unforeseeable events, crisis-oriented immediate decisions on long-term viability and preservation are required. Uncertainty must be tolerated and managers be prepared to make best guesses on available data (Menges 1991). Although the MVP concept is easier to apply to animal populations, it is still useful also in connection with plants for several reasons (Gilpin and Soulé 1986). It reminds us that the focus of conservation must be on populations rather than individual trees, and that only minimum requirements for persistence during a minimum period of time can be planned. Still today, not all managers consider genetics an important factor in planning preservation of fragmented forests. This is partly due to ignorance in a field of biology conventionally felt to be difficult.

9.4 Recommended Literature In his classic paper, Ledig (1992) discussed the various types of human impact on forests and addressed their genetic implications. During the past 25 years many books on conservation have analyzed the impact of fragmentation of plant and animal populations. Six introductory chapters of the book edited by Young and Clarke (2000) provide an overview of genetic and demographic issues related to fragmentation. Two among the subsequent chapters covering case studies deal with tropical forest trees. In the book edited by Brigham and Schwartz (2003), many authors analyzed the preconditions for persistence in various plant populations and designed management procedures. This book demonstrates that the biological characteristics of the respective species must be accounted for, if planning is to be reliable. In the book edited by Laurance and Bierregaard (1997b), many aspects of the peculiarities of tropical forest fragments and their appropriate management are viewed and discussed. A contribution of Young and Boyle (2000) to the book edited by A. Young et al. (2000) presents a brief but rigorous discussion of the genetic aspects of fragmentation.

127

Genetic Aspects of the Management of Natural Tropical Forests

10.1 Introduction Research that is somehow related to the genetics of tropical forest trees has long been largely confined to breeding of a few tree taxa that are suitable for plantation forestry. Owing to the restriction of forest genetics to breeding work, the genetic aspects of the management of natural forests in the tropics have received only little attention and relevant studies are scarce. Numerous reports on progress achieved through breeding of these taxa have been published but only a few publications deal with genetic implications of the management of natural forests. Research on silvicultural systems for the sustainable management of tropical forests and their conversion to ecologically valuable species-rich production forests has for several decades been supported by, for instance, German development agencies (Lamprecht 1986). However, genetic aspects have been neglected in this context. Although genetic diversity is one of the keystone components of biological diversity, forest genetics research is still considered to be of secondary importance. Only fairly recently has the preservation of biological diversity been accepted as an integral part of sustainable management systems for tropical forests (p. 119 ff. in Deutscher Forstverein 1986; p. 50 f. in Enquete Kommission “Vorsorge zum Schutz der Erdatmosphäre” des deutschen Bundestages 1990; Ministerial Conferences on the Protection of Forests in Europe, Strasbourg 1990: Resolution S2; Helsinki 1993: Resolution H2). Since then, there has been growing public awareness of the fact that the destruction of tropical forests leads to an irreversible loss of genetic information of many species. Few studies on the genetic impact of the management of natural forests were made in view of the need for the conservation of genetic resources of tropical tree species (Chap. 14). More recently, the number of genetic research projects on an international level has increased. Some of the results will be addressed in the present chapter. The subject deserves our interest, since much the surviving part of a tropical forest is likely to be timber-production forest (Lamprecht 1986; Whitmore 1998).

10

130

C HAPTER 10 Genetic Aspects of the Management of Natural Tropical Forests

10.2 Selective Logging A widely practiced silvicultural management technique in tropical forests is various ways of selective logging and in particular target-diameter felling, i.e., removing all trees above a certain diameter limit (between 45 and 70 cm) with or without restrictions, such as a sufficient number of adolescent trees left and sufficient numbers of regenerates (p. 131 ff. in Montagnini and Jordan 2005). It is, of course, problematic to define which numbers are sufficient from the genetic point of view (Wickneswari and Boyle 2000). Although the amounts of harvested wood volume are relatively low (Chap. 2 in Bruenig 1996; Chaps. 7, 10 in Whitmore 1998), the extraction of a proportion of trees of merchantable size has a twofold genetic impact. It reduces the number and density of reproducing trees in the populations of the respective species and may imply dysgenic selection. This impact is stronger, if logging is repeated in several successive cycles (polycyclic systems of logging; p. 136 f. in Montagnini and Jordan 2005), and the shorter the intervals are in connection with a lower diameter limit (p. 179 ff. in Bruenig 1996). As Murawski and Hamrick (1992) have shown, the density of flowering trees in a population varies over time. Much at variance from this is the effect of repeated logging of a species, since its population density cannot recover anymore. Planting in overlogged forest that once appeared to be suitable for enrichment planting is problematic. Summarizing various pertinent silvicultural projects, Bruenig (1996, p. 205 ff.) came to the conclusion that the results were by no means encouraging. This precludes genetic enrichment at least in tropical rain forests. From the genetics aspect, natural recovery might be a slow process. Things are, of course, different in extremely overlogged forests. Selective logging has been practiced for a long time on a small-scale basis but has drastically increased with the increased demands for timber and with the progress made in improving logging technology (p. 87 ff. in Bruenig 1996). Wickneswari et al. (1999) and Wickneswari and Boyle (2000) reviewed the results of a comprehensive study in Peninsular Malaysia. In two regenerated stands with 14 and 41% disturbance there was no adverse change of genetic diversity after a single low-intensity logging event in comparison with unlogged lowland dipterocarp forest. However, in a ridge forest the average loss of allelic variants in five species with different life-history strategies after logging ranged from 8 to 25% and their genetic diversity was reduced to values ranging from 5 to 23%. This reduction closely corresponded to the reduction in tree numbers caused by tree harvesting itself. The loss of 25% of the alleles and 9% of genetic diversity owing to logging was higher in species of low abundance (Shorea leprosula) than in those of high abundance (Scaphium macropodum) with 8% of the alleles and 5% of the diversity. One of the observed species, the climbing

10.2 Selective Logging

palm Daemonorops verticillaris, exhibited 14% allele loss owing to mortality caused by the logging operation and the trampling of saplings. In Central Africa Malcolm and Ray (2000) have found that the area of skidding trails, secondary roads (for removing locally skidded logs), and primary roads (frequently traveled by vehicles) themselves was still smaller than the area of adjacent forest where the canopy was damaged. The diversity of tree species among the saplings in the adjacent forest was reduced in comparison with that in unlogged forest. Canopy damage was identified to be more hazardous than clearing and eventually grading the ground for roads. With increasing time since the logging operation, edge effects continued. These authors also devised reduced-impact procedures in logging operations and road construction in order to minimize damage to the remaining tree vegetation. Sinuous roads help to take advantage of existing gaps and to avoid toppling trees. Dispersed felling helps to avoid tangling of trees and to prevent too-large gaps. Last but not least, reducing the number of trees harvested per hectare involves obvious genetic benefits, although this procedure has economic constraints. Besides describing the ecological impact of logging, Bruenig (1996, Chap 2) and Montagnini and Jordan (2005, p. 139 ff.) have shown ways towards considerate procedures for this type of rain forest use. Among several elements of low-impact logging Whitmore (1998, p. 137) recommends the preplanning of snig-tracks, directional felling to minimize damage to the canopy and facilitate log removal, and winching. Logging is the major form of human disturbance, but it is still just one form. A gradual reduction in vigor may result from practices that affect tree health without tree removal, such as harvesting bark or resin (in dipterocarps or in conifers such as Pinus and Agathis). The genetic impact strongly depends on the product collected. Wickneswari and Boyle (2000) analyzed the impact of harvesting wood and collecting other forest products such as fruits and seed. Collection of fruits, nuts, and other nontimber products has long been practiced by the indigenous human populations (Prance 1994). This author also pointed out that in certain dioecious species felling the trees for fruit collection leaves only male trees. 10.2.1 Effective Population Density After Logging

The reduction of tree numbers causes an immediate reduction of genetic variation. Just as adverse is another aspect. Owing to the reduced density of individuals in the reproductive age class, the mating contact between the trees in the population and even more so that between more remote parts of a population is expected to decrease (Finkeldey 2002). Eventually the effective population density is only loosely related to the physical density, since the number of trees per unit area is hardly appropriate

131

132


to describe the totality of conditions for the dispersal of sufficient pollen (Sect. 5.2). This is more relevant in species-rich tropical moist forests than in other climatic zones, because the pollen of tropical trees is dispersed by pollinators adapted to food trees occurring at a certain density. Increased distances to conspecific neighbors are then less likely to be bridged by pollinators. Logging is also likely to have an effect on the species spectrum of pollen vectors at least of the keystone species, since these are the primary target of logging operations. Logging is expected to have an influence on the mating system of the remaining trees in general, since the balance between cross-pollination and self-pollination is changed in favor of the latter. If the genetic barriers against self-pollination are not fully efficient, self-fertilization as the most severe form of inbreeding is implied. Since the largest trees produce most flowers, their extraction has the most severe impact on the quantity of seed produced and its genetic structure. Reduced mating contact between trees within populations due to selective cutting results in decreased fertility of dioecious and self-incompatible species (Ghazoul et al. 1998) and promotes inbreeding in species with a mixed mating system. This is not just an expectation, as Murawski et al. (1994b) have shown in Shorea megistophylla, a canopy-emergent species of the Sri Lankan dipterocarp forest. These authors compared the estimated outcrossing rate in undisturbed forest with a density of ten trees of 20 cm or greater breast height diameter per hectare and on plots that were logged 20 years before with a density of only two trees per hectare. This reduction is equivalent to an increase in average tree distance from about 30 to about 70 m. It was unclear whether trees that were left at the time of logging have since then entered the reproducing part of the population. A genetic survey of seed at several polymorphic allozyme loci provided for a comparison of an important parameter of the mating system on the two plots. The multilocus estimate of the outcrossing rate t (Sect. 6.3.2) amounted to 0.86 in the primary and 0.71 in the less dense forest; hence, the outcrossing rate was significantly reduced owing to the disturbance. Lee (2000) studied the impact of reducing tree density on the mating system of another canopy-emergent dipterocarp species, Dryobalanops aromatica, in two different types of forest in Peninsular Malaysia. In response to lower density and greater distance to the nearest conspecific neighbor (Table 10.1), the Table 10.1. Mating system parameters of Dryobalanops aromatica Gaertn. f. in primary and logged forest. (Adapted from Lee 2000)

Primary forest Logged forest

Population density (ha–1)

Average distance to nearest neighbor (m)

tm

rp

15 7

26 38

0.92 0.77

0.39 0.11


outcrossing rate in the logged forest was remarkably reduced. Since even less cross-fertilization in this species was estimated elsewhere (Chap. 13), it may be concluded that self-incompatibility is presumably only partial, so the difference of roughly 15% more self-fertilization after logging reflects the effect of lessefficient pollen dispersal. If there once existed a family structure in the primary forest, the consanguineous neighbors have been extracted, so after logging less biparental inbreeding arises in comparison with that in the primary forest. Part of the inbreeding due to the increased proportion of self-fertilization would then be compensated by this condition. The striking difference in rp, the probability that seeds produced by the same seed tree are full-sibs, is difficult to explain. The reduced density would be expected to reduce also the number of male mates of the seed trees (i.e., the reciprocal of rp) instead of raising it from less than 3 to 10 (Table 10.1). Eventually the increase in atmospheric turbulence was favorable to pollinator movement and thus pollen dispersal. The author also pointed out that the density of the flowering trees rather than that of all trees had a major influence. Kitamura et al. (1994) did not find a significant reduction in the outcrossing rate of this species in Borneo. These authors reported estimates of t = 0.86 in primary forest as compared with t = 0.79 in secondary forest, where the trees larger than 60 cm in diameter had been logged 20 years before. However, the physical population density may not simply be taken as the effective density, as was indicated at the beginning of this subsection. Consequently, the estimates of tm vary among populations in response to their numerous other determinants. The authors interpreted the results by indicating high flowering density, although in the secondary forest this was around 7 ha–1, i.e., half of that in the primary forest. The results of Lee (2000) show that the increased intertree distances implied by logging disturb the sensitive balance between mating system parameters at least in particular situations. Besides pollination possibly being less complete, the proportion of biparental inbreeding is supposedly lower but the proportion of self-fertilization is raised. In a heavily logged population of Pinus caribaea, Zheng and Ennos (1997) did not only find a marked reduction in the outcrossing rate (tm = 0.894) but also substantial incidence of consanguineous biparental mating. This was obviously due to highly irregular spacing of the remaining trees. The impact of logging was also studied by Dayanandan et al. (1999) in three Costa Rican populations of Carapa guianensis, a species of the Meliaceae. In a biological reserve used as a reference this species had a density of ten reproductively mature trees per hectare. In two managed forests approximately four and three large trees, respectively, had been removed about 10 years earlier. The density of medium-sized trees (diameter at breast height greater than 35 cm) after logging was still 11 and 15 per hectare, respectively. One of these two

133

134


managed forests was surrounded by degraded forests and pastures; the other was part of a large contiguous block of forest. Adult genotyped trees had an age of more than 100 years; saplings were estimated to be between 5 and 10 years old. In all three populations the genetic distance between the adult and sapling cohorts was small; however, the average distances among the three sapling cohorts (D = 0.15) was twice that among the adult cohorts (D=0.08), indicating reduced influx of external effective pollen prior to the establishment of the sapling cohorts. Since no other reason for an allele frequency change of microsatellites is thinkable, the hypothesis of reduced gene flow must be accepted. It indicates incipient genetic divergence. The degree of logging must be considered mild, anyway. There was no evidence of inbreeding in any of the three populations. Once the density of a population has been strongly reduced, the distances among the trees approach or even exceed the distances of seed dispersal. This induces spatial genetic structures in the form of family structures, i.e., neighborhoods of related trees. The existence of family structures is most likely to exist in barochorous species (Sect. 5.3.1) with their inefficient mechanisms of seed dispersal. Hence, by logging close neighbors are removed. However, new and more pronounced family structures are expected to arise, since seed shadows do not overlap. The seeds in the same seed shadow are also sired by fewer pollen parents and are expected to be closely related. This implies future inbreeding owing to preferential mating of related neighboring trees. Indications of a family structure were found in the dioecious species Altingia excelsa on the basis of surveys of morphological traits and isozymes (Sakai 1985). The negative consequences of inbreeding on phenotypic trait expressions, i.e., inbreeding depression (Sect. 6.4.2), are difficult to observe in natural forests; however, there is no doubt that inbreeding affects naturally regenerated populations just as much as it affects planted species. Inbreeding depression was demonstrated by numerous studies also in tropical trees. The mating contact among unrelated trees of predominantly outcrossing species must not be unduly interrupted by management practices. This obviously requires not only keeping minimum sizes but also keeping minimum densities of tree populations and preserving the food basis of pollen vectors. Owing to our restricted knowledge of the mating system and of patterns of gene flow in most tropical tree species, it is difficult to estimate threshold sizes and threshold densities of populations. The long-term genetic effects of lowered density during subsequent generations is largely unknown. More research in this field is urgently needed. Wickneswari and Boyle (2000) stress the desirability of research on the response of tree species with different life-history strategies to logging disturbance in general. Bruenig (1996, p. 166 ff.), Montagnini and Jordan (2005, p. 133 ff.), and Wickneswari and Boyle (2000) describe various logging regimes mainly developed in Southeast Asia that presumably differ in their genetic impact.


10.2.2 The Question of Dysgenic Selection

Under regular silvicultural management the economically most desirable trees are kept in the stand in order to improve the economic value of the stand increment. This practice hopefully involves also some genetic improvement in view of subsequent natural regeneration. Harvestable trees in target-diameter felling systems are selected on the basis of a certain minimum diameter and stem form, so that during exploitation the most vigorous, straight trees are removed. This means that from the time of logging onwards, trees with superior phenotypes are excluded from reproduction. Conversely, the remaining population consists mainly of young or slow-growing trees with inferior value for use. Their reproduction is favored; therefore, the concomitant mass selection is dysgenic. Depending on the degree of the heritability of traits this contributes to inferior expressions of the selected traits in subsequent generations. Both the ecological and the economic value of the population regenerated from the remaining trees may still be reduced. Ziehe and Hattemer (2002) reported that target-diameter felling as practiced in old beech stands in central Europe had little effect on genetic structures at marker gene loci. However, owing to heterozygote advantage at some gene loci the most heterozygous trees were removed first. This had a minor effect on the loss of rare alleles. However, at a gene locus involved in controlling diameter growth target-diameter felling would substantially diminish the potential for diameter growth in subsequent generations. The effect of dysgenic selection for growth rate and stem form has often been feared but has rarely been verified by using rigid methodology (Finkeldey and Ziehe 2004). This condition is expected to persevere over generations as pointed out by Lindquist (1946, Chap. 2) in his early book on the role of genetics in Swedish silviculture. Phenotypic trait expressions may not only be potentially affected by inbreeding depression due to selective logging in tropical forests but also by unintentional selection. The heritability of economically important phenotypic traits is presumably low in natural tropical forests owing to high environmental heterogeneity. Furthermore, the correlation between the age and the size of trees is frequently weak in natural tropical forests, i.e., the tallest trees are eventually but not necessarily the oldest. Thus, if fast-growing trees are preferentially removed from the population by selective logging, the risk of dysgenic selection is not necessarily high. The magnitude of the damage due to dysgenic selection is still open to debate. Whitmore (1998, p. 134) describes it as anybody’s guess. In order to minimize the risk of dysgenic selection, trees with superior phenotypic trait expressions should be able to contribute to the production of the

135

136


progeny generation in proportions that are at least equivalent to their frequency in the parent generation. Repeated logging involves repeated selection, provided that notably many trees of the respective species are extracted during subsequent cycles. The potential danger of dysgenic selection depends on the time of regeneration. If the progeny generation was established prior to the main logging operation and natural regeneration was kept protected during extraction of the logs, its genetic structures are not directly influenced. In this case the genetic information of the harvested trees is represented in the progeny generation.

10.3 Natural Regeneration The assumed ideal of naturally regenerated forest being maximally stable, maximally diverse, and minimally susceptible or maximally resistant to unanticipated threats, is a commonly held paradigm. The three properties are assumed to be simultaneously achieved by the normal processes of forest regeneration without human intervention (Namkoong 1999).

The sporadic exploitation of a forest is usually not followed by artificial regeneration. Even though a logged or exploited forest may be naturally regenerated, it is still not a natural forest anymore. By definition, natural forests have never been influenced by humans. Since this refers in particular to their reproduction, natural forests have to be discriminated from those that were naturally regenerated after silvicultural treatment, particularly after selective logging. Since the intensity of human influences varies, we have to consider degrees of naturalness. Other human influences exerted as man-made environmental change of the global water, energy, and nitrogen budget should not be considered marginal. Namkoong (1999) and Wickneswari and Boyle (2000) state that virtually all forests in the world are directly or indirectly affected by human activities. However, it is difficult to decide, whether a given forest has escaped human influence completely. The three properties mentioned in the prefatory quotation of the present subsection are likely to be exacerbated with increases in human intervention. Therefore, both climax forests and secondary forests are still considered natural, if they were predominantly or completely regenerated without planting or sowing. They are then likely to possess a family structure (Sect. 5.3.2). Loyn and MacAlpine (2001) used the term “natural” to indicate lack of discernible impact of modern technological society. Of particular importance is that many if not all mature trees have contributed their share to the subsequent generation. Since not all trees reproduce at a given time, the extension of the regeneration period over several years up to several decennia helps to achieve this condition. However, maximizing diversity does not necessarily lead to maximizing stability (Altukhov 2006).

10.4 Genetic Aspects of the Manipulation of Dynamic Processes in Natural Forests

Since only few tree populations might reproduce panmictically, genetic equilibria might be largely absent, so even without human interference some genetic change during reproduction represents the rule rather than the exception (p. 175 ff. in Hattemer et al. 1993). The genetic structures of natural populations have usually been shaped by adaptational processes over many generations, even though periodic disruptions of benign environmental conditions and shifts in the selective environment may be required for regeneration (Namkoong et al. 2000). Nevertheless, the generation cycle is completed on the same site. This stands for some continuity and is much at variance from artificial regeneration that is normally linked to a transfer of reproductive material of the population. A population that has never been transferred by humans and that has undergone an uninterrupted sequence of natural regeneration is called autochthonous. Such a population is expected to be adapted to the locally prevailing environmental conditions. This does not mean that autochthonous populations could not lack adaptational optima. Conversely, if a population has ever been transferred by humans (in the seed stage), it is called allochthonous. It is difficult to decide whether a given stand is autochthonous. It is normally not expected to represent optimum conditions for current or future generations. If it does not fit into a geographic variation pattern of other, hopefully natural stands (Chap. 12), its autochthony is doubtful. If artificial regeneration of forests has created a mosaic of autochthonous and allochthonous stands, the attribute of being autochthonous has to be understood quantitatively in view of the influx of genes from allochthonous neighboring stands. Genetic aspects of artificial regeneration of forests will be addressed in Chap. 13 in more detail.

10.4 Genetic Aspects of the Manipulation of Dynamic Processes in Natural Forests Von Gadow and Kleinn (2005) distinguish just two archetypes of forest management besides sporadic exploitation. Under rotation forest management the totality of the standing volume is removed at the end of the rotation period (Fig. 10.1). The stand is then artificially regenerated (Sect. 11.7), whereby genetic structures are changed completely. This system is typical of intensive silviculture as found primarily in the southern hemisphere. Under a continuous-cover system the standing volume is never harvested completely (Fig. 10.1). The age of the stand is undefined and forest development does not follow a cyclic harvest-and-regeneration pattern. Instead, it oscillates around a certain level. Natural regeneration is possibly going on periodically depending on overwood retention.

137

138

C HAPTER 10 Genetic Aspects of the Management of Natural Tropical Forests V

RFM systems

V

CCF systems

age 0

R

R

time t1

t2

Fig. 10.1. Development of biomass (v) over age or time for two archetypical systems of forest management. Left: rotation management (RFM) systems; R denotes the end of a rotation period. Right: continuous-cover forestry (CCF); the biomass is largely a function of the time elapsed since the most recent thinning or harvest. (From von Gadow and Kleinn 2005)

Both archetypes may be sustainable (Sect. 10.5), although more risks are involved in rotation forestry. A wide variety of silvicultural regimes have been developed for the two systems. A forest development path or treatment schedule (von Gadow and Kleinn 2005) comprising a sequence of operations related to thinning, harvesting, and regeneration on the basis of growth models or experience can be designed for any forest type or individual forest (Bruenig 1996). Continuous-cover forestry in species-rich tropical forests challenges refined methods of controlling density, the species composition, the allocation of trees belonging to the various species, and the regeneration at least of the keystone species (Sect. 7.6 in Whitmore 1998). Thinning as such in conifer stands without applying any phenotypic criteria of tree removal such as described by Buchert et al. (1997) and El-Kassaby (1999, 2000) is expected to lead to a reduced number of genetic variants and reduced diversity. This reflects only the decrease in the number of individuals left on a given area. As Hosius (1993) has shown in Picea abies, thinning from above and thinning from below may have different implications for the genotypic structure of the remaining stand. Density and mating system parameters such as outcrossing rates are to be considered sensitive indicators of a changed mating system owing to the removal of trees for regeneration. Referring to western North American conifers, El-Kassaby (1999) reported the surprising result that the incidence of mating among relatives was higher in stands treated under the shelterwood system (around 200 trees per hectare and 30% of the basal area are removed) than in old growth (used as a “control”). One should have expected that the removal of possibly related neighbors would force mating among less related trees. Under this system also the correlation of paternity (rp) exceeded that of old growth. However, under a patch-cut system with patch sizes ranging from 1.5 to 2 ha the incidence of correlated mating was typically reduced. Later analyses of the same replicated field experiment (El-Kassaby et al. 2003) yielded the result of small, significant treatment effects in only a few cases, although the

10.4 Genetic Aspects of the Manipulation of Dynamic Processes in Natural Forests

most drastic reduction was to fewer than 25 adult trees per hectare. Number, density, and spatial arrangement of the seed-bearing trees are important determinants of genetic parameters of the offspring. Glaubitz et al. (1999) found no difference of genetic diversity and multiplicity between native stands of Eucalyptus sieberi and their natural regeneration achieved under the shelterwood system with 10% of the adult trees left. In a study of regeneration in the closely related E. considensiana Glaubitz et al. (2003) concluded that clear-felling and aerial resowing with local seed was more efficient in genetic diversity conservation than “natural” regeneration under the seed tree system. The number of seed trees must have been much smaller than the number contributing genetic information to the seed collected. According to Finkeldey (2002) continuous-cover forestry has the prime impact on the population density of target species and on kinship of conspecific neighboring trees. Experimental results from forests in both the temperate zone and the humid tropics suggest that the latter are more sensitive because of their a priori lower population densities; therefore, the consequences for inbreeding and lowered heterozygosity are more pronounced in tropical forests. Information on threshold densities is urgently needed. Silvicultural management practice may considerably change environmental conditions to an extent that populations lose their adaptedness. Adaptational processes are then required to ensure the survival of the population. Evolutionary adaptations are a manifestation of changed genetic structures (Sect. 7.2). Genetic structures of populations are likely to be affected by silvicultural practices although it is often difficult to prove adaptive or nonadaptive changes of genetic structures in response to the human impact on forests. The genetic dynamics during the reproduction of species-rich tropical forests is rather complex. Frequently a succession of different dominant plant species is observed (p. 377 ff. in Richards 1981). The reoccupation of a habitat by species of late successional stages is only possible from surrounding populations if early stages of forest regeneration are dominated by species of primary forests. The life cycle of tropical forests is characterized by the disappearance not only of trees but also of local populations. This holds in particular if forests are simultaneously regenerated naturally in a large area, for example, after a catastrophic event. Extinction and recolonization are natural events both for pioneer species and for species of later successional stages, including species of the climax forest, although some species with less efficient systems of seed dispersal and without a good seed or seedling bank are less tolerant to local extinctions. Partially isolated populations experiencing a dynamics of local extinctions followed by recolonization are defined as metapopulations (Hanski and Gilpin 1991). The population genetics consequences of a metapopulation structure cannot be analyzed if only single populations are observed (McCauley 1991). A habitat

139

140


is often recolonized by progeny of only a few individuals. Effects of founder events were studied in the Neotropical tree Corythophora alta on the basis of frequencies of haplotypes of maternally inherited chloroplast DNA (Hamilton 1999; Sect. 5.3.2). Reproduction-effective population sizes are usually small, resulting in genetic drift, loss of genetic variation, and inbreeding (Gilpin 1991). Genetic drift shaped genetic structures in a disturbed habitat of the tree Symphonia globulifera owing to the reproductive dominance of a few pasture trees (Aldrich and Hamrick 1998; Example 9.1). The genetic differentiation among populations is likely to increase owing to drift effects. Metapopulation structure retards adaptive genetic differentiation of subpopulations in response to the selection by microsite conditions but favors the capacity of the total population to adapt to an environmentally heterogeneous habitat. The functioning of a metapopulation depends on the relative sizes and densities of the subpopulations as well as on their pattern, since these two factors have a strong impact on the intensity and the directions of gene flow. The genetic consequences of a metapopulation structure have rarely been experimentally investigated in natural or managed forests of the tropics. The temporal dynamics of genotypic structures of Cecropia obtusifolia, a pioneer species possessing a population structure resembling a metapopulation (Alvarez-Buylla et al. 1996), were discussed in Sect. 6.2.2. Selective logging results in the extinction of local populations of at least some species in species-rich forests of the humid tropics. Recolonization of an area via migration starts out from surrounding populations possessing a metapopulation structure. If under rotation forestry all the mature trees are harvested before reproduction, artificial regeneration is obligatory. Since this is mostly done by using reproductive material collected elsewhere, an artificial process of migration is induced. Its genetic implications will be discussed in Sects. 11.5 and 12.6.

10.5 Genetic Aspects of Sustainability in Natural Tropical Forests With increasing knowledge of the complex structure and functioning of forest ecosystems, the meaning of the term sustainable forestry has changed repeatedly. This is due to frequent redefinitions of human relationships with nature. Also, definitions of sustainability presented in the literature depend very much on the view taken by the respective authors. Principles derived for sustainable management must be ecologically viable, economically feasible, and socially desirable (Salwasser et al. 1993). Bass (2001) considers sustainable forestry in terms of the security of specific forest goods and services. The book by Ferguson (1996) represents a successful attempt to bring ecology and economy together. Gregorius

10.5 Genetic Aspects of Sustainability in Natural Tropical Forests

(2001) explained the sustainable treatment of resources under the aspect of systems theory and drew conclusions on some basic forestry operations. Essential elements of the definition presented by Bruenig (1996, p. 270) are “Sustainability is the capacity of a system in its entirety to persist and survive. Within the time dimension of human perception, it relates to permanence and steadiness of states and processes in contradiction to the fact that there is no steady-state in ecosystems. In forestry, sustainability is a comprehensive ethical principle, substantiated according to agreed criteria which are monitored by agreed indicators of certain forest properties and judged by recognized standards of ecological, technical, economic and sociological state or flow variables.” The capacity to persist and survive implies the capacity of self-regulation and adaptive change. This author also addresses the complexity of the relation between sustainability and biodiversity. The latter term refers to “. . . the variety and variability among living organisms and the ecological complexes in which they occur. Diversity can be defined as the number of different items and their relative frequency. For biological diversity, these items are organized at many levels, ranging from complete ecosystems to the chemical structures that are the molecular basis of heredity. Thus, the term encompasses different ecosystems, species, genes and their relative abundance.” This comprehensive definition made in 1987 by the US Government’s Office of Technology Assessment has been discussed by Boyle (1994). The need for the integration of genetics and ecology was stressed by Hattemer and Gregorius (1996), because regulatory adaptation and thus the stability of an ecosystem are unthinkable without the genetic variation of the species spectrum (Noss 1993). Guidelines for sustainable management can be derived from the definition of sustainability. Ecological sustainability has been overshadowed by the aspect of sustained production in view of human demands of wood, timber, and other forest products. With regard to the future of tropical rain forest ecosystems, Honadle (1993) analyzed causes of tropical deforestation and discussed ways to get away from current “human predation on tropical forests.” This author proposed strengthening of certain nongovernmental organizations and international processes in order to promote sustained tropical forestry. Absolute requirements for sustainable management of tropical rain forests are effective measures to diminish the impact of uncontrolled logging practice (p. 222 f. in Whitmore 1998). Bruenig (1996, Chap. 3) explained the concept of sustainable tropical rain forests and derived strategies for their sustainable management. Bruenig (1996, p. 131 f.) and von Gadow (2005, Chap. 4.5) described the historical roots of the scientific concept of sustainable forestry in central Europe. However, there is no doubt that among forest-dwelling people in traditional societies the attitude towards the forest has been sustainability-

141

142


oriented the world over. Unfortunately, the rural poor in the respective parts of the world suffer most from the forest decline. Today, the ethical imperative of intergenerational fairness and justice deserves most weight among the motivations for sustainable forest management. This anthropocentric vision has been accepted by the Brundtland Commission (World Commission on the Environment and Development 1987). The present generation of men and women should not impair the potential of future generations to cover their demands. The idea of intergenerational justice might well be extended to intercontinental justice, since the per capita consumption of natural resources still varies greatly among continents but the world’s climate relying strongly on globally sustainable forestry is shared by all living beings. Humans are not only an integral part of forest ecosystems but play an ever more prominent role in them. Primack and Corlett (2005) analyzed the decline of tropical rain forests and identified its complex causes. They described the international economic connections and stressed the responsibility of countries outside the tropics for the decline. Among possible solutions the authors discussed also certification of forests and forest products (see also p. 223 in Whitmore 1998; p. 144 f. in Montagnini and Jordan 2005; Bennett 2001; Putz and Romero 2001). In future, genetic fingerprinting methods might play an important role in the prevention of illegal felling in tropical forests, forest certification, and observation of the chain-of-custody. The development of molecular genetics tools to test statements concerning the origin of tropical wood from certified forest enterprises is ongoing (Finkeldey et al. 2006). Recent definitions of sustainable forest management account for the dynamics of the systems and try to consider all factors necessary for the long-term preservation of a balanced status. An assessment of the stability of a forest ecosystem and the sustainability of wood and timber production requires not only detailed analysis of nutrient cycling (Ulrich 1981) but also of the information budget of tree populations as triggered by their genetic system. Sustainable management of forests implies not only the maintenance of a balanced nutritional status and other chemical aspects. The functioning of a forest ecosystem rests also on information, which is predominantly stored as genetic information of the animals and plants that are part of the ecosystem. Sustainability does not imply the preservation of a fixed status, for example, of certain genetic structures. The objective is the continued existence of the ecosystem itself rather than the preservation of certain components such as individuals, genetic variants, or species. A necessary condition for the longterm persistence of an ecosystem is the preservation of the evolutionary adaptability of its keystone species (Sect. 7.2). Thus, apart from economic necessities, the sustainable management of a forest requires the conservation of the genetic resources at least of some keystone species (Ziehe et al. 1989; Chap. 14).


Most tropical tree species harbor considerable amounts of genetic variation (Sect. 3.4) and are thus capable of evolutionary change. Preliminary experimental results indicate flexibility of the reproduction systems and temporal dynamics of genetic structures of tropical forest trees similar to those of tree species of the temperate and boreal zone. Therefore, sustainable silvicultural management techniques must not reduce the genetic variation and the adaptive potential of keystone species. Prabhu et al. (2001) explained the value of a useful concept for defining sustainable forest management as well as for assessing and for monitoring the sustainability of forest management systems. Criteria and indicators are to outline conditions that should be met if forest management is to be deemed sustainable (Boyle 2000). Criteria, standards by which a thing is judged, are not direct measures but refer to large-scale combinations of information. An example may be “processes including human intervention that contribute to the maintenance of genetic variation.” Indicators are variables used to infer attributes of the sustainability of a forest and its utilization. They must be relevant to the goal of assessment and must deliver precise, meaningful, and reliable information. They must also be easy to detect. They may be either absolute, rankings, or conditions relative to a threshold. An example may be “ongoing directional change of genotype or allele frequencies due to selection.” This indicator can be verified by repeated genetic surveys (Boyle 2000). Another example, “distances between forest fragments can be bridged by pollen transfer,” can be verified by the presence of nonlocal private alleles in the progeny of neighboring populations. It is difficult to assess the sustainability of tropical forest ecosystems and their adequate management, particularly if economic and also sociological aspects are considered; therefore, the description of a condition by a few indicators was once criticized for being reductional. On the other hand, as Loyn and MacAlpine (2001) put it, “sustainable forest management demands that we manage to sustain the whole ecosystem, and monitor appropriate subjects which indicate how well we are doing. We have no choice but to use indicators” (Raison et al. 2001). If only genetic sustainability of ecosystems and their management is concerned, things become easier. Because of the widespread implementation of population genetics models that are appropriate for causal interpretation of empirical findings, forest geneticists are better off. Nonetheless, genetics may only contribute its share to the analysis of sustainability jointly with other fields. Adaptive and nonadaptive changes of genetic structures are caused by the evolutionary factors drift, gene flow and migration, selection, and mutation, and the mating system. As a rare event, mutation is not an essential factor during overseeable time periods and is neglected in this context. Selective logging, grazing, forest fires, utilization of forest by-products, and conversion of forest to other

143

144


land uses are examples of human influences on forests that are also expected to change genetic structures (Namkoong et al. 2002). These authors proposed the following four general indicators in the context of genetic variation as the basis of adaptability: genetic drift, selection (in the widest sense), migration, and the mating system. Some relationships are presented in Fig. 10.2. The most fundamental ecological indicator is the amount of available habitat in the landscape. Prevalent forms of forest destruction are conversion for food production by clearing and fire. They are not considered selective in the genetic sense. Loyn and MacAlpine (2001) have proposed the formulation of ecological indicators of fragmentation. Their discussion is very helpful for describing the physical conditions for gene flow with due detail. The mating system, although part of the reproduction system, cannot be observed directly but must be reconstructed from findings on the latter by routine methodology. It includes also selection processes. Therefore, the closeness of the reproduction system to panmixia is presented as an indicator of the mating system. The difference between the census number of populations and their parental pool size was addressed in more detail in Chap. 9. Sexuality refers to sexual function of monoecious genotypes and is thus used as a genetic verifier. On the other

Human influences

Genetic processes

Forest destruction

Selection

Directional genetic change

Verifiers genetic

demographic

Gene frequencies

Population differentiation

Habitat fragmentation

Logging

Indicators

Drift

Level of variation

Migration

Mating isolation

Phenotypic means Age/size class distribution

Multiplicity,

Census no.

Diversity etc.

Parental pool size

Gene flow (Effective pollen transfer)

Physical isolation Seed dispersal

Non-timber harvesting

Mating system

Closeness of reproduction system to panmixia

Sexuality Genotype frequencies Correlated matings Homozygote excess Spatial genetic structures

Sex ratio Pollinator abundance Seed germination Mortality

Fig. 10.2. Genetic processes in tree populations as influenced by forest events, their indicators, and their verifiers. Dashed arrows represent only weak relationships. For some further explanations see the text. (Adapted from Namkoong et al. 2002)


hand, the morphological sex ratio of dioecious populations is used as a demographic verifier (Sect. 4.2.3). Not all relationships are indicated in Fig. 10.2. The genetic verifiers presented are but examples. Demographic verifiers are difficult to show in their relation to indicators by this type of presentation. For instance, seed germination and mortality have a close relationship to selection rather than to the mating system. The genetic indicators and their response to forest management practices can be assessed by periodic surveys of the variation at gene markers and of quantitative traits. Since this genetic monitoring using appropriate markers (Weising et al. 2005) requires much effort, only exemplary surveys can be made in some populations. Additional indicators are necessary in order to assess the success of measures aimed at the conservation or restoration of genetic variation by forest management practices. Naturally, genetic indicators are valuable for monitoring only to the extent that the respective processes can be managed (Loyn and MacAlpine 2001). According to their economic and/or ecological value, keystone tree species possessing a variety of genetic characteristics and life-history traits should be selected for monitoring. They may be suitable indicators representative of many other organisms of the ecosystem. The assessment of a whole group of species is worthwhile in view of the cost and the amount of information to be gained (Loyn and MacAlpine 2001). Quantitative phenotypic traits as well as biochemical and molecular genetics markers should be observed in order to assess the dynamics of genetic variation of different parts of the genome at adaptive and nonadaptive gene loci. Continuous refinement of methods in response to the increase in knowledge gained is imperative. The set of indicators presented in Fig. 10.2 may be specified and improved considerably. If the results of periodic assessments of indicators are treated as repetitions of an observational study, they may be integrated into something like adaptive management (Prabhu et al. 2001). The lay public seriously devoted to biological conservation looks at any management of tropical forest as a “disturbance.” It is true that any human interference somehow disturbs natural processes but not every disturbance is alarming in view of the sustainability of forests. Also, tree species and forest ecosystems differ in their ability to regenerate. We may not forget that even though selective harvests of one or a few species may leave the forest environment physically undisturbed, mating partners for the harvested species may be so diminished that regeneration is at least locally unlikely (Namkoong et al. 2000). Nevertheless, disturbances occur also in nature and challenge adaptive processes in populations. Furthermore, in view of immense human pressure on forests all over the globe, the appropriate management of forests is a necessity just as is their utilization. Unless at least some regional economic value is assigned to forests, human societies will hardly be convinced that forests are

145

146


something worth conserving at all. Not much tropical forest will survive that is not logged sooner or later. Therefore, the integration of appropriate logging regimes into forest management will be of more help to forests than mere concern about the inevitable. However, the avoidance of major disturbances is imperative. If the new environments differ much from the predisturbance environments, then adaptation to novel environments requires a large adaptive capacity. This can be carried only in a large pool of potential parents available for reproduction. In this case, natural selection may still not be able to increase ecological security or economic productivity rapidly enough (Namkoong et al. 2000). It should have become clear by now that it is not justified to consider any change in genetic structures at any marker gene locus induced by a certain management practice as a violation of the principle of genetic sustainability (Finkeldey and Ziehe 2004). Genetic sustainability in the sense of MüllerStarck (1996) is impaired only if the state of adaptedness or the adaptive capacity (Ziehe et al. 1999) is endangered.

10.6 Recommended Literature Stern and Roche (1974, p. 233 ff.) have presented a fairly general description of how humans affect forest ecosystems. Finkeldey and Ziehe (2004) reviewed a large number of studies made on the genetic impact of forestry in both temperate and tropical forest. The focus of their paper is on forestry operations such as artificial regeneration, soil amelioration, thinning and harvesting, and the regulation of forest density and species composition. They come to the conclusion that forestry induces most severe and detrimental genetic change in tropical forests. Finkeldey (2002) has formulated a basic hypothesis on the genetic implications of continuous-cover forestry. Its implied assumptions are discussed and some pieces of experimental evidence from both moderate and tropical forests are presented. This hypothesis applies to any type of continuous-cover forest. A joint test of all of these assumptions in one or several representative examples is encouraged.

Provenance Research

11.1 Introduction During many generations the genetic structures of autochthonous populations have developed under the influence of the evolutionary factors in conjunction with historical events such as catastrophes. Since the totality of these effects is likely to differ between the habitats of conspecific populations, the genetic structures of these populations are likely to diverge. This differentiation is more or less pronounced depending on the intensity of those effects, particularly selection, and the length of the time period during which they have been active. The model-supported study of the genetic differentiation of populations at marker gene loci has opened the way to its causal interpretation. However, populations usually differ only little at biparentally inherited marker loci. This contrasts with often much higher differentiation at adaptive and economically important traits (Kremer et al. 1999). The study of the notable phenotypic differentiation of populations is closely connected to their adaptation to the different environmental conditions of their native habitats. Much regional variation in day length, average temperature, and precipitation enforcing adaptation exists in the temperate and boreal zone. In the tropics and subtropics the rainfall regime varies more and takes precedence (p. 55 ff. in Morgenstern 1996). The main objective of ecological genetics is to learn about the evolutionary genetic change in plant populations brought about by ecological factors in their natural habitat (Sect. 2.5). Provenance research has also a direct application: not only the expression of adaptive traits but also other traits determining the value of populations for use may be tested in experiments. On the basis of the results, populations may be transferred to be planted elsewhere in order to increase stability and production. Hence, provenance research can also be seen as a first step towards breeding and domestication. Only those populations can be bred that justify the subsequent investment. In quite a few instances, arbitrary reproductive material has been introduced to other countries. Only later it was found out that by prior systematic sampling and testing, large vain expenditure may have

11

148

C HAPTER 11 Provenance Research

been avoided. Owing to the rapid destruction of natural forests, there is an urgent need to accelerate provenance collections (Ladrach 1998).

11.2 Definitions The geographic site where an autochthonous population has undergone its evolutionary development is called its origin. With reference to origin, the autochthonous population itself is commonly called a provenance and a test for measuring its phenotypic traits in comparison with those of other populations and in a controlled environment is, consequently, called a provenance test. The term “provenance” is also used for a sample (drawn at the seed stage) from an autochthonous population and the planting material raised therefrom to be included in a provenance test. Unfortunately, the term “provenance” is further used for the origin of an autochthonous population. In legislation on forest reproductive material the term is used even more generally for a “place in which any stand of trees is growing” (EU 2000). This latter definition does not discriminate whether this place is not only the growth site but also the origin of the respective population that is thus autochthonous or whether it is allochthonous there. Although the latter usage is certainly appropriate for a legal text, it has led and still leads to regrettable misunderstandings in the discussion of its genetic foundations. Zobel and Talbert (1984, p. 80) pointed to the existing ambivalence and proposed denoting by “provenance” a place and by “seed source” the trees growing there and from which seed is collected. Also Zobel and van Buijtenen (1989, p. 33) and Evans and Turnbull (2004, p. 138) used this definition that avoids misunderstandings about a site and the trees growing there. This distinction helps to cure but part of the problem, since “provenance” neither indicates the origin of a population nor indicates “seed source,” whether a population is autochthonous or not (Jones and Burley 1973). Ladrach (1998) emphasized the need for clear definitions in this field. An allochthonous population is regarded as introduced or exotic, if it grows outside the natural distribution range of the respective species. An exotic plant species is also denoted as a neophyte. A population that has been transferred by humans within its natural distribution range is neither exotic nor autochthonous anymore. Unfortunately, the attribute “indigenous” is not helpful in relation to autochthony. For instance, in EU (2000) it denotes a stand or seed source that is either autochthonous or that has been “raised artificially from seed, the origin of which is situated in the same region of provenance.” The latter term refers to a group of areas subject to sufficiently uniform ecological conditions, in which seed sources show similar phenotypes or genetic characters. This term

11.2 Definitions

is again spongy under a genetic aspect and may illustrate the need for clear definitions that might help to advance the discussion on the choice of appropriate forest reproductive material. A race is an aggregate of several similar populations that are differentiated from other races of the same species genetically or by particular expressions of genetically controlled phenotypic traits. As an intraspecific division, races are still connected by common descent. Just as populations, also races are difficult to delineate. This is particularly true if they occupy the same habitat and are thus sympatric. Owing to the effect of gene flow, also allopatric races occupying different territories present problems of clear-cut delineation. The phenotypic traits used for the description of races should be under strict genetic control, i.e., their expression should show no or only minor environmental influence. In plant taxonomy, races are often described morphologically as subspecies or varieties. In the northern temperate and boreal zones, tree races often originated in glacial refuges and have persisted during their postglacial reimmigration. Also strains of animals or plants developed under the control of humans are called races (Sect. 12.2). An allochthonous population that after its transfer to its new growth site has developed particular genetically controlled trait expressions owing to adaptation is commonly called a land race (p. 89 in Zobel and Talbert 1984). The genetic change is attributed to natural selection, artificial selection of trees with the best performance as in breeding, or silvicultural selection. Eldridge et al. (1994, p. 251) consider it essential that the respective population is exotic. Owing to the low number of generations since then, it is hardly justified to expect substantial adaptation of the genetic structure of a land race to its new environment. Zobel and Talbert (1984), Eldridge et al. (1994), and Ladrach (1998) consider a population with the quantity of genetic change as achieved during even only one generation a land race. Since a population originating from seed collected in its progenitor population elsewhere can then have achieved improved adaptedness only through viability selection, the customary definitions imply strong effects of this type of selection that are linked to high mortality. The aforementioned authors stress that the quality of the provenance used in the first planting is crucial for the quality of the developing land race. This means that virtually all allochthonous populations, particularly all plantations of exotic tree species, deserve the attribute “land races.” For instance, first-generation stands of North American conifers such as Douglas fir (Pseudotsuga menziesii) and eastern white pine (Pinus strobus) planted early in the twentieth century in central Europe have been called land races if they show good growth and their offspring show better health. In field tests with Pinus sylvestris, an exotic species in the Netherlands, Squillace et al. (1975) found that the progeny of trees from foreign sources planted before 1900 grew faster and were more resistant against the needle cast disease than the progeny of trees from Dutch sources planted

149

150


later. A geographic pattern also existed, since offspring of trees in the northwest of the country grew faster and were more resistant than those from the south. The authors considered it highly likely that the combined effect of natural adaptation and artificial selection was responsible for this gradual geographic gradient. Eldridge et al. (1994, p. 31 f.) discussed the status of exotic eucalypt plantations and stated that “there is not much evidence in writing about the improvement in eucalypts and other trees one generation after introduction from natural stands.” They attributed the superior growth of offspring of Eucalyptus urophylla planted in the Congo in comparison with the growth of the offspring of the progenitor population in Timor to the eventuality that the neighborhood inbreeding effects of the natural stand had been broken down in the plantation. Ladrach (1998) discussed several reports on land races in tropical tree species. The critical factor in their formation seems to have been artificial selection rather than natural adaptation. Consequently, some land races that have arisen on one continent showed superiority over natural provenances elsewhere. In some instances, the poor seed production at the new growing site had to be compensated by intervention of humans. In future the importance of land races will certainly increase. The close connection between their emergence and domestication will be discussed in Sect. 12.2. Processes inducing immediate adaptedness may without doubt yield arguments for the development and the existence of land races in trees.

11.3 Historical Development The performance of forest tree plantations critically depends on the origin of the reproductive material used for their establishment. Uncritical use of reproductive material regardless of its origin has led to adverse experiences made with conifers planted in several European countries. This insight gained considerable attention from scientists and foresters concerned with the establishment of conifer plantations. The foundation of the Verein deutscher forstlicher Versuchsanstalten (Association of German Forestry Research Institutes) comprising the German, Austrian, and Swiss forestry research institutes in 1892 and later, in 1929, the formation of the International Union of Forestry Research Organizations (IUFRO) have been intimately linked to research on the growth of different provenances of economically relevant conifer species. From its very beginning this research field has strongly relied on international co-operation and the exchange of research material among countries. The history of the research on geographic variation in forest trees was described by Morgenstern (1996, Chap. 1). The first internationally co-ordinated provenance trials for tropical forest tree species were established during the 1960s. Their main focus again was on

11.4 Provenance Tests

conifers. For example, the Oxford Forestry Institute co-ordinated a worldwide series of provenance trials of Pinus caribaea (Barnes and Burley 1990). Provenance trials and breeding populations were also established for other tropical pines from Central America and Southeast Asia as well as for some other conifers, e.g., Araucaria spp. (see Burley and Nikles 1972, 1973 for details). At the same time the Food and Agriculture Organization (FAO) of the United Nations organized and co-ordinated provenance trials of Eucalyptus species in co-operation with Australian agencies (Example 11.1). Extensive work has been done on E. camaldulensis. A series of teak (Tectona grandis) provenance trials followed in 1973 and 1974 (Example 11.2). The Danish DANIDA Forest Seed Centre co-ordinated these trials (Hedegart 1976; Kjær et al. 1995). More internationally co-ordinated trials were established following changed priorities for preferred plantation species in tropical countries. Provenance trials were established for Gmelina arborea, Acacia spp., Dalbergia sissoo, Casuarina equisetifolia, Cedrela spp., Leucaena spp. and other Central American legumes, Calycophyllum spruceanum, and neem (Azadirachta indica) (Burley 1973; Awang 1994; Kundu and Tigerstedt 1998; Sotelo Montes et al. 2003). Endo (1994) and Ladrach (1998) reported an impressve number of provenance studies in Middle America made by CAMCORE. Many more institutions and organizations have been active in provenance research and subsequent breeding of tropical trees or have supported projects in this field, but they cannot possibly be listed here.

11.4 Provenance Tests A survey of genetic marker gene loci covering different tree populations can readily be made using plant parts or seeds collected in those populations themselves. Comparisons between populations in phenotypic trait expressions are not as easy to accomplish. The differentiation in phenotypic traits, particularly the differential adaptedness, of provenances can be tested either in long-term field experiments or in short-term tests under controlled conditions in phytotrons, greenhouses, or nurseries. The results o the latter type of experiments are valid only if some overriding environmental factors can be readily simulated in the controlled environment. For instance, Bekessy et al. (2003) tested Chilean and Argentinean provenances of Araucaria araucana in the greenhouse and found much higher drought resistance in those from the drier regions east of the Andes. Field trials require much more time, cost, and effort. They are nevertheless indispensable experiments for genetically efficient selection of populations

151

152


suitable for the establishment of plantations and for the utilization of genetic variation within populations by breeding. The main practice-oriented objective of the single field trial is to compare average trait expressions of the experimental entries in economically relevant traits in a uniform environment. As will be further described in Sect. 11.4.3, exposure of the experimental entries to a variety of environmental conditions increases the information gained considerably. Provenance tests represent also a collection of material that can be utilized for more intensive breeding purposes. Provenance trials are only one type of genetically motivated field trials. Their establishment is an integral part of most breeding programs (Chap. 12). Their methodology is not much different from that in forest tree breeding, and many basic considerations of planning apply to any type of genetically motivated field trials. Methods for their design, establishment, and analysis have been covered more thoroughly in books such as those by Wright (1976) and Williams et al. (2002). The latter publication contains worked examples and reports on several experiments in tropical trees. Davidson (1995) prepared a useful manual for this type of experimentation with tropical trees. It is open to debate and to empirical research with model plants whether part of the observed differentiation of provenances is the result of imprinting, since the seed used for the experiments was produced under possibly very different environmental conditions. In experiments requiring great effort, Skrøppa and Johnsen (1999) found that offspring of Scandinavian Norway spruce, Picea abies, possessed considerably less tolerance to frost if the seeds were produced in a heated greenhouse (Johnsen et al. 2005a, b) . Although this species is adapted to a cool climate, it obviously loses its adaptedness owing to raised temperatures during seed maturation. Conversely, it acquires cold resistance only during embryo development under lower temperatures in the field. Hence, the behavior under different temperature regimes is controlled by the expression of the genes responsible rather than the accumulation of adaptive genes during many generations. The observed change is achieved during a few months at a crucial stage of the life cycle. This result is undoubtedly relevant also for trees in warmer climates, because drought stress (Vornam et al. 2003) had a similar effect as frost stress. Since corresponding experiments with tropical trees have not yet been made, genetic variation has still to be considered the main basis of their evolutionary adaptability. The conservation of adapted, genetically variable and thus adaptable populations is of particular importance, since rapid and pronounced environmental change is predicted. Examples of expected environmental change in forest ecosystems are air pollution and emissions in general, global warming, the conversion of pristine forests to production forests, and other forestry practices. The conservation of genetic variation within and among populations is


a necessary component of sustainable forestry in view of the unforeseeable future of forest ecosystems. Provenances of most tree species that have been tested so far differ in the expression of many phenotypic traits, however, this differentiation is rarely paralleled by findings on genetic structures at many marker loci (Bekessy et al. 2003). A wealth of results of various provenance trials in tropical tree species have been published during the past 30 years. They confirm that strong differentiation among populations is the rule for almost all species studied in many economically relevant traits. These results seem to contradict surveys of marker gene loci that frequently have proved only little differentiation among populations. Typically, the proportion of the total genetic diversity due to differentiation among populations (FST = GST; Sect. 3.3.2) at marker loci is less than 20% (Sect. 3.3). However, one has to consider the different modes of gene action in genetic markers and in those involved in the control of growth traits (Lewontin 1984). Also, different evolutionary factors might have shaped the genetic structures at currently available marker gene loci and at those gene loci that contribute to particular trait expressions. The only slight differentiation of tree populations at marker loci is attributed to their life-history traits and, in particular, widespread gene flow (Austerlitz et al. 2000; Chap. 5). The contrasting condition in phenotypic traits must be explained by selection; however, selection was shown to cause strong differentiation at an adaptive trait but only low differentiation at the controlling gene loci if many genes, each with only minor effects on the phenotype, are involved and if gene flow and diversifying selection are high (Kremer et al. 1999; Le Corre and Kremer 2003). If selection resulted in pronounced differentiation in spite of ongoing gene flow, the selection against immigrants is expected to be strong. These immigrants are presumably maladapted and represent a severe genetic load in the receptor populations. This condition has some bearing on the choice of provenances for the establishment of plantations (Sect. 11.6). It has also some bearing on the choice of populations as genetic resources (Sect. 14.5). 11.4.1 Types of Field Experiments in Provenance Research

In view of the cost, provenance trials aimed at the study of ecological genetics problems might also serve practical goals. It has to be kept in mind that provenance tests are planted at regular, wide spacing on cleared land without the protection of a canopy in order to create an environment as uniform as possible. These highly artificial conditions are exactly the same as in plantations; therefore, the results are expected to yield information primarily on the behavior of trees in plantations rather than in the forest communities to which they belong. Although provenance research ranks high in priority, many foresters

153

154


are only little aware that the implementation of its results ultimately leads to stands to be established with allochthonous populations. Several types of practice-oriented field trials have evolved: 1. The objective of a species trial is to assess the general suitability of a species as a plantation species for a specific region. This practice-oriented type of experiment is also called a species-elimination trial, since it is aimed at testing a large number of species potentially suitable for artificial regeneration. It serves also for the exclusion of those from further experiments that would obviously fail in plantations. Owing to the fast juvenile growth of tropical tree species, some rough information can be gained after only 3–4 years, so the duration of such trials can be rather short (p. 62 ff. in Davidson 1995). Ideally, several species are identified as potentially suitable for plantation establishment at the end of the trial. These species will be included in further test plantations. Species elimination trials are pilot studies for practice-oriented provenance research. 2. The performance of several populations of species that have been identified as potential plantation species is compared in provenance trials. Their objective is the identification and selection of provenances that are particularly suitable for plantations on the basis of the expression of economically relevant quantitative traits. In “first-generation” tests populations covering possibly a large part of the natural distribution range of a species are included. Occasionally, locally well tried land races are also included as a reference. Populations that proved to be most suitable as seed sources for plantations are entered into restricted “second-generation” provenance trials. All of these experiments are practice-oriented. Barnes and Simons (1994) even state that before embarking on a breeding project both the full natural range and the exotic range of the species should be sampled. In fact, the progeny of exotic plantations performed among the best-growing entries in several provenance experiments. 3. Some types of field trials may be combined in order to maximize information gain and to save time. For example, species-provenance trials have been established containing a few species each represented by several provenances. These are particularly useful if decisions on the introduction of exotic species are to be made. Also progenies of single seed trees may be kept separate in provenance trials, unless the number of experimental entries exceeds given limits. If the identities of single tree progenies are maintained in a provenance trial, information on both genetically based differentiation among populations and genetic variation within populations may be obtained. In tropical countries some combined provenance-progeny trials have been established (Kanowski and Nikles 1989; Chap. 12).


11.4.2 Traits Studied in Field Trials

In view of the cost involved in field work, periodically several traits determining plantation yield might be observed simultaneously. Several traits are usually recorded that characterize growth speed, such as tree dimensions reached during given periods of time. At a young age the most important trait is plant height. In later stages of the experiment, tree diameter (at breast height) is recorded for the sake of easier measurement. Viability and tree growth indicate harvestable yield per unit area, if not even stem volume or the total biomass produced above ground can be measured directly. Occasionally traits characterizing the health condition of the trees such as leaf or needle color or the resistance against an insect or pathogen are relevant. The presence or absence of symptoms of disease makes sense only if there is notable infection pressure. The same applies to resistance against insect attack. Other traits codetermine the use of provenances for commercial plantations, i.e., the quality of the harvested product. Examples are morphological traits such as stem straightness and branching habit (number, size, or angle of branches). Depending on the planned utilization of plantations, wood characteristics such as fiber length and cellulose content are measured. Sprouting ability is an important trait if future plantations are to be coppiced (Quaile and Mullin 1984). Palmer (1994) has presented a refined catalogue of desirable properties of tropical forest trees. Only some of these traits are measured on a continuous scale and are quantitative in the close sense of the word. Most traits characterizing growth (height and diameter growth etc.) are quantitative; however, the viability is a trait with only two possible trait expressions in the individual tree (survived or dead). Many traits characterizing the health status are also qualitative with two expressions only (infected or not) or are recorded in discrete classes. Some traits such as growth form and stem straightness are actually not measured but scored and are therefore subject to personal bias. Some of the traits recorded in field trials are closely associated with the state of adaptedness of the trees to the environmental conditions at the trial site. This obviously holds true for viability and for traits characterizing tree health, particularly under environmental stress or exposure to infection. Growth traits are not necessarily positively correlated to fitness. Other economically important traits, for example, stem straightness and length of branch-free bole, are hardly associated with fitness. Trees showing superior growth may even possess lower viability. Fitness components connected to fertility and reproduction are only rarely observed in field trials. Thus, the provenances that are identified as superior in field trials are not necessarily the best-adapted to the environmental conditions at the experimental site.

155

156


11.4.3 Design, Conduct, and Analysis of Provenance Trials 11.4.3.1 Collection of Material for Provenance Trials

The modalities of seed collection are a largely neglected element of the methodology of provenance trials. Principles of seed collection must be observed in order to prevent the induction of undesirable genetic change. Many details of the reproduction system that are of relevance for seed collection have been studied by means of genetic marker investigations. Above all, the genetic structures of the seeds produced by a tree population vary over time. In order to achieve a seed sample that possesses genetic structure representative of its basic material, seeds produced in several years have to be mixed. This in turn would require seed storage until material from all seed sources could be collected repeatedly, an impossible undertaking. The seed of many tropical tree species, for example, the dipterocarps, cannot be stored for long since they are recalcitrant (Schmidt 2000). Consequently, seed should be collected in a year of abundant seed production or just after most of the trees have flowered abundantly, as Hodgson (1977) recommended with regard to seed orchards (Sect. 12.8.1). The seed lots collected from individual seed trees are genetically differentiated. This is due to the variation among the individual genotypes of the seed trees themselves and to the differentiation of their individual effective pollen clouds (Sect. 6.2.1). The less the seed lots differ, the easier it is to arrive at a representative sample of the genetic structure of the reproductive output of the seed source. Figure 11.1 shows factors responsible for the differentiation of the individual effective pollen clouds. The higher the density of the population, the more efficient the system of pollen dispersal, and the more homogeneous the physical structure of the population are, the larger is the expected effective number of trees involved in cross-fertilization and the less are effective pollen clouds expected to be differentiated (Sect. 6.2.1). Excessive self-fertilization that might have an impact on phenotypic trait expressions may easily be tested in small seed samples before a major collection. Genetic incompatibility (Sect. 6.5.1) and the degree of self-sterility (Sect. 6.5.2) are beyond human control. However, sufficient size of the seed source helps to capture many S-alleles and provides for high fertility of stands to be derived from a seed source. The effect of the genetic differentiation of neighborhoods conditioned by spatial genetic structure, correlated matings, and hence a low number of effective male mates of seed trees contributes directly to genetic differentiation among the seed lots. The more pronounced the mass flowering of the trees, the less may flower phenology contribute to this differentiation.

11.4 Provenance Tests Population density

Effective distance to potential male mating partners

Wind speed and wind directions Pollinator behavior and sedimentation speed of windborne pollen

Influx of external effective pollen

Differentiated individual effective pollen clouds

Physical stand structure

Spatial genetic structures Flower phenology

Incompatibility and self-sterility

Correlated matings Excessive self-fertilization

Fig. 11.1. Causes of genetic differentiation of the pollen that becomes effective in individual trees of a population (individual effective pollen clouds)

In forests consisting of just one dioecious tree species the differentiation of the seed lots produced by single trees can be just as pronounced as in speciesrich forests. For instance, the 35 seed lots sampled in each of two populations of the dioecious, wind-pollinated conifer Araucaria angustifolia shared only 70% of their alleles at six enzyme gene loci (Sousa et al. 2005; Sect. 5.2.2). On the basis of correlated matings, the average effective number of male mating partners was estimated to be only 3 or 4. Obviously, the female trees represented by half of the population greatly impair pollen dispersal. However, the sedimentation speed of the pollen in this species is the highest hitherto found in conifers, which may make this species unique. The differentiation of the seed lots produced by single trees calls for seed collection from many trees. If seed collected from individual trees is bulked arbitrarily, as is done in many population genetics surveys, genetic structures depend very much on the incidental amounts contributed by individual trees to the total sample. Consequently, it is hard to predict whether a close or only loose genetic relationship between a population and its reproductive output is detected. Furthermore, the modalities of sampling play a role in whether panmictic equilibrium (Sect. 6.2.2) is reached. Approximately equal shares of the individual seed trees are desirable. Since the vastly different amounts of seed produced by individual trees are primarily due to their different crown sizes reflecting their different competition situations

157

158


rather than genetic effects, it hardly makes sense to collect amounts that are roughly proportional to their seed production in the given year. A minimum distance between the seed trees (at least double tree height or 100 m) helps to avoid harvesting seeds from related trees. This recommendation is particularly important if the gene flow system suggests the existence of a family structure in a native population. Seed collection from close neighbors also increases the risk of sampling progeny from reciprocal matings. Owing to scarce reproduction in parts of the distribution range of the species, the seed required for a provenance trial cannot be collected in 1 year. Storage until completion is not possible for many tropical tree seeds. An expedient solution of this conflict is to split up the experiment and compare subsets of the provenances on the basis of the same age reached in different years. The storage of seeds as practiced for many tree species of the temperate zones is impossible for many tropical trees with recalcitrant seeds. One has to be aware that the losses of germinability that are unavoidable even in conifers of the temperate and boreal zones lead to genetic change of the stored seed. The mortality during seed storage is selective as shown by Melchior (1986a) for aspen (Populus tremula) seeds. Very common in the past have been “mail-order collections.” In these cases foresters in the region of origin of the seed sources were contacted and were asked for shipment of local seed lots. The colleagues sent a sample of seed, mentioning at best the number of seed trees, the distances between those trees, and the proportions of seed collected from them. Many provenance experiments with conifers of the northern hemisphere started out from seed lots representing unspecified mixtures of seed from normally “at least ten” trees. It is easy to comprehend that trait expressions observed in such “provenances” are possibly biased considerably. The moderate demand for seed tempts us into seed collection from a small number of trees only. After all, seed must be harvested from a minimum number of trees. Several authors recommended at least ten to 25 seed trees as sufficient to represent the genetic structures of a population (p. 30 in Eldridge et al. 1994; p. 10 in Williams et al. 2002; p. 84 in Davidson 1995). Population genetics considerations suggest that these numbers are insufficient to obtain a representative sample of the genetic structure of a population, since not only the genotypes of the seed trees but also the genetic structures of their effective pollen clouds are differentiated. Certain moderately frequent alleles and allele combinations are expected to be severely underrepresented or overrepresented if seeds are harvested from only a small number of seed parents. Consequently, these numbers represent absolutely lower bounds. However, as Williams et al. (2002) note, even 25 seed trees might be hard to find in small local populations. It is then open to debate whether populations of this size are eligible for entry into provenance experiments.


Selection of the seed trees on the basis of phenotypic superiority should be avoided. The attempt to combine seed collection for a provenance experiment with something like plus-tree selection (Sect. 12.5) may lead to a biased sample. As a measure of precaution, small trees should not be favored as seed parents, although it may be easier to collect seeds from them. Further recommendations apply to the necessary documentation of all operations and the deposition of copies of all records in several safe places. Particularly, the information on the geographic location of the seed sources is important. Good test results of a provenance will be exploited by later mass collection of seed. In this case the seed source sampled for the test must closely correspond to that of later mass collection. For short-term tests it is worthwhile mapping and possibly labeling the seed trees. In most tropical tree species, seed is harvested directly from standing trees. In view of major differences in germinability, keeping single-tree seed lots separate until sowing is advisable. Separation of single-tree progenies is, of course, necessary if the provenance test is to be combined with a progeny test. Last but not least, the seed collected from an autochthonous seed source is more representative of its genetic structure, the less the external influx from large, presumably differentiated, allochthonous stands in the vicinity is to be expected. 11.4.3.2 Production of Planting Stock

In view of the environmental influence on planting stock the seed must be sown out in the same year in the same nursery. It is an absolute precondition of experiments that all experimental material has received equal pretreatment. Particular care is devised during the germination phase. Since the plant material of the experimental entries must receive the same nursery treatment, seed lots and later the seedlings must be carefully kept separate. Losses of plant material have to be minimized, since any such loss might be selective. The resulting genetic change induced by environmental conditions that are atypical of those that the trees are later exposed to may bias the data not only on survival. In some provenance experiments aftereffects of differential nursery treatment have been observed. 11.4.3.3 Experimental Design

Regrettably enough, even many years after the principles of the planning of experiments as conceived by R.A. Fisher had been introduced in research, costly provenance trials with forest trees have been established without paying

159

160


attention to appropriate design. Although older provenance trials were laid out without adequate design, they are still an important source of information on the growth of trees of different origin. This is particularly true for the first provenance trials set out by P.A. de Vilmorin in Pinus sylvestris as early as between 1820 and 1840, and early trials established with tropical trees such as teak (T. grandis) (Coster and Eidmann 1934; Coster and Hardjowasono 1935). However, small differences between provenances can hardly be entrusted and generalization of the results requires caution. Only some general suggestions for the design and the analysis of genetically motivated field trials in forestry will be made. This topic is covered with due detail in other publications (e.g., Williams et al. 2002). Principal elements of experimental design now considered indispensable are replication of the entries, the random layout of the experimental units (plots), and the control of the experimental error. The plots may comprise just one tree but mostly comprise many more trees. Single-tree plots are suitable for shorttime experiments only. Larger plots provide for later thinning when the individual trees require more growth space. They are the method of choice for medium-term tests. Commonly used are square plots with 4, 9, 16, or 25 trees. Larger square plots reduce the effect of competition between adjacent plots with possibly widely different growth speed of the trees growing there. In fourtree plots every tree has more neighbors in other plots than in its own plot. In a square plot with 25 trees, nine of them have no neighbor belonging to an adjacent plot. In row plots the effects of competition between adjacent plots are maximized. Four replicates of an experiment with a plot size of 25 trees leads to 100 trees per provenance. Cotterill (1990) found that for estimating the provenance means with sufficiently low error a total number of 200 would be required. Depending on the heritability of the trait (Chap. 12), five to ten families should be represented among those. Considerations of genetic representativity call for larger numbers of families. The randomized complete block design as proposed by R.A. Fisher 80 years ago has become the most frequently used design for genetically motivated field experiments in forestry, if not even for just any experiments. In this design, a block contains just one plot of every entry. The spatial arrangement of the plots of the various entries within the block is random; therefore, the design is called random and complete. The design comprises several complete blocks (or replicates). Statistical considerations call for a certain number of replicates. If p provenances are to be tested in r replicates, the experiment contains pr plots. If on every plot t trees are planted at a spacing of k meters, say, the experiment requires an area of prtk2 square meters plus the space for one or two border rows. If, for instance, only 20 provenances are to be compared in a short-term experiment containing four replicates with nine trees per plot at


a spacing of 2 m × 2 m, an area of 2,880 m2 is required. After including two border rows, the area is 0.34 ha. Testing the provenances in five replicates with 16 trees planted in every plot requires an area of a little more than 0.7 ha. Testing larger numbers of provenances soon leads to areas that are difficult to get with adequate homogeneity. The same is true with larger plot sizes planned for a longer duration of the experiment and accounting for thinnings. Larger plot sizes reduce the risk of losing plots owing to mortality. Trials in randomized complete blocks as the simplest design are insensitive to the implied increase in complexity and have also for this reason been used widely. The statistical evaluation of other designs requires complicated methodology, if data are missing. However, the recent improvement in computational facilities allows us to analyze also incomplete data, so incidentally missing plots do no longer represent a problem. Now as before, the conformity with the principles of experimental design is a precondition for adequate analysis. The blocks of an experiment may contain only part of the entries. In these incomplete block designs the size required for one block is reduced, which provides for better control of the experimental error than the design using complete blocks. There exist large numbers of incomplete block designs. They are recommended by Williams et al. (2002) as efficient standard designs for experiments with forest trees. 11.4.3.4 Choice of Experimental Sites

The choice of an appropriate experimental site is frequently the most difficult practical problem. The research institutions conducting field experiments often have no access to land that is most desirable for test plantations. The site should be as homogeneous as possible in order to keep experimental error low. The size should not comprise a systematic trend in edaphic or climatic conditions as they occur on a slope. The consequences of a systematic change of environmental conditions can only partly be reduced by an appropriate modification of the experimental design (p. 51 f. in Eldridge et al. 1994). Also sites with obvious heterogeneity of growth conditions such as small-scale variation are unsuitable for experimentation. The trial site should represent the ecological conditions of the region of future mass cultivation of the selected provenances. If the test environment is atypical of the environmental conditions of future use, the transfer of the experimental results to practical forestry is less efficient or may even lead to failure. The establishment of field trials on inadequate sites is one of the most frequent and foreseeable reasons for disappointing results or even complete failures of experiments in tropical countries. The site should also be easily accessible and be safe from destruction.

161

162


11.4.3.5 Number and Distribution of Locations of an Experimental Series

The ecological conditions prevailing at a given site induce certain growth differences between the provenances. Apart from the genetically conditioned growth potential of the provenances, these differences may be larger or smaller, since the provenances respond to the environmental conditions in different ways. Therefore, it is strongly recommended to plant an experiment on more than one site as an experimental series. Statistical considerations suggest that the design of the various trials should be identical. According to experience with many series of experiments, not only the degree of differentiation changes with the planting site but also the ranking of the provenances. For instance, a provenance displaying fastest growth in one location may be among the slow-growing provenances in another (Example 11.1). This phenomenon is described as interaction between provenances and environments. The existence of interaction between genetically defined experimental entries and the environment or G × E interaction is commonly experienced. Interaction in the close sense, i.e., changes in the ranks, calls for the selection of different provenances at different locations. The importance of interactions relative to the variation between provenance means can be measured by the relative sizes of variance components estimated by analysis of variance. Since only changes in the ranks are crucial, the consistency of the rankings may be measured by the concordance statistic (Hattemer 1969). In recent years the evaluation of interactions has been considerably improved with the help of refined statistical models. Raymond and Lindgren (1990) have discussed procedures of deploying provenances over planting sites. They proposed a new approach based on the reaction of provenances to given changes in ecological severity between the origin and the planting site. The locations of the experimental series should have climatic and edaphic conditions covering the typical spectrum of those in the region of future planting. In order to avoid serious setbacks, only those provenances with good performance on several sites should be recommended for planting. The superior performance of such provenances is considered environmentally stable. Recommendations of provenances for planting involve a high risk of failure if they are based on only one or a few field trials. This obviously holds even more true, the more the environmental conditions at future planting sites differ from the environmental conditions at the experimental sites. The establishment of an experimental series requires more planting stock and more area; however, the number of replicates in the single experiments can be reduced. Experimental series require also much more effort in coordination and conduct.

11.5 Provenance Differentiation and Geographic Variation Pattern

11.4.3.6 Observation Period of Field Experiments

The effort required for recording the data from a single tree can be low (e.g., measurement of breast-height diameter) or considerable (e.g., certain physiological variables such as water-use efficiency). In most cases data are recorded repeatedly in order to get both early results and information on their temporal dynamics. The estimated means of the various traits are used for ranking the provenances. The ranks may change considerably during the duration of a trial. For example, in a trial in northern Thailand a land race of Pinus merkusii from Papua New Guinea proved to grow much faster than provenances from Thailand (Example 14.1). However, with increasing age of the trial this land race lost its superiority in height growth over provenances from northeasternThailand. No significant relation between the height of trees at the age of 3 and 13 years was observed in this experiment (Changtragoon 1984). Thus, with increasing age of the trial, conclusions on provenances with superior growth had to be repeatedly revised. The length of the observational period depends on the objective of the experiment, the species, and the traits to be assessed. Species-elimination trials in tropical countries are often restricted to a duration of 3–4 years. This time is considered sufficient to assess the suitability of a species for plantation forestry. Provenance trials should have a duration of at least half of the planned rotation period. The fast juvenile growth of most tropical plantation species and their short rotation periods allow the length of the observation period to be reduced considerably in comparison with that for field trials with temperate or even boreal species. For instance, in the context of selecting provenances of E. grandis for high pulpwood production, Ladrach (1998) mentions a rotation period of 6 years and a test period of 3 years.

11.5 Provenance Differentiation and Geographic Variation Pattern Two basic types of geographic variation have been described (Langlet 1963; Stern 1963; p. 89 ff. in Stern and Roche 1974). Since the differences in trait expressions mainly reflect the results of adaptation to their habitats (origins), the pattern of selective environmental factors has to be considered. Some of these factors, such as day length and average annual temperatures, vary continuously. Edaphic or biotic factors such as the occurrence of competing species may vary more or less abruptly. The totality of these effects finds its common expression in the variation pattern of the populations.

163

164


Clinal variation, a term coined by Huxley (1938), describes a continuous gradient in the frequency of genes or in average expressions of a trait in populations that closely follow an ecological gradient. For example, Langlet (1936) described a close relationship between the relative dry matter content of the needles of 1-year-old seedlings of Swedish Pinus sylvestris provenances and the latitude and elevation of their origin. The relative dry matter content increases with increasing latitude and elevation of the origin and thus with the gradually decreasing temperatures. Ecotypic variation, a term proposed by Turesson (1922), denotes discrete trait expressions of populations or races of a species in adaptive traits following a discontinuous variation pattern of the selecting environmental conditions. Different ecotypes may be found in proximity to each other. A relationship between the spatial proximity of populations and their genetic similarity does not necessarily exist. Clinal variation is commonly described by regressions of trait expressions of the different provenances on the expression of ecological factors of their origins. In some parts of the world ecological factors are closely correlated to the geographic coordinates. Average temperatures and rainfall may be correlated with latitude and altitude. Day lengths are determined by latitude; therefore, the geographic coordinates are commonly used as a replacement for the longterm averages of meteorological data, since no measurement is required. Both types of variation patterns are often simultaneously observed and may overlap in forest tree species (Example 11.1). The evaluation of the geographic variation pattern and its interpretation by ecological data requires much more intensive, preferably rangewide, sampling of populations. For practice-oriented provenance tests the sampling might concentrate on a region with environmental conditions similar to those for the prospective use of well-performing provenances. Garnier-Géré and Ades (2001) used several easily measured variables to statistically explain the variation in average diameter growth of more than 60 provenances of E. delegatensis on seven Australian sites. Two groups of variables were (1) bioclimatic data such as solar radiation, temperature, and precipitation and (2) data on the ecological site quality of the origin. The models explained high percentages of both the variation among provenances and the provenance by environment interactions. They represent an excellent means for locating adapted provenances for given sites. Interaction between provenances and environments may cause heterogeneity of the geographic variation patterns estimated in different locations. To expect regressions describing clines also in land races would mean overestimating the speed of response to selection. A geographic pattern fitting into one of the types is not always found. Differentiation patterns among provenances of E. camaldulensis in economically relevant traits are described in Example 11.1. The use of this species in plantation forestry has been greatly promoted by results of provenance trials.

11.6 Choice of Provenances in Tropical Forestry

Example 11.2 describes a series of internationally co-ordinated provenance trials in teak. Pronounced geographic variation has been found in some other eucalypt species (Eldridge et al. 1994). Notable differentiation of provenances was also found in Casuarina equisetifolia (Pinyopusarerk and Williams 2000). Superiority of land races of Casuarina equisetifolia was found in at least some traits related to growth or growth form. However, Chanshama et al. (1999) reported that land races of E. tereticornis ranked low in provenance tests. Kundu and Tigerstedt (1998) found notable physiological differentiation of ten provenances of neem (Azadirachta indica). Laboratory tests revealed three major groups related to annual rainfall at the origin. Offspring of exotic plantations were among the experimental entries with medium rainfall. The pattern of differentiation among four provenances of neem did not correspond with that of the allelic structures at six allozyme loci (Kundu 1999).

11.6 Choice of Provenances in Tropical Forestry If large-scale plantations of a species are planned in a region, recommendations are best based on the results of a series of experiments. The establishment of experimental series admittedly takes time, effort, and money. The long-term observation of field trials is nonetheless indispensable for making reliable decisions on provenance choice. The results of provenance tests can be implemented in different ways. Mass collection of seed for plantations in the field test forbids itself because of too few parent trees and their lack of isolation. However, Ladrach (1998) described the conversion of a provenance test into a seed production area (see also Sect. 13.5) as a rapid procedure of using the test results. Given that the experimental trees were planted at a spacing of 3 m × 3 m in six-tree row plots, 30 evaluation trees were left unless mortality exceeded 15% at the end of the observation period of 3 years. At this age the best trees are marked as seed trees. If half of the provenances are rejected, about 100 trees are expected to be left at a fairly regular spacing. Presumably the bulk of the seed produced there arises from matings between trees of different provenances. Provenance hybrids are likely to occur if the flowering periods of predominantly outcrossing species overlap. The proportion of hybrids thus depends on the flowering phenology and other parameters of the mating system. Phenotypic trait expressions of provenance hybrids are usually unknown. Such a seed production area was considered to be the beginning of a land race. Another way is collection of more seed in those stands where the material for the provenance tests had been collected. However, according to common experience the parent stands often no longer exist when the provenance test is

165

166


evaluated. This is particularly true with provenances of exotic species. Once a provenance has been tested in the field, the material under test may be integrated into a base population for breeding. The choice of the “right” provenance is of particular relevance in exotic species. In this case reproductive material is transferred into possibly different environmental conditions. However, the choice of provenances should always be considered carefully at the planning stage. Several criteria for the selection of provenances are used. They differ according to the status of knowledge on patterns of genetic variation and differentiation. If no information on geographic patterns is available from field tests, seed from autochthonous populations growing in a similar environment close to the planned growth site is recommended. These populations are likely to be adapted to the prevailing environmental conditions. Their recommendation is much more reliable than that of collecting seed just in any, possibly allochthonous, stands in the vicinity. The choice of reproductive material is often done on the basis of ecological similarity of the seed collection site and the prospective planting site alone. It may be more reliable to compare the prospective planting site with the environment at the origin of the respective population. The environment at the origin of a population is at least as relevant for its behavior at new sites than only the phenotype of the autochthonous collection stands. To many forest managers “provenance” has become the overriding concern in the establishment of plantations. Unfortunately, many foresters do not distinguish between the growing site and the origin of populations. They are in fact concerned only about the site of seed collection. However, knowing where seed was collected without making reference to the origin of the respective population does not make a better plantation. Unfortunately, the origin of populations of tropical trees is only rarely documented properly. Ladrach (1998) reported some general guidelines for provenance choice derived by Zobel et al. (1987). They might apply to seed collections in autochthonous populations. Collection at high elevations for plantations at low elevations is advised against. Also a shift of reproductive material from maritime origins to more inland regions should be avoided. Reproductive material from seed sources in regions with uniform climate should not be transferred to regions with strong fluctuations. Also, seed of populations on basic soil should not be chosen for planting on acid soils. Likewise, the transfer from sandy soils to clay soils is not recommended. When reporting on experience with plantations of various exotic pine species, Marsh (1970) came to the conclusion that the origin and the growing site should match as closely as possible. The study of environmental matching in terms of the temperature regime, the amount and the seasonal distribution of rain fall, etc. (Davidson 1995)


might be completed by considering also edaphic and biotic conditions (Zobel et al. 1987). Particularly in exotic species the suitability of a plantation for seed collection cannot be assessed on the basis of a few trees of unknown origin and unknown history (Jones and Burley 1973; Hattemer 1987). When discussing the adaptedness of populations of E. globulus in various parts of the world, Eldridge et al. (1994, p. 31 f.) raised the question, “whether one should be satisfied with a land race as a current source of seed and as the genetic base for future selection.” Mentioning casual introductions of seed to South Africa as early as before 1860, the authors doubt that the land race is better than newly introduced seed of the best provenances. Comparisons between the offspring of land races and their natural progenitor populations in designed field trials are needed for an objective judgment. In view of the situation in eucalypts, Eldridge et al. (1994, p. 2) also state that the collection of seed from natural populations has led to considerably more improvement than the uncritical use of land races. Seed collection in small plantations for the establishment of other small plantations opens a sequence of populations with decreasing genetic variation without the option of improving yield. The choice of a certain provenance is often determined by the availability of sufficient quantities of reproductive material in tropical plantation forestry; however, there is no reason to believe that incidentally available material is also the most suitable for the prospective growing sites. Clines existing in economically relevant traits greatly facilitate the selection of provenances. On the basis of the studies by Langlet (1936) on clinal variation, Eriksson et al. (1980) and Mátyás and Yeatman (1992) have even formulated transfer rules of provenances on the basis of the ecological difference between the origin and the planting site. Westfall (1992) developed them further to the construction of seed transfer zones. Two methods were discussed by Raymond and Lindgren (1990). In some tropical countries the establishment of field trials has become an integral part of more comprehensive programs for plantation forestry. Most plantation species in tropical countries have been exotics (Chap. 13). The majority of eucalypt and pine plantations as well as acacia plantations in the humid and semihumid tropics are outside the natural range of the species. T. grandis, Swietenia spp., Leucaena spp., and other agroforestry and multipurpose tree species are planted on all continents. Only recently the awareness of the multiple uses of indigenous species has grown (Leakey and Newton 1994a; Margraf and Milan 2004). Unfortunately, it is often difficult to obtain sufficient quantities of seed from autochthonous or even natural stands of a species. Seed centers responsible for the collection, storage, and deployment of tree seed do not exist in all countries. The situation is different for Australian tree species. Small quantities of Eucalyptus and Acacia seed for field testing provenances can be procured

167

168


through the Australian Tree Seed Centre (Arnold and Midgley 1995). The availability of reproductive material for both field-testing and large-scale planting has significantly contributed to the wide use of Australian tree species as exotics throughout the tropics. Evans and Turnbull (2004 p. 113 ff.) discussed some examples of successful species introductions. Programs for afforestation and plantation forestry have often been promoted by international organizations and/or bilateral development agencies that also supported the establishment of field trials (Barnes and Burley 1990; Palmberg and Esquinas-Alcazar 1990; Arnold and Midgley 1995; Ladrach 1998). The respective provenance trials and other long-term experiments were often initiated with considerable effort and input (Examples 11.1, 11.2). However, the long-term maintenance of the experiments and their adequate analysis sometimes became impossible, when the respective projects expired. However, meaningful outcomes take their time. Provenance research in tropical forest tree species is more complex than in trees of the temperate zone even though rotation and observation periods are shorter. Conifer plantations have not achieved the important role once forecasted by the FAO and other international agencies during the 1960s. Thus, the considerable amount of results available on provenance differentiation of pines is of limited use. However, for certain species of pines and some eucalypts as well as a few valuable hardwood species such as teak, results of long-term observation of provenance trials are available. Provenance research is closely associated with plantation forestry and concentrates on important plantation species and a few others with potential importance. Practice-oriented provenance research in a tree species should therefore be made only if it is of prospective use as a plantation species. New provenance trials have recently been established for agroforestry species and multipurpose tree species. There exists a strong argument for exploring provenance variation in regular tests before major plantation projects are taken up. If provenances of a species are differentiated in spite of inferred widespread gene flow, immigrants must be part of the genetic load. Provenances that are not adapted to the planting sites do not only imply the risk of failures. If they are able to reproduce, the respective populations increase the genetic load of the surrounding autochthonous populations through their emitted effective pollen. Therefore, the large-scale transfer of arbitrary, possibly maladapted, reproductive material adversely interferes with ecological genetics principles. Field trials with several temperate conifers yielded the information that the local autochthonous source is not always the fastest-growing or otherwise “best” provenance (Namkoong 1969). For example, local sources of Pinus taeda and Pinus palustris were frequently inferior to provenances transferred over large distances (Wells and Wakeley 1966, 1970). A frequent experiencee in reciprocal transplant experiments is that the local seed source is always among


the top-ranking but is not necessarily the best-adapted nor the most productive. There may exist obvious adaptation lags. However, limits of seed transfer may be derived from mortality data of transferred provenances (Mátyás and Yeatman 1992). Example 11.1: Provenances of Eucalyptus camaldulensis and Their Growth Performance The natural distribution of E. camaldulensis is confined to Australia but the species covers the largest area of all eucalypts on the continent (Fig. 11.2). The species is the most important plantation species for dry and semidry climates. At least 500,000 ha of forest of this species have alone been established in the Mediterranean region, mostly in Spain and Morocco. The species is also increasingly used in semiarid regions of the tropics with summer rain. E. camaldulensis grows naturally in proximity to rivers and lakes. The species is closely related to E. tereticornis, E. brassiana, and E. exserta. Natural hybrids between these species and some other eucalypts of the subgenus Symphyomyrtus occur. Morphological and physiological differences between northern and southern races of E. camaldulensis are obvious. The two races are even regarded as subspecies by some authors. Genetically controlled variation among provenances of E. camaldulensis has been described for the following traits: growth, wood properties, salt tolerance,

Petford

Lake Albacutya

Fig. 11.2. Natural distribution of Eucalyptus camaldulensis and approximate locations of two better-known provenances. (Adapted from Eldridge et al. 1994)

169

170


drought tolerance, frost tolerance, lignotubers and coppicing, and concentration of leaf oils (p. 60 ff. in Eldridge et al. 1994). Provenance trials established in areas with predominant rainfall in summer proved the superiority of northern provenances. The observed variation is predominantly clinal; however, this general pattern is blurred by ecotypic variation on a smaller geographic scale. For example, the provenance Petford performed well in almost all experimental locations with tropical summer rain. The difference in volume production between the best-growing and the poorest-growing among 16 provenances tested in a trial in Nigeria was 300% (Otegbeye 1985). Provenance trials in a Mediterranean climate with winter rain showed a superiority of southern provenances but also ecotypic variation was superimposed on the general clinal pattern. Some provenances proved to be fast-growing in most trials established in this type of climate. For example, provenance Lake Albacutya is slow-growing in locations with tropical summer rain but performs well in a Mediterranean climate. The difference in volume production among provenances was up to 800% at an age of 10 years in a trial established in Israel. The differences in the growth potential of provenances are striking. Provenances from northern Australia, an area characterized by tropical summer rains, perform well in a similar climate throughout the tropics. Provenances from the southern winter rain region perform better in a Mediterranean climate. Particularly fast growing provenances have been identified for both climatic regions. The geographic variation in other traits is frequently ecotypic. Owing to the focal interest in growth they have been studied with less intensity. The observation made in provenance trials that certain populations are particularly suitable for areas characterized by tropical summer rain and a long dry season greatly contributed to the current importance of the species for afforestation in dry and semidry climates of the tropics (From Pryor and Byrne 1969; Midgley et al. 1989; p. 60 ff. in Eldridge et al. 1994). Example 11.2: International Provenance Trials in Teak (Tectona grandis) Teak is an important tree species in monsoon forests of South Asia and Southeast Asia. Its natural distribution ranges from India to Myanmar, Thailand, and Laos. Whether teak forests in Java are indigenous is doubtful. The total area of natural forests dominated by teak is more than 30 × 106 ha. Teak is also a valuable species for plantation forestry. Most plantations exist in Asia, but teak has been successfully planted in all main regions of the tropics (Tewari 1992). The artificial regeneration of teak forests has a particularly long tradition on Java. The wood of teak has excellent properties. It is suitable for furniture making, interior decoration, and as a veneer and plywood. Thus, the value of teak


plantations depends not only on growth rates but also on the quality of the stems produced. Recognizing the importance of teak as a plantation species, the DANIDA Forest Seed Centre organized a series of international provenance trials (Kjær and Verapong 1995). Forty-eight provenance trials were established in 1973–1974. A total of 75 provenances and land races from India, Laos, Thailand, Indonesia, Africa, and Latin America were tested. The provenances were assigned to eight main “provenance regions.” In the single trials only subsets of provenances were represented. Unfortunately, material from Myanmar was not available for the trials. Only 21 out of the 48 once-planted trials were included in a first internationally co-ordinated evaluation at an age of 7–9 years. A second evaluation was made at the age of 17 years (Kjær and Verapong 1995). By that time the number of trials evaluated had decreased to only 8. The main planting region for teak is tropical Asia. However, only one trial in Asia, located at Pah Nok Kau in northeastern Thailand, was evaluated at the age of 17 years. This trial comprised 25 provenances from seven provenance regions. No significant differences among provenance regions were observed in survival and growth (measured as basal area) at the age of 9 and 17 years. Small but significant differences were observed in stem form and stem quality at age 17. The stem form of provenances from Thailand is slightly superior to that from most other provenances. Provenances from Indonesia showed inferior stem form. The results of African and Latin American provenance trials indicated preferable provenance regions. However, pertinent recommendations must be specified for the observed traits. Since fast-growing provenances may have poor stem form, it is impossible to generally recommend certain provenances on the basis of these international trials. No particular provenances or provenance regions seem to be superior to others in the main planting regions. The results of the series of internationally co-ordinated provenance trials on teak are of limited value for immediate application for the following reasons: ●

●

Provenances from the large, continuous natural distribution area in Myanmar were not available for the trials. Approximately half of the 30 × 106 ha of natural forests dominated by teak were found in Myanmar (Tewari 1992). Even though this area is declining owing to ongoing deforestation, Myanmar is frequently regarded as a diversity center for teak, and teak from Myanmar is often believed to be “superior” to that from other sources (Gyi 1993). Only eight trials out of the originally planted 48 experiments were evaluated at the age of 17 years, i.e., less than one third of the regular rotation period of teak plantations. At this stage the other trials either did not exist

171

172


●

●

anymore or could not be evaluated for some reason. The remaining trials that could be evaluated are distributed unevenly. Three trials are located on Puerto Rico, and only one represents the main growing region of teak in Southeast Asia. No trials were evaluated at the age of 17 years in southern Asia (India). In many traits no significant differences among provenances or provenance regions were found. Statistically significant patterns of variation seem to be ecotypic. Neighboring provenances often showed contrasting trait expressions. The recommendation of “superior” provenances or provenance regions is complicated by the contrasting variation patterns found in economically relevant traits. Provenances showing good survival and growth are often inferior in quality traits such as stem quality (From Kjær et al. 1995).

11.7 Recommended Literature Stern and Roche (1974) analyzed the genetic variation in tree populations and their genetic systems mainly in the context of their adaptation to the environment. Also the genetic implications of the influences of humans on tropical, temperate, and boreal forest ecosystems have found the interest of those authors. The book by Wright (1976) on forest genetics covers all aspects of this field as seen 30 years ago, with emphasis on breeding. Methods and results of provenance research are presented in three of the 19 chapters. The book by Eldridge et al. (1994) deals with provenance research, breeding, propagation, and resource conservation in eucalypt species. Morgenstern (1996) has produced the first book to concentrate on geographic variation of tree populations in phenotypic traits. The results for the amount of variation and variation patterns are implemented in silviculture and breeding in the major climatic regions of the world, although the emphasis is on temperate and boreal species. Williams et al. (2002) concentrated on the design, conduct, and analysis, including data processing, of experiments in provenance research and breeding. Much other literature on this subject has been published in languages other than English.

Domestication and Breeding of Tropical Forest Trees

12.1 Introduction Genetic variation is not only expressed as differentiation among populations, but is also observed within populations. The objective of breeding is to exploit genetic differences between individuals in populations in order to change the expressions of economically relevant traits for increasing plantation yield. The basic approach to conventional forest tree breeding is artificial selection of individuals and progeny families. These are subsequently propagated sexually or vegetatively by the breeder. Breeding of tropical forest tree species has been and will be confined to a comparatively small number of taxa that are suitable for plantations. Main species and genera are those for which large-scale provenance trials have been established (Chap. 11). Provenance testing as a lower-intensity activity yields the basic material that close-sense breeding projects can build upon. Since even provenances of good performance contain trees of low value for use, only further breeding is likely to achieve the “quantum leap” in the quality of plantations (Mesén et al. 1994). The most comprehensive breeding programs have so far been conducted on tropical eucalypts and pines. Most of the traits to be improved by breeding are influenced by both genetic and environmental factors. Many of these traits that were mentioned previously (Sect. 11.3.2) are metric, i.e., they can be measured on a continuous scale. This holds specifically for traits influencing quantity and quality of yield. The conventional concentration of research and training in tropical forest genetics on tree improvement has led to some successful breeding projects; however, it has also contributed to failures and diverted interest from essential characteristics of tropical forest trees. Costly and time-consuming breeding projects must be grounded on previous studies of key aspects of the genetic system of the respective species or must be accompanied by such studies. An approach aimed at a short-term increase of plantation yield harbors considerable risks if it is not accompanied by population genetics studies of the genetic

12

174

C HAPTER 12 Domestication and Breeding of Tropical Forest Trees

system of the target species. Breeding of forest tree species is still an important application of forest genetics research. The decision to embark on a breeding project can best be substantiated if adequately sized, well-adapted, and productive populations with known history are available as base populations. A breeding goal must then be defined. Starting selection right away in a population that happens to be locally available is not necessarily promising and hardly justifies the substantial future investment. Other preconditions are trained personnel, sufficient land, and continued financing. Species of agricultural crop plants have long been subjected to breeding. Inadvertent selection practiced by humans has been an integral part of their evolution. Today, more deliberate and refined methods are in use. The main principles of breeding were successfully used long before the molecular basis of inheritance was discovered and even before genetics became a science. It is likely that humans have influenced the genetic structures of forest tree populations for several thousand years. However, systematic breeding of forest trees is a much more recent development. European and North American conifers and poplars were the first forest trees subjected to intensive breeding in the twentieth century. Methods involving gene transfer and other biotechnology cannot be addressed in this introduction.

12.2 Domestication Publications on the domestication of forest trees mainly deal with the genetic consequences rather than the process itself (El-Kassaby 2000). Harvesting seeds from the wild, natural state, germinating them, producing planting stock, and planting stands is just the first step. Further and more crucial steps follow in breeding. In the introduction to his comprehensive account of domestication Darwin (1875) wrote, “From a remote period in all parts of the world, man has subjected many animals and plants to domestication or culture.” In the past, foresters have taken seeds from the wild and used them without selection, thereby practicing inadvertent sampling of tree populations. Libby (1973) considered as essential elements of domestication the selection of desired trees in variable populations and their mainly vegetative propagation. Forest trees have been subjected to a series of selection stages as the essential phases of the domestication process. These involve an increase in frequency of desirable physical attributes with economic relevance such as those listed in Sect. 11.4.2. Certain procedures of seed collection, seed extraction, and storage and, last but not least, the methods of planting stock production are linked to breeding.

12.2 Domestication

These procedures are necessary elements of bringing trees under the control of humans. Allard (1960, p. 28) defines domestication as the bringing of a wild species under the management of humans. Sooner or later the natural reproduction system is brought under control and converted into something highly artificial. Darwin (1875) did not mention changes in the reproductive system in our modern sense. He emphasized that the principle of selection is very important: “No doubt man selects varying individuals, sows their seeds, and again selects their varying offspring. But the initial variation on which man works, and without which he can do nothing, is caused by slight changes in the conditions of life, which often must have occurred under nature. Man, therefore, may be said to have been trying an experiment on a gigantic scale; and it is an experiment which nature during the long lapse of time has incessantly tried. Hence it follows that the principles of domestication are important for us.” and continues: “Domestic races of animals and cultivated races of plants often exhibit an abnormal character, as compared with natural species; for they have been modified not for their own benefit, but for that of man.” One may doubt that domestication of forest trees will ever progress to a stage where they totally rely on humans and would become extinct if left alone, although some authors consider the inability to survive in natural ecosystems as part of the definition of full domestication (Clement and Villachica 1994). These authors proposed also a nomenclature of domestication. Only semidomesticates, although they are significantly modified from the wild state, can survive if abandoned. The lowest degree of domestication is that of species that are only managed. Darwin (1875, Chap. 23) emphasized, “Length of time is all important; for as each character, in order to become strongly pronounced, has to be augmented by the selection of successive variations of the same kind, this can be affected only during a long series of generations.” This judgment is supported when considering the long way from species-elimination trials to provenance trials of the promising species, breeding experiments, and the development of methods for propagation and mass cultivation. The natural system of reproduction of populations is replaced step by step by artificial matings following certain designs (Sect. 12.5) and the propagation of the individuals selected. De Vecchi Pellati (1970) made the important point that in conjunction with artificial selection just this change is an essential element involved in domestication. She also stated that the new populations are normally propagated in isolation, while native populations receive easily new contributions from different original populations. Hence, three crucial elements in domestication and breeding to increase fitness for purpose or domestic fitness (p. 94 f. in Eriksson and Ekberg 2001) are sampling, selection, and the change of the reproduction system. The domestication process has a strong social involvement. For this reason, Evans (1980)

175

176


defined it as the improvement of the harvest index, the proportion of humanly useful material of plants.

12.3 Genetic Controlledness of Phenotypic Traits The success of artificial selection for a phenotypic trait critically depends on its genetic control. This is an absolute biological precondition for starting a breeding project and one may only then predict some response of the population to artificial selection. If trait expression proved to be different among provenances of a species in a provenance experiment, then we may safely conclude that this trait is under genetic control. If the test material has received equal treatment, we cannot explain the observed phenotypic differentiation other than by differentiation at the controlling gene loci. Planned experiments are the standard procedure for inferring genetic controlledness. The genetic controlledness of adaptive traits is also crucial in the context of genetic conservation (Chap. 14). The statement of genetic controlledness alone is reliable only if the experiment was based not only on an environmental design (Sect. 11.4) but also on a genetic design (Sect. 12.5). Since an experiment takes effort and time, we have a problem in testing the genetic controlledness of traits that are expressed only late in the ontogeny of long-lived forest trees. Hattemer and Ziehe (1997) have presented a review of pertinent procedures. 12.3.1 Individuals Related by Descent

Similarity between parents and their progeny has been extensively exploited in the theory and practice of breeding. The enormous breeding progress made by early humans was hardly based on comparison tests with a sophisticated design but rather on the experience that a plant or animal with desirable properties gave rise to progeny with the same or even better value for use. The phenotypic similarity between parents and progeny must be based on the transmittance of genes and implies genetic control; however, only little can be said about the mode of inheritance. Regression of offspring on a parent can be tested in seed trees and their progeny from natural regeneration in the forest. Any observed similarity in trait expression is then primarily due to the contributions of the female gametes to the progeny. The seed parents and their progeny are compared in the same local environment where reproduction took place. However, even if the peculiarities of seed chorology admit the attribution of young trees to only one seed parent, it is essential to check the descent of the progeny by genetic markers.

12.3 Genetic Controlledness of Phenotypic Traits

12.3.2 Pair Comparisons in the Field

Whether a phenotypic trait observed within a stand has a substantial genetic component may not be directly testable by parent–offspring regression. If genotypes and environmental conditions are associated, the variation of phenotypic trait expressions may be controlled uniquely by environmental conditions, and genetic patterns could also be correlated with phenotypic patterns. Hence, a precondition for reliable conclusions is stochastic independence between genotypes and environmental situations. This is safely fulfilled in artificial stands but it is not necessarily so in stands that originated by natural regeneration. Most artificial stands are also even-aged. This is relevant if age modifies trait expression. Gregorius (1989) devised a method to find out whether one or several phenotypic traits are under genetic control. For purposes of comparison, pairs of trees are chosen in the forest depending on the expression of a phenotypic trait (Fig. 12.1). When a tree with phenotype X is found, its nearest neighbor with phenotype Y is chosen. This sampling procedure is supposed to guarantee local stochastic independence between genotypes and environmental situations. Herewith, it is essential that the partners could have emerged from the same

Fig. 12.1. The method of pairwise sampling. In this example, stress symptoms are tested for genetic controlledness. (From Ziehe and Gregorius 1996)

tolerant or not infested sensitive or infested

comparison of genetic structures

177

178


seed pool. Hence, the contrasting phenotypes should be located within seed transport distances from a single potential seed parent. After genotyping many pairs of trees at a set of marker gene loci, a statistical test of homogeneity between the genetic structures of phenotypes X and Y is undertaken. If the null hypothesis is rejected, phenotypes X and Y are associated with different genotypes and the trait must be under genetic control; however, gene loci with significant differences are not necessarily involved in controlling this trait – they may merely be stochastically associated with controlling gene loci. Such stochastic associations can originate under a system of mixed self-fertilization and random mating, which results in a correlation of heterozygosity over the gene loci (Ziehe 2006). The difference in genetic structure between the phenotypic groups is quantified by genetic distances. Gene loci that are stochastically associated with controlling gene loci are more likely to show substantial distances if they carry a larger amount of variation. If carriers of contrasting phenotypes tend to show similar frequency deviations in different populations, we may call the genetic differences consistent. Then either stochastic associations with the controlling gene loci are of a similar type, or the marker gene locus surveyed is involved in the genetic control of the trait. In predominantly outcrossing species, an increasing number of populations with different genetic backgrounds can substantially strengthen this evidence. The extension of pair comparisons to traits with more than two expressions is straightforward. Several slight modifications of the method were discussed by Hattemer et al. (1993, Sect. 16.4). The genetic control of traits in trees at any age can efficiently be tested in the forest environment even after the stand has been thinned repeatedly. Since the method may also be applied to nonreproducing trees, it has essential practical advantage. The method has been applied to European beech (Fagus sylvatica L.) but did not support the hypothesis that certain stem forms of this species are under genetic control (Ziehe and Hattemer 2002). In various instances, as reviewed by Geburek (2000), the method helped to present evidence that air pollution induced selection in European forest trees. In stress-related traits also degrees of heterozygosity in the tolerant and sensitive groups were compared. This method is straightforward and simple to apply but does not yield any information on the rigidity of the genetic control. Ritland (1996b) has devised a different approach to inferring genetic control of phenotypic traits. The method is based on phenotypic similarity of pairs of (not necessarily neighboring) individuals measured in the field. The phenotypic similarities of tree pairs are regressed on the degrees of their pairwise relationship as estimated from the fraction of alleles shared by the two individuals at a set of marker gene loci. This yields a quantitative measure of genetic control (narrow-sense heritability; Sect. 12.4).

12.4 Linear Model of Genetic Effects on a Phenotypic Trait

12.3.3 Response to Natural Viability Selection

A trait that possesses special adaptive relevance and, therefore, responds to natural selection is early survival. If early mortality affects genetic structures at marker loci, its variation can also be attributed to the control by either these marker gene loci or stochastically associated gene loci. However, if the response of several populations to similar selection regimes is consistent, then the observed marker gene locus may be considered to be among the gene loci that control the trait under selection. Since the trait considered is viability, the controlling gene loci must then be considered to be adaptive (Ziehe et al. 1999).

12.4 Linear Model of Genetic Effects on a Phenotypic Trait 12.4.1 Phenotypic Trait Expressions

Both genotype and environment contribute to phenotypic trait expressions (Sect. 2.6), and a simple linear additive model postulates the following interrelation P = G + E,

(12.1)

where P denotes the phenotypic and G the genotypic value of an individual, and E is an environmental deviation. The supposed additivity of genotypic and environmental effects on the expression of a phenotypic trait is intuitively appealing and has proven its heuristic value in many empirical studies; however, there is no obvious biological justification for the presumed additivity. Multiplicativity or other, more complex interrelations may also be used to explain phenotypic trait expressions apart from simple additivity. The environment does not only modify the expression of phenotypic traits. It also determines the fitness of genotypes (Gregorius and Kleinschmit 2001). The phenotypic trait expressions of different genotypes may respond to different environmental conditions in dissimilar ways. This observation has been made in many experimental series with genetically defined entries in different environments. Consequently, the simple additive model is expanded by an interaction term IG ¥E in order to account for this fact: P = G + E + IG ¥E.

(12.2)

However, the separation of genotypic and environmental effects is not possible by introducing an interaction term. The supposed interactions between

179

180


genotypic and environmental effects are now no longer independent variables and cannot be measured separately. The study of reaction norms of genotypes helps to solve part of this dilemma (Gregorius and Namkoong 1986). It was also addressed by Hattemer et al. (1993, Sect. 16.2). The production of sexual offspring is required in most breeding strategies. Thus, the impact of a genotype on the phenotypic trait expression of the individual itself is of less importance for the trait expression of its progeny. However, genotypes are not directly inherited. Only gametes are transmitted from parents to offspring following segregation and recombination. Interdependencies of the expression of the alleles at a single gene locus are the reason for the occurrence of dominance deviation D. In a similar way, interactions between gene loci are responsible for phenotypic effects on certain traits that cannot simply be explained by additivity of the effects of several controlling loci. These nonadditive effects of gene loci are called epistatic. Epistatic effects are responsible for an interaction deviation I that just like D influence the genotypic value of an individual. Neither D nor I can be transmitted to the offspring. These considerations make it necessary to expand the linear model by splitting up the genotypic value into three components, i.e., the additive effect or breeding value A, dominance deviation D, and interaction term I, in order to account for epistatic effects of gene loci controlling the phenotype such as G = A + D + I.

(12.3)

In Eq. 12.3 the quantities D and I depend also on gene frequencies; therefore, they are not only measures of dominance and epistasis but are also properties of the population (pp. 116, 119 in Falconer and MacKay 1996). This conclusion is of crucial importance in quantitative genetics. The genetic component to phenotypic trait expressions is determined by genotypes at a number of gene loci. In diploid organisms each genotype is composed of two alleles. Figure 12.2 illustrates the phenotype of a metric trait as controlled by a single gene locus in the absence of environmental or epistatic effects. The average trait expression of the heterozygous genotype is not necessarily exactly half way between the average trait expressions of the two different homozygotes. d measures the deviation of the average trait expression of the heterozygote from the intermediate value of the average trait expressions of the two homozygotes. Intermediarity is characterized by d = 0. Positive or negative dominance deviations are defined as partial dominance (d < a) or overdominance (d > a). The dominance is complete if the heterozygote possesses the same genotypic value as A1A1 and d = a, for instance. From Fig. 12.2 we may derive some simple conclusions for the increasing complexity of artificial selection for greater trait expression with increasing d:

12.4 Linear Model of Genetic Effects on a Phenotypic Trait Phenotype a

d

µ −a

A1 A1

A 1 A2

A2 A2

Genotypes

Fig. 12.2. Partitioning of the genotypic value. m is considered the average phenotypic value of the trait produced by the homozygous genotypes in the absence of environmental effects. Phenotypic values can therefore be called genotypic values. a is the effect of the genotype possessing the advantageous allele A1 in homozygous condition. Consequently, – a measures the effect of the genotype possessing only A2. d is the dominance deviation. Note that a and d are defined relative to m. (Adapted from Becker 1993)

●

●

●

Dominance, partial (0 < d < a) or absent (d = 0). If the breeder selects individuals with the greatest trait expression and excludes all others from reproduction, the population will contain only allele A1 in the subsequent generation. Owing to complete selection against A2 the advantageous allele is fixed after one generation. The breeding goal has been reached. Complete dominance A1 > A2 or d = a. In this situation the selected part of the population contains genotypes A1A1 and A1A2. In order to find out which individuals with the greatest trait expression have the desirable genotype A1A1 and which are only A1A2, the breeder has to cross them with individuals possessing the smaller trait expression and to test their progeny. This involves a certain effort and also a serious setback, since all progenies produced contain the undesirable allele A2. Overdominance or d > a. Reproduction of the selected part of the population will consist of heterozygotes only and will contain A1 and A2 with equal frequency. The subsequent generations will contain also homozygotes with nonoptimal phenotypes. The balanced selection equilibrium prevents further breeding progress. A possible solution is to vegetatively propagate individuals with the greatest trait expression.

It is also easy to see in Fig. 12.2 that any environmental effects would blur the phenotypic differences between the three genotypes, so also individuals with genotype A2A2 may be selected, although they possess an undesirable phenotype. This can only reduce breeding progress.

181

182


12.4.2 Genetic Variance Components and Heritability

For the following considerations we admit environmental effects but do not account for the existence of genotype × environment interaction. If we modify Eq. 12.1 by partitioning the genotypic value G into its components (Eq. 12.3), we get P = A + D + I + E.

(12.4)

If we now consider random individuals in the population, the quantities in Eq. 12.1 become random variables. Assuming their mutual independence, we may derive an important property of their variances, VP = VG + VE,

(12.5)

and, according to Eq. 12.4, VP = VA + V D + V I + V E.

(12.6)

VA is the component of the genetic variation that can be captured also under sexual reproduction and its concomitant recombination. VD measures the variation due to the effect of alleles masking others that have no apparent effect. It is possible to estimate these variation components from appropriate experiments following an adequate statistical design involving related individuals with known pedigree. These may have both parents in common and are then full-sibs, or may have only one parent in common and are then half-sibs (Sect. 12.5). The phenotypic similarity among relatives is the basis for estimating the “rigidity of genetic control.” In fact, the proportion of the phenotypic variation in the population attributable to genetic differences among the individuals or VG/VP is the heritability in a broad sense or hw2 . As can be seen from Eq. 12.6, it is never smaller than the heritability in a narrow sense or h2 = VA/VP (p. 123 in Falconer and MacKay 1996) that measures the degree of genetic control by additive effects only. The term “degree of genetic control” must be understood in the sense of statistics rather than causation. The heritability in the narrow sense is of main interest for breeders, since it is involved in quantifying the expected change of average trait expression by mass selection. This heritability quantifies the relative amount of “useful” variation that may be exploited by breeding, because only the fraction of the genotypic values due to additive effects is inherited. Heritabilities range from zero to unity. A high heritability value h2 implies rapid improvement of trait expressions by breeding methods based on sexually produced offspring. h2 = 0 means that no genetic variation exists which can be used in breeding, or that breeding is not possible for the observed trait. The maximum value of unity would mean that

12.5 Estimation of Breeding Parameters; Progeny Testing

the observed phenotypic variation is completely determined by additive genetic variation. In consequence, breeding is expected to be highly efficient. The efficiency of breeding methods and the breeding progress or genetic gain to be expected from artificial selection for a trait increase with increasing values of phenotypic variation, narrow-sense heritability, and the intensity of selection. Although the analysis of variance requires some assumptions to be fulfilled, it is a widely used standard method for the interpretation of experimental results in the breeding of animals and plants in spite of these restrictions. It is also in frequent use in forest tree breeding and has proven its practical value in forest trees just as for other plants. The results of an experiment strongly depend on the environmental conditions at the site (Nienstaedt and Riemenschneider 1985), the experimental design, and the sample of genotypes studied (p. 332 in Hattemer et al. 1993). Estimators of trait heritability frequently exhibit also a temporal dynamics (Nienstaedt and Riemenschneider 1985; Jayasekera et al. 1994).

12.5 Estimation of Breeding Parameters; Progeny Testing Progeny tests have a dual function. They open the way to estimating genetic components of the variation in economically relevant phenotypic traits. Furthermore, it is possible to evaluate selected trees and to select them on the basis of the performance of their offspring, or to select families and/or individuals within those families (Sect. 12.5.2). However, many related individuals are produced by progeny testing. The degree of their relationship (usually fullsibs or half-sibs) must be known for interpretations in terms of genetic variance components. A variety of mating designs have been worked out and are being practiced. They differ in efforts to be taken, and in the type and amount of information on breeding parameters (van Buijtenen 1976; Otegbeye 1998). Otegbeye (1998) included descriptions of the (environmental) design of the respective field tests. 12.5.1 Progenies of Open-Pollinated Trees

The simplest form of a progeny test is the comparison of progenies of single seed trees after open pollination. This used to be the most frequently applied procedure. If seed trees are assumed to have mated randomly, the allelic structure of their individual effective pollen clouds represents that of the population; hence, they are not differentiated. The variation between the families is then due to the variation in the breeding values of the seed trees.

183

184


In this basic experimental design the progeny means can be used to estimate the breeding values of the seed trees, i.e., the means of their offspring under the conditions of random mating. Analysis of variance of the trait observations allows then estimation of the additive genetic variance and heritabilities. Finally, the breeding progress expected from selection of the seed trees or that expected from the selection of trees within the progenies can be estimated. However, these ideal conditions are rarely realized. Before sufficiently variable genetic markers were available, the mating system of trees could not possibly be analyzed in detail; therefore, it was considered random for the sake of convenience. However, the pollen clouds differ, because only moderate numbers of male mating partners become effective for particular seed trees (Sect. 6.2.1; Fig. 6.1), which induces a drift effect. If also a family structure exists in the population, the differentiation of effective pollen clouds is enforced by differentiation of the neighborhoods (Sect. 11.5). Under random mating all offspring of a family are ideally sired by different pollen parents; hence, they share only their female parent and may be considered half-sibs that are expected to hold one quarter of their genes in common. However, in many species the families might contain also offspring from selffertilization. Owing to restrictions in the effective number of male mating partners, certain portions of the family members are in fact full-sibs, since they share both of their parents. Hence, the progeny of a tree after open pollination is a mixture with different but unknown degrees of relationship. This condition is highly likely to be the rule for forest trees. Thus, the genetic relationship among the offspring of single trees is closer, i.e., they share more than one quarter of their genes (Squillace 1974). Consequently, the variation among the families is inflated. If it is used for estimating VA, heritabilities are overestimated. The estimates of VA and h2 are also not reliable. Selecting the trees on the basis of the performance of their open-pollinated progeny would be only a rough approximation, since the offspring of a seed tree have a more or less unique group of male parents. Dvorak and Wright (1994) took the complicated kinship relations of open-pollinated progenies into account by assuming the average degree of relatedness to be one third. Eldridge et al. (1994, p. 203) reported problems in interpreting the results of testing open-pollinated progeny.

12.5.2 Progeny Tests After Controlled Pollination

In order to define sibship more precisely, trees must be control-pollinated following a mating design. For instance, in an experiment denoted as polycross, seed trees are artificially pollinated by a pollen cocktail, i.e., a mixture of the pollen of many trees. Since a great deal of the production of full-sibs is pre-

12.5 Estimation of Breeding Parameters; Progeny Testing

vented, the degree of relationship is then more homogeneous and closer to what is expected from half-sib families. If pollen of the respective seed trees is not contained in the otherwise identical pollen mix, self-fertilization is excluded. It has turned out that the male mating partners contributing equal amounts to the pollen mix still do not have equal proportions of offspring. The progenies arising from this experiment are, nonetheless, expected to yield unbiased estimates of breeding values of the seed trees and the additive genetic variation. Many other mating schemes have been developed (Otegbeye 1998). The production of full-sib and half-sib families in the same experiment offers manifold options for analysis and use of genetic variation. In a factorial mating design (Fig. 12.3, top) a moderate number n of seed trees are pollinated by a larger number m of other trees used as males. The objective is to estimate the general combining abilities of those m trees, i.e., the average performance of their offspring when mated with the – three or four – testers. Also specific combining ability can be estimated. It reflects the ability of the trees to transmit genotypic effects when crossed with specific other trees. The group of n full-sib progenies of each tree are half-sibs among themselves, since they have just their male parent in common. Consequently, the variation among all m half-sib families yields an estimate of 0.25VA. The means of these groups reflecting their general combining abilities are estimates of their breeding values in this experimental population. Furthermore, the variation among the families having both parents in common yields an estimate of 0.5VA + 0.25VD, so also the dominance variance can be estimated separately and can be exploited by selection.

1 2

1 2 3 4 ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻

... ... ...

m ⫻ ⫻

⫻ ⫻ ⫻ ⫻

...

⫻

1 2 3 4 ⫻ ⫻ ⫻ ⫻ ⫻ ⫻

... ... ... ... ...

n ⫻ ⫻ ⫻ ⫻

. . . n

Fig. 12.3. Complete mating schemes for progeny testing. Top: Factorial design comprising m trees used as males and n trees used as testers. Bottom: Half-diallel comprising n trees without selfs and reciprocal crosses

1 2 3 4 . . . n −1

⫻

185

186


In the diallel cross all possible progenies are produced among a group of n cosexual trees. In Fig. 12.3 (bottom) a variant of this design is shown (a halfdiallel) that includes neither selfed progenies nor the progenies resulting from reciprocal crosses. If, for instance, trees 3 and 4 have to be crossed, either tree 3 is pollinated by tree 4 or tree 4 is pollinated by tree 3, or the cross is produced in either direction and the seeds are pooled. Just as in the factorial design both VA and VD as well as general combining abilities can be estimated. Pollination in most zoophilous tree species requires much effort. Various methods used in eucalypts have been described by Hodgson (1976b), van Wyk (1977), and Eldridge et al. (1994, p. 205 f.). Species with different floral structure require very different methods (Sedgley et al. 1992; Ouédraogo 1997). Since work in tree crowns is an additional obstacle, pollinations are best made in a clone bank or seed orchard. In view of the large number of progenies to be produced [n × m in the factorial design and n × (n – 1)/2 in the halfdiallel], there exist numerous incomplete designs requiring less effort (p. 234 ff. in Namkoong 1981). Progeny tests on the basis of complete designs are hardly applicable in just any tropical forest tree species. If not all crosses are successful at the same time, missing families require advanced computational methods for evaluation. The same is true if mortality in the field experiments leads to missing plot values or even family means. The single-pair mating design is recommended for avoiding relationships between families (Otegbeye 1998). The expected breeding progress is much faster if a large number of full-sib families are tested as compared with the test of single tree progenies (Wright 1976, p. 176). Among tropical tree species, this approach has been confined to the intensively bred genera Eucalyptus (Bouvet and Vigneron 1996) and Pinus (Nikles 1995).

12.6 Methods of Selection The decision for a method of selection is based on experimental evidence of its breeding efficiency and on whether it fits into the given breeding strategy. For instance, on the basis of the results of a progeny test, the genetic parameters are estimated and used for exploring the breeding progress to be expected from various selection regimes such as the selection among families or of single trees within families. If the variation within families is much larger than that between them, it is advised to retain more families and select with higher intensity within families in a combined procedure. The variant maximizing expected gain (Shelbourne 1992) is finally chosen as the most promising one. For estimating the breeding progress expected from various selection regimes, the heritabilities mentioned in Sect. 12.3 have to be adjusted to the particular

12.6 Methods of Selection

experimental design and the particular selection procedures. McKinley and van Buijtenen (1998) have concisely reviewed the methods of selection in tree populations. The closer the age at selection is to the rotation age, the more reliable is selection in view of the desired target expression of the trait. However, the performance at earlier ages, possibly at half of the rotation age, is often sufficiently closely correlated to the performance at the rotation age. Periodic measurements yield information on ongoing changes in the ranks. For breeding programs in several tree species of the northern temperate zone continued over several generations, the selection age has been optimized. 12.6.1 Selection of Plus Trees

The selection of plus trees is a method of mass selection, i.e., the selection of individuals in populations in the field for their subsequent propagation. The selection intensity may be high. Together with provenance testing, selection of trees in the wild (Ledig 1973) represents the first step in many breeding projects. It differs much from the selection in plantations. In natural forests the phenotype hardly indicates the genotype of old trees closely (Sect. 10.2.2). Heritabilities are supposed to be low owing to environmental heterogeneity (competition, etc.). Plantations consist of more or less equally spaced individuals possessing the same age and primarily belonging to the target tree species; therefore, the neighbors can be used as comparison trees. In natural tropical forests, at best a regional average may be used as a reference. However, plus-tree selection is defined as a method based on the observation of the phenotypes of the selected trees themselves rather than those of their progenies. The types of traits used for plus-tree selection are basically those already mentioned in Sect. 12.4.2. Trees are selected on the basis of their superiority in certain traits, for example, because of their good stem form and/or volume growth. Practical guidelines for the identification of plus trees have been summarized by Schmidt (1993a). The selection is often based on indices which combine several traits characterizing the growth and quality of trees (Zabala 1994; McKinley and van Buijtenen 1998). Studies in European temperate conifers proved the efficiency of positive plus-tree selection for increased height growth and even more so volume growth and stem form (Cornelius 1994; Simpson 1998). Other investigations on North American conifers detected only a small effect or failed to disclose any effect of plus-tree selection (Wright 1976, p. 165). Eldridge et al. (1994) attribute the spectacular improvement of eucalypts to the selection of outstanding

187

188


individuals in plantations in combination with their large-scale vegetative propagation (Sect. 12.8.3). Selection of individual trees is a very simple procedure also practiced in standard silvicultural management systems. The previously described selection of plus trees is “positive” since trees with superior phenotypes are selected. Selection can also be “negative,” i.e., trees with unwanted phenotypes are removed from a population first and the remaining trees form the reproducing population. Thus, selective thinning favoring superior phenotypes by removing inferior ones is a form of phenotypic selection (Wright 1976, p. 165) with low intensity. It is the reverse of dysgenic selection as discussed in Sect. 10.2.2. The efficiency of this type of selection is likely to be low, i.e., the yield of forest tree plantations is supposed to be only moderately increased and/or improved. This holds in particular if most of the observed phenotypic variation is environmentally caused by marked heterogeneity of environmental conditions. The observed variation of growth traits is also caused by a variation of the developmental stage, i.e., the different age of trees. An analysis of stand structure, its history, and environmental variation gives hints on the importance of genetic variation for an observed phenotypic variation (Davidson 1995, p. 9). 12.6.2 Selection in Progeny Tests

In view of the generation interval of trees, selection of individuals according to the performance of their offspring is a clumsy procedure. That the parents cannot be selected before their offspring have been measured is a serious drawback. It is helped by proceeding with the propagation of the selected individuals and later removing those with poor progeny. Much more efficient is the selection of families or of individuals within families, whereby the family mean is used as a reference. Most efficient is the combination of these two. Family selection is considered efficient also for traits with low heritability, since unlike plus-tree selection the average performance of entire families is evaluated. In view of the conservation of genetic information, a major disadvantage is the elimination of entire families or even all offspring of certain parents. If this operation is repeated during subsequent generations, it leads to an increase in inbreeding; therefore, it is operated with low intensity, so that only the poorest families are eliminated. Moderate numbers of families contained in a field test clearly restrict selection intensity. Selection of individual trees disregarding their family membership is the opposite of family selection. Inbreeding increases at a slower rate but smaller breeding progress may be expected. Therefore, individual trees are selected in conjunction with family selection. A general disadvantage of progeny testing and the selection in these tests is the existence of a large proportion of related individuals, as can be derived from


Fig. 12.3. Following evaluation, the field test may be converted into a seedling seed orchard (Sect. 12.8.2). 12.6.3 Selection of Clones

Clonal propagules derived from an individual are not progeny in the close sense. They can nevertheless be tested in field experiments. Selection of clones in field tests still leads to single individuals but their performance can be evaluated much more efficiently than that of plus trees, even though only a small number of ramets per clone are required. During clonal propagation the multilocus genotypes are conserved. Breeding progress is thus faster than in conventional selection based on sexual progeny, because in the selection of clones also all nonadditive genetic variation, i.e., dominance and epistatic effects (D and I; Sect. 12.4), is used. The advantages of clonal selection become more prominent with increasing significance of nonadditive genetic variation (Burdon 1989). However, only one or a few selection steps are possible in a given finite aggregate, since the number of individuals would get smaller at every step. The selection of clones is part of the most successful breeding programs with eucalypts that also involve species and provenance trials, hybridization among species, and progeny testing. 12.6.4 Marker-Assisted Selection

Certain types of highly variable molecular markers can be identified in large numbers (Sect. 2.3.1). These markers are useful for the construction of genetic maps and for localizing the chromosomal position of gene loci with a direct impact on economically important phenotypic traits. These latter gene loci are called quantitative trait loci, commonly abbreviated as QTLs. Particularly suitable for such studies are simple sequence repeats, restriction fragment length polymorphisms, random amplified polymorphic DNA s (RAPDs), amplified fragment length polymorphisms, and single nucleotide polymorphisms (p. 282 ff. in Becker 1993; Chap. 21 in Falconer and MacKay 1996). Prior to the identification of QTLs, a linkage map with a large number of variable gene loci is constructed. Genes located on different chromosomes are inherited independently. The genes located on one chromosome form a linkage group. They tend to be transmitted to the offspring jointly in the gametes. However, parts of homologous chromosomes are occasionally exchanged at meiosis owing to crossing over. The probability of crossing over between two genes located on the same chromosome, i.e., their recombination rate, increases with increasing distance between the genes. Thus, the probability of joint transmission of two genes is higher if the genes are located close to each other.

189

190


The most common strategy is to identify a large number of highly variable DNA markers in one or a few segregating full-sib families. The frequencies of certain multilocus genotypes allow linkage groups to be identified. The number of these linkage groups is expected to equal the gametic number of the chromosomes. Further studies would allow the linkage groups to be assigned to certain chromosomes; however, this step is not required for QTL identification. Within each linkage group the relative position of many marker gene loci can be estimated on the basis of their recombination frequencies. The result is a genetic map which illustrates the relative position of polymorphic gene loci to each other (see an example in Fig. 12.4). Ideally, many markers distributed regularly over the complete genome at approximately equal distances allow a “saturated” map to be set up. At least 100 polymorphic gene markers and preferably many more are required. Variation in quantitative traits must be observed on the same individuals within the GROUP 1

GROUP 2

GROUP 3

GROUP 4

GROUP 7 GROUP 8/12 GROUP 9/11 GROUP 10

GROUP 5

GROUP 6

GROUP 13/11 GROUP 14

CM 20 40 60 80 100 120 140 160 180 200 FRESH WEIGHT OF MICROPROPAGATED SHOOTS (FWS)

# STUMP SPROUT CUTTINGS (#CUTT)

% ROOTING OF CUTTINGS (%ROOT)

Fig. 12.4. Linkage map and localization of quantitative trait loci in a Eucalyptus urophylla clone. (Adapted from Grattapaglia et al. 1995)


family. Statistically significant associations between quantitative traits and the variation at certain gene markers can then be found. A frequently used method is to relate the genetic markers of individuals with contrasting phenotypic trait expressions such as particularly fast or slow growth. A QTL showing an association to particular trait expressions is then presumed to be located close to those markers. Finally, the approximate location of genes influencing economically relevant traits can be identified on the map (Fig. 12.4). It is highly unlikely that any of the gene markers is part of the nucleotide sequence of the QTL itself. Thus, the function and mode of gene action of this locus are not characterized. Only the location of the QTL can be inferred. In marker-assisted selection, individuals are then no longer evaluated only on the basis of the traits themselves but also on the basis of the variants at closely linked marker gene loci. Once identified, these loci can be observed more easily and at a much younger age. However, the applicability of markerassisted selection in forest tree breeding is restricted for some reasons: ●

●

●

The results for linkage among gene loci clearly apply to the respective species but the results for the association of certain alleles at the marker loci and QTLs apply to the particular full-sib family only. Since forest trees are selected within populations rather than within one family only, it must be possible to transfer the results to other families or even populations. Since in outcrossing tree species linkage disequilibria even within genes rapidly decrease with physical distance, even the transfer between families would be possible only for closely linked markers. The effect of single gene loci on quantitative trait expressions is often small and difficult to detect by QTL mapping techniques. For instance, the suitability of the markers must be tested by regressing the expressions of the phenotypic trait on the presence of no, one, or two marker alleles. Other test procedures involve the phenotypic difference between carriers and noncarriers of certain allelic variants. Also the phenotypic differentiation between offspring carrying a codominant allelic variant in a homozygous or heterozygous condition can be tested for significance. These methods were used by Gailing et al. (2005) in several families of central European oaks for a study of QTLs for flushing date, an adaptive trait in temperate climates. In two species of the family Fagaceae, Quercus robur L. and Castanea sativa Mill., QTL positions for this trait but not for the likewise adaptive isotope discrimination were conserved (Casasoli et al. 2006). Putative candidate genes for bud burst were at least located on the same chromosomes. As follows from the test procedure, the identification of QTLs is complicated if the observed quantitative variation is mainly due to nonadditive gene effects.

191

192


Linkage maps have been developed for many agricultural species and breeding techniques based on marker-assisted selection are in use. Saturated linkage maps have also been developed for a few temperate conifers. A genetic map on the basis of RAPDs has been developed for some eucalypt families within the framework of the intensive breeding program of Aracruz Florestal (Grattapaglia and Sederoff 1994). A linkage map was developed from the full-sib family of a hybrid cross Eucalyptus urophylla × E. grandis (Grattapaglia et al. 1995). Eleven linkage groups corresponding to the gametic chromosome number of eucalypts were identified. QTLs were mapped for three different quantitative traits: three QTLs located in different linkage groups have an influence on the rooting percentage of cuttings, four QTLs influence the fresh weight of micropropagated shoots, and two QTLs have an impact on the sprouting ability of stumps. Since full-sib families are difficult to produce in some forest trees, also open-pollinated progenies have been successfully used for detecting QTLs influencing the volume yield of E. grandis families (Grattapaglia et al. 1996). More recently, a linkage map was constructed for E. globulus, and candidate genes (Sect. 7.2) for wood, fiber, and floral traits were mapped (Thamarus et al. 2002). The identification of QTLs and efforts to achieve marker-assisted selection are also ongoing for tropical acacias (Butcher and Moran 2000; Butcher 2004). The individuals selected in this way are integrated into a population for further breeding operations or for propagation.

12.7 Heterosis Breeding in Forest Trees Heterosis or hybrid vigor is defined as the superior trait expression of heterozygous genotypes in comparison with either corresponding homozygote. It is the opposite of inbreeding depression. Positive effects of heterozygosity are widespread in forest trees (Namkoong 1985; Ziehe and Hattemer 1998). There are various ways to achieve increased heterozygosity, the precondition for heterosis. Within a population, genetic differentiation of individuals can be inferred from a large genetic distance between their multilocus genotypes at a large number of markers. If these markers are representative of the totality of the genome, the individuals may be expected to possess different alleles also at the gene loci controlling economically relevant traits; hence, after crossing those individuals, a large proportion of heterozygotes may be expected. Villaincourt et al. (1995) found in E. globulus that genetic distance estimated by RAPDs was of little value in predicting heterosis; however, these authors stated that genetic distance may be helpful in predicting inbreeding depression from mating closely related trees.

12.7 Heterosis Breeding in Forest Trees

A method of heterosis breeding that is widely practiced in annual crop plants is recurrent self-pollination until the degree of heterozygosity is substantially reduced. After three generations it amounts to at most 6.25% and after three more generations to less than 1% at any gene locus. On crossing those “selfed lines,” a certain proportion of highly heterozygous individuals show heterosis in traits related to yield and stress resistance. This procedure is hardly applicable in tree species owing to the length of their generation cycle, the increase of mortality, and the retarded onset of reproduction with increased homozygosity. Experiments of Durel and Kremer (1995) in Pinus pinaster, a species less sensitive to inbreeding depression, showed no heterosis in vigor after crossing trees that arose from first-generation selfing. Hybridization in the true sense is the formation of progeny by the fusion of two gametes produced by individuals belonging to different botanical species. Interspecific hybrids can be produced if there is a certain degree of mating propensity between the respective (mostly allopatric) species. Species hybrids of several forest tree genera proved to be particularly fast growing or show other beneficial trait expressions, for instance, improved resistance against pathogens (Khurana and Khosla 1998). Hybridization between species belonging to the genera Larix, Pinus, Populus, and Salix gained practical importance in forestry of the northern temperate region. Most experience regarding species hybrids of forest trees that are suitable for a tropical climate has been gained from eucalypts (Barnes and Simons 1994; Example 12.1) and North American pines (Francis et al. 1984; Nikles 1989). Meanwhile, hybridization among species also attracts attention as part of breeding programs of other tropical tree taxa. Hybridization has been proposed as a way to overcome Heteropsylla attacks on Leucaena leucocephala (Hughes 1989; p. 323 in Smith et al. 1992; Sect. 12.11). Hybridization is also widely used in breeding of Acacia species that are suitable for the humid and semihumid tropics. Good performance of the species hybrid Acacia mangium × Acacia auriculiformis has been observed (Sedgley et al. 1992). Hybridization in connection with vegetative propagation has gained practical importance for poplars and willows in the temperate zone and for eucalypts in the tropics. The vegetative propagation of eucalypt hybrids has strongly contributed to the success of breeding programs in Aracruz, Brazil (Example 12.1) and in the Republic of Congo. It has also proved to be efficient in several other programs (see van Wyk et al. 1989 for an example from South Africa). Some species hybrids show heterosis, others do not. The genetic causes of superior trait expressions in hybrids between certain forest tree species are still dubious. A combination of genetic information responsible for favorable trait expressions realized only in species hybrids is a likely cause (p. 122 in Namkoong et al. 1988). An increased individual heterozygosity of hybrids at some particular or many gene loci must also have a positive impact on

193

194


economically relevant traits owing to overdominance (p. 57 f. in Namkoong et al. 1988. In some instances progenies are regarded as hybrids, although the parent trees belong to the same biological species but to different subspecies, varieties, or provenances. Interprovenance hybrids of several conifer species (e.g., Araucaria cunninghamii, Pinus pinaster, Pinus sylvestris) were produced. However, interprovenance hybridization currently plays a minor role in forest tree breeding (Nikles 1993). Bolstad et al. (1991) reported on intermediate growth of interprovenance hybrids in Pinus banksiana. The expression of phenotypic traits in provenance “hybrids” could not only show whether there exists heterosis that could be used in breeding, but also whether a reduction in yield has to be expected in the subsequent generation. This question has gained importance, since plantations containing mixtures of populations (Stern 1963) are used for seed production. Example 12.1: Breeding Eucalypts in Aracruz, Brazil More than 130,000 ha of eucalypt plantations have been established since 1967 in Aracruz, Brazil, by the private company Aracruz Florestal, a subsidiary of Aracruz Cellulose. The wood produced is used for the production of pulp and paper in the paper mill of the parent company. During the early years of the program, an average annual volume yield of 12.4 m3 ha−1 was achieved under favorable climatic and edaphic situations. These first plantations were established from seedlings and were repeatedly harvested by coppicing. Since 1979, the original plantations have successively been replaced by clonal plantations. The clonal plantations grow extraordinarily fast. Their average annual volume production amounts to 60 m3 ha−1. Air-dried pulp yield increased from 3.3 (in 1969) to 15.9 t ha−1 year−1 in 1990. Furthermore, the wood produced is particularly well suited for pulp production and the clones are highly resistant against insects and diseases. Clonal plantations have been harvested since 1986. Apart from the increase in volume yield, the cost of harvesting and wood extraction as well as the amount of wood required per unit weight of pulp was reduced. The first seedlings for plantation establishment were Brazilian land races of the species E. grandis, E. saligna, and E. alba”. The origin of these land races is unknown. In 1973 a comprehensive breeding program was initiated using both sexual and vegetative propagation. During 1989 more than 1,500 provenances or families of 55 eucalypt species were under test in Aracruz. Provenances of E. grandis from Queensland and E. urophylla from Flores and Timor (Indonesia) proved to be particularly suitable for plantation establishment in Aracruz. More than 130 ha of seed production areas and over 50 ha of seed orchards were established for the production of sexual offspring.

12.7 Heterosis Breeding in Forest Trees

Full-sib families of E. grandis and the hybrid E. grandis ¥ E. urophylla were produced. Controlled pollination for artificial hybridization required the development of techniques for the long-term storage of pollen, because different eucalypt species flower at different times in Aracruz. Partial diallels were used as mating schemes. Breeding success considerably increased by the vegetative propagation of elite trees. Seventy-five clones originating from crosses between E. grandis, E. urophylla, and E. pellita were selected in 1989. These clones proved to be completely resistant against cancerous diseases and, when coppiced, showed high volume growth and desirable wood properties. The trees performing best in progeny tests are continuously selected according to a set of criteria and are mass-multiplied in clonal archives in order to steadily increase the number of clones used in the wood-producing plantations. The multiple-population concept (Sect. 12.9) has been integrated with 20 subpopulations of E. grandis and E. urophylla consisting of 20 trees each. The best-performing trees in ten fullsib families produced in each subpopulation are used for making hybrid crosses and selecting among those hybrids for clonal testing. Principles of the breeding strategy of Aracruz Florestal are presented in Fig. 12.5. Breeding is supported by research on in vitro techniques for the Introduction of species and provenances

V

Vegetative propagation

Best species and provenances Seed collection Best trees in best provenances

Progeny test

Unimproved plantations (with hybrids)

V

Selected trees Best trees in best families V

Controlled hybrids

Progeny test

Breeding orchard

Individual tree selection Seed orchard 1st generation

Progeny test

Seed orchard 1 1/2 generation

Individual tree selection

Selected trees Best trees in best families

Seed orchard V

2nd generation

V

Clonal test V

Clonal multiplication area

V

Clone bank

V

Clonal test V

Clone bank

V

V

Clonal multiplication area

Routine plantation

Fig. 12.5. Breeding strategy for eucalypts in Aracruz. (Adapted from Campinhos and Ikemori 1989)

195

196


multiplication of clones and by the use of molecular markers to identify QTLs (Sect. 12.6). The breeding strategy of Aracruz Florestal is integrated into a production system which also comprises components such as pest management, maintenance of balanced plant nutrition by minimizing nutrient extraction, and, if necessary, fertilization, careful logging operations, and preservation of biological diversity. More than 70,000 ha of natural forests (i.e., 35% of the total area) has been preserved between the eucalypt plantations for the maintenance of biodiversity (From Campinhos and Ikemori 1989; Campinhos 1991, 1993a, b, 1999; Foster and Bertolucci 1994).

12.8 Propagation of Breeding Products A breeding project includes both the selection of superior trees and their subsequent multiplication in order to capitalize on the breeding progress achieved; hence, a certain progress achieved in breeding makes sense only if it is possible to mass-propagate the selected material with as little genetic change as possible. Several options for sexual and vegetative propagation exist in tropical tree species that were hitherto subjected to domestication. Seed production areas are addressed in Sect. 13.5. 12.8.1 Clonal Seed Orchards

Propagating plus trees by harvesting their seed is hardly adequate for producing the quantity and quality of reproductive material required for the establishment of plantations. For the breeding progress involved in plus-tree selection to become manifest, the individuals have to be brought into mating contact in a clonal seed orchard. The fertilization of ovules by external pollen (pollen contamination) is avoided by maintaining sufficient distance to conspecific populations in the region and by planting large orchards. Remote pollen is then highly unlikely to prevail even if the density of the orchard is low. In young seed orchards one has to envisage incomplete pollination by the orchard clones, since the trees do not yet produce abundant pollen and are planted with wide spacing. Consequently they are eventually flooded by foreign pollen produced by unselected trees in surrounding stands. The reproduction in seed orchards is rarely random (Sect. 6.2) just as is the case with forest stands. However, an equilibrium resulting in Hardy–Weinberg structures (Sect. 6.2.2) and allele frequencies representing the frequencies

12.8 Propagation of Breeding Products

within the deme of the plus trees selected may be approximated by some provisions made in the design of the orchard. The clones are as a rule represented by the same number of copies (ramets). In order to overcome implicit setbacks due to high variation in reproductive output of the clones, additional ramets for the less prolific clones have been recommended; however, this condition possibly induces additional kinship among the seed produced depending on the variation in male mating success. Vegetative propagation by cuttings is difficult if not impossible because of the usually advanced physiological age of the plus trees selected (Hong 1975). The trees selected are usually first incorporated in a clonal garden, a multiplication garden, or a clonal archive. The scions for grafting are taken therefrom. Many techniques for the production of grafts are in use (van Wyk 1977; p. 399 ff. in Hartmann and Kester 1983). It must be checked repeatedly that the grafts in fact develop from the scion rather than the unselected root stock (Example 11.2). The number of clones represented in the orchard is important. In order to maximize breeding progress, the selection intensity is high and the number of retained clones moderate. However, if the orchard contains only a few clones, the risk of phenological mismatch among them increases. The risk of unintentional genetic change or even loss of alleles is obvious (Hattemer et al. 1982). The size of an orchard is best determined by the number of clones rather than the number of ramets per clone. Many old seed orchards contain only 20 or even fewer clones. Meanwhile, larger numbers of clones (up to 100) and, consequently, lower numbers of ramets per clone are common (cf. Example 11.2). Problems related to losses of genetic variants due to small numbers of orchard clones can be solved by reducing the intensity of plus-tree selection that per se is very high. Ideal would be the reduction of the number of ramets per clone to a few or just one grafted scion. Each clone is represented with a limited number of ramets (grafts). Increasing the number of ramets of a clone is genetically equivalent to the enlargement of the crown of an individual tree; thus, the chance for self-pollination increases. With reference to the conditions in a seed orchard, we have to consider pollinator movement not only within but also between the crowns of the ramets; hence, both individual and clonal self-fertilization are possible unless prevented by a system of incompatibility. Both types are identical with regard to the genetic consequences (Sect. 6.4). The necessary spatial isolation from other populations of the same species obviously depends on the system of pollen dispersal, and in particular on the efficiency of the pollen vectors (Sect. 5.2). Clonal seed orchards are designed in order to maximize the number and proportion of offspring from outcrossing with pollen from other selected clones. Keeping the average distance between ramets of the same clone large helps to lower the risk of clonal self-fertilization. Many orchards are designed in completely randomized blocks. Some subsequent

197

198


corrections are made in order to prevent close proximity of ramets of the same clone. Certain systematic designs are even more efficient in decreasing the proportion of clonal self-fertilization (Nester 1994). The density of seed orchards is much lower than in normal plantations in order to get larger tree crowns. The trees are often sheared in order to develop a short trunk and a wide crown to facilitate easy seed collection. Early and gregarious flowering is promoted. Clonal seed orchards were established on a large scale for conifers of the northern hemisphere (Pinus spp., Picea spp., and Pseudotsuga spp.) and for pines, eucalypts, and teak in the tropics (Eldridge 1975; Feilberg and Søegaard 1975; Example 13.2). The mating systems of clonal seed orchards have been studied in both temperate and boreal conifers. Estimates of the proportion of self-fertilization in conifer orchards are usually low. They are thus comparable to the situation in plantations and natural populations (Muona and Harju 1989). Deviations from random mating were proven in orchards of Pinus sylvestris (MüllerStarck 1982) and of teak (Finkeldey 2006). El-Kassaby and Askew (1998) described some genetic characteristics of the reproductive process in clonal seed orchads. An excess of self-fertilization and other deviations from random mating as well as pollen contamination have an impact on the genetic structures of the progeny produced; thus, the genetic gain attainable by a production of seeds in orchards is likely to be overestimated if it is based on the assumption of random mating among the clones and their isolation (Müller-Starck 1982, 1991). Estimates of the contamination rate with foreign pollen from outside the seed orchard are often surprisingly high (Paule et al. 1993; Pakkanen et al. 2000; Sect. 5.2.2). In South Korea, Pinus × rigitaeda, the hybrid between Pinus rigida and Pinus taeda, has been mass-produced by handcrosses made on grafts. The equally successful hybrid between Larix europaea and Larix kaempferi can be massproduced in specially designed seed orchards containing many grafts of a selfsterile clone and many clones of the other species. Seed is then collected only from the self-sterile clone. Example 12.2: Clonal Seed Orchards of Teak (Tectona grandis) Among tropical tree species the largest area of clonal seed orchards has been established for teak. More than 1,800 ha was planted from 1966 to 1978 in Thailand (Fig. 12.6). More than 1,000 ha exists in India, more than 500 ha in Indonesia, and smaller areas in Bangladesh, Myanmar, China, and several other countries. The establishment and maintenance of the clonal seed orchards of teak in Thailand is under the supervision of the Teak Improvement Center established


Fig. 12.6. A clonal seed orchard of teak (Tectona grandis) in Thailand. (Photo: R. Finkeldey)

in Lampang in northern Thailand. In total more than 300 clones were selected on the basis of the phenotypic superiority of plus trees. A clonal multiplication garden for the multiplication of selected trees was established at the Teak Improvement Center. Single buds were grafted on stumps (Fig. 12.7). This grafting method (“budding”) is quick and yields a high proportion of successful grafts (Keiding and Boonkird 1960). Planting distances varied from 3 m × 3 m to 12 m × 12 m. The oldest orchard contains only 16 clones. Younger seed orchards contain up to 100 clones; some of the clones are planned to be selectively removed depending on the results of clonal and progeny tests.

199

200


a b Fig. 12.7. Grafted teak (T. grandis) plants produced by the budding technique. (Photos: R. Finkeldey)

Many seed orchards were established far from other teak forests in order to ensure their reproductive isolation; however, several of the sites chosen were unsuitable for growing teak. Thus, not all of the orchards established survived and are productive. Seed harvest is below expectation in the 18–30-year-old orchards. In all orchards physiological incompatibilities between scion and root stock are frequently observed on older grafts (Fig. 12.8). The low amounts of fruits currently harvested in clonal seed orchards provide only a minor proportion of the total reproductive material for teak plantations in Thailand. Genetic marker studies based on isoenzyme gene loci were used to check the clonal identity of the ramets in the clonal multiplication garden of the Teak Improvement Center and in two clonal seed orchards. Clonal identity has not been maintained for many trees of the multiplication garden and, consequently, neither has it been maintained in the seed orchards. Possible explanations are a mix-up of ramets during the establishment of the clonal garden and the seed orchards as well as the occurrence of adventitious sprouts from the stump which eventually develop into the main stem of the supposed graft. High levels


Fig. 12.8. Physiological incompatibility between root stock and scion in a grafted teak (T. grandis) tree. (Photo: R. Finkeldey)

of self-fertilization (Example 6.1) and restricted pollen dispersal (Example 5.1) were observed in the seed orchards (From Hedegart et al. 1975; Finkeldey 2006). 12.8.2 Seedling Seed Orchards

Also seedling seed orchards established with sexually produced trees serve the production of seeds. In most cases seedling seed orchards arise from the conversion of a progeny test aimed at the identification of superior families.

201

202


Those families showing inferior trait expressions are rogued by selective thinning that stimulates also crown enlargement. Also inferior trees within well-performing families are removed from the field test. After this combined selection step, seeds are harvested that are used for the establishment of plantations. More details related to the establishment of seedling seed orchards have been summarized by Schmidt (1993b). Seedling seed orchards were successfully established for many tropical tree species. Examples are Cordia alliodora in Costa Rica (Boshier and Mesén 1989), several eucalypt species (Eldridge 1975), and Acacia mearnsii (Jones and Burley 1975). Inbreeding depression is likely to result from a large proportion of mating within families which may counteract the positive effects of selection on phenotypic traits. Selection among families must not be very intensive in order to leave a sufficient number of families to avoid strong inbreeding. If later conversion of a progeny test is planned, certain precautions in the layout contribute to the success of producing superior-quality seed. The test should be planted so as to maintain some degree of isolation. This is not very easy to accomplish, since the progeny tests must be planted on typical sites of the respective species where conspecific stands are possibly common. Both plot size and spacing should allow sufficient numbers of trees to be kept during their reproductive phase. As Fig. 12.9 shows, the layout in conventional plots may lead to a highly uneven distribution of the remaining trees unless the selection intensity is fairly low. Randomization of all trees within the blocks (single-tree plots) helps to cure this problem and also helps to avoid an immediate neighborhood of full-sibs. Since it is then very difficult to keep track of the trees belonging to certain families, the production of a map and permanent tagging of the trees during the experimental phase is recommended. Ladrach (1998) warns that errors in installation and labeling of trees may ruin an experiment. In tree species pollinated by either wind or animals the age, size, and density of an orchard determine the pollen production and hence the risk of the orchard being flooded by the pollen of the surrounding forest. Of course, distance isolation is important. In contrast to clonal orchards, the system of effective pollen dispersal can be analyzed in this type of orchard where every tree has another genotype, so paternities of particular trees and not only clones can be assigned to offspring (Sect. 5.2.3). A 15-year-old field test with 12 open-pollinated progenies of E. grandis in Madagascar had been planted with single-tree plots in 40 replicates (Chaix et al. 2003). Following evaluation, it was converted into a seed orchard comprising 349 trees. The orchard trees and seedlings raised from their seed were genotyped at six microsatellite loci. Thirty-nine percent of the effective pollen was not produced inside the orchard or by trees in the two border rows. Effective pollen received by 30 seed trees was transferred by introduced honey


Fig. 12.9. Conversion of a progeny test with 16 families (left) into a seedling seed orchard (right). One block containing four trees per family is shown. The eight families with the numbers 4–11 are selected. In a two trees and in b only one tree per four-tree plot is retained. In c and d the plots are not contagious and the trees are randomized within the block. In c two trees per family are retained and in d only one tree per family is retained. Note the uneven distribution of the remaining trees in a and b

bees over distances ranging between 32 and 85 m with no mating preference for neighboring trees. The outcrossing rate of 96.7% was much higher than reported from natural forests (Table 6.2; Example 6.2). Seed trees of this species growing close to the edge of the orchard showed the strongest deviations of the realized matings from those expected. The reproductive success of the families deviated moderately from equilibrium expectations. Pertinent ratios ranged from 0.32 to 2.02, i.e., the families contributed between one third of and double the expected number of genes to the seed produced.

203

204


When comparing the mating system in an isolated seedling seed orchard with that in primary or logged forest of Dryobalanops aromatica (Sect. 10.2.1), Lee (2000) found much reduced outcrossing rates (ts = 0.436) and a moderately reduced number of male mates for single seed trees (rp = 0.428 vs. 0.389 in primary forest). The author attributed this finding to a lack of pollinators and the transfer of much pollen by gravity alone, since the orchard was established with only one species. This may have created problems with the subsistence of pollinators in this irregularly flowering tree species. The difference between the multilocus (tm) and the single-locus (ts) outcrossing rate is an estimator of the level of inbreeding due to mating other than selfing, i.e., mating among relatives or consanguineous mating. A large difference tm – ts = 0.115 indicated the use of related planting stock when the orchard was planted 70 years earlier. This result was all the more surprising, since the density of the orchard with 20 trees per hectare exceeded that in the primary forest with 15 tress per hectare. In all, this seed orchard showed increased inbreeding by both self-fertilization and consanguineous biparental mating. 12.8.3 Mass Multiplication of Clones

During clonal propagation the entire genotypes of the trees selected are captured for the establishment of plantations, since no recombination is involved. This provides for one-to-one exploitation of an achieved level of breeding and is a particular advantage for rare outstanding trees. Zobel and Talbert (1984, Chap. 10) and Zobel (1992) strongly advocated operational vegetative propagation in tropical forest trees. Park et al. (1998) reviewed the advantages and explained the integration of clonal selection and vegetative propagation into breeding programs. Many forest tree species are easy to propagate vegetatively (Sect. 4.3.1). Thus, single or a few superior trees may be selected in the forest or in a field test and identically mass-multiplied. Asexual propagation is the basis of clonal forestry (Chaperon 1984). Among practical advantages, clonal propagation is useful, since the reproductive phase need not be awaited; however, the maturation phase must be arrested by hedging. Clonal forestry decreases the heterogeneity of chemical and physical wood properties. A homogeneous product can be beneficial in itself for certain end uses such as pulp and paper making. Greater uniformity of plantations might be of particular value, where wood is used in industry. Also, specific management procedures may be developed for individual clones (Burdon 1989). Particularly clonal forestry critically depends on specific production and management systems. The establishment of largescale clonal plantations requires intensive forest management techniques, including forest protection and the maintenance of soil fertility, for example, by fertilization (Example 12.1).

12.9 Multiple Population Breeding

The advantages listed previously once gave rise to clonal forestry. In contrast to sexual propagation, clonal propagation reduces the genotypic multiplicity of plantations drastically unless the heterozygosity of both parents is extremely low. It is highly unlikely that two trees even of the same full-sib family share an identical multilocus genotype. The recombination during sexual reproduction results in a very large number of multilocus genotypes. In clonal forestry the number of multilocus genotypes equals the number of clones represented in a plantation. Although the spectacular yields of selected clones in plantations have created much enthusiasm about clonal forestry in tropical countries, the loss of genotypic variability within plantations is associated with certain risks (Sect. 12.11). Also in vitro propagation has received much interest (Khuspe et al. 1994). These authors report on adapting techniques to a variety of tropical trees and to the unpredictably flowering bamboos. Summarizing experience with eucalypts in clonal forestry, Eldridge et al. (1994, p. 2) stated that the mass vegetative propagation of plus trees selected in plantations has been most successful. Clonal forestry based on an intensive breeding program may be comparable to the breeding of high-yielding varieties of agricultural crop plant species during the 1960s and 1970s. The success of the “green revolution” has not simply been due to isolated breeding programs but resulted from their integration into intensified agricultural production systems in tropical countries. In species hybrids vegetative propagation has a particular advantage, since recombination during sexual reproduction would reduce an effect of heterosis. Some species hybrids proved to be particularly suitable for vegetative mass multiplication and the establishment of clonal plantations. Probably the most successful breeding program for tropical trees was initiated in Aracruz, Brazil (Example 12.1). A similarly comprehensive breeding and multiplication program which provides clones for large-scale plantations of eucalypt hybrids was initiated in the Republic of Congo with French support during the 1970s (Delwaulle 1989). Clonal eucalypt plantations owned by private enterprises exist also in India and other countries. The production and selection of clones in these plantations is reported to be based on less intensive breeding programs than those in Aracruz and the Republic of Congo (Lal 1993).

12.9 Multiple Population Breeding The hitherto mentioned breeding methods are hierarchical since certain individuals or genotypes are selected from a base population, and these individuals or their offspring are assembled to a breeding population of again smaller size, etc. This approach has been adopted with certain modifications from traditional

205

206


breeding methods for agricultural crop plants. The expected breeding progress is accompanied by a gradual loss of genetic variation in the breeding population which does not only affect the gene loci selected but owing to possible drift effects the totality of the genome. The multiple population breeding concept of Namkoong is nonhierarchical and aims at the preservation and utilization of genetically differentiated subpopulations (Namkoong et al. 1980, 1988, p. 70 ff.; Namkoong 1984). In 1994, Namkoong was awarded the Marcus Wallenberg Prize “for his pathbreaking contributions to quantitative population genetics, tree breeding, and management of genetic resources, which form a solid scientific basis for the maintenance of biological diversity in forests all over the world.” The objective is no longer breeding of a single population accompanied by a reduction of genetic variation or the selection of only a few genotypes that are expected to produce superior phenotypes in different environments. Instead, the breeding population consists of a variable number of separate subpopulations. Multiple population breeding is directed towards different objectives, i.e., different traits are selected at different sites. Even if the selection criteria are the same, sets of controlling genes are expected to be differentiated. Both the adaptedness and the adaptability to different and changing environmental conditions are promoted by the maintenance or enhancement of a differentiated population structure. The concept sacrifices the maximization of breeding progress in a few economically important traits. An important objective is the enhancement of the adaptedness of different subpopulations to diverse environmental conditions. Adaptability is maintained by the resulting differentiated population structure. The concept is highly flexible but requires skill for managing the genetically effective population size, including pedigree control. It is possible to adjust the breeding program to changed economic priorities since different breeding objectives are defined in different subpopulations. Both the number of replicate populations and the number of parents in populations, their diversity and their differentiation can be used to control the total genetic variation available for managers to use in deploying trees in plantations (Namkoong 1999). Now as before, the total size of the subdivided breeding population is crucial under the aspect of inbreeding avoidance. However, the genetic variation can be managed more effectively. Multiple population breeding is particularly suited for potentially important plantation species of the tropics which still are in an early state of domestication. A main advantage of the concept is its flexibility both with regard to the necessary expenditure and the possibility for modifications during its implementation (Namkoong et al. 1980). For instance, also heterosis breeding can easily be integrated. Furthermore, breeding and conservation objectives are regarded as equally important and may be profitably combined in a single program (Namkoong 1984).

12.10 Breeding Strategy

Hierarchical breeding strategies lead to a reduction of population sizes and thus to increased genetic drift in shaping genetic structures. Genetic drift affects not only selected loci, but principally all gene loci. The main effects of genetic drift are an increased risk for the loss of genetic variation, random fluctuations of genetic structures at single loci, and modified associations of alleles at many loci. The breeding strategy of several tree species in Zimbabwe has followed the multiple population breeding concept since the early 1980s (Barnes and Mullin 1989). Although the concept has hitherto gained only limited practical importance in spite of its many advantages, it is hopefully going to be implemented in more projects in future. Simons et al. (1994) have developed concepts of a strategy shaped to nonindustrial tropical trees.

12.10 Breeding Strategy The breeding progress made in a single generation is hardly ever substantial enough to justify the expenditure for a single selection cycle. Depending on what has been achieved, selection continues to a second, third, etc. generation. For instance, Reddy and Rockwood (1989) reported on breeding progress made in tree volume during four generations of selection in a land race of E. grandis. Volume per tree increased from 7.04 dm3 in the first generation to 21.5, 25.9, and 40.2 dm3 in the subsequent generations. The ultimate goal is the long-term breeding gain. In the framework of a breeding strategy, it is planned how to manage those various populations (White 1987), the methods of selection and propagation, and the introduction of unrelated material. The first elements of planning are the evaluation of material available as a base population and its size. Quite a few breeding projects were taken up that were too small from the very beginning. Others were taken up without properly sampling the distribution range of the species in order to set up a suitable base population (Barnes and Simons 1994). In the first generation, mass selection prevails. Later operations include a combination of family and individual selection. The available information on genetic parameters increases and may be integrated. Libby (1973) has set up and discussed a fairly general strategy for tree breeding. Important items are comparisons of expected and realized breeding progress, the monitoring of the genetic variation at marker gene loci, and not too intensive or even aggressive selection (Sect. 12.11). Intensive selection reduces the population size and favors the increase of the inbreeding coefficient. A solution can be found by selection within several breeding groups. Subsequently more diverse outcrossed progeny for deployment are produced in combined orchards.

207

208


The main objective of forest tree breeding is the modification of average trait expressions of economically relevant traits in plantations. However, breeding and propagation of breeding products is usually done in small populations. Thus, it is meaningful to distinguish the following populations as part of a breeding strategy (Libby 1973; van Buijtenen 1975; Kang et al. 1998): ●

●

●

●

Both natural populations and plantations may be used as base populations from which individual trees are selected both at the beginning of the breeding program and during later stages. It must neither be modified nor be manipulated by artificial selection or even replaced. Its genetic structure can still not be maintained in its original condition but is subject to evolutionary change. Its production function has low priority. Selected trees or their progenies are merged into breeding populations such as progeny tests or clonal tests. They are mainly used for evaluating and selecting individuals and must allow for future responses to selection. They represent the core of breeding projects, they must be carefully designed and monitored, and they must be able to persist. Propagation populations contain subsets of the trees of the breeding populations and serve the multiplication of selected genetic information (not necessarily genotypes). Examples are seed production areas and seed orchards. The breeding products harvested from propagation populations are used to establish production populations (plantations).

The populations serve different functions. For instance, a progeny test (breeding population) may later be converted to a seedling seed orchard (propagation population). Under a multiple population breeding strategy, the base, breeding, and propagation populations are primarily separated, although there are relationships between the populations (Kang et al. 1998). The breeding program has the improvement of yield (in the widest sense) and the conservation of genetic variation as two goals possessing equal priority. It is integrated into the domestication process of the respective tree species. The decrease of additive genetic variation (Young and Boyle 2000) in breeding populations as closed systems is inevitable. One may think of a somewhat simple relationship requiring little theory of genetic variances and selection plateaus. In a given population a selection trait is assumed to be controlled by several biallelic gene loci without dominance or epistatic effects. If the favorable alleles possess low frequency in the base population, the genetic diversity is going to increase under selection. Maximum diversity is reached when the alleles possess equal frequency. The more the breeding goal is further approached, the smaller becomes the diversity. The situation is, of course,

12.11 Genetic Consequences of Domestication and Breeding

fundamentally different if new individuals, families, or even other populations with other allelic variants are integrated into the breeding population. Breeding of forest trees is time-consuming even in tropical species, since only a long-term approach to breeding is successful. The progress expected from breeding can only be higher than that from operations such as plus-tree selection alone. A breeding strategy combines several methods of breeding and propagation in order to both satisfy the immediate need for reproductive material and reach long-term breeding objectives. The breeding strategy of Aracruz Florestal for the breeding of eucalypts was discussed in Example 12.1. The fundamental importance of improving our knowledge of the genetic system of tropical forest trees and the equal priority of breeding and conservation objectives are not sufficiently appreciated by conventional breeding hierarchical strategies. Management plans for the genetic resources of a species will put more emphasis on population genetics research to understand and possibly manipulate the genetic system of a species. The final objective will be to integrate breeding and conservation objectives into a coherent and flexible strategy. The multiple population breeding concept of Namkoong (Sect. 12.9) deserves attention in this context.

12.11 Genetic Consequences of Domestication and Breeding 12.11.1 General Considerations

The global process of domestication in conjunction with breeding serves the welfare of mankind. It neither shall be nor can be hindered. On the other hand, at some stages of this process there exists the chance to retard the tendency towards impoverishment of genetic variation. Domestication may also help to rescue genetic variants. Artificial selection changes genetic structures of populations in manifold ways. Successful selection for a trait changes the allelic structures at the controlling gene loci and has some correlated effects on other traits. Artificial selection is hardly identical to natural selection both with regard to the traits concerned and with regard to the preferred trait expressions. For this reason the adaptedness of a population or its adaptability to environmental change (Sect. 7.2) can be adversely affected by breeding. Guries (1990) put is this way, “Human desires may be counter to those that have guided evolution over a long period of time, and compromises may have to be struck between the intensity of breeding activities and biological factors that work to ensure

209

210


the long-term survival of the species.” However, the actual danger of reduced adaptedness through breeding has not been proven by experimental studies in tropical forest trees. Some results have been reported for temperate and boreal species. For example, selection for fast growth during early stand development of Norway spruce (Picea abies) negatively affects the resistance against wind-break at later age (p. 399 in Hattemer et al. 1993), an important element of stability. In this case, breeding would lead to instability, i.e., the departure of realized from predicted performance (Namkoong 1999). Among all organisms subjected to breeding work, the environment of forest trees can be controlled least. The less intensive the finite breeding populations are selected, the less is population size reduced. Part of the breeding progress should be relinquished, for instance, by retaining a few individuals also of culled families in the breeding populations and the propagation populations such as seedling seed orchards. The breeder must find a best compromise between selection objectives and selection intensity. Another form of renunciation is restricting the time during which seed is collected from a given seed orchard. In view of the money spent for establishment, there is a tendency to use the offspring for planting on vast areas. These considerations may look uneconomic and paradoxical at first sight but they represent sustainability imperatives. Immediate maximum economic benefit from tree breeding implies setbacks during future generations of both trees and humans. The technologies of breeding and selection offer opportunities for managing the genetics and guiding the evolution of forests that neither decrease nor increase risks over that of uncontrolled, natural regeneration. Technology involved in domestication and breeding is neither the problem nor the solution. It is a tool that humans have available. It can be used well or poorly (Namkoong 1999).

12.11.2 Monitoring of Breeding Projects by Use of Genetic Markers

There is ample opportunity for the use of gene markers in breeding programs (Yeh 1989; Forrest 1994; Butcher et al. 1999). A highly rewarding application is the verification of the validity of crosses, as an example may show. A tree breeder had established a clonal seed orchard. When it started to flower, crosses among the clones were made for progeny testing. Years later it was detected by the application of genetic markers that the progeny test was a complete mess, because a muddle had happened already during production of the grafts. Part of the financial loss may have been avoided by checking the clonal identity of the ramets and by analyzing seed samples in the laboratory before sowing. Since possibly


high investment is made into pedigree breeding continued over generations, keeping track of ancestries and coancestries is highly recommended. These markers are widely unspecific and help to infer and monitor variation in the genome, i.e., the genetic background of, for instance, breeding populations. It is not necessarily likely that the moderate number of markers that we are able to identify will allow inference of the adaptive variation or even the possession of adaptive variants at certain gene loci. Goto et al. (2001) developed an efficient method to check the identity of ramets in clonal orchards on the basis of RAPDs. According to Park et al. (1998), keeping track of clonal identities also in vegetative mass propagation is of minor importance but might depend on the management of deployment. Applications of isoenzymes were discussed by El-Kassaby and Ritland (1998), Bergmann and Hattemer (1998), and Mitton (1998). Acheré et al. (2004) and Eliott et al. (2005) gave examples illustrating that meanwhile all of these objectives can be reached by the use of more variable or uniparentally inherited markers.

12.12 Recommended Literature The textbook by Wright (1976) covers forest genetics and tree breeding, including provenance research. An introduction to quantitative genetics that has found wide circulation is the book by Falconer and MacKay (1996). It deals with population genetics and all aspects of quantitative genetics, including simultaneous selection for multiple traits and the use of QTLs. Namkoong et al. (1980, 1988) presented comprehensive accounts of breeding forest trees on the basis of population and quantitative genetics. The book by Zobel and Talbert (1984) covers all aspects of tree breeding, such as methods and results. The book by Zobel and van Buijtenen (1989) is on wood and deals with the genetic aspects, such as variation and breeding. The book edited by Fins et al. (1992) covers all aspects of forest tree breeding and presents numerous examples from temperate species. Eldridge et al. (1994) cover breeding, with emphasis on eucalypts. In the book edited by Mandal and Gibson (1998), topics with significant current interest in forest genetics and tree breeding are discussed. As stated at the end of Chap. 11, the book by Williams et al. (2002) deals with experimental design and analysis also in close-sense tree breeding.

211

Genetic Aspects of Plantation Forestry in the Tropics

13.1 Introduction In tropical countries ever more land is becoming available for tree plantations. This is due to the destruction and degrading of natural forests or failures of their management and growing problems of access to existing forest. Plantations represent either reforestation, i.e., changing deforested land back to forest land, or afforestation, i.e., creating forest where it “historically” did not exist. The high productivity of tropical and subtropical plantations and their numerous ecological, social, and economic benefits make plantations highly attractive. Much attention has recently been devoted to research and planting of nitrogen-fixing hardwoods such as Leucaena spp. (Brewbaker and Sorensson 1994) and Sesbania spp. (Owino et al. 1994). Planting is urgently required on eroded sites or on sites threatened by erosion. It is also practiced with the objective of restoring forest ecosystems. Planting is done by private companies, in the framework of special developmental projects, and in national afforestation programs. Planting projects are also promoted by private companies and nongovernmental organizations (Evans and Turnbull 2004). The total area of forest tree plantations in tropical and subtropical countries has been estimated to be close to 100 × 106 ha in 2000 (Table 13.1) and may have exceeded this area since then. The annual loss of forest area is still estimated to exceed by far that of newly established plantations (FAO 2002). A comprehensive discussion of tropical tree plantations has to take into account ecological, genetic, silvicultural, and socioeconomic aspects (Evans and Turnbull 2004). The genetic aspects of man-made forests established by planting or sowing are centered around their establishment. Therefore, the most significant genetic aspect is the choice of reproductive material used. Plantations are much less competitive than natural forests. The competition among adult trees for light, nutrients, and water is not accompanied by as much mortality as during the early phases of the life cycle in natural populations. These critical phases are replaced under human intervention by procedures designed towards

13

214

C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics

Table 13.1. Areas of planted forests including woodlots in tropical and subtropical regions and their average annual increase from 1965 to 2000. Numbers refer to thousands of hectares. (Adapted from Evans and Turnbull 2004) Region

1965

1980

1990

2000

Annual average

Africa Asia (including southern China) Northern Australia and Pacific Islands Central America and Caribbean South America Total

1,378 421 70 219 579 6,667

2,724 13,046 269 486 4,448 20,973

3,773 29,245 420 786 8,470 42,694

4,566 73,444 480 1,311 8,634 88,435

170 2,110 16 44 263 2717

maximum amounts of planting stock to be derived from a given quantity of seed. This is typical of all forms of artificial regeneration. Many tropical plantations consist of very few if not only one tree species and are therefore perceived as ecologically unstable moncultures. However, the real cause for concern should rather be the eventual genetic uniformity of the respective species that often accompanies monoculture (Guries 1990). This is due to the restricted number of parent individuals of the planting stock and is most pronounced in uniclonal plantations. Visual uniformity of even-aged plantations established at regular spacing is attractive to some people and “unnatural” to others but of minor importance. A moderate number of tropical tree species are used in plantations (Table 13.2). The species differ much in growth and ecological demands. Their choice depends on the purpose of the plantations. A large proportion of them are used Table 13.2. Species used in tropical planted forest. Percentages refer to the main groups only. Adapted from Evans and Turnbull (2004, p. 37) Genus/group

Percentage

Species

Eucalyptus

50

Acacia Tectona Pinus

17 10 23

grandis, camaldulensis, globulus, saligna, deglupta, tereticornis, robusta, citriodora, exserta, urophylla, hybrids, and others nilotica, mangium, auriculiformis, crassicarpa grandis caribaea, patula, elliotti, oocarpa, kesyia, merkusii, massoniana, and others Cunninghamia lanceolata, Araucaria cunninghamii, Araucaria angustifolia, Cupressus lusitanica, and others Gmelina, Terminalia, Albizzia, Leucaena, Grevillea, Prosopis, Triplochiton, Meliaceae spp. (e.g., Swietenia, Azadirachta), Paraserianthes, Cordia, Casuarina, and others

Other conifers Other hardwoods

13.2 Plantations of Exotic Tree Species

as exotics. The relative plantation area of tropical pines is likely to decrease, while several fast-growing hardwoods (e.g., Acacia spp., Gmelina arborea, several Meliaceae) are gaining importance. Teak is the only plantation species planted predominantly within its natural distribution range; however, significant exotic plantations of teak also exist. As Evans and Turnbull (2004, p. 11) noted, the distinction between forestry and agricultural tree crops is not clear. For instance, rubber plantations are classed as forests, while plantations of Acacia mearnsii that are also used for bark contents are not. Numerous nonindustrial tree species are planted in agroforestry, the mixture of tree-growing with food crops or livestock. These species are grown for timber, oils, fruits, etc.

13.2 Plantations of Exotic Tree Species When an exotic species is to be planted, the forester has the choice among a very wide spectrum of species that have the advantage of being free of their native herbivores and parasites. It may take some time until local insects or fungal pathogens become adapted to the new food source, or until the native pests and diseases are transferred. At present, exotic tree species are of prime importance for plantation forestry throughout the tropics. Managers can build on international experience with these well-known species that have been successfully used for plantations worldwide. Pure plantations using only a single or a few commercial species are the rule. Particularly various fast-growing, mainly exotic, pioneer tree species such as eucalypts, acacias, Gmelina, and pines are grown in singlespecies plantations. The plantations existing in more than 100 countries are used for pulp, woodchips, sawtimber, fuel wood, and other products. Owing to extremely short rotation periods, ecological risks are expected to be low. If plantations are grown on suitable sites, the wood production is extremely high. However, the demand for nutrients and water may also be high. More and more such plantations are established with breeding products (Chap. 12), which provides for sometimes spectacular rentability and makes plantation forestry attractive. These plantations cannot simply be considered to be substitutes of natural forest (Chap. 8 in Bruenig 1996). Instead of replacing existing primary or secondary forest, they should be planted on deforested land and remove pressure from natural forest. The introduction of exotic species has frequently happened in several steps. The widest distribution originated from locations far away from natural populations. For example, both the coffee (Coffea arabica) and rubber (Hevea brasiliensis) trees were originally planted as exotics in fairly small populations

215

216


in botanical gardens in Europe. Large-scale exotic plantations of these crops were later established throughout the tropics using material from only a few plants growing in botanical gardens (Monaco 1977; pp. 22ff., 213 f. in Smith et al. 1992). Botanical gardens have also played an important role in the artificial deployment of some forest trees. The provenance or even the origin of reproductive material for exotic plantations is often poorly documented. Maternally inherited markers strongly suggest long-distance seed transfer prior to the establishment of Dalbergia sissoo plantations in Nepal (Pandey et al. 2004; Example 2.2). Uncontrolled transfer of seeds of forest trees even among continents has been practiced for centuries. Small test plantations were established and reproductive material was harvested in large quantities from those small test plantations or purely incidental introductions, once the species turned out to be suitable for plantation establishment in a region. Other species, such as Leucaena leucocephala in tropical Asia, proved to be well adapted to the environmental conditions prevailing at their numerous exotic planting sites. They regenerate naturally and quickly disseminate their genetic information as colonizing species. Mahogany (Swietenia spp.) has become one of the most important timber species in the Philippines. Several documents suggest that no reproductive material has ever been imported directly from tropical America. The first generation of mahogany on the Philippines was probably established with material imported from a few exotic plantations in Sri Lanka. The uncontrolled distribution of reproductive material and the small number of founder individuals that sometimes were even related have several genetic implications. The planted material does not always originate from populations most suitable for plantation establishment in a region. The distribution of species and provenances by humans is more often an outcome of random events rather than following a sound plan for the evaluation of genetic resources. Inbreeding is a likely result of harvesting sexually produced progeny in small trial plantations of a newly introduced exotic species and may further increase in advanced generations. A decrease of vitality, viability, and growth observable in the second and later generations is a likely consequence of inbreeding depression (Example 13.1). Another negative effect of small sizes of founding populations has been the loss of genetic variation due to drift. Many exotic plantations exhibit reduced levels of genetic variation as compared with levels in natural populations because of random loss of genes. The “genetic base” of the plantations is said to be “narrow.” Low levels of genetic variation impair processes of evolutionary adaptation to changing environments. Populations with low genetic variation are not suitable as base populations for breeding programs. Many attempts to introduce plantation species have failed owing to environmental mismatch, pests or diseases, and the disregard of some simple principles of provenance choice.

13.2 Plantations of Exotic Tree Species

Example 13.1: The Origin of Early Plantations of Acacia mangium in Sabah Acacia mangium is currently one of the preferred species for plantation establishment in the humid and subhumid tropics. The plantation area has rapidly increased, in particular in Southeast Asia. Acacia mangium was first introduced to Sabah (Malaysia) in 1967. The first introduction was the offspring of a single open-pollinated seed tree in Australia. Two small stands of 34 and approximately 300 trees, respectively, of its openpollinated progeny were planted. More than 15,000 ha of plantations was established in Sabah during the period from 1971 to 1984. Seeds for the establishment of these plantations originated exclusively from these first plantations and their advanced generations. Fructification of Acacia mangium is prolific a few years after planting. Up to five consecutive generations of Acacia mangium were produced until 1984. Not to forget, the ancestors of all these trees are the initially introduced progeny. A simple nursery trial was established comparing seedlings from the first to third generations, a phenotypically selected second generation, and, for comparison, a later-introduced provenance mix (Table 13.3). The growth of seedlings harvested from the second and third generations was reduced compared with that for the first generation and the provenance mix. Accumulating inbreeding effects in advanced generations are a likely explanation for this result. The seeds harvested from the first-generation stands must have been affected by inbreeding. However, seedling growth was identical to that of the provenance mix. A possible explanation is the generally low level of heterozygosity of Acacia mangium (Table 3.1), leading to only moderate inbreeding depression. Furthermore, the provenance mix was likely to contain some genotypes that are not adapted to the conditions in the test. The fastest-growing seedlings were harvested in a stand of the second generation which was transformed to a seed production area by phenotypic mass selection. Sim (1984) explained this result by the development of a land race particularly adapted to the conditions in Sabah, and by the efficiency of mass selection. (From Sim 1984)

Table 13.3. Average height of Acacia mangium seedlings in a nursery trial in Sabah. For further explanations see the text. (Adapted from Sim 1984) Seed harvest in stand

Seedling height (cm)

1st generation 2nd generation 2nd generation, selected 3rd generation Provenance mix

32.5 20.7 35.2 18.1 32.5

217

218


13.3 Plantations of Indigenous Species Given the threat to the forest gene pool, the promise of a renewed private planting programme, and our ignorance of propagation techniques and growth potential of many potentially important species, it is more than ever necessary to co-ordinate national germplasm collection and breeding efforts on native hardwood trees . . . (Lawson 1994).

The species spectrum of indigenous species of a country is, of course, not as broad as the worldwide spectrum of exotic species. Indigenous species still possess the advantage of being adapted to the regional climatic conditions. Local consumers are familiar with the properties of their wood and the various nontimber products that they yield. The utilization of species for plantation forestry is frequently impeded by the low level of knowledge of their basic biological, ecological, and silvicultural features. Evans and Turnbull (2004, Chap. 8) discussed arguments for the choice of a given species and presented examples of indigenous species that have recently been used for plantations. Many of these species are rare in natural ecosystems. Their use in plantations containing a few or only one species involves a drastic change in ecological conditions, particularly the pressure of infestation by fungi and insects. The importance of indigenous species is still low in plantation forestry of the tropics (Table 13.2). However, the anticipated increase of the plantation area is expected to be based on a larger number of species, including indigenous species and “lesser-known” species. An increased number of species used for plantation forestry promotes the establishment of mixed plantations (FAO 1992). The growth of species in natural forests of the region gives some indication of their value for use. Subsequent species-elimination trials allow us to assess whether a species is suitable for being grown in plantations. The observation of a few trees, for example, in a botanical garden or an arboretum alone, is not a sufficient basis for the identification of potentially important plantation species. A test of the suitability of a species or population for plantation establishment is only meaningful if the population from which the seeds were harvested is conserved at least until the end of the trial. This represents a genetic resource as a necessary component of any program for the domestication of a species. Procedures of seed procurement of exotics used during the early days of tropical forest plantations may easily be avoided in indigenous species. This provides for more genetic variation being retained in plantations. Nevertheless, some basic principles of seed harvest and of experimental design must be followed. Above all, seeds should never be harvested from isolated trees since inbreeding is likely to occur and the genetic structures differ from those of natural populations.

13.3 Plantations of Indigenous Species

The most important exotic plantation species are characterized by their gregarious production of seeds. The copious small seeds of species such as pines, eucalypts, acacias, and teak are easy to harvest in large quantities and to store for several months or even years, i.e., the seeds are orthodox. This feature and the ease of the production of seedlings in nurseries contributed to the importance of the species for plantation forestry. Problems in seed procurement for afforestation projects are frequently underestimated. It is difficult to produce and store suitable reproductive material in sufficient quantities. This is due to irregular and sporadic seed production and/or recalcitrant seeds, i.e., seeds which cannot be stored for long periods. Other species have a seed dormancy that is difficult to break for largescale production of seedlings. The main problem with those species is frequently not the choice of genetically suitable reproductive material but rather the production of sufficiently large quantities of planting stock. Seed orchards (Sect. 12.8) help to solve some of these problems. Other help is seen in vegetative propagation. Leakey et al. (1982) emphasized the potential use of numerous species for plantation forestry in the tropics. They described the options for an improved seed procurement and relevant research in this context using the example of the West African species Triplochiton scleroxylon, a species with an erratic pattern of reproduction. By adequate treatment and the use of appropriate facilities, the storage period of seeds could be extended. Besides the development of methods for controlled pollination, an integrated procedure allows for vegetative propagation and the early selection of tree clones with good stem form. Ladipo et al. (1991) reported further on this research. In West Africa, where the discrepancy between annual deforestation and total tree plantations is particularly wide, Lawson (1994) identified an urgent need for domestication of indigenous tree species such as Terminalia ivorensis, Terminalia superba, Triplochiton scleroxylon, and Lovoa trichilioides. The clonal propagation option is preferred in view of the problems involved in the procurement of seed of good quality, and the increased yield. This may be an efficient measure where forests disappear fast. Milimo et al. (1994) listed species of semiarid West Africa that are worthy of breeding. In view of various difficulties involved in propagation by seed, these authors stressed the importance of vegetative propagation and reported on experimental results. In their comprehensive review of domestication of the genus Swietenia and other mahoganies, Newton et al. (1994) described the development of molecular genetics techniques, provenance research, selection, and the development of both conventional and in vitro propagation. On this basis, appropriate silvicultural systems accounting for the Hypsipyla problem could be developed. The worldwide use of eucalypts may create the impression that these are the only important tree species in Australia. Nevertheless, projects in provenance

219

220


research, breeding, and vegetative propagation of lesser-known species of Casuarina, Melaleuca, Grevillea, and Sesbania are of importance also to countries where these species are of potential use as exotics (Booth and Turnbull 1994). Enrichment planting, underplanting valuable species in partially cleared natural forest, is an efficient means of conserving the target species and of introducing tested planting stock of the respective indigenous species (Lawson 1994). Research on domestication of agroforestry tree species is unfortunately concentrated on a few exotic species, although the number of indigenous species that are suitable for cultivation in this agricultural system is very large (Sinclair et al. 1994). As for the establishment of mixed forests, increased planting of indigenous, frequently denoted as “lesser-known,” species is desirable. The quality of reproductive material of lesser-known species is often poor: genetic differentiation patterns are unknown and consequently no information is available concerning the choice of optimum provenances for plantation establishment in a region (Chap. 11). Also the physiological quality of harvested seeds is frequently poor. Seeds are often infected by pathogens because they were picked up from the ground rather than from standing trees or because they were stored using inappropriate facilities (Kamra 1986). Leakey and Newton (1994b) chose the name “Cinderella tree species” for these species that have not contributed significantly to reafforestation, possibly because they are rare in natural forests. Many of them yield also fruit, nuts, and other nontimber products (Clement and Villachica 1994; Okafor and Lamb 1994; Prance 1994). These species do not necessarily grow less fast than other species (Maghembe et al. 1994). Many fruits and nuts have been collected from these species in the wild but numerous species are also well suited for agroforestry. Trees of indigenous species planted in agroforestry ecosystems do not necessarily possess less variation than those in natural forests (Hollingsworth et al. 2005).

13.4 Basic and Reproductive Material The success of artificial regeneration of forests critically depends on the choice of suitable reproductive material. Seeds, seedlings produced in nurseries or collected in forests (wildings), or vegetative propagules (e.g., stem or root cuttings) are reproductive material. Reproductive material must be adapted or adaptable to the environmental conditions prevailing at the planting sites. The objective is to establish a sufficient number of trees that are capable of surviving until the end of the rotation period. Farmers attribute much weight to the choice of suitable varieties of agricultural crops; however, foresters involved in the establishment of plantations

13.4 Basic and Reproductive Material

frequently show a surprising lack of interest in the selection of suitable reproductive material. Once unsuitable material has been chosen, the choice cannot be rectified later. The consequences of selecting inadequate reproductive material become obvious only after a few years. Forestry is a low-input enterprise. It can hardly compete with tree plantations for producing export crops such as oil palm, or rubber. Forest plantations are characterized by large temporal and spatial heterogeneity of the environment. Thus, the identification of optimum planting stock is of more crucial importance than in most agricultural crops. The situation is aggravated by the still poor knowledge of the genetic systems and genetic variation patterns of most tree species. Detailed legal regulations on the marketing of forest reproductive material exist in the member countries of the OECD (Nanson 2001) and the EU (2000), i.e., mostly countries of the temperate and boreal zones. The forest owner is, of course, free to produce reproductive material in his or her own forest in order to meet his or her own demand. Comparable regulations do not exist or are of only little practical importance in most tropical countries. Designated authorities control collection, processing, storage, seed testing, and marketing of forest reproductive material. Genetic methods of certification are not included. The legislation refers to forest reproductive material but contains very little on the reproductive material itself. The bulk of the regulations deals with basic material. The implicit assumption is that reproductive material possesses the same genetic structure as the basic material that it was derived from (Hattemer 1987). Types of basic material are stands, seed orchards, and clones. There are several categories of reproductive material. It may be classified as selected if it was derived from basic material that meets criteria such as superior phenotypic quality, uniformity, and isolation. The stands must be officially approved for seed collection. This accounts for the certainly most widely practiced method of selection in tree species, the selection of stands without progeny testing. Foresters the world over collect seed for routine plantations in stands selected for their superior phenotype. In the member countries of the OECD and the EU this is even a legal precondition for the marketing of the respective reproductive material. This method is not based on the estimation of genetic parameters such as expected breeding progress. It has neither been proven whether the plantations established with selected reproductive material outperform other stands. Not even exemplary progeny tests have been made in support of this procedure. For several reasons it is unlikely that this procedure leads to appreciable, if any, breeding progress. The selection is done at moderate intensity among stands scattered over large regions where the effects of the environment and possible silvicultural treatment are unknown but presumably large. According to what we know on the reproduction system of stands, genetic equilibrium is at best realized approximately. Last but not least, since in EU and OECD countries forest reproductive material to be put

221

222


on the market is restricted to approved basic material, only about 1% of the area covered by the main tree species contributes offspring to plantings in some member countries. The area where seed is collected changes with time, so there is some change in the collective of approved basic materials. The percentage of the area covered by the respective species is nonetheless very small. However, the given degree of isolation hardly prevents influx of effective pollen. If no additional cost is involved, seed may be collected in selected stands, although expectations of far-higher value for use are not necessarily justified. Source-identified reproductive material does not meet the criterion of being derived from selected base material. It is assured to have been produced in a given region of provenance. Qualified reproductive material must be produced by parent trees or in a seed orchard the components of which were selected on an individual basis. It may also be propagules of a clone or clonal mixture. Tested material is not just derived from phenotypically superior basic material but has been tested for improved value for use in experiments that meet certain criteria of design. The improved value is defined as the superiority over officially prescribed standard entries included in the experiments. Until the present, only a very small fraction of reproductive material has been marketed in this category. The amounts of reproductive material collected in selected, qualified ,or tested base materials are not restricted. The advantage of the legislation is certainly the concomitant documentation of the seed collection site, even if this may not represent relevant information on reproductive material (Jones and Burley 1973).

13.5 Production and Collection of Seed Natural forests as a seed source may present problems for seed collection. Particularly the collection from tall trees in less prolific species is laborious in spite of modern equipment. Fruits of many species are dehiscent. Fruit and seed parasitism and fruit-eating animals mitigate harvested seed. Melchior (1986b) described problems related to seed procurement for three important and promising South American tree species in Peru and Venezuela (Cedrelinga catenaeformis, Juglans neotropica, Bombacopsis quinata). 13.5.1 Seed Production Areas

Seed production areas arise either by establishment of a stand with the objective of seed production or by phenotypically superior stands being converted

13.5 Production and Collection of Seed

to seed production areas by thinning. The main objective of the thinning is to enhance pollen and seed production of the remaining trees. Since competing trees belonging to other species are likely to impede pollen dispersal of the target species, they are removed just as phenotypically minor trees. This procedure is similar to that leading to the approval of stands for seed collection in OECD and EU member countries. If the converted stands were chosen according to phenotypic criteria only, it is recommended including the progeny in provenance tests in order to get some indication as to their merit. Also, the conversion to seed production areas is recommended in regions that have been proved to be suitable for seed procurement on the basis of the results of provenance trials. Practical considerations regarding seed production areas were summarized by Schmidt (1993c). If stands possess some breeding status, they are really selected in the close sense of the word and may be used for establishing seed production areas. These stands may be used for the production of better-quality seed and may only later be replaced by seed orchards or other more sophisticated techniques for the deployment of a breeding product. Wellendorf and Kaosa-Ard (1988) described the role of seed production areas in the breeding program for teak in Thailand (Example 11.2). Species-rich tropical natural forests with low density of the target species are more difficult to convert but are still widely used as seed production areas. In temperate conifers the density of seed production areas is of no or minor relevance (El-Kassaby et al. 2003). Zheng and Ennos (1997) compared the outcrossing rate in a seed production area with that of a natural population of Pinus caribaea in a mixed stand. The two estimates were absolutely the same but in the seed production area there were also indications of biparental inbreeding. These authors made the interesting point that the milder form of inbreeding implies more inbreeding depression in the plantations established with this seed, since the offspring arising from self-fertilization would hardly germinate at all. They recommended properly designed clonal seed orchards that lack a family structure. In an orchard of this type neither form of inbreeding was observed. In a seed production area of Eucalyptus citriodora in Brazil, Yeh et al. (1983) found outcrossing rates around 0.82, i.e., markedly above those in most natural stands of eucalypts (Table 6.2). 13.5.2 Provenance Resource Stands

Provenance resource stands are a particular type of seed production area. They are frequently established simultaneously with the provenance trials, i.e., during an early stage of a breeding program. They represent a mixture of the progeny of numerous seed trees of a single promising population and should

223

224


be reproductively isolated. A main objective of such stands is the dynamic ex situ conservation of genetic resources (Chap. 14). When matured, stands of superior provenances are used for the production of seeds for plantations. The genetic structure of the progeny also of provenance resource stands is not necessarily identical to that of the trees representing a provenance in the field trial but is similar to it. The cost is much less than that for establishing the field trial, since no design is required. This type of stand is highly recommended but rarely practiced. Nikles and Newton (1984) described their establishment with several Pinus caribaea var. hondurensis provenances in Queensland, Australia. 13.5.3 Seed Orchards

Seed orchards are primarily established in order to produce large quantities of seed from a limited number of selected trees (Sect. 12.8). Selected trees are cloned and copies assembled to a propagation population. The trees may also just have been scattered remnant individuals in degraded forest so that mating contact can be reestablished. Once an orchard exists, seed collection is much cheaper than anywhere else. In the present context, increased seed production and greater ease of seed collection is of prime importance. Not only grafting but also rooted cuttings may be used for the establishment of clonal seed orchards, provided that they flower early at convenient height above the ground. Seeds of B. quinata can easily be produced in clonal seed orchards, since a simple method for the vegetative propagation of the species was found. Melchior (1965) observed how the indigenous population erected living fences. They used long and heavy branches rather than cuttings of the size of a pencil. The resulting multisprout bushes flowered abundantly in the second year. Just as in the crowns of adult trees, certain bees operated as nectar thieves during the day. The comparison of the genetic quality of seed produced in natural and domesticated populations is complex and implies the risk of oversimplification. El-Kassaby (1999) pointed out the genetic variation in orchard seed may easily be higher than in the seed collected from a stand, if many clones were derived from trees selected in several genetically differentiated populations. Care is required when choosing the stands to be used as a reference. A wellfounded example was presented by Godt et al. (2001): 47 and 48 trees were sampled in seven and five populations of Picea glauca and Pinus banksiana, respectively, scattered over a vast area of forest in Canada. Forty and 31 seed orchard clones selected in this area, respectively, showed only minor reductions in variation assayed at 18 allozyme loci. Although numerous rare alleles were not encountered in the seed orchards, the genetic distance was negligible.

13.5 Production and Collection of Seed

Plus trees are possibly older and more heterozygous at certain gene loci than other trees in the same population (Bergmann and Ruetz 1991). They may have more variable progeny. Comparisons of the inbreeding coefficient of the seed are possibly less problematic. Since any comparisons are still subject to the vagaries of sexual reproduction, repeated sampling of the seed output is required. Also the less expensive seedling seed orchards are mostly linked to breeding programs. Lee (2000) pointed to the aspect of pollination in seed orchards. When studying the mating system of Dryobalanops aromatica, a dipterocarp species with irregular flowering, he found an indication for a lack of pollinators in a seed orchard being responsible for increased self-fertilization (see also Sects. 10.2.1, 12.8.2). If a seed orchard does not consist of selected families to be propagated under isolation, precautions may be taken for improving the situation by locating orchards where they are surrounded by extensive forests of dipterocarps or nondipterocarps providing pollinators. The appropriate spatial configuration can help to minimize other consanguineous matings. The incidence of the same species might also provide genes. Lee (2000) also warned against placing seed orchards into other ecosystems in order to avoid undesirable molecular imprinting and maladaptation. Therefore, this author proposed delineating planting zones on the basis of biological and ecological data and establishing plantations only in their own planting zone. This procedure is practiced in many countries of the northern temperate zone. It represents an alternative to the principle of transfer rules as mentioned in Sect. 11.6. The aspect of pollination was also discussed by Moncur et al. (1995). In seed orchards of several eucalypt species these authors observed a considerable improvement in both the amount and the quality of seed after setting up beehives. The average outcrossing rate was raised from 0.76 to 0.91. In Europe, beehives are also placed in remote stands of rare trees of the rose family in order to increase the seed crop. Also Moran et al. (1989b) and Chaix et al. (2003) (Sect. 12.8.2) reported considerably lower rates of self-fertilization in seedling seed orchards than in natural stands of eucalypts. The former authors emphasized the reduced neighborhood inbreeding by less spatial clumping of relatives. In a seed orchard containing both grafts and open-pollinated progeny of plus trees in Douglas fir, Ritland and El-Kassaby (1985) found only little self-fertilization but indications of consanguineous matings primarily due to the large and varying size of the families. In a seedling seed orchard of Pinus merkusii Siregar and Hattemer (2001) observed a relatively low outcrossing rate (0.87) with little incidence of consanguineous biparental mating; however, during the extended flower period the phenology differed between families and the individual effective pollen clouds were genetically differentiated. In a far-distant natural stand the outcrossing rate of 0.98 was similar to that of other pines and

225

226


much higher than in natural stands of the species in Thailand (Changtragoon and Finkeldey 1995a; Sect. 6.3.2). If the orchard produces an excess of seed, collection should not be concentrated on part of the clones or families. Better, although more costly, is to collect less seed from all trees so that more effective neighborhoods are represented in the seed.

13.6 Collection and Storage of Seed Seed is either collected off the trees or picked up from the ground. Seed collection by climbing is expensive and dangerous in spite of modern equipment. The seed should be collected in an appropriate way so as to minimize losses of genetic variants. The genetic implications of seed collection from trees were discussed in Chap. 11. Too few trees might be available for seed collection from the crowns of felled trees. Seeds of other species are caught in nets laid out on the forest floor. If a restricted number of nets are laid out under highly prolific trees only, a spatial genetic structure becomes manifest in the seed crop and induces genetic change. In stands with a family structure this seed is expected to be markedly different from seed picked up by hand from the total area of the stand as was shown by Ziehe et al. (1998) in a study of beech (Fagus sylvatica). The seed of many tropical trees such as Triplochiton scleroxylon or Araucaria spp. is also collected from the ground. Also when collecting wildings, their genetic variation should be accounted for. A patchy distribution of saplings indicates that they might represent offspring of a restricted number of trees. The recalcitrant seeds of many tropical species are sensitive to desiccation (mangrove species, neem, Araucaria angustifolia and other Araucariaceae, most dipterocarps) and are thus rather short-lived. Seeds of many tree species of the temperate zones are orthodox and can be stored for a considerable time. The seeds of most Acacia spp., Eucalyptus spp., and teak are orthodox. Seed treatment such as the careful reduction of the moisture content and the storage in a strictly controlled environment can be used to extend the life span of the seeds of most species (Schmidt 2000). Air-conditioned cold-storage rooms are used for the long-term storage of seeds in gene banks. Optimum conditions for long-term storage depend on the seed physiology of the species. Seeds of some tropical species retain germinability even at temperatures below the freezing point of water and thus can be frozen without damage. The life span of seeds of other species is longest at low temperatures just above the freezing point of water. The humidity in storage rooms is also controlled and usually reduced compared with that outside. More

13.7 Planting Stock Production

details on optimum storage conditions for the seeds of tropical trees have been summarized, for example, by Albrecht (1993), Tompsett (1994), and Schmidt (2000, p. 223 ff.). The germinability of many recalcitrant seeds cannot be extended to more than a few months. The principal distinction between orthodox and recalcitrant seeds is maintained even under optimum storage conditions. The long-term storage of seeds is not an option for the conservation of genetic resources of most species with recalcitrant seeds. In any event, the loss of germinability represents a reduction of the population size. As some studies in temperate tree species have shown, this reduction is also connected with selection (Melchior 1986a; p. 340 f. in Hattemer et al. 1993; El-Kassaby 1999). The selection going on during germination and nursery operations rids the planting stock of adaptively weak individuals arising from inbreeding. Seeds of certain families survive for different periods, so the variation decays during storage. It is an open question whether this also implies adaptation to the storage conditions. Evans and Turnbull (2004, p. 134), Schmidt (2000), and Midgley (1996) reported new developments in seed collection and seed technology of tropical trees. Research on seed germination of fruit-bearing trees is under way (Maghembe et al. 1994).

13.7 Planting Stock Production Artificial regeneration does not rely on the natural seed fall and the concomitant seed dispersal. Seeds are collected off the trees, processed, and eventually stored. The seed is either sown out directly or planting stock is raised in a nursery or greenhouse under intensive treatment so that mortality and thus the chance of adaptive viability selection are minimized. Outplanting or seeding is hardly ever done on the site of the parent stand. The implied habitat change interrupts the long-term process of adaptation and enforces selection with changed intensity and direction. The emerging artificial stand no longer has a family structure. 13.7.1 Seedlings

The seeds are germinated in nurseries or greenhouses under favorable conditions in order to achieve the maximum planting stock. The prevention of mortality implies that only little viability selection can take place at this early stage of the life span. In tree species of the northern temperate zone, it was shown that the early viability selection represents an important element of the genetic

227

228


system of trees (Ziehe et al. 1999). This type of selection is then postponed to the period after outplanting. Also grading of plant material has a genetic implication. Under highly favorable conditions of planting stock production, Eriksson and Lindgren (1975) found less inbreeding depression in both Picea abies and Pinus sylvestris and inferred that there may be more inbred offspring among tall plants. In experiments done by Konnert and Schmidt (1996), the tallest plants in planting stock of Picea abies turned out to possess the lowest degree of heterozygosity–a property that is otherwise an advantage for field survival. In European white fir the result was the reverse. In both species, the taller and shorter plants were genetically differentiated. 13.7.2 Clonal Multiplication

Plants are regenerated directly from propagules without generative processes being involved (Sect. 4.3). The complexities of the natural systems of sexual reproduction in tropical trees (Sect. 4.2) are avoided. So are problems involved in parts of the production of planting stock. Consequently, ever more tropical plantations are established with clonal material. Some authors (Felker 1994) advocate clonal propagation for the sole reason of the self-sterility of a species and the implied difficulty in capturing the genotype of individuals with desirable properties in sexual progeny. Numerous techniques are used for the vegetative propagation of tropical forest trees. Smits et al. (1994) have reported on the production of planting stock by cutting propagation of dipterocarps, a family with erratic flowering, recalcitrant seed, and much seed parasitism. The different techniques for macrovegetative and microvegetative propagation of forest trees are not part of forest genetics research and cannot be discussed in this introduction. Laboratories and culture rooms for microvegetative propagation are not available everywhere. Besides the capture of breeding progress (Sect. 12.8.3) the use of vegetative propagation techniques solves problems in seed supply. The selection of “superior” clones for plantation establishment is often of only secondary importance (cf. the example of Triplochiton scleroxylon described by Leakey et al. 1982; Sect. 13.3). The erratic flowering and seed production of dipterocarps of the humid tropics was the principal motivation for the development of vegetative propagation techniques for this important family (Smits 1993). Some tropical tree species are potentially important plantation species owing to certain silvicultural features or wood properties; however, seeds are available only sporadically and/or in insufficient quantities. The seeds of these species are often recalcitrant, and the production of seedlings is sometimes

13.8 Establishment and Development of Plantations

difficult. The inevitable loss of genetic variation associated with the shift from sexual to asexual propagation techniques is accepted in view of the pragmatic advantages of vegetative multiplication techniques such as the uniformity of the plantations. Seed production of other species such as most eucalypts is prolific, seeds are orthodox, and sexual progenies, i.e., seedlings, are easily produced using standard nursery techniques. In these species asexual propagation is more laborious and expensive. Vegetative propagation is applied only for the mass multiplication of certain selected trees. The term “clonal forestry” is associated with an anticipated increase of the yield by planting these selected clones. The risks associated with the unavoidable reduction of genetic variation in clonal plantations occur, of course, independently from the motivation for the use of vegetative propagation techniques. The often-neglected genetic implications of clonal plantations are addressed in Sect. 13.9. However, vegetatively propagated plants are not necessarily superior to generatively produced progenies. For example, the growth of vegetatively propagated E. grandis plants was compared with that of seedlings from the same trees after open pollination in Brazil. For a variety of reasons, the growth of the seedlings was superior to that of the vegetative offspring 30 months after planting (Kageyama and Kikuti 1989).

13.8 Establishment and Development of Plantations From the genetic point of view the choice of the reproductive material is the most important decision before plantation establishment; however, genetic structures of artificially regenerated forest tree populations may change in time during the reduction of the number of individuals owing to viability selection. A temporal dynamics of genetic structures was observed in experimental populations of some temperate forest trees during early development of the seedlings (Kim 1985; Ziehe et al. 1999). In order to provide ample room for adaptive change, large numbers of individuals should be planted (Langner 1966; Ledig and Kitzmiller 1992; Müller-Starck 1996). In rotation forestry with short rotation periods, this safeguard is of less importance than in long-lived tree species. Forest reproductive material can be transferred over large distances. The capacity of artificial migration is obviously unlimited. The ecological gradients overstepped by this transfer matter greatly. Sometimes reproductive material is totally ill adapted. Uncritical use of just any reproductive material has long been and still is hazardous to forestry; therefore, the careful choice of populations must be integral part of adequate forest management. The number of plants per area or tree spacing is an important aspect of planting. In tropical plantations mostly fewer than 1,000 trees are planted depending

229

230


on the purpose of the plantation. The potential for selective change of genetic structures and the purging of physiologically ill adapted individuals depends on the genetic variation of the planted population, the number of trees planted, and subsequent mortality. Evolutionary adaptations (Sect. 7.2) avoiding critical population sizes are more likely in plantations established with a large number of different genotypes planted per unit area. No evolutionary adaptation on the basis of viability selection is possible if only a single clone is planted. The temporal dynamics of the genetic structures of a population is not only influenced by natural mortality. Silvicultural operations and particularly selective thinning have also the potential to change genetic structures. Different thinning methods affect genetic structures in different ways as was shown for Norway spruce (Picea abies) by Hosius (1993). Comparable experimental studies for plantations of tropical forest trees are not yet available. The temporal dynamics of genetic structures in planted forests deserves particular attention if it is planned to harvest reproductive material in a plantation either after its conversion to a seed production area or if it was planned to be naturally regenerated. Changes of genetic structures in production populations that will never give rise to a subsequent generation are of no evolutionary significance although they may have an impact on economically relevant trait expressions.

13.9 Natural Regeneration of Plantations If an indigenous or introduced plantation species turns out to grow obviously well and to be adapted to the new site and sets seed freely, it can be considered “naturalized” (p. 11 in Evans and Turnbull 20041). Even after natural regeneration it is still a man-made forest. It may be on the way to becoming a land race, since at least a whole generation cycle has been closed under field conditions on the new site. Land races with particular morphological and physiological trait expressions have possibly evolved from exotic plantations within a few generations (Sect. 11.2). The genetic differentiation of land races from natural populations is likely to increase the adaptedness to environmental conditions prevailing at the new planting sites; however, genetic differentiation between exotic and natural populations may also be due to genetic drift in the exotic plantations and thus a consequence of evolutionary change that is not necessarily adaptive. At the time of planting it may be left open whether a plantation might be regenerated naturally in future. Plantations are not expected to have a spatial genetic structure except through selection (Eldridge et al. 1994). The density should not fall short of a critical lower limit below which pollination is incomplete and outcrossing is reduced in favor of self-fertilization. The advantage of

13.10 Use of Breeding Products

the natural seed fall is the larger number of seeds and seedlings produced during the regeneration period comprising several years. In ever-more countries a mosaic of autochthonous and allochthonus stands exists, so the incidence of hybridization between trees from stands of different origin must be expected to increase. Particular features of natural regeneration were dealt with in Sect. 10.3.

13.10 Use of Breeding Products All plantations are perceived to be in danger to some degree. They are threatened by a massive buildup of pests and diseases (p. 387 in Evans and Turnbull 2004). Examples of devastating outbreaks have been reported from the tropics, the temperate zone, and the boreal zone. Plantations are still of growing importance. They are the only means of deploying breeding products and obtaining return from earlier investment. It was pointed out in Sect. 12.11 that the collection of unrestricted quantities of seed in a given seed orchard or seed production area implies that the progeny of the respective trees cover an area exceeding that of the collection stand by a factor of several hundred. This is the experience in countries with official approval of seed collection stands and provenance recommendations. It is more costly to establish many orchards and many seed production areas and to collect only moderate amounts of seeds there; however, it is more in line with sustainability principles. Even the selection of provenances is a form of breeding that usually precedes more intensive selection steps. The large-scale planting of introduced species, provenances, or other breeding products counteracts the preservation of locally adapted populations that may contain elsewhere rare alleles or even unique alleles or allele combinations in high frequencies. Thus, important genetic resources for future adaptation processes and breeding programs can get lost. The loss of local genetic resources can be either direct by substitution or indirect through gene flow from breeding products and other allochthonous populations into surrounding natural populations of the same species. The dangers associated with the loss of locally adapted populations through their substitution by a few high-yielding varieties were first recognized for cultivated species. They have been the main motivation for the establishment of gene banks of many agricultural crop species. Tree plantations using breeding products currently cover only a small part of tropical forests. The danger of the loss of locally adapted forest tree populations is still much less severe than damage caused by forest destruction. Reduced genetic variation is a consequence of most breeding programs. This increases the risk of severe attacks by insects and diseases that may rapidly adapt to a homogeneous source of nutrients offered to them during several

231

232


generations. A well-known example is the destruction of most plantations of Leucaena leucocephala by cicadas of the genus Heteropsylla. The attacks were particularly severe in tropical Asia during the late 1980s and virtually terminated the use of the species for industrial plantation establishment (p. 323 in Smith et al. 1992). The calamity was promoted by the rapid large-scale expansion of plantations using a few varieties of “giant leucaena.” Recent breeding work in Leucaena is directed towards improving the resistance against Heteropsylla. This requires the test of genetically more variable Leucaena leucocephala varieties and hybridisztion with related species. Extraordinarily low amounts of genetic variation are also the likely cause for the destruction of Paulownia taiwaniana plantations in Taiwan. Paulownia taiwaniana plantations were established on a large scale during the 1970s. Just like other species of this genus, Paulownia taiwaniana is mainly propagated by root cuttings and other forms of vegetative propagation. Genetic variation within and among these plantations was extremely low (Finkeldey 1992). In the late 1970s almost all plantations were severely attacked by a witches-broom disease caused by mycoplasma-like organisms. There exists empirical evidence that long-term risks can be raised by breeding for resistance to present threats. By breeding and propagation, the number of genotypes required for buffering the population against unknown future threats is likely to be reduced. The consequence would be increased instability owing to recurrent artificial selection (Namkoong 1999). The risk of a large-scale planting of breeding products with reduced genetic variation is particularly high in clonal forestry; however, Park et al. (1998) pointed out that the diversity of progeny from intensively rogued seed orchards is not too different from that of clones that were derived from intensive selection in families produced by certain regular mating schemes. The production of genetically identical trees either by conventional cutting propagation or advanced biotechnology such as somatic embryogenesis and their planting offer chances and bear risks that must be balanced by appropriate management systems. Clones occur also in nature as the result of adaptation. Also, the cultivation of clones in Cryptomeria japonica, Cunninghamia lanceolata, and other conifers has a long history in eastern Asia. Further to Sect. 12.8.3, some guidelines are presented that may help to alleviate the problems associated with the large-scale planting of clones or other material of low genetic variation. They are admittedly based on hypotheses and models: ●

In view of increased vulnerability of less variable plantations as production populations, the use of clonal mixtures is recommended. Hundreds of hectares planted with a single clone are all but desirable. The variation in a mixture of given clones is maximized if all clones are represented in equal

13.10 Use of Breeding Products

●

proportions. This need not imply loss of uniformity of the wood and timber produced. The effect of mixture is the larger, the more the clones are genetically differentiated from each other at adaptive gene loci or in resistance. This quality of the mixture is more likely to be achieved if the clones are of diverse descent. This spreads the risk in view of adverse environmental factors and retards contagion during infestation by diseases and by defoliating or other insects. The question of the necessary number of components in mixtures has been tackled by several authors. For instance, Hühn (1986) considered the risk of losing the plantation through the loss of an unacceptable fraction q of trees prior to final harvest. The risk is a function of this fraction, the number n of clones, the susceptibility x, and the intensity a of contagion. For given q and a the risk decreases with increasing n and decreasing x. There are also special cases of x where the risk increases with increasing n. In many situations the risk function is asymptotic for larger n. Increasing this n beyond 30–40 equally represented clones is of little effect. Libby (1982) used a binomial approach for constructing the risk function and recommended between seven and 25 clones.

The genetic control of susceptibility was not considered in these models. In the approach taken by Roberds et al. (1990) a single agent active in the plantation and a single biallelic locus controlling resistance was assumed; hence, the frequency of either a dominant or a recessive allele for susceptibility could be integrated. These authors also assumed a particular mode of selecting given numbers of clones from progenies of the base population for propagation and found rather complex behavior of the risk function. It depends on the mode of gene action, since under dominance the proportion of susceptible genotypes is always larger. It also depends on the intensity of infestation whether increasing numbers of clones reduce the risk. If the intensity of pest attack increases, the number of clones necessary decreases. Simulation studies by Bishir and Roberds (1999) revealed that situations requiring more than 40 clones are not prevalent. Campinhos (1999) reported the use of 80 clones per production population in a given cycle of clonal mass cultivation in Aracruz (Example 12.1). Several EU and OECD countries have issued regulations on minimum numbers. One of the strong simplifying assumptions for the validity of these models is the static nature of susceptibility, i.e., the clones are either susceptible or not and this property does not change during the single rotation period considered. However, it is our experience that insects or pathogens, organisms with shorter generation cycles and more flexible genetic systems, evolve mechanisms to exploit the homogeneous substrate offered by genetically uniform plantations for years. The clonal aggregates are in turn hardly able to develop evolutionary adaptation in response to these attacks. This condition introduces uncertainty.

233

234


Consequently, if biotic agents are concerned, the problem of damaging agents cannot be solved in the long term by planting resistant clones. The evolving nature of the agents themselves cannot be incorporated into breeding trees for immediate resistance. It is not known whether the resistance breeder is always able to keep a step ahead or whether coevolving organisms with a more flexible genetic system are likely to take over (Namkoong 1999). It is still open to debate whether stability of host–pathogen and host–herbivore relationships can be achieved under mass cultivation of resistant clones. ●

●

●

●

●

●

Planting of clones and sexually produced offspring in intimate mixture or on the scale of medium-sized plots helps to increase overall variation. Since planting the clones in intimate mixture may still interfere with management and industrial use, their arrangement in a mosaic of monoclonal blocks is the preferred layout in Aracruz. In view of the dispersal of pathogens and insects, the establishment of such a mosaic in large connected tracts of plantations may be of minor relevance for contagion. The question of intimate mixture or monoclonal blocks does, of course, have relevance for yield and management (Foster and Bertolucci 1994). The use of clones might be restricted to fractions of the total forest area of a region. Campinhos (1999) reported on large tracks of native vegetation in-between clonal eucalypt plantations in Brazil (Example 12.1). Different performance of clones under different environmental conditions calls for the mass cultivation of different clonal mixtures in different regions as practiced by Aracruz Florestal. This procedure also increases total diversity. Shorter rotation cycles shorten the time available for herbivore and parasite populations to become genetically adapted to particular clones, unless the plantation is coppiced or the area is replanted by the same clonal aggregate. Short rotation periods are the great advantage of tropical tree plantations if compared with trees of other climatic zones. Many genotypes should be preserved in large clone banks (clonal archive), so that the spectrum of the mass-cultivated clones can be changed permanently. It is crucial to keep track of the identities of the clones deployed by means of highly variable genetic markers. The absolute maximum number of copies of a single clone ever deployed should be limited. In Sweden this number has been limited by law. On the other hand, the owners of small estates tend to propagate trees on their neighbors’ fields, making their choice on grounds of good growth, thereby selecting for rootability. This is hardly possible to change by legislation but leads to widespread deployment of few clones if not even a single clone.

Risk considerations are not intended to question or even disparage the magnificent success of clone breeding in the tropics and subtropics. Some of the


world’s most productive tree plantations, primarily poplars and eucalypts, use clones. Crucial advantages of clonal forestry are the independence of seed shortages, the uniform stands, and the ease of their management. Problems are seen in the deployment of clones. The concern is motivated by the growing need for intensive measures of plant protection in agricultural crop plants, not least in perennial crops such as avocado, banana, citrus, coffee, coconut, mango, oil palm, papaya, rubber, tea, and vine. Sooner or later in all plant cultures with reduced genetic variation a slow but steady increase in the incidence of insects and diseases has to be envisaged and the pressure on tree plantations will increase. Since forest trees are not a cash crop, intensive chemical control of pests and diseases rules itself out for economic, not to speak for ecological reasons; therefore, the matter is of importance both economically and environmentally. It is therefore preferable to rely on prevention. Foster and Bertolucci (1994) have discussed several independent approaches to the maintenance of some level of genetic variation in large-scale clonal plantations. For the time being, intensive monitoring is advised and is done (Campinhos 1999). The economic returns from highly productive plantations allow for this active monitoring of pests and diseases (Evans and Turnbull 2004). Repeated reference made to the clonal program of Aracruz Florestal should not be perceived as hidden criticism. This program is under way in the most widespread exotic tree genus and is also the most advanced and successful. In 1984, Leopoldo G. Brandao, Edgard Campinhos Jr., Ney M. Dos Santos, and Yara K. Ikemori were awarded the Marcus Wallenberg Prize “for their pioneering work leading to significant scientific and technological breakthroughs in developing commercial forests based on cloned Eucalyptus. Their methods have stimulated world-wide emulation. The high productivities achieved will reduce pressures on natural forests.”

13.11 Recommended Literature The book by Evans and Turnbull (2004) covers all aspects of tropical tree plantations. It represents comprehensive information on this subject and in Part III, devoted to plantation silviculture, gives answers related to genetics that can only be touched on in the present publication. Subjects related to genetics are also covered in the book by Zobel et al. (1987).

235

Conservation of Genetic Resources in Tropical Forest Trees

There is ample evidence that genetic impoverishment of many tree species has occurred and is occurring as a result of the destruction of forest ecosystems of which these species are a part. Furthermore, it is certain that in many countries forest genetics and tree improvement programs will not be operative in time and on a sufficient scale to insure the preservation in clone banks and seed orchards of an adequate sample of the spectrum of variation within the important tree species. Numerous foresters are aware of this situation, and of the need for the conservation of forest gene resources by insurance of the survival of the forest ecosystems in which the species are found, and by the establishment of seed banks, clone banks, and seed orchards. Furthermore, united action is occurring at an international level that has a direct bearing on this problem (p. 274 in Stern and Roche 1974).

14.1 Introduction About one quarter of the earth’s land surface has been converted to food production. Only 10% of the original landscapes is estimated to have remained unchanged. This condition and other forms of human impact on plant and animal populations have led to the conservation imperative accepted by ever-more people. The importance of the conservation of biological variation is undisputed. Genetic variation is a crucial component of biological diversity, since it is a necessary precondition for the continuation of biological evolution. Furthermore, genetic variation within populations is a preconditon for heterozygosity, which according to empirical studies is correlated to fitness at the population level. Losses of heterozygostity have a deleterious effect on fitness (Reed and Frankham 2003). Although this correlation depends on the methods of assessment used and is only moderately strict, it can be considered a second argument for the commitment to genetic conservation as maintaining the basis of both the long-term stability and the short-term productivity of forest

14

238

C HAPTER 14 Conservation of Genetic Resources in Tropical Forest Trees

ecosystems. Genetic conservation is clearly a strong contributor to the conservation of general biodiversity and according to Altukhov (2006) implies preserving genetic diversity of population systems that still exist but are degraded by utilization (e.g., logging) and artificial propagation (nonexhausting management of natural resources), restoring systems whose structures have already been damaged, and creating new population systems in regions with appropriate natural, historical, and economic conditions. It is self-evident that conservation genetics has a population genetics basis. Practical recommendations for the conservation of genetic resources are based on the understanding that only certain types of genetic information qualify as a genetic resource. A gene resource is defined as the biological material either known to or expected to contain either specific or extensively variable genetic information (Ziehe et al. 1989). Therefore, the ultimate target of genetic conservation is neither an individual tree nor a taxonomic unit but the variation of nucleotide sequences in a genetic resource. Genetic conservation is understood as the preservation of genetic resources in a condition allowing for their regeneration. The need for the conservation of plant genetic resources was first recognized for agricultural species. The Russian plant geneticist N.I. Vavilov (1887–1943) identified centers of diversity possessing a wealth of varieties (genes) of cultivated plants during the 1920s and 1930s (Vavilov 1992, 1996; Altukhov 2006). Tropical forests are centers of diversity of wild species. Namkoong et al. (2002) and Boyle (2000) analyzed the rationale of conserving forest genetic resources and developed a system of making pertinent decisions. The discussion for a conservation strategy in this chapter is based on the sequence of operations as suggested by Ziehe et al. (1989). The concept has been previously presented by Finkeldey and Hattemer (1993), Hattemer (1995, 1997), and Finkeldey (1996). There exists a wide range of motivations for conserving forest genetic resources. Some authors committed to resource conservation of widespread forest trees of the northern temperate zone feel it is their general responsibility to maintain variable and adaptable tree populations. To other authors actively involved in the domestication of tropical forest trees, resource conservation is primarily an integral element of their efforts in domestication (Leakey and Newton 1994b). Domestication of trees implies that they are assigned higher value. Different genetic systems of trees and different goals pursued in different parts of the world call for a variety of measures to be taken. Common species are usually not endangered. Many rare (endemic) species are threatened, particularly those occurring with a unique population. An increasing number of tropical forest tree species are at the fringe of extinction. Their loss would be final, since extinction is forever.

14.2 Development of Tree-Conservation Genetics

Once certain priorities (Sect. 14.3) have been agreed upon, a conservation project for a given tree species starts with the clear identification of the objectives (Sect. 14.4). Genetic resources are subsequently selected mainly on the basis of the available knowledge of spatial patterns of genetic variation (Sect. 14.5). The choice of the method of physical preservation of the genetic information in the selected organisms (Sect. 14.6) is related to the final step, the regeneration of the resource (Sect. 14.7).

14.2 Development of Tree-Conservation Genetics Since 1960s, programs for the genetic conservation of a few species have been initiated in the framework of planned or ongoing breeding programs. Conservation of forest genetic resources was considered an item of breeding strategies also in more recent publications (Cossalter 1989; Eldridge 1990). However, the significance of conserving genetic resources of tropical forest tree species not (yet) incorporated in breeding programs was also realized (Kemp et al. 1976). Scientific results and practical guidelines for the conservation of forest genetic resources were first summarized in a report edited by Roche (1975). During the 1980s, conservation biology emerged to provide the underlying science needed to slow down the process of extinction by conserving biological diversity at all levels, including the genetic diversity within species. This crisis-oriented discipline (Soulé 1991) builds on the fields of ecology, biogeography, and genetics applied to small and declining populations. The crucial role of tropical forests for the global conservation of biological diversity has been repeatedly pointed out by conservation biologists (Myers 1986). Genetic resources of forest trees are endangered for several reasons, including forest destruction and fragmentation, global climate change, inappropriate management practices, selective logging in natural forests, the conversion of speciesrich forests to plantations of low diversity, and inappropriate methods of breeding. The destruction of tropical forests due to unsustainable management, shifting cultivation, and the conversion to other land uses, for example, for grazing, are regarded as the most serious threat to biological diversity on a global scale. Estimates of the annual loss of tropical forest area are alarmingly high (Chap. 10.8 in Whitmore 1998; FAO 2002). The extinction of tree species as a result of forest destruction is poorly documented (Sayer and Whitmore 1990). The causes of the destruction of tropical forests by humans are manifold, complex, and closely related to the demographic and socioeconomic development of human societies in tropical countries. Measures for the conservation of biological diversity will be successful only if attention is paid to the socioeconomic situation of the local human population as the main target group for

239

240


long-term beneficial effects of conservation programs. Local people must receive compensation for the loss of their utilization rights in protected forests and should possibly benefit from a conservation project on a community basis. Conservation of genetic resources and utilization of forest products are not incompatible. The objective of genetic conservation is to preserve genetic information rather than just any material of a species. The individual carriers of genetic information, the trees, are transient and may be harvested as long as their genetic information itself is conserved. Thus, the establishment of reserves in situ does not rule out the management of the respective forest, including harvest of forest products (Cossalter 1989). Intensive production systems of low biological diversity such as monocultures, including clonal plantations, may also have potential benefits for the conservation of genetic resources. When genetically diverse forests are conserved in a given region, urgently needed products can be produced elsewhere (Gladstone and Ledig 1990). The ability of tree populations to evolve is a foundation stone of sustained forest management (Hattemer and Melchior 1993; Boyle 2000).

14.3 Defining Priorities Owing to their rarity or importance, certain types of genetic information are in more urgent need of conservation than others. Biodiversity conservation targets exist at three independent levels: ecosystems, species, and genes (Sherwin and Moritz 2000). The choice of priority targets for conservation programs is based on the ecological and/or economic importance of populations, species, or species groups and the potential threats to their gene pools and their hierarchical, historically formed structures (p. 371 ff. in Altukhov 2006). Present thinking aims at the conservation of species and concentrates on efforts to conserve as many of the huge number of endemic species in tropical forests as possible (Andersen et al. 1997). A brief description of important and particularly endangered trees and shrubs has been published by the FAO (1986). This list is periodically updated by the Panel of Experts on Forest Genetic Resources of the FAO. In the present text the emphasis is on the conservation of genetic variation within the taxa. Conservation measures for a species are initiated because of its economic and silvicultural importance or its ecological significance as a keystone species for the stability of an ecosystem. In view of the species richness in tropical forests, intensive conservation programs can be conducted only for a certain proportion of species. The intensity of a conservation program must be adjusted to the importance of a species (Namkoong 1986). An intensive genetic management including co-ordinated measures for the evaluation, conservation, and utilization of genetic resources is primarily possible for the commercially important target species. A multiple-

14.4 Conservation Objectives

population breeding strategy combining breeding goals and the ex situ conservation of genetic resources is particularly recommended for plantation tree species of the tropics (Sect. 12.9). The principal objective of a genetic management for the vast number of tree species without immediate commercial importance is the preservation of their genetic variation. This requires an improved understanding of the genetic system of at least some keystone or indicator species. Most rare and threatened species exist as small isolated populations. It is urgent that these species be conserved. Viana et al. (1997) listed some nongenetic indicators of vulnerable species that conservation efforts should be targeted at. Bierregaard et al. (1997) extended this to a brief discussion of important aspects of fragmented tropical ecosystems and relevant research priorities. Species that were once common approach this condition and become increasingly fragmented; therefore, considerate handling of important features of the genetic system and of sufficient population size (Sect. 9.3) is devised for any tree species exposed to human interference (Young et al. 1999; A. Young et al. 2000). That their populations are eventually harbored in small forest fragments might help managers to adopt a holistic approach rather than limit their attention to large forest tracts or simply the most pristine areas (p. 279 in Laurance and Bierregaard 1997b).

14.4 Conservation Objectives The objectives of a genetic conservation program depend on the target species and their ecological and/or economic importance. The objective of any conservation measure should be clear from the beginning (Eriksson et al. 1993). The following objectives are similar to those distinguished by Ledig (1986a). 14.4.1 Objective 1: Preservation of the Potential for Particular Trait Expressions

The preservation of the genetic information controlling superior expressions of economically relevant traits supports breeding projects. It involves uncertainties, since quantitative traits are controlled by multiple gene loci, the expression of which is modified by environmental conditions. During drastic environmental change due to management or climate change, the maintenance of trait expressions themselves is not certain even if the genetic structure of a population is preserved unchanged. In view of the predicted environmental change on a local, regional, and global scale, the universal significance of this objective is therefore questionable, although the objective is widely practiced.

241

242


14.4.2 Objective 2: Preservation of Maximum Variation

The conservation of genetic multiplicity is the main objective of gene conservation of agricultural crop plants in gene banks. Even a single copy (allele) at a gene locus is of potential value for breeding. Genetic diversity is considered a value in itself. An important allele can be incorporated into breeding populations by a variety of mating designs and can thus be multiplied indefinitely. In its strict sense the term “gene conservation” applies to conserving a maximum number of alleles in a genetic resource. According to Marshall and Brown (1975), genetic conservationists pursuing this goal are primarily interested in preserving at least one copy of each of the different alleles of the target species and concentrate on the preservation of maximum genetic multiplicity in the totality of the genome comprising genes for both production and adaptation. The preservation of specific rare or even unique genes is expected to gain further importance in view of molecular methods for the transfer of single genes. The transfer of single genes from widely separated taxa to trees has become routine in modern molecular genetics laboratories (Rautner 2001). For plant species, including forest trees, that are not as intensely cultivated as agricultural crop plants the preservation of a maximum number of alleles is less important. Very rare or even unique alleles do not enhance the capacity of populations to survive in a particular environment or after rapid environmental change (Finkeldey 1993). Furthermore, methods for the incorporation of single rare alleles into breeding populations of forest trees will hardly become available in the foreseeable future. A prominent example of this objective is genetic incompatibility conditioned by an S-locus (Young et al. 1999; A.G. Young et al. 2000). Allele loss at an S-locus represents an immediate threat to population viability. For instance, assuming random pollination in a population possessing a gametophytic system of incompatibility (Sect. 6.5.1) with n equally frequent alleles, a fraction of 2/n of the pollen is expected to be eliminated in the styles. The term “self ”incompatibility, is misleading, because also a fraction of cross-pollen gets lost and for small n this fraction may prevail over the pollen involved in selfpollination. This means severe mate limitation and a fertility reduction if n is small; therefore, the conservation of a maximum of variation is devised. There exists experimental evidence that incompatibility as a postpollination event occurs also in tropical forest trees (Boshier 2000; Sect. 6.5.1). Since it takes much effort and time to identify its genetic control and to survey variation of populations at the S-locus, one must trust that the larger the population, the more S-alleles may be encountered.

14.5 Selection of Genetic Resources

14.4.3 Objective 3: Preservation of Adaptability

Trees experience large variation of environmental conditions both in time and in space. In consequence, forest trees must tolerate high environmental variation and forest tree populations must be able to produce a large number of genotypes capable of survival and reproduction under various environmental conditions. Reduced variation may compromise the evolutionary potential, particularly the ability of a population to respond to changing selection pressure. Since forestry is a typical low-input production system, the physical, chemical, and biological environment of forest trees cannot be controlled as strictly as in the more intensive agricultural production systems. Genetic variation is a prerequisite of processes of evolutionary adaptation. The main ecological function of genetic variation is the ability of populations to respond to changed and unpredictable environmental conditions with adaptive changes of their genetic structures (Sect. 7.2). The preservation of adaptability or adaptive capacity has been pointed out as the main objective for the conservation of forest genetic resources (Gregorius 1991). Genetic variation within keystone species is of particular importance for the maintenance of the structure and function of an ecosystem if rapid or strong environmental change occurs or is expected. In contrast to objective 2, the target is the adaptable population rather than the single gene. Hardly much evidence exists on the decline of tree species for lack of adaptive capacity. This, however, is presumably due to the fact that in most documented cases of species extinction human impact was much faster than the gradual loss of species occurring with genetic impoverishment. Genetic variation at adaptively relevant gene loci and the possession of genetic variants with positive adaptive value are still an insurance for a population to persist in a changing world. It must be hoped that at least some of those genetic variants will contribute to adaptation also in the future. The analysis of genetically controlled adaptive phenotypic traits can only be helpful in planning conservation programs pursuing this goal.

14.5 Selection of Genetic Resources The first step in any conservation project is to find out how the diversity we wish to conserve is distributed in space (Bawa and Krugman 1990). We can but preserve adequate samples of the total diversity and ecosystems, including both centers of diversity and endemites. Since they run the risk of remaining

243

244


undetected in our sampling, it is of fundamental importance to find populations most worthy of conservation. This means giving priority to natural populations. Populations should not be regarded as a genetic resource if their origin is uncertain or if genetic bottlenecks that happened during the recent past are inferred. If the observed disturbance had little or no effect on variation, they may be considered equivalent (Soonhuae et al. 1995). The selection of genetic resources of a species must also consider the protection of the respective ecosystem. Unfortunately, nature reserves receiving high-priority protection are not necessarily established where the populations are the most desirable from the conservation point of view. Species-rich sites possess advantages for species conservation because of saving on costs, but are not necessarily adequate for genetic resource conservation. It is not known whether geographic regions with high species diversity can also be expected to host much genetic diversity within those many species. Pursuing a certain conservation objective refers to particular genetic information that is preferably to be conserved. It is involved in the expression of superior phenotypes in economically relevant quantitative traits (objective 1), is particularly diverse (objective 2), or is important for the preservation of evolutionary adaptability (objective 3). Some knowledge of spatial patterns of genetic variation is required for selecting genetic resources on a biologically sound basis. For instance, peripheral populations exposed to different ends of the ecological amplitude of the species range under certain conditions possess conservation value when pursuing objective 3 (Lesica and Allendorf 1995). Genetic data help to make better use of the available populations by maximizing the evolutionary-response potential of a set of genetic resources. Thus, genetic inventories are essential components of conservation strategies. A nongenetic approach by Burgman et al. (2001) allows target areas and their size to be identified on the basis of types and degrees of threat. It takes much less effort and is highly suitable if urgent decisions are required. Spatial patterns of genetic variation are complex and cannot be described by any of the previously mentioned methods alone. The combination of several methods for the assessment of variation patterns is recommended for the selection of genetic resources. If the objective is the preservation of superior trait expressions (objective 1), hints on particularly valuable genetic resources are mainly obtained from field trials. Genetic marker studies and possibly early testing are particularly recommended for capturing maximum variation or evolutionary adaptability (objectives 2 and 3; p. 154 in Finkeldey 1993). A genetic resource must justify later efforts spent on its conservation. Choosing arbitrary populations or stands for their incidental phenotypic superiority in volume yield or stem form (as in approval for the production of selected reproductive material) may be common practice in many countries but does not necessarily serve the purpose of genetic conservation. The phenotypic


appearance of stands or their variation in phenotypic traits can at best be the basis for the selection of genetic resources for pursuing objective 1. The close connection to breeding goals is evident. However, the success of this selection criterion is doubtful unless the degree of genetic control of the observed traits is known. The selection mainly or even exclusively by phenotypic traits is not suitable for conservation objectives 2 and 3. In most instances the choice among several candidate collectives requires our decision. In some extremely rare species only one option of conservation exists, i.e., the rescue of what there is. Genetics and demography are then challenged (Schwartz et al. 2000; Yates and Broadhurst 2002). No sampling is relevant in these cases. It becomes very important to keep track of identities and consanguinities (Krauss et al. 2002) during the process of physical conservation. This is particularly true with tree species that have gone extinct in the wild and survive only as some ex situ resources. Remnants of tree populations spared because of poor site quality for agricultural use no longer represent the species as such (Brown and Hardner 2000; Young and Boyle 2000). The identity of the species in a genetic sense is likely to be lost under these conditions. 14.5.1 Inventory of Genetic Marker Loci Most efforts towards defining evolutionarily significant units have focused on genetic distance measures based on markers to determine when populations are sufficiently distinct to warrant separate conservation. These methods, however, work poorly for many tree species, as they are based on selectively neutral markers and thus reflect mutation, genetic drift and migration, but not adaptive differences (Namkoong et al. 2000).

It has become common practice to assess variation at numerous markers and average those estimates for making comparisons between populations that might be eligible as gene resources because of their variation. Inventories of molecular genetic marker loci provide some insight into mutation events (Vornam et al. 2004), gene flow (Chap. 5), the reproduction system (Chaps. 4, 6), metapopulation structure, as well as past and present population size (Chap. 9). Many molecular markers are unlikely to be of any relevance for fitness. Past bottlenecks can be traced (Example 3.1), since the risk of allele loss is the same regardless of adaptive significance of the respective alleles. Although there is no strict relationship between the size of a population and its genetic variation (Sect. 9.3), the size of a population is still a crucial selection criterion. Very small populations have to be assumed genetically depauperate. Neutral markers are not the target of selection and may therefore not be considered surrogates of the traits in which variation is adaptive. They are not

245

246


suitable tools for the detection of a geographic genetic pattern indicating the selection response of tree populations; therefore, genetic variation of adaptive significance may not be appropriately reflected at neutral marker loci. When discussing the properties of alloenzyme loci, Finkeldey and Mátyás (1999) found that even though some of these gene loci under certain environmental conditions possess relevance for fitness (Müller-Starck 1985; Ziehe et al. 1999; Geburek 2000), they are still unsuitable for directly inferring the adaptive capacity of populations and species. Reed and Frankham (2001) consider molecular markers including alloenzyme loci unsuitable for inference on adaptive variation and differentiation (see also Navarro et al. 2005 for an example in Cedrela odorata and Yang et al. 1996 for an example in a temperate pine species). Two main objectives are evident with regard to the application of genetic marker studies for the selection of genetic resources: 1. Centers of genetic diversity or populations or regions with large amounts of genetic variation can be identified. The identification of a center of genetic diversity of Acacia auriculiformis in Papua New Guinea (Wickneswari and Norwati 1993) was described in Example 3.1. 2. It is possible to identify populations containing otherwise rare or even unique alleles in high frequency. Also the occurrence of localized common alleles points towards valuable genetic resources (Brown 1978). Genetic variation within populations and their differentiation has long been observed most easily and reliably at alloenzyme loci. Their limited number is a disadvantage for the selection of genetic resources. Many surveys have shown remarkably low amounts of genetic differentiation among populations at these loci (Sect. 3.4), although genetic differences in quantitative, adaptive traits is suggested by the results of field trials (Chap. 11). These contrasting patterns of genetic diversity are frequently explained by the presumed selective neutrality of variation at all alloenzyme loci (Muona 1989). Recently developed molecular tools allow the study of more types of genetic markers (Sect. 2.3.1). Not all types of molecular markers are useful for the selection of genetic resources (Szmidt and Wang 2000). Essential advantages are access to large numbers of gene loci, the potential use of different markers with different modes of inheritance (uniparental, biparental), and different degrees of selective relevance. Approaches to relate variation at molecular markers to adaptive variation patterns were mentioned in Sect. 7.2. In view of the existing scarcity of available experimental data, an assessment of the general potential of molecular genetics methods for the selection of forest genetic resources would be premature. A very large number of parameters have been developed that are suitable for condensing the information gained from genetic inventories (Sect. 3.3).


Different parameters express different properties of the variation in populations and their differentiation at a gene locus. None is alone suitable for analysis, as ongoing research shows. Petit et al. (1998) have discussed some of these parameters in the context of selecting populations as genetic resources. Unfortunately, existing software tends to give preference to a certain type of marker only. Of importance is also the way of combining the data from several markers for statements on the genome. Although neutral markers are unlikely to be involved in future processes of adaptation, they are appropriate for inferring the recent evolutionary past of populations, which in turn represents key information on their adaptive capacity. Just as in seed collection for provenance experiments, seed samples for assessing variation must not only be large enough but must be collected in a judicious way (Sect. 11.4.3). The same applies to sampling adult trees themselves for genotyping (Murillo Gamboa 1997) and to collecting seeds for conservation stands (Sect. 14.6.2). Kim et al. (1994) devised a method of designing sample sizes by finding the point where the cumulative frequency curves of detected genic variants flatten out. Favoring allozyme loci, Brown and Hardner (2000) derived appropriate sample sizes that are relevant under various conditions. These authors also stress the need for sampling a relatively large number of sites that are distributed according to ecological conditions of the range and the local frequencies of a species. If seed is analyzed, the results of repeated sampling are much more reliable. 14.5.2 Inventory of Adaptive Trait Expressions and Adaptive Markers

Phenotypic traits are expressions of genetic information at the level of the transcribed genes. Phenotypic variation deliberately excludes variety at lower levels of organization, such as within DNA base sequences where these differences have no effect on phenotype. The observation of phenotypic traits characterizing the health status, growth, and reproduction of a population in a field experiment allows us to draw conclusions on its adaptedness to the given environment (Sect. 7.2). The adaptive potential, i.e., the spectrum of environments that a population is able to get adapted to, can be assessed in a series of field trials. In both situations a wellperforming population is usually considered adapted or adaptable, respectively, although adaptation may be of a purely physiological nature. Strictly speaking, inference on adaptedness in environments not included in the experimental series cannot be made. Neither can adaptability to evolutionarily completely novel environmental factors be foreseen. Nevertheless, populations performing

247

248


well in experimental series are inferred to possess not only advantageous genetic variants at relevant gene loci but possibly also enough variants that may locally function as a genetic load but are advantageous under changed environmental conditions. Several options exist for assessing the genetic controlledness of phenotypic traits (Sect. 12.3). Populations become prime targets for genetic conservation measures if their superior average expressions of economically relevant traits has a genetic basis. Traditional field trials in forestry are time-consuming. Results from planned or recently established trials are not available for the selection of genetic resources if the initiation of conservation measures is urgent. Early tests in the greenhouse, climate chamber, laboratory, or nursery are conducted to observe quantitative variation of seedlings. The conditions can be modified as required. Environmental variation can be minimized or even controlled. These experiments usually take only several months. It is usually impossible to identify the gene loci controlling a quantitative trait, their mode of gene action, and their genetic structures; therefore, only very few marker genes with relevance for survival or reproduction have so far been analyzed (Sherwin and Moritz 2000). Some authors consider it possibly misleading to use data on neutral genetic variation as the exclusive basis of decisions on selecting populations as gene resources (Bekessy et al. 2003). The following case studies support this view: 1. Larsen (1986) observed quantitative variation patterns of seedlings of European silver fir (Abies alba) provenances and identified a center of genetic diversity in the southernmost region of the distribution range in Calabria in southern Italy. The assumed refuge population could not go on postglacial northward migration and was therefore not forced through a bottleneck when remigrating around the Alps that run from east to west. This hypothesis was later confirmed by Bergmann et al. (1990) by genetic surveys. Since Calabrian firs were never confronted with the climate even at medium elevations in central Europe, they suffer from frost when transferred. However, apart from the large variation at alloenzyme loci, they have very valuable properties (Larsen and Mekic´ 1991; Geburek 2000) even when they are planted far north of the natural distribution range of the species. Their consideration as a gene resource would be based both on their history and on laboratory tests. 2. Bekessy et al. (2003) studied the variation of random amplified polymorphic DNA (RAPD) markers and two phenotypic traits related to drought resistance (root mass ratio and isotyope discriminating) in Araucaria araucana, a conifer endemic in the zone of prevailing westerly winds in South America. Its distribution covers a wide elevational range from 600 to 2,000 m. Annual precipitation decreases from 4,000 mm west of the Andes in


Chile to less than 200 mm in their wind shade in Argentina. In a greenhouse, seedlings of nine study populations showed significant variation and differentiation that was uncorrelated between markers and quantitative traits. The populations east and west of the Andean range clustered together only in the quantitative traits, whereas clustering based on RAPD data indicated grouping of populations across the Andes. It is fairly obvious that the eastern populations had acquired higher water use efficiency regulated by the selective rainfall gradient. However, only the level of differentiation in RAPDs was significantly correlated to geographic distance (r = 0.56), indicating that this was shaped through limited gene flow and genetic drift. Once an adaptive trait has been identified, it may serve as a target variable in the selection of populations. However, not just the variation (including the genetic load) alone is crucial but also the possession of favorable variants in notable frequencies that may increase the chance for persistence of the populations in the foreseeable future. 3. Pastorino et al. (2004) studied Austrocedrus chilensis, another dioecious South American conifer occurring a little further south. Several hundred trees from 15 Argentinean populations sampled along three east – west transects were surveyed at 12 alloenzyme loci, half of which were virtually monomorphic. The populations differed moderately with a tendency towards the distribution of genetic variation along a latitudinal gradient, the northerly populations being more variable. The most variable were found in the steppe. This condition may reflect the effect of remigration from several glacial refuges in more northerly locations. Again, genetic and geographic distances ranging from 10 to 450 km were significantly correlated (r = 0.47). Although mean annual precipitation ranged from 2,500 mm in the Andes to 330 mm in the steppe, this was not reflected by the geographic variation pattern. The selection pressure exerted by the steep precipitation gradient was not paralleled by alloenzyme variation. In other instances such a selection pressure might not be quite as obvious. 4. In contrast to the two South American conifers possessing a small range and more or less patchy distribution, Eucalyptus delegatensis (Sect. 11.5) is widespread in southeastern Australia and Tasmania and forms many large populations. Garnier-Géré and Ades (2001) found a close correlation between diameter growth and survival in this species and studied geographic variation of this adaptive trait in a comprehensive provenance experiment. The variation between populations and their interaction with seven experimental locations could be explained by various properties of the physical environment at the origins and biological qualities of the respective forest communities in analyses of covariance (factorial regression). Data on solar radiation allowed adaptedness and adaptability to the experimental sites to be predicted with high degrees of determination.

249

250


The variations in temperature and rainfall were typically better predictors of adaptability than the respective averages. The advantage of analyzing geographic patterns is the predictability of the conditions at nonsampled sites. Adaptedness clearly means adaptedness to past selection pressure. In summary, it can be stated that genetic marker variation provides insight into the genetic system, the Quarternary past of populations, and their centers of variation. This indirect information of adaptability hardly possesses a causal association with environmental variation but still makes a valuable contribution to the development of conservation strategies and might be even more useful in combination with data on adaptive traits derived from experiments done in judiciously chosen environments (Eriksson et al. 1993; Eriksson 1995; Ennos 1996). Inventories of the variation at marker gene loci should then be made if objective 2 deserves the highest consideration. Inventories of the variation in adaptive phenotypic traits should be made if objective 3 is to be pursued (McKay and Latta 2002). The integration of genetic and phenotypic information takes time and effort; therefore, intensive studies have to be confined to species of prime ecological or/and economic relevance. Results may be transferred to less important species with similar life-history traits. Immediate action for the protection of the genetic resources of a species often becomes urgent in the absence of any data on spatial patterns of genetic variation and without the possibility to conduct any of the experiments outlined previously. This situation applies to many lesser-known species in tropical forests. The conservation of at least one autochthonous population in each of the main distribution areas of a species is recommended in this case. Populations growing under both optimum and extreme environmental conditions should be preserved. Both the geographic center of the species distribution and marginal populations should be represented in order to conserve genetic information of populations that are possibly differentiated from each other at adaptive loci. Once a certain population has been found to be worthy of conservation, it must have a sufficient size to be maintained over generations.

14.6 Conservation Methods Appropriate methods used for conservation are required to preserve genetically variable populations. The survival of particular individuals is usually of minor concern.

14.6 Conservation Methods

Genetic resources can be preserved in situ in natural, autochthonous populations. Ex situ conservation applies to all methods requiring the transfer of reproductive material to preservation sites outside the natural environment. Another fundamental distinction of conservation measures refers to the temporal dynamics of the genetic structures of a resource. Methods that allow for changes of genetic structures or that even promote adaptive changes of genetic structures in response to newly experienced or modified environmental conditions are called dynamic. The in situ preservation of genetic resources is by definition a dynamic conservation method. Conservation is static, if it is aimed at keeping the genetic structure of the resource unchanged. The ex situ conservation of genetic resources uses both dynamic and static approaches. By monitoring using appropriate markers or trait expressions, it can be found out whether the criteria applied at selection are still fulfilled. If these genetic surveys require too much effort, at least the maintenance of an appropriate reproduction-efficient population size should occasionally be checked. 14.6.1 Dynamic Conservation in Situ

Both conceptual and practical considerations call for an important role of in situ conservation of tropical forest trees because of its dynamic character. This is different from conservation of many agricultural species (Sect. 14.6.3). Their genetic resources are preserved mainly in gene banks and by other static methods. If the environment changes drastically and rapidly, the preservation of evolutionary adaptability can be achieved only by dynamic conservation methods. The dynamics is induced by natural mortality and by the process of natural regeneration. The genetic resources of tropical tree species that depend on interactions with other species at least during particular stages of their life cycle can best be conserved by dynamic methods. Their conservation is usually most easy in situ. For example, the static preservation of many figs (Ficus spp.) would not pose any problems owing to the ease of vegetative propagation. However, static conservation is likely to result in the (local) extinction of the pollinating wasps (Example 7.2). A sexually produced progeny generation cannot be produced in this case and, by definition, the conservation method would fail to preserve a population. Although species interactions in tropical forests are often complex and poorly understood, they are likely to be of much more significance for the conservation of genetic resources in tropical forests than in temperate ecosystems. The protection of certain forested areas is necessary for the in situ conservation of genetic resources of trees. Thus, in situ gene reserves must be incorporated into overall plans for the establishment of nature reserves and other

251

252


protected areas of a region or a country. The enforcement of protection measures and the acceptance by local communities is more difficult and important than the legal declaration of a protected area in most tropical countries. Tree populations in nature reserves, national parks, etc. are not necessarily typical genetic resources in the close sense, since they were not chosen for their genetic variation. These reserves are still important sites for the genetic conservation of tree populations in their natural forest community. They may be considered more or less undisturbed and, last but not least, they are protected against loss. Restrictions on the removal of plants or plant parts may represent a hindrance of the occasional utilization of genetic resources and of silvicultural management. An important advantage of in situ conservation is the possibility to combine the conservation of the genetic resources of a target species with the conservation of other, associated species in the same block or belt of conservation forest (Cossalter 1989). In situ conservation is the only feasible option for the conservation of genetic resources of species of minor silvicultural importance with unknown reproductive biology. However, protection of a certain area is not necessarily in accordance with the specific requirements for the genetic conservation of a target species. Silvicultural management may be needed to direct the dynamics of the ecosystem or to counteract human influence. Management techniques are as simple as the establishment of buffer zones and firebreaks and the removal of competing vegetation, or more laborious, such as the promotion of regeneration by sowing or planting locally harvested material. An increasing number of in situ resources are located in somehow managed landscapes, which adds one more dimension to the complexity involved in maintaining resources (Boyle 2000). The presence of humans in forest reserves may have positive effects on protection (Schwartzman et al. 2000). Forest genetic resources have nothing in common with museums. Many tree species of tropical forests experience strong fluctuations of their population size or density in an area of given size. This applies primarily to species confined to particular successional phases. Local extinction is a natural process for species characterized by a metapopulation structure (Sect. 9.2). However, the fragmentation and isolation of protected areas may prevent their recolonization by many species (Ashton 1981). Human intervention by silvicultural management such as vine control and enrichment plantings is unavoidable in order to ensure the survival of populations of the target species and thus the conservation of the genetic resource in many small gene reserves. It has still to be found out which interventions are incompatible with the conservation function of forest (Putz et al. 2001). Complexity induced by allochthonous populations within reach of the pollen transfer distance of gene resources (Ennos 1996; Lindgren and Lindgren 1996; Leinemann and Hattemer 2006) raises problems. If the resource is large


enough, the delineation of buffer zones helps to prevent undue contamination in natural regeneration. A small resource may be rescued by vegetative propagation (Sect. 14.6.4). 14.6.2 Dynamic Conservation ex Situ in Man-Made Forests

Dynamic ex situ conservation of genetic resources is the preservation of populations in man-made forests established outside their natural habitat with generative planting stock. It is an important complementary method (Li et al. 2002) for the conservation of genetic resources primarily of plantation species (objectives 1 and 3). In ex situ conservation, more management options may exist as to size, density, and design. The same principles as described in Sect. 9.3 apply to ex situ conservation. The mating system and gene flow are influenced by the regulation of population density, stand structure, and the spatial arrangement of the trees at planting. Planting stock from several populations of a region may be combined to a single ex situ conservation stand or kept separate. The establishment of conservation stands under different environmental conditions eventually leads to adaptive changes of genetic structures. Changes including losses of genetic variation due to genetic drift can be minimized by maintaining large effective population sizes in conservation stands. Provenance resource stands can be regarded as conservation stands. They can be combined with other forests for the production of reproductive material or converted to seed production areas (Guldager 1975). In principle, the dynamic ex situ conservation of genetic resources is compatible with breeding. Conservation of genetic resources is mainly achieved in base populations. Since both breeding and propagation populations are selected for their own superior phenotype or that of their progeny, they should not be considered genetic resources for the preservation of the evolutionary adaptability of a species. A nonhierarchical multiple-population breeding strategy (Sect. 12.9) coincides with the establishment and management of ex situ conservation stands (p. 143 ff. in Finkeldey 1993). Genetically differentiated populations are established or maintained in different environments with the aim of enhancing their adaptive differentiation. Variable selective regimes with both regard to selection intensity and breeding goals are applied. Directional selection and genetic drift potentially lead to a loss of genetic variation in single populations; however, no or only slight losses of genetic multiplicity at adaptive loci are expected within the overall resource composed of all populations. Genetic monitoring is recommended in order to assess the loss of genetic variation within populations and to take appropriate countermeasures by promoting gene flow through the controlled exchange of planting stock.

253

254


Genetic variation within populations can be maintained by a moderate level of gene flow and/or migration among genetically differentiated, semi-isolated subpopulations, thus adding dynamic processes involving several subpopulations. The evolutionary adaptability of the populations comprising a complex of ex situ conservation stands can be maintained by limited gene flow even if the sizes of the subpopulations are small. Thus, population sizes of single subpopulations may be chosen to prevent excessive inbreeding and very strong drift. Both colonization processes and gene flow by pollen in artificial metapopulations might include still-existing populations. It depends on the respective species to what extent this system can be fragmented. In any event, populations are ephemeral. The natural regeneration of these resources is difficult unless they are colonized by a disperser fauna providing for mating contact and for mixing genotypes in the subsequent generation. Conservation forests have a long-term perspective only if the system of reproduction becomes effective as in forests comprising several complementary species of trees and animals. They require more than just minimal management. On the basis of a cluster analysis of larger populations of E. albens at several alloenzyme loci, Prober and Brown (1994) devised a conservation strategy for a species without major geographic differentiation. They recommended a wide geographic spread of reserved populations scattered across the region and their augmentation with conservation stands for maximizing overall genetic diversity. The seed for these stands was to be collected from at least ten trees in a large local population with at least 500 reproductive individuals. This number appears to be low (Sect. 9.3). The efficiency of different sampling strategies in two small and one large population of Pinus tecunumanii was assessed by Dvorak et al. (1999). They concluded that seeds from ten to 20 seed trees are sufficient to sample all alleles with frequencies of more than 5% in each population. It is true that under favorable conditions of pollen flow these few seed trees sample most of the alleles in the population. Since the seed trees contribute at least half of the genetic information of the seed, their low number might imply an undue shift in allele frequencies. Practical advantages of ex situ conservation stands are not limited to the options of a genetic management as outlined earlier. The location for the establishment of ex situ conservation forests can be chosen freely. This allows ex situ conservation stands to be protected more easily and efficiently as compared with in situ gene reserves. The decision about planting either a single large or several small (SLOSS) stands ex situ is largely academic (Hattemer 2006; Sect. 9.3). Forest genetic resources can only be established where appropriate land is available. Consequently there is little room for this type of planning. The distribution of given material over several sites has the advantage in view of safety of the resource. The emphasis must be on the larger of those copies. However, if some


planning can be made, both the size and the design of genetic resources can be chosen. It has to be kept in mind that the establishment of several small resources is related to habitat fragmentation that might enhance the risk of losing the resource (Wilcox and Murphy 1985); therefore, the issue is not splitting up a given resource but rather establishing several of them in only partial isolation. Long-term conservation focuses on persistence at spatial scales larger than that of the single population (Thrall et al. 2000). As mentioned in Sect. 9.3, a system of interconnected populations of moderate to large size has an optimal chance of persistence. Persistence of a resource over generations must account for other tree species providing a continuous food basis for pollen and seed dispersers, so that natural regeneration works and currently unoccupied habitat fragments in the vicinity may be colonized. Thrall et al. (2000) reviewed both the theory of metapopulations and its application to genetic conservation. 14.6.3 Conservation of Seeds in Gene Banks

Seeds can be transported easily and large numbers of genotypes can be stored in a small facility. Their long-term storage in gene banks is the most important static method for the conservation of crop plant genetic resources. The conservation of seeds is static inasmuch as the genetic structures of the resource do not change during storage. The seed physiology of a tree species sets limits to the potential of long-term storage of seeds in gene banks. Long-term storage of forest seeds as a genetic resource is often combined with the establishment of forest seed centers in tropical countries. The main task of seed centers is the procurement and provision of seeds for afforestation (Albrecht 1993). A concentration of tasks related to forest tree seed handling and storage is meaningful in view of the limited personal and financial resources available. Long-term storage is impossible for recalcitrant seeds. The seed physiology and optimum storage conditions of many tropical tree species are unknown. In comparison with the lifetime of trees or their rotation periods, the maximum duration of seed storage is short even for species with orthodox seeds. The seeds of annual agricultural crop plants are periodically sown out and the seed of the new generation is put back into the gene bank. This cannot be done for trees, so the germinability decays, which induces genetic change. Since the main objective of a conservation programs for forest trees is the preservation of evolutionary adaptability, static conservation methods are less suitable than dynamic methods. Establishment and operation of a gene bank are costly. Last but not least, gene banks are sensitive to short-term interruptions of their operation, such as failure of the electricity network, which is frequent in some areas.

255

256


The long-term storage of seeds is a complementary method to dynamic conservation of forest genetic resources for some important tree species. The main advantage is the preservation of a large number of genotypes in a fairly small facility. 14.6.4 Vegetative Propagation for the Conservation of Forest Genetic Resources

Vegetative propagation allows us to identically reproduce and multiply genotypes. It is a method to rescue moderate numbers of remnant trees of rare species in danger of extinction and thus to preserve moderate numbers of individual genotypes. It is therefore not necessarily suitable for the conservation of forest genetic resources in general. Since in asexual propagation recombination is excluded, the number of conserved genotypes is usually much lower than in sexually reproducing populations. Nevertheless, vegetative propagation is of great interest in domestication and represents an important technique for conservation programs for tropical forest trees (Longman 1976; Ladipo et al. 1991; Smits 1993; Khuspe et al. 1994; Leakey et al. 1994; Newton et al. 1994; Smits et al. 1994). Trees and their genotypes become mobile by vegetative propagation. This mobility is of particular importance for emergency measures to protect highly endangered genetic resources (Finkeldey and Hattemer 1993). Many populations or even species are facing immediate extinction in their natural habitat, for example, owing to large-scale forest destruction. Vegetative propagation offers the opportunity to escape extinction and conserve genetic resources by moving genotypes to areas which can more easily be protected than the natural habitat. Disturbance of forests possibly interrupts the mating contact among plants by decreasing population densities below a critical threshold. Vegetative propagation allows artificial populations to be established by planting together propagules from remnant trees scattered over the landscape and provides for their reproduction with tolerable levels of inbreeding. Assembling vegetatively propagated genotypes in clonal conservation orchards is a typical emergency measure to protect genetic resources from immediate extinction (Rotach 2000). Clonal plantations established as part of a breeding program such as clonal seed orchards, clonal archives, and clonal gardens preserve the genetic information of particular phenotypically superior genotypes, although they contribute only little to the preservation of genetic variation and evolutionary adaptability (objective 3). 14.6.5 Biotechnological Methods of Preservation

The distinction between traditional and modern biotechnological methods is somehow arbitrary. Macrovegetative propagation of plants is not regarded as a

14.7 Regeneration of Forest Genetic Resources

biotechnology but tissue culture is a typical biotechnological method. However, the genetic consequences of macrovegetative and microvegetative propagation techniques are very similar. Biotechnological methods have so far been of minor importance for the conservation of tropical forest genetic resources. The application of modern and sophisticated technologies fascinates research workers in many countries. However, better understanding of the spatial and temporal dynamics of genetic structures of the target tree species should have research priority. The application of biotechnological methods such as in vitro culture, slow-growth techniques, and cryopreservation to the physical conservation of tree material is less urgent and can hardly be considered a main focus of forest genetic research (p. 151 ff in FAO 1994). Tissue culture allows us to conserve and vegetatively propagate trees in vitro. It is an important alternative to the conservation of seed in particular for species with recalcitrant seeds. Tissue cultures of many species can be maintained with moderate efforts over long periods of time if kept in suitable conditions, i.e., at low temperature and light and in appropriate media. The long-term in vitro preservation of genotypes is practiced for some agricultural crops such as cassava but hardly for tropical forest trees (Engelmann 1994; p. 31 f. in FAO 1994). A virtually indefinite preservation of numerous plantlets in a small facility is feasible by cryopreservation, i.e., the preservation of biological tissue at a temperature of –196˚C in liquid nitrogen. Embryos, pollen, and other tissue from temperate forest trees were frozen in liquid nitrogen and could be successfully regenerated (Jörgensen 1990; Park and Son 1996). Molecular genetics techniques allow us to isolate, multiply, and conserve fragments of DNA outside of living cells of the respective species. A gene library, gene bank, or genomic bank is a collection of cloned DNA fragments which represent parts or even the complete genome of an organism. The fragments are inserted into plasmids as vectors that are preserved in living cells of a host organism. The host organisms, usually bacteria, are capable of multiplication by cell division. Not only the host DNA, but also the foreign DNA is transmitted during cell division. DNA fragments of any organism can be preserved and amplified indefinitely in genomic banks.

14.7 Regeneration of Forest Genetic Resources A program for the conservation of a genetic resource is frequently considered successful if genetic information has been preserved for a certain time period. However, the main objective of conservation programs for forest genetic resources is not the preservation as such. It is rather the establishment of a

257

258


population that is both adapted to the environmental conditions prevailing during the regeneration of the resource and adaptable to future environmental change. The success of a conservation program can be assessed only after the resource has been regenerated; thus, the regeneration of the resource is an integral part of the conservation strategy and has to be planned during an early stage of a conservation program. This applies, for example, to the natural regeneration of an in situ resource. If static conservation methods have been employed and, in particular, if the resource has been conserved by biotechnological methods, the regeneration is strictly separated from the conservation phase. Genetic resources conserved in situ are regenerated naturally. Natural regeneration depends on numerous ecological, silvicultural, and socioeconomic conditions in tropical forests (see the case study by Neil 1990 on Agathis macrophylla in Vanuatu). If natural regeneration is insufficient, it may be completed by planting stock raised from local seed. Dynamic evolutionary processes are promoted if an ex situ conservation stand is regenerated naturally. After several generations of natural regeneration a land race is likely to develop. Alternatively, the resource can be regenerated by planting or sowing material that was harvested there. It is possible to mix reproductive material from several genetically differentiated ex situ conservation stands derived from the same population in order to simulate migration and to increase genetic multiplicity. Germinability of seeds is obviously required for the regeneration of a resource after storage of material in a gene bank. The planting stock derived from these seeds is used either for the establishment of ex situ conservation stands or for enrichment planting. A general weakness of static conservation methods becomes evident during the regeneration of a resource. No adaptive changes of genetic structures are possible during the storage period. A genetic resource may have lost its adaptedness if drastic and large-scale environmental change occurred during this phase. In this case evolutionary adaptation is necessary during the regeneration. The adaptation process is likely to be accompanied by strong viability selection and consequently a severe reduction of population size. Hence, the survival of the population or at least the establishment of a genetically variable ex situ conservation stand is at risk (Ziehe et al. 1989). Genetically variable progeny is produced in clonal conservation orchards. Seeds harvested there may be used for the establishment of ex situ conservation stands or enrichment planting in logged forests. Too small numbers of clones induce losses of genetic multiplicity due to genetic drift. In anemophilous tree species, stored pollen may be used to support seed formation. If biotechnological methods were used for the physical preservation of material, particular techniques and abilities are required to regenerate complete plants. Regeneration from tissue cultures requires hormonal treatment,


different types of media, and modified culture conditions. The regeneration of material conserved by cryopreservation in liquid nitrogen is even more difficult and requires in most cases an in vitro phase. The regeneration of a complete plant from the DNA fragments of a gene library is likely to be impossible even in the distant future. Molecular genetics methods do not have practical importance for the physical conservation of genetic material. The application of other biotechnological methods is impeded by difficulties to regenerate trees from the conserved material or to establish a genetically variable population. Genetic monitoring at marker loci of a resource is indicated during the regeneration in order to periodically assess genetic structures. Countermeasures are indicated if genetic variation is lost. Example 14.1: Conservation of the Genetic Resources of Pinus merkusii in Thailand On the Southeast Asian mainland, the natural distribution of P. merkusii stretches from Myanmar to Thailand, Laos, Cambodia, and Vietnam. The species is also naturally distributed in Indonesia on Sumatra and in the Philippines on Luzon and Mindoro. Large-scale plantations have been established in Indonesia, mainly on Java, with provenances from Sumatra. The main distribution area of the species in Thailand is split into three isolated regions in northern, central, and northeastern Thailand. In Thailand natural stands of P. merkusii (Fig. 14.1) are affected by forest destruction, illegal logging, and resin tapping. P. merkusii is regarded as a potential plantation species in Thailand and other countries on the Southeast Asian mainland in particular for resin production. However, plantation establishment is difficult owing to the occurrence of the “grass stage” (Fig. 14.2), a long-lasting growth depression of seedlings as an adaptation to periodic ground fires (Fig. 14.3). A provenance trial revealed fast growth, good stem form, and a comparatively short duration of the grass stage in provenances from northeastern Thailand (Sect. 11.4.3). The average level of genetic variation at isozyme gene loci is low for the species in Thailand (He = 0.058; Table 3.1). Extraordinarily low outcrossing rates were observed in some stands of P. merkusii in Thailand (Sect. 6.3.2) and were explained by a combination of low density, overaging, and poorly synchronized flowering (Changtragoon and Finkeldey 1995a). Genetic resources of P. merkusii are primarily conserved in two in situ conservation areas in northeastern Thailand since these provenances appear to be most suitable for seed harvest for plantation establishment. Logging is strictly forbidden in the gene reserves, firebreaks were laid out, and the reserves are frequently patrolled by the local forest service. However, the protection of the two in situ conservation areas is extremely difficult owing to the pressure of

259

260


Fig. 14.1. Natural Pinus merkusii forest in northern Thailand. (Photo: R. Finkeldey)

the local population on the few remaining forests. Almost all pines in the gene reserves are severely damaged by illegal resin tapping and/or the removal of parts of the lower stem for the production of charcoal (Fig. 14.4). Ground fires are frequent because the rural population wants to enhance the growth of grasses for forest pasture. Seed production of P. merkusii is poor in the gene reserves owing to overaging, poor health, and low population density. Natural regeneration is almost completely absent in the gene reserves. Young seedlings rarely survive owing to yearly ground fires during the long dry season. In view of the absence of natural regeneration, seeds were harvested and sown in temporary nurseries. Threeyear-old seedlings of local origin were planted in the gene reserves. These plants survive light ground fires (Fig. 14.3).


Pinus kesiya

Pinus merkusii

Fig. 14.2. An approximately 3-year-old seedling of P. merkusii in the grass stage (front). In the background is a seedling of P. kesiya, another pine native to Thailand. (Photo: R. Finkeldey)

The establishment of at least two more in situ gene conservation areas in northern and central Thailand, respectively, is recommended in addition to the already existing gene reserves. Because of the less dense human population, protection of an area is easier in these regions. Spontaneous natural regeneration is abundant in the north. Light felling and other silvicultural techniques can be used to naturally regenerate a gene reserve in northern Thailand. Ex situ conservation measures are important complementary activities to conserve the genetic resources of P. merkusii in Thailand. Seeds cannot be stored long at ambient temperatures and facilities for the long-term storage of seeds are not available. Vegetative propagation by means of cuttings is difficult. The establishment of dynamic ex situ conservation areas is currently the only feasible and meaningful measure in addition to the in situ gene reserves.

261

262


Fig. 14.3. A P. merkusii seedling after a ground fire. The long, fleshy needles are burned, but the terminal bud survived and a new sprout has developed. (Photo: R. Finkeldey)

Ex situ conservation areas should be established in each of the three main distribution areas of P. merkusii in Thailand. Changtragoon and Finkeldey (2000) recommend establishing one ex situ conservation area in each region. Areas should be selected according to the general suitability for the species and the ease of protection. Material should be harvested from at least three different stands in each region. At least ten trees should be harvested in each stand following the general recommendations for seed harvest in provenance trials as


Fig. 14.4. Heavily damaged P. merkusii tree in an in situ gene reserve in Thailand. (Photo: R. Finkeldey)

outlined in Sect. 11.4.3. Seeds from the same region may be bulked, but the reproductive material should not be exchanged among regions. The ex situ conservation stands should be planted as pure forests in high density. Outcrossing is expected to be promoted in the plantations owing to the high population density and strong flowering of young trees after reaching maturity. The double purpose of the plantations will be to conserve the genetic resources of P. merkusii in Thailand and to serve as seed production areas for the production of seeds eventually required for commercial plantings (From Sa-Ardavut et al. 1989; Changtragoon and Finkeldey 1995a, 2000).

263

264


14.8 Programs for the Protection of Forest Genetic Resources National programs for the conservation of forest genetic resources are usually planned and executed jointly by scientists and practical foresters trained in tree breeding. Few researchers are experienced in the application of population genetics and conservation-specific issues in tropical forests. Conservation of forest genetic resources is often mistaken as subordinate elements of tree improvement programs and the complex biological background is overlooked. Apart from tree breeding, the conservation of genetic resources in tropical forests is closely connected to the silvicultural management of forests (e.g., natural regeneration), nature conservation (conservation areas), legal regulations (declaration of protected areas), political and socioeconomic issues (execution of protection measures), and overall development planning. The biological and genetic considerations outlined earlier must be incorporated into these subject matters in order to initiate promising conservation programs. This applies in particular to all planned activities for the in situ conservation of forest genetic resources. The local population is not only directly affected by many conservation measures, but also by the main target group for the long-term beneficial effects of a conservation program; thus, the rural populations must be integrated in programs for the protection of forest genetic resources. Land is needed for in situ and ex situ gene reserves and the utilization of forest products from conservation areas is restricted. The planning of conservation measures must take into account legal and customary rights of forest owners and traditional users. At an international level, several documents from the United Nations Conference on Environment and Development (UNCED) during 1992 (the “Rio Summit”) cover aspects of relevance for the conservation of tropical forest genetic resources. The Forest Principles, the Convention on Biological Diversity, and Agenda 21 are of particular importance in addition to the 27 general principles of the Rio Declaration (Finkeldey 1996). The importance of international scientific co-operation and financial aid for the protection of natural forests in developing countries is stressed in the Forest Principles and the Convention on Biological Diversity. They emphasize the sovereign right of countries to utilize their genetic resources [see Article 8(g)]. The potential use of genetic resources by modern biotechnologies was a controversial issue during the formulation of the Convention of Biological Diversity. Technology transfer to developing countries is mentioned as being of particular importance in this field. Also Chap. 11 (Combating Deforestation), Chap. 15 (Conservation of Biological Diversity), and Chap. 16 (Environmentally Sound Management of


Biotechnology) of Agenda 21 deal with forest genetic resources. The particular role of forests for the conservation of biological diversity is recognized, the significance of the of in situ conservation of forest genetic resources is stressed, the complementary sides of the conservation and utilization of genetic resources are emphasized, and the importance of regional and international co-operation for the conservation of forest genetic resources is highlighted. However, it is still a long way ahead before the ambitious objectives of the “Rio Documents” concerning the conservation and sustainable utilization of forest genetic resources are achieved.

14.9 Recommended Literature A systematic treatment of the principles of genetic conservation is given in the book by Frankel and Soulé (1981). Two books with various contributions on animals and plants, including some on tropical forests, are those edited by Schonewald-Cox et al. (1983) and Soulé (1986). Nowadays, various journals and many books are devoted to genetic conservation. The book edited by A. Young et al. (2000) is devoted to forest trees. In several contributions reference is made to tropical trees. Very recently, Geburek and Turok (2006) have edited a book on forest genetic resource conservation. The emphasis is on European trees but many contributions deal with items of general relevance to genetic resources of trees, including tropical forest trees. The booklet edited by the FAO (1993) gives a general treatment of genetic conservation in tropical forest trees and presents some case studies. Three booklets edited by the International Plant Genetic Resources Institute (Anonymous 2004a–c) describe both general principles and methods of genetic conservation in woody plant species.

265

References

Acheré V, Faivre Rampant P, Pâques LE, Prat D (2004) Chloroplast and mitochondrial molecular tests identify European × Japanese larch hybrids. Theor Appl Genet 108:1643–1649 Ackerly DD, Rankin-de-Merona JM, Rodriguez WA (1990) Tree densities and sex ratios in breeding populations of dioecious Central Amazonian Myristicaceae. J Trop Ecol 6:239–248 Ackerman JD, Mesler MR, Lu KL, Montalvo AM (1982) Food-foraging behaviour of male Euglossini (Hymenoptera: Apidae): vagabonds or trapliners? Biotropica 14:241–248 Adams WT (1992) Gene dispersal within forest tree populations. New For 6:217–240 Adis J (1990) Thirty million arthropod species – too many or too few? J Trop Ecol 6:115–118 Albrecht J (1993) Forest seed handling. In: Pancel L (ed) Tropical forestry handbook, vol 1. Springer, Berlin Heidelberg New York, pp 381–462 Aldrich PR, Hamrick JL (1998) Reproductive dominance of pasture trees in a fragmented tropical forest mosaic. Science 281:103–105 Aldrich PR, Hamrick JL, Chavarriaga P, Kochert G (1998) Microsatellite analysis of demographic genetic structure in fragmented populations of the tropical tree Symphonia globulifera. Mol Ecol 7:933–944 Allard RW (1960) Principles of plant breeding. Wiley, New York Altukhov YP (2006) Intraspecific genetic diversity. Monitoring, conservation, and management. Springer, Berlin Heidelberg New York Altukhov YP, Salmenkova EA (2002) DNA polymorphism in population genetics. Russ J Genet 38:989–1008 Alvarez-Buylla ER, Chaos Á, Piñero D, Garay AA (1996) Demographic genetics of a pioneer tropical tree species: patch dynamics, seed dispersal, and seed banks. Evolution 50:1155–1166 Andersen M, Thornhill A, Koopowitz H (1997) Tropical forest disruption and stochastic biodiversity losses. In: Laurance WF, Bierregaard RO (eds) Tropical forest remnants: ecology, management, and conservation of fragmented communities. University of Chicago Press, Chicago, pp 281–291 Andrianarimisa A, Bachmann L, Ganzhorn JU, Goodman SM, Tomiuk J (2000) Effects of forest fragmentation on genetic variation in endemic understory forest birds in central Madagascar. J Ornithol 141:152–159 Anonymous (2004a) Forest genetic resource conservation and management: vol 1: overview, concepts and some systematic approaches. FAO-FGR-FLD-IPGRI, Rome Anonymous (2004b) Forest genetic resource conservation and management: vol 2: in natural forests and protected areas (in situ). FAO-FGR-FLD-IPGRI, Rome

268

References

Anonymous (2004c) Forest genetic resource conservation and management: vol 3: in plantations and gene banks (ex situ). FAO-FGR-FLD-IPGRI, Rome Apsit VJ, Hamrick JL, Nason JD (2001) Breeding population size of a fragmented population of a Costa Rican dry forest tree species. J Hered 92:415–420 Armstrong WP (1996) Plant sexuality and political correctness. Wayne’s word 4. http://waynesword.palomar.edu/ww0404.htm Arnold RJ, Midgley SJ (1995) Conserving biodiversity – the work of the Australian tree seed centre. Commonw For Rev 74:121–128 Ashton PS (1969) Speciation among tropical forest trees: some deductions in the light of recent evidence. Biol J Linn Soc Lond 1:155–196 Ashton PS (1981) Techniques for the identification and conservation of threatened species in tropical forests. In: Synge H (ed) The biological aspects of rare plant conservation. Wiley, Chichester, pp 155–164 Ashton PS (1982) Dipterocarpaceae. In: van Steenis CGGJ (ed) Flora Malesiana, series 1, Spermatophyta, vol 9, part 2. Nijhoff, The Hague pp 237–552 Ashton PS (1988) Dipterocarp biology as a window to the understanding of tropical forest structure. Annu Rev Ecol Syst 19:347–370 Austerlitz F, Mariette S, Machon N, Gouyon P-H, Godelle B (2000) Effects of colonization processes on genetic diversity: differences between annual plants and tree species. Genetics 154:1309–1321 Austerlitz F, Dick CW, Dutech C, Klein EK, Oddou-Muratorio S, Smouse PE, Sork VL (2004) Using genetic markers to estimate the pollen dispersal curve. Mol Ecol 13:937–954 Avise JC (2000) Phylogeography. The history and formation of species. Harvard University Press, Cambridge Awang K (1994) F/FRED’s species network: results, constraints, achievements In: Finkeldey R, Havmoller P (eds) Proceedings of the regional workshop on species improvement networks (SPINs), Chiang Mai, Thailand, July 1994. RAS/91/004 field document no 19. UNDP/FAO Forest Tree Improvement Project, Los Baños, pp 10–18 Baker HG (1960) Apomixis and polyembryony in Pachira oleaginea (Bombacaceae). Am J Bot 47:296–302 Baker HG (1970) Evolution in the tropics. Biotropica 2:101–111 Ballal SR, Foré SA, Guttman SI (1994) Apparent gene flow and genetic structure of Acer saccarum subpopulations in forest fragments. Can J Bot 72:1311–1315 Bänfer G, Fiala B, Weising K (2004) AFLP analysis of phylogenetic relationships among myrmecophytic species of Macaranga (Euphorbiaceae) and their allies. Plant Syst Evol 249:213–231 Banik RL (1995). Domestication and breeding of bamboos. RAS/91/004 working paper no 9. UNDP/FAO Forest Tree Improvement Project, Los Baños Barlow BA (1958) Tetraploid twins and apomixis in Casuarina nana Sieb. Aust J Bot 7:230–237 Barnes RD, Burley J (1990) Tropical forest genetics at the Oxford Forestry Institute: exploration, evaluation, utilization and conservation of genetic resources. For Ecol Manage 35:159–169 Barnes RD, Mullin LJ (1989) The multiple population breeding strategy in Zimbabwe. Five year results. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling

References

forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 148–158 Barnes RD, Simons ASJ (1994) Selection and breeding to conserve and utilise tropical tree germplasm. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 84–90 Barth FG (1982) Biologie einer Begegnung. Dioe Partnerschaft der Insekten und Blumen. Stuttgart, Deutsche Verlagansanstalt Bass S (2001) Policy inflation, capacity constraints: can criteria and indicators bridge the gap? In: Raison RJ, Brown AG, Flinn DW (eds) Criteria and indicators for sustainable management. CABI, Wallington, pp 19–37 Bawa KS (1973) Chromosome numbers of tree species of a lowland tropical community. J Arnold Arbor 54:422–434 Bawa KS (1974) Breeding systems of tree species of a lowland tropical community. Evolution 28:85–92 Bawa KS (1980) Evolution of dioecy in flowering plants. Annu Rev Ecol Syst 11:15–39 Bawa KS (1990) Plant-pollinator interactions in tropical rain forests. Annu Rev Ecol Syst 21:399–422 Bawa KS, Krugman SL (1990) Reproductive biology and genetics of tropical trees in relation to conservation and management. In: Gomez-Pompa A, Whitmore TC, Hadley M (eds) Rain forest regeneration and management. UNESCO, Paris, and The Parthenon Publishing Group, Carnforth, pp 119–136 Bawa KS, Opler PA (1975) Dioecism in tropical forest trees. Evolution 29:167–179 Becker H (1993) Pflanzenzüchtung. Ulmer, Stuttgart Bekessy SA, Ennos RA, Burgman MA, Newton AC, Ades PK (2003) Neutral DNA markers fail to detect genetic divergence in an ecologically important trait. Biol Conserv 110:267–275 Bennett EL (2001) Timber certification: where is the voice of the biologist? Conserv Biol 15:308–310 Berg CC (1989) Classification and distribution of Ficus. Experientia 45:605–611 Berg EE, Hamrick JL (1997) Quantification of genetic diversity at allozyme loci. Can J For Res 27:415–424 Bergmann F (1974) The genetics of some isoenzyme systems in spruce endosperm (Picea abies). Genetika 6:353–360 Bergmann F, Hattemer HH (1998) Isozymes in forest genetic research. In: Mandal AK, Gibson GL (eds) Forest genetics and tree breeding. CBS, New Delhi, pp 227–238 Bergmann F, Ruetz W (1991) Isozyme genetic variation and heterozygosity in random tree samples and selected orchard clones from the same Norway spruce population. For Ecol Manage 46:39–47 Bergmann F, Gregorius HR, Larsen JB (1990) Levels of genetic variation in European silver fir (Abies alba). Are they related to species’ decline? Genetica 82:1–10 Bermingham E, Dick C (2001) The Inga – newcomer or museum antiquity? Science 293:2214–2216 Bierregaard RO, Lovejoy TE, Kapos V, dos Santos AA, Hutchings RW (1992) The biological dynamics of tropical rain forest fragments. Bioscience 42:859–866 Bierregaard RO, Laurance WF, Sites JW, Lynam AJ, Didham RK, Andersen M, Gascon C, Tocher MD, Smith AP, Viana VM, Lovejoy TE, Sieving KE, Kramer EA, Restrepo C,

269

270

References

Moritz C (1997) Key priorities for the study of fragmented tropical ecosystems. In: Laurence WF, Bierregaard RO (eds) Tropical forest remnants: ecology, management, and conservation of fragmented communities. University of Chicago Press, Chicago, pp 515–525 Billington HL (1991) Effect of population size on genetic variation in a dioecious conifer. Conserv Biol 5:115–119 Bishir J, Roberds JH (1999) On numbers of clones needed for managing risks in clonal forestry. For Genet 6:149–155 Blattner FR, Weising K, Bänfer G, Maschwitz U, Fiala B (2001) Molecular analysis of phylogenetic relationships among myrmecophytic Macaranga species (Euphorbiaceae). Mol Phylogenet Evol 19:331–344 Bolstad PV, Bawa KS (1982) Self incompatibility in Gmelina arborea L. (Verbenaceae). Silvae Genet 31:19–21 Bolstad SB, Kang H, Guries RP, Marty TL (1991) Performance of inter- and intraprovenance crosses of jack pine in central Wisconsin. Silvae Genet 48:261–266 Booth TH, Turnbull JW (1994) Domestication of lesser-known tropical tree species: the Australian experience. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 189–194 Boshier D, Mesén JF (1989) Breeding population of Cordia alliodora in Costa Rica. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 406–407 Boshier DH (2000) Mating systems. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford, pp 63–79 Boshier DH, Chase MR, Bawa KS (1995) Population genetics of Cordia alliodora (Boraginaceae), a neotropical tree. 2. Mating system. Am J Bot 82:476–483 Bouvet JM, Vigneron P (1996) Variance structure in Eucalyptus hybrid populations. Silvae Genet 45:171–177 Boyle TJ (2000) Criteria and indicators for the conservation of genetic diversity. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford, pp 239–251 Boyle TJB (1994) Introduction to the NATO advanced research workshop. In: Boyle TJB, Boyle CEB (eds) Biodiversity, temperate forest ecosystems and global change. Springer, Berlin Heidelberg New York, pp 1–7 Brewbaker JL, Sorensson CT (1994) Domestication of lesser-known species of the genus Leucaena. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources, Her Majesty’s Stationery Office, London, pp 196–204 Brigham CA, Schwartz M (2003) Population viability in plants. Springer, Berlin Heidelberg New York Bronstein JL, McKey D (1989) The fig/pollinator mutualism: a model system for comparative biology. Experientia 45:601–604 Bronstein JL, Gouyon P-H, Kjellberg F, Michaloud G (1990) The ecological consequences of flowering asynchrony in monoecious figs: a simulation study. Ecology 71:2145–2156

References

Brown AHD, Hardner CM (2000) Sampling the gene pools of forest trees for ex situ conservation. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford, pp 185–196 Brown AHD (1978) Isozymes, plant population genetic structure and genetic conservation. Theor Appl Genet 52:145–157 Bruenig EF (1996) Conservation and management of tropical rainforests. CABI, Wallingford Brunner AM, Busov VB, Strauss SH (2004) Poplar genome sequence: functional genomics in an ecologically dominant plant species. Trends Plant Sci 9:49–56 Bryndum K, Hedegart T (1969) Pollination of teak (Tectona grandis L.). Silvae Genet 47:155–158 Buchert GP, Rajora OP, Hood JV, Dancik BP (1997) Effects of harvesting on genetic diversity in old-growth eastern white pine in Ontorio, Canada. Conserv Biol 11:747–758 Burdon RD (1989) When is cloning on an operational scale appropriate? In: Gibson GI, Griffin AR, Matheson, AC (eds), Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 9–27 Burgess IP, Williams ER, Bell JC, Harwood CE, Owen JV (1996) The effect of outcrossing rate on the growth of selected families of Eucalyptus grandis. Silvae Genet 45:97–100 Burgman MA, Possingham H, Lynch AJJ, Keith DA, McCarthy MA, Hopper SD, Drury WL, Passioura JA, Devries RJ (2001) A method for setting the size of plant conservation target areas. Conserv Biol 15:603–616 Burley J (1973) Sources and distribution of seedlots in the CFI international provenance trial of Cedrela odorata (including C. mexicana and C. tubiflora). In: Burley J, Nikles DG (eds) Tropical provenance and progeny research and international cooperation. Commonwealth Forestry Institute, Oxford, pp 234–237 Burley J, Nikles DG (eds) (1972) Selection and breeding to improve some tropical conifers, vol 1. Commonwealth Forestry Institute, Oxford, and Department of Forestry, Brisbane Burley J, Nikles DG (eds) (1973) Selection and breeding to improve some tropical conifers, vol 2. Commonwealth Forestry Institute, Oxford, and Department of Forestry, Brisbane Butcher PA (2004) Genetic mapping in acacias. In: Kumar S, Fladung M (eds) Molecular genetics and breeding of forest trees. Haworth, New York, pp 411–427 Butcher PA, Moran GF (2000) Genetic linkage mapping in Acacia mangium. 2. Development of an integrated map from two outbred pedigrees using RFLP and microsatellite loci. Theor Appl Genet 101:594–605 Butcher PA, Moran GF, Perkins HD (1998) RFLP diversity in the nuclear genome of Acacia mangium. Heredity 81:205–213 Butcher PA, Glaubitz JC, Moran GF (1999) Applications for microsatellite markerks in the domestication and conservation of forest trees. For Genet Resour 27:34–42 Butcher PA, Moran GF, Bell R (2000) Genetic linkage mapping in Acacia mangium. 1. Evaluation of restriction endonucleases, inheritance of RFLP loci and their conservation across species. Theor Appl Genet 100:576–583 Byrne M (2000) Disease threats and the conservation genetics of forest trees. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford, pp 159–166

271

272

References

Caballero A (1994) Developments in the prediction of effective population size. Heredity 73:657–679 Campinhos E (1991) Plantations of fast-growing species for tropical areas. Rev For Fr Hors Ser 5:111–120 Campinhos E (1993a) A Brazilian example of a large-scale forestry plantation in a tropical region: Aracruz. In: Davidson J (ed) Proceedings of the regional symposium on recent advances in mass clonal multiplication of forest trees for plantation programmes. RAS/91/004 field document no 4. UNDP/FAO Forest Tree Improvement Project, Los Baños, pp 46–59 Campinhos E (1993b) The use/place of clonal multiplication in tree breeding and propagation programmes. Prerequisites for success and reasons for failure. In: Davidson J (ed) Proceedings of the regional symposium on recent advances in mass clonal multiplication of forest trees for plantation programmes. RAS/91/004 field document no 4. UNDP/FAO Forest Tree Improvement Project, Los Baños, pp 9–30 Campinhos E (1999) Sustainable plantations of high yield Eucalyptus trees for production of fiber: the Aracruz case. New For 17:129–143 Campinhos E, Ikemori YK (1989) Selection and management of the basic population Eucalyptus grandis and E. urophylla established at Aracruz for the long term breeding programme. In: Gibson GI, Griffin AR, Matheson, AC (eds), Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 169–175 Cao CP, Finkeldey R, Siregar I, Siregar U, Gailing O (2006) Genetic diversity within and among populations of Shorea leprosula Miq. and Shorea parvifolia Dyer (Dipterocarpaceae) in Indonesia detected by AFLPs. Tree Genet Genomics (in press) Cardoso MA, Provan J, Powell W, Ferreira PCG, De Oliveira DE (1998) High genetic differentiation among remnant populations of the endangered Caesalpinia echinata Lam. (Leguminosae – Caesalpinioideae). Mol Ecol 7:601–608 Casasoli M, Derory J, Morera-Dutrey J, Brendel O, Porth J, Guehl JM, Villani F, Kremer A (2006) Comparison of quantitative trait loci for adaptive traits for oak and chestnut based on an expressed sequence tag consensus map. Genetics 172:533–546 Cascante A, Quesada M, Lobo JJ, Fuchs EA (2002) Effects of dry tropical forest fragmentation on the reproductive success and genetic structure of the tree Samanea saman. Conserv Biol 16:137–147 Cavers S, Navarro C, Lowe AJ (2003) Chloroplast DNA phylogeography reveals colonization history of a Neotropical tree, Cedrela odorata L., in Mesoamerica. Mol Ecol 12:1451–1460 Céspedes M, Gutierrez MV, Holbrook NM, Rocha OJ (2003) Restoration of genetic diversity in the dry forest tree Swietenia macrophylla (Meliaceae) after pasture abandonment in Costa Rica. Mol Ecol 12:3201–3212 Chaix G, Gerber S, Razafimaharo V, Vigneron P, Veghaegen D, Hamon S (2003) Gene flow estimation with microsatellites in Malagasy seed orchard of Eucalyptus grandis. Theor Appl Genet 107:705–712 Chalmers KJ, Waugh R, Sprent JI, Simons AJ, Powell W (1992) Detection of genetic variation between and within populations of Gliricidia sepium and G. maculata using RAPD markers. Heredity 69:465–472

References

Chamberlain JR, Galway NW, Simons AJ (1996) Population structure in Gliricidia sepium (Leguminosae) as revealed by isozyme variation. Silvae Genet 45:112–118 Chamshama SAO, Mugasha AG, Wate PA (1999) Variation in performance of Eucalyptus tereticornis provenances at Michafutene, Mozambique. Silvae Genet 48:261–266 Changtragoon S (1984). Variation in growth, seed and pollen charcteristics of Pinus merkusii Jungh. and de Vriese. Master’s thesis, Kasetsart University, Bangkok Changtragoon S, Finkeldey R (1995a) Patterns of genetic variation and characterization of the mating system of Pinus merkusii in Thailand. For Genet 2:87–97 Changtragoon S, Finkeldey R (1995b) Inheritance of isozymes phenotypes of Pinus merkusii. J Trop For Sci 8:167–177 Changtragoon S, Finkeldey R (2000) Der Nutzen von Genmarkern für die Erhaltung forstgenetischer Ressourcen in Thailand. For Snow Landscape Res 75:153–162 Chaperon H (1984) Influence of propagation by cuttings on the breeding strategy of forest trees. In: Barnes RD, Gibson, GL (eds) Provenance and genetic improvement strategies in tropical forest genetics. Commonwealth Forestry Institute, Oxford, and Forest Research Centre, Harare, pp 135–148 Charlesworth D (2004) Plant evolution: modern sex chromosomes. Curr Biol 14:R271–R273 Chase M, Kesseli R, Bawa K (1996a) Microsatellite markers for population and conservation genetics of tropical trees. Am J Bot 83:51–57 Chase MR, Moller C, Kesseli R, Bawa KS (1996b) Distant gene flow in tropical trees. Nature 383:398–399 Clarke G, Young AG (2000) Introduction: genetics, demography and conservation of fragmented populations. In: Young AG, Clarke GM (eds) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge, pp 1–6 Clement CR, Villachica H (1994) Amazonian fruits and nuts: potential for domestication in various agroecosystems. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestification and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 230–238 Colinvaux PA, De Oliveira PE (2001) Amazon plant diversity and climate through the Cenozoic. Palaeogeogr Palaeoclimatol Palaeoecol 166:51–63 Collevatti RG, Brondani RV, Grattapaglia D (1999) Development and characterization of microsatellite markers for genetic analysis of a Brazilian endangered tree species Caryocar brasiliense. Heredity 83:748–756 Comps B, Gömöry D, Letouzey J, Thiébaut B, Petit RJ (2001) Diverging trends between heterozygosity and allelic richness during postglacial colonization in the European beech. Genetics 157:389–397 Compton SG, Ellwood MDF, Davis AJ, Welch K (2000) The flight heights of chalcid wasps (Hymenoptera, Chalcidoidea) in a lowland Bornean rain forest: fig wasps are the high fliers. Biotropica 32:515–522 Connel JH, Lowman MD (1989) Low-diversity tropical rain forests: some possible mechanisms for their existence. Am Nat 134:88–119 Cook AG, Janetos AC, Hinds WT (1990) Global effects of tropical deforestation: towards an integrated perspective. Environ Conserv 17:201–212 Cook JM (1993) Sex determination in the Hymenoptera: a review of models and evidence. Heredity 71:421–435 Cook JM, Power SA (1996) Effects of within-tree flowering asynchrony on the dynamics of seed and wasp production in an Australian fig species. J Biogeogr 23:487–493

273

274

References

Cordeiro NJ, Howe HF (2001) Low recruitment of trees dispersed by animals in African forest fragments. Conserv Biol 15: 1733–1741 Cornelius J (1994) The effectiveness of plus-tree selection for yield. For Ecol Manage 67:23–34 Cossalter C (1989) Genetic conservation: a cornerstone of breeding strategies. In: Gibson GI, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 28–38 Coster C, Eidmann FE (1934) Selectie-onderzoek van den Djati (Tectona grandis L.f.). Tectona 27:1–45 Coster C, Hardjowasono MS (1935) Selectieonderzoek van den Djati (Tectona grandis L.f.) II. De groei gedurende het tweede levensjaar. Tectona 28:3–21 Cotterill PP (1990) Numbers of families and progeny required for provenance testing. Silvae Genet 39:82–83 Cox PA, Elmqvist T (2000) Pollinator extinction in the Pacific islands. Conserv Biol 14:1237–1239 Crawford AJ (2003) Huge populations and old species of Costa Rican and Panamanian dirt frogs inferred from mitochondrial and nuclear gene sequences. Mol Ecol 12:2525–2540 Crow JF, Kimura M (1970) An introduction to population genetics theory. Harper and Row, New York Cruz A (1981) Bird activity and seed dispersal of a montane forest tree (Dunalia arborescens) in Jamaica. Reprod Bot 34–44 Cruzan MB (1998) Genetic markers in plant evolutionary ecology. Ecology 79:400–412 Curtis JT (1956) The modification of midland-grasslands and forests by man. In: Thomas WL (ed) Man’s role in changing the face of the earth. University of Chicago Press, Chicago, pp 721–736 Darby HC (1956) The clearing of the woodland in Europe. In: Thomas WL (ed) Man’s role in changing the face of the earth. University of Chicago Press, Chicago, pp 183–216 Darwin C (1859) On the origin of species by means of natural selection. Murray, London Darwin C (1875) The variation of animals and plants under domestication. Two volumes. (2nd edn, Pickering and Chatto, London, 1988). Murray, London Davidson J (1995) Training manual on tree breeding and propagation. RAS/91/004 field document no. 21. UNDP/FAO Forest Tree Improvement Project, Los Baños Dawson IK, Powell W (1999) Genetic variation in the Afromontane tree Prunus africana, an endangered medicinal species. Mol Ecol 8:151–156 Dawson IK, Waugh R, Simons AJ, Powell W (1997) Simple sequence repeats provide a direct estimate of pollen-mediated gene dispersal in the tropical tree Gliricidia sepium. Mol Ecol 6:179–183 Dayanandan S, Dole J, Bawa K, Kesseli R (1999) Population structure delineated with microsatellite markers in fragmented populations of a tropical tree, Carapa guianensis (Meliaceae). Mol Ecol 8:1585–1592 Dellaporta SL, Calderon-Urrea A (1993) Sex determination in flowering plants. Plant Cell 5:1241–1251 Delwaulle JC (1989) Plantation clonales an Congo point des recherches sur le choix des clones dix ans après les premières plantation. In: Gibson GL, Griffin AR, Matheson

References

AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 431–434 de Nettancourt D (2001) Incompatibility and incongruity in wild and cultivated plants. Springer, Berlin Heidelberg New York den Njis HCM, Oostermeijer JGB (1997) Reproductive biology and gene flow in a fragmented landscape. Bocconea 7:153–165 Deutscher Forstverein (1986) Erhaltung und nachhaltige Nutzung tropischer Regenwälder. Weltforum, Munich De Vecchi Pellati E (1970) Evolution and importance of land races in breeding. In: Proceedings of the 2nd world consultation on forest genetics and forest tree breeding. FAO, Rome, pp 1266–1278 Dick CW (2001) Genetic rescue of remnant tropical trees by an alien pollinator. Proc R Soc Lond Ser B 268:2391–2396 Dick CW, Etchelecu G, Austerlitz F (2003) Pollen dispersal of tropical trees (Dinizia excelsa:Fabaceae) by native insects and African honeybees in pristine and fragmented Amazonian rainforest. Mol Ecol 12:753–764 Diniz-Filho JAF, Telles MPC (2002) Spatial autocorrelation analysis and the Identification of operational units for conservation in continuous populations. Conserv Biol 16:924–935 Dobzhansky T (1968) Adaptedness and fitness. In: Lewontin RC (ed) Population biology and evolution. Syracuse University Press, Syracuse, pp 109–121 Dobzhansky T, Ayala FJ, Stebbins GL, Valentine JW (1977) Evolution. Freeman, San Francisco Doligez A, Joly HI (1997) Mating system of Carapa procera (Meliaceae) in the French Guiana tropical forest. Am J Bot 84:461–470 Donaldson J, Nänni I, Zachariades C, Kemper J (2002) Effects of habitat fragmentation on pollinator diversity and plant reproductive success in renosterveld shrublands of South Africa. Conserv Biol 16:1267–1276 Dow BD, Ashley MV (1998) High levels of gene flow in bur oak revealed by paternity analysis using microsatellites. J Hered 89:62–70 Durel CE, Kremer A (1995) Hybridization after self-fertilization: a novel perspective for maritime pine breeding. For Genet 2:117–120 Dvorak WS, Wright JA (1994) Offspring-parent correlation for wood density of Pinus tecumanii between native and exotic environments. Can J For Res 24:1593–1596 Dvorak WS, Hamrick JL, Hodge GR (1999) Assessing the sampling efficiency of ex situ gene conservation efforts in natural pine populations in central America. For Genetics 6:21–28 Eguiarte LE, Perez-Nasser N, Piñero D (1992) Genetic structure, outcrossing rate and heterosis in Astrocaryum mexicanum (tropical palm): implications for evolution and conservation. Heredity 69:217–228 Eguiarte LE, Burquez A, Rodriguez J, Martinez-Ramos M, Sarukhan J, Piñero D (1993) Direct and indirect estimates of neighborhood and effective population size in a tropical palm, Astrocaryum mexicanum. Evolution 47:75–87 Eldridge KG (1975) Broadleaved Seed Orchards. Part C – Eucalyptus species. In: Faulkner R (ed) Seed orchards. Her Majesty’s Stationery Office, London, pp 134–139

275

276

References

Eldridge KG (1990) Conservation of forest genetic resources with particular reference to Eucalyptus species. Commonw For Rev 69:45–53 Eldridge KG, Davidson J, Harwood C, van Wyk G (1994) Eucalypt domestication and breeding. Oxford University Press, Oxford Eliott F, Shepherd M, Henry RJ (2005) Verification of interspecific pine hybrids using paternally inherited chloroplast microsatellites. For Genet 12:81–87 El-Kassaby YA (1999) Impacts of industrial forestry on genetic diversity of temperate forest trees. In: Mátyás C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 155–169 El-Kassaby YA (2000) Effects of forest tree domestication on gene pools. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford, pp 197–213 El-Kassaby YA, Askew GR (1998) Seed orchards and their genetics. In: Mandal AK, Gibson GL (eds) Forest genetics and tree breeding. CBS, New Dehli, pp 103–111 El-Kassaby YA, Ritland K (1998) Isozymes: techniques, interpretation and application to forest genetics. In: Mandal AK, Gibson GL (eds) Forest genetics and tree breeding. CBS, New Dehli, pp 213–226 El-Kassaby YA, Dunsworth BG, Krakowski J (2003) Genetic evaluation of alternative silviculture systems in coastal montane forests: western hemlock and amabilis fir. Theor Appl Genet 107:598–610 Ellstrand NC (1992) Gene flow among seed plant populations. New For 6:241–256 Endo M (1994) Twelve years of contribution to reforestation in the Andean region of Colombia. For Ecol Manage 63:219–233 Engelmann F (1994) In vitro systems for the conservation and utilization of germplasm. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 49–52 Ennos RA (1996) Utilising genetic information in plant conservation programmes. In: Hochberg ME, Clobert J, Barbault R (eds) Aspects of the genesis and maintenance of biological diversity. Oxford University Press, Oxford, pp 278–291 Enquete Kommission “Vorsorge zum Schutz der Erdatmosphäre” des deutschen Bundestages (1990) Schutz der Tropenwälder. Eine internationale Schwerpunktaufgabe. Economia und Müller, Bonn Eriksson G (1995) Which traits should be used to guide sampling for gene resources? In: Baradat P, Adams WT, Müller-Starck G (eds) Population genetics and genetic conservation of forest trees. SPB, Amsterdam, pp 349–358 Eriksson G, Ekberg I (2001) An introduction to forest genetics. SLU Uppsala, Uppsala Eriksson G, Lindgren D (1975) Några genetiska reflexioner kring plantsortering. Sver Skogsvårdsförb Tidskr 73:413–426 Eriksson G, Andersson S, Eiche V, Ifver J, Persson A (1980) Severity index and transfer effects on survival and volume production of Pinus sylvestris in Northern Sweden. Stud For Suec 156:5–32 Eriksson G, Namkoong G, Roberds JH (1993) Dynamic gene conservation for uncertain futures. For Ecol Manage 62:15–37 Erwin TL (1988) The tropical forest canopy. The heart of biotic diversity. In: Wilson EO, Peter FM (eds) Biodiversity. National Academy Press, Washington, pp 123–129 Erwin TL (1997) Biodiversity at its utmost: tropical forest beetles. In: Reaka-Kudla ML, Wilson DE, Wilson EO (eds) Biodiversity II. Understanding and protecting our biological resources, Henry, Washington, pp 27–40

References

EU (2000) Council Directive 1999/105/EC of 22 December 1999 on the marketing of forest reproductive material. Official Journal of the European Communities L 11/17 of January 15, 2000 Evans J, Turnbull JW (2004) Plantation forestry in the tropics. The role, silviculture, and use of planted forests for industrial, social, environmental, and agroforestry purposes. Oxford University Press, Oxford Evans LT (1980) The natural history of crop yield. Am Sci 68:388–397 Falconer DS, MacKay TFC (1996) Introduction to quantitative genetics. Pearson, Edinburgh FAO (1986) Databook on endangered tree and shrub species and provenances. FAO forestry paper no 77. FAO, Rome FAO (1992) Mixed and pure forest plantations in the tropics and subtropics. FAO forestry paper no 103. FAO, Rome FAO (1993) Conservation of genetic resources in tropical forest management. Principles and concepts. FAO forestry paper no 107. FAO, Rome FAO (1994) Biotechnology in forest tree improvement with special reference to developing countries. FAO forestry paper no 118. FAO, Rome FAO (2002) Global forest resources assessment 2000. Main report. FAO forestry paper no 140. FAO, Rome Fedorov AA (1966) The structure of the tropical rain forest and speciation in the humid tropics. J Ecol 54:1–11 Feilberg L, Søegaard B (1975) Historical review of seed orchards. In: Faulkner R (ed) Seed orchards. Her Majesty’s Stationery Office, London, pp 1–8 Felker P (1994) Capturing and managing the genetic variation in Prosopis spp. for economically useful characters. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 183–188 Ferguson IS (1996) Sustainable forest management. Oxford University Press, Oxford Finkeldey R (1992) The hybrid origin of Paulownia taiwaniana Hu & Chang. Evidence from isozyme gene markers. Silvae Genet 41:278–282 Finkeldey R (1993) Die Bedeutung allelischer Profile für die Konservierung genetischer Ressourcen bei Waldbäumen. Göttingen research notes in forest genetics 14. Institute of Forest Genetics and Forest Tree Breeding, Göttingen Finkeldey R (1994) A simple derivation of the partitioning of genetic variation within subdivided populations. Theor Appl Genet 89:198–200 Finkeldey R (1995) Homogeneity of pollen allele frequencies of single seed trees in Picea abies (L.) Karst. plantations. Heredity 74:451–463 Finkeldey R (1996) Conservation of forest genetic resources in tropical Asia. In: Park YG, Sakamoto S (eds) Biodiversity and conservation of plant genetic resources in Asia. Japan Scientific Society, Tokyo, pp 117–135 Finkeldey R (2002) Reproduction in continuous cover forests–the geneticist’s perspective. In: von Gadow K, Nagel J, Saborowski J (eds) Continuous cover forestry. Assessment, analysis, scenarios. Kluwer, Dordrecht, pp 67–79 Finkeldey R (2006) Genetische Untersuchungen zur Reproduktion von Teak (Tectona grandis L.f.) in Thailand. Schr Forstl Fak Univ Gött Niedersächs Forstl Versuchsanst (in press) Finkeldey R, Hattemer HH (1993) Gene resources and gene conservation with emphasis on tropical forests. FAO/IBPGR Plant Genet Resour Newsl 94/95:5–10

277

278

References

Finkeldey R, Mátyás G (1999) Assessment of population history and adaptive potentials by means of gene markers. In: Mátyás C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 91–104 Finkeldey R, Ziehe M (2004) Genetic implications of silvicultural regimes. For Ecol Manage 197:231–244 Finkeldey R, De Guzman N, Changtragoon S (1998) Inheritance of isozyme phenotypes in Pterocarpus indicus Willd. For Genet 5:165–170 Finkeldey R, De Guzman N, Changtragoon S (1999) The mating system of Pterocarpus indicus Willd. at Mt. Makiling, Philippines. Biotropica 31:525–530 Finkeldey R, Rachmayanti Y, Gailing O (2006) Molecular genetic tools for the identification of the origin of wood. In: Kües U (ed) Wood production, wood technology and biotechnological impacts. Universitätsverlag Göttingen, Göttingen (in press) Fins L, Friedman ST, Brotschol JV (eds) (1992) Handbook of quantitative forest genetics. Kluwer, Dordrecht Forrest GI (1994) Biochemical markers in tree improvement programmes. For Abs 55:123–153 Foster GS, Bertolucci FLG (1994) Clonal development and deployment: strategies to enhance gain while minimising risk. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 103–111 Foster PF, Sork VL (1997) Population and genetic structure of the West African rain forest liana Ancistrocladus korupensis (Ancistrocladaceae). Am J Bot 84:1078–1091 Francis PJ, Nikles DG, Goschnick DN (1984) Early development of an F1 hybrid of Pinus caribaea var. hondurensis x Pinus oocarpa in Queensland. In: Barnes RD, Gibson, GL (eds) Provenance and genetic improvement strategies in tropical forest genetics. Commonwealth Forestry Institute, Oxford, and Forest Research Centre, Harare, pp 510–515 Frank SA (1985) Hierarchical selection theories and sex ratios. II. On applying the theory, and a test with fig wasps. Evolution 39:949–964 Frankel OH, Soulé ME (1981) Conservation and evolution. Cambridge University Press, Cambridge Franklin IA (1980) Evolutionary change in small populations. In: Soulé ME, Wilcox BA (eds) Conservation biology: an evolutionary perspective. Sinauer, Sunderland, pp 135–150 Fuchs EJ, Lobo JA, Quesada M (2003) Effects of forest fragmentation and flowering phenology on the reproductive success and mating patterns of the tropical dry forest tree Pachira quinata. Conserv Biol 17:149–157 Gabriel WJ, Garrett PW (1984) Pollen vectors in sugar maple (Acer sacharum). Can J Bot 62:2889–2890 Gailing O, Kremer A, Steiner W, Hattemer HH, Finkeldey R (2005) Results on quantitative trait loci for flushing date in oaks can be transferred to different segregating progenies. Plant Biol 7:516–525 Galindo-González J, Guevara S, Sosa VJ (2000) Bat- and bird-generated seed rains at isolated trees in pastures in a tropical rainforest. Conserv Biol 14:1693–1703 Gamage DT, de Silva M, Yoshida A, Szmidt AE, Yamazaki T (2003) Molecular phylogeny of Sri Lankan Dipterocarpaceae in relation to other Asian Dipterocarpaceae based an chloroplast DNA sequences. Tropics 13:79–87

References

Garnier-Géré PH, Ades PK (2001) Environmental surrogates for predicting and conserving genetic variability in tree species. Conserv Biol 15:1632–1644 Gautier-Hion A, Michaloud G (1989) Are figs always keystone resources for tropical frugivorous vertebrates? A test in Gabon. Ecology 70:1826–1833 Geburek T (2000) Überlegungen zur Auswahl von in situ-Ressourcenpopulationen im Rahmen der forstlichen Generhaltung. For Snow Landscape Res 75:121–136 Geburek T, Turok J (2006) Conservation and management of forest genetic resources in Europe. Arbora, Zvolen (Slovakia) Gentry AH (1986) Endemism in tropical versus temperate plant communities. In: Soulé ME (ed) Conservation biology. The science of scarcity and diversity. Sinauer, Sunderland, pp 153–181 Gentry AH (1992) Tropical forest biodiversity: distributional patterns and their conservation siginificance. Oikos 63:19–28 Gerber S, Chabrier P, Kremer A (2003) FAMOZ: a software for parentage analysis using dominant, codominant and uniparentally inherited markers. Mol Ecol 3:479–481 Gernandt DS, López GG, García SO, Liston A (2005) Phylogeny and classification of Pinus. Taxon 54:29–42 Ghazoul J, Liston KA, Boyle TJB (1998) Disturbance-induced density-dependent seed set in Shorea siamensis (Dipterocarpaceae), a tropical forest tree. J Ecol 86:462–473 Giebernau M, Hossaert-McKey M, Anstett M-C, Kjellberg F (1996) Consequences of protecting flowers in a fig: a one-way trip for pollinators? J Biogeogr 23:425–432 Gillespie JH (1998) Population genetics. A concise guide. Johns Hopkins University Press, Baltimore Gillet E, Hattemer HH (1989) Genetic analysis of isoenzyme phenotypes using single tree progenies. Heredity 63:135–141 Gillet EM (1997) Maximum likelihood estimators of the gametic contributions to single-plant progenies. Biometrics 53:504–523 Gillies ACM, Cornelius JP, Newton AC, Navarro C, Hernández M, Wilson J (1997) Genetic variation in Costa Rican populations of the tropical timber species Cedrela odorata L., assessed using RAPDs. Mol Ecol 6:1133–1145 Gilpin M (1991) The genetic effective size of a metapopulation. Biol J Linn Soc Lond 42:165–175 Gilpin ME, Soulé ME (1986) Minimum viable populations: processes of species extinction. In: Soulé ME (ed) Conservation biology. The science of scarcity and diversity. Sinauer, Sunderland, pp 19–34 Gladstone WT, Ledig FT (1990) Reducing pressure on natural forests through highyield forestry. For Ecol Manage 35:69–78 Glaubitz JC, Strk J, Moran GF (1999) Genetic impacts of different silvicultural practices in native eucalypt forest. In: Mátyás C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 183–195 Glaubitz JC, Murrell JC, Moran GF (2003) Effects of native forest regeneration practices on genetic diversity in Eucalyptus consideniana. Theor Appl Genet 107:422–431 Godfray HCJ (1988) Virginity in haplodiploid populations: a study on fig wasps. Ecol Entomol 13:283–291 Godt MJW, Hamrick JL, Edwards-Burke MA, Williams JH (2001) Comparisons of genetic diversity in white spruce (Picea glauca) and jack pine (Pinus banksiana) seed orchards with natural populations. Can J For Res 31:943–949

279

280

References

Goldstein DB, Pollock DD (1997) Launching microsatellites: a review of mutation processes and methods of phylogenetic inference. J Hered 88: 335–342 Goossens B, Chikhi L, Jalil MF, Ancrenaz M, Lackman-Ancrenaz I, Mohamed M, Andau P, Bruford MW (2005) Patterns of genetic diversity and migration in increasingly fragmented and declining orang-utan (Pongo pygmaeus) population from Sabah, Malaysia. Mol Ecol 14:441–456 Goremykin VV, Hellwig FH (2005) Evidence for the most basal split in land plants dividing bryophyte and tracheophyte lineages. Plant Syst Evol 254: 93–103 Goto S, Miyhara F, Die Y (2001) A fast method for checking the genetic identity of ramets in a clonal seed orchard by RAPD analysis with a bulking procedure. Silvae Genet 50:271–275 Grattapaglia D, Sederoff R (1994) Genetic linkage maps of Eucalyptus grandis and Eucalyptus urophylla using a pseudo-testcross: mapping strategy and RAPD markers. Genetics 137:1121–1137 Grattapaglia D, Bertolucci FL, Sederoff RR (1995) Genetic mapping of QTLs controlling vegetative propagation in Eucalyptus grandis and E. urophylla using a pseudotestcross strategy and RAPD markers. Theor Appl Genet 90:933–947 Grattapaglia D, Bertolucci FLG, Penchel R, Sederoff RR (1996) Genetic mapping of quantitative trait loci controlling growth and wood quality traits in Eucalyptus grandis using a maternal half-sib family and RAPD markers. Genetics 144:1205–1214 Gregorius H-R (1978) The concept of genetic diversity and its formal relationship to heterozygosity and genetic distance. Math Biosci 41:253–271 Gregorius H-R (1989) The attribution of phenotypic variation to genetic or environmental variation in ecological studies. In: Scholz F, Gregorius H-R, Rudin D (eds) Genetic effects of air pollutants in forest tree populations. Springer, Berlin Heidelberg New York, pp 3–15 Gregorius H-R (1991) Gene conservation and the preservation of adaptability. In: Seitz A, Loeschcke V (eds) Species conservation: a population-biological approach. Birkhäuser, Basel, pp 31–47 Gregorius H-R (2001) Sustainable treatment of resources: the genetic basis. In: MüllerStarck G, Schubert R (eds) Genetic response of forest systems to changing environmental conditions. Kluwer, Dordrecht, pp 203–222 Gregorius H-R, Kleinschmit JRG (2001) Norm of reaction and adaptational value considered in a tree breeding context. Can J For Res 31:607–616 Gregorius H-R, Namkoong G (1986) Joint anslysis of genetic and environmental effects. Theor Appl Genet 72:413–422 Griffin AR (1990) Effects of inbreeding on growth of forest trees and implications for management of seed supplies for plantation programmes. In: Bawa KS, Hadley M (eds) Reproductive ecology of tropical forest plants. Parthenon, New Jersey, pp 355–374 Griffin AR, Cotterill PP (1988) Genetic variation in growth of outcrossed, selfed and open-pollinated progenies of Eucalyptus regnans and some implications for breeding strategy. Silvae Genet 37:124–131 Griffiths AJ, Miller JH, Suzuki DT, Lewontin RC, Gelbart WM (2000) An introduction to genetic analysis. Freeman, New York Guldager P (1975) Ex situ conservation stands in the tropics. In: Roche L (ed) The methodology of conservation of forest genetic resources. Report on a pilot study. FAO, Rome, pp 85–92

References

Guries RP (1990) Forest genetics and forest tree breeding. In: Young RA, Giese RL (eds) Introduction to forest science. 2nd edn. Wiley, New York, pp 99–118 Gyi MKK (1993) Teak in Myanmar. In: Wood H (ed) Teak in Asia. FORSPA publication no 4. Technical document GCP/RAS/134/ASB. FAO, Bangkok, pp 51–62 Haffer J, Prance GT (2001) Climatic forcing of evolution in Amazonia during the Cenozoic: On the refuge theory of biotic differentiation. Amazoniana 16:579–605 Hall BG (2001) Phylogenetic trees made easy: a how- to manual for molecular biologists. Sinauer, Sunderland Hall P, Chase MR, Bawa KS (1994a) Low genetic variation but high population differentiation in a common tropical forest tree species. Conserv Biol 8:471–482 Hall P, Orrel LC, Bawa KS (1994b) Genetic diversity and mating system in a tropical tree, Carapa guianensis (Melicaceae). Am J Bot 81:1104–1111 Hall P, Walker S, Bawa K (1996) Effect of forest fragmentation on genetic diversity and mating system in a tropical tree, Pithecellobium elegans. Conserv Biol 10:757–768 Hamilton MB (1999) Tropical tree gene flow and seed dispersal. Nature 401:129 Hamrick JL, Godt MJW (1989) Allozyme diversity in plant species. In: Brown AHD, Clegg MT, Kahler AL, Weir BS (eds) Plant population genetics, breeding, and genetic resources. Sinauer, Sunderland, pp 43–63 Hamrick JL, Godt MJW (1996) Effects of life history traits on genetic diversity in plant species. Philos Trans R Soc Lond Ser B 351:1291–1298 Hamrick JL, Godt MJW (1997) Allozyme diversity in cultivated crops. Crop Sci 37:26–30 Hamrick JL, Murawski DA (1991) Levels of allozyme diversity in populations of uncommon Neotropical tree species. J Trop Ecol 7:395–399 Hamrick JL, Godt MJW, Sherman-Broyles SL (1992) Factors influencing levels of genetic diversity in woody plant species. New For 6:95–124 Hanski I, Gilpin M (1991) Metapopulation dynamics: brief history and conceptual domain. Biol J Linn Soc Lond 42:3–16 Hart TB (1990) Monospecific dominance in tropical rain forests. Trends Ecol Evol 5:6–11 Hart TB, Hart JA, Murphy PG (1989) Monodominant and species-rich forests of the humid tropics: causes for their co-occurrence. Am Nat 133:613–633 Hartmann HT, Kester DE (1983) Plant propagation. Principles and practices. Prentice Hall, Englewood Cliffs Hattemer HH (1969) Versuche zur geographischen Variation bei der japanischen Lärche. Silvae Genet 18:1–23 Hattemer HH (1987) Are the EEC directives on forest reproductive material genetically adequate? Silvae Genet 36:94–102 Hattemer HH (1991) Genetic analysis and population genetics. In: Fineschi S, Malvolti ME, Cannata F, Hattemer HH (eds) Biochemical markers in the population genetics of forest trees. SPB, The Hague, pp 5–22 Hattemer HH (1995) Concepts and requirements in the conservation of forest genetic resources. For Genet 2:125–134 Hattemer HH (1997) Gene conservation in tropical forests. Nat Resour Dev 45/46:42–53 Hattemer HH (1999) Wald aus der Retorte? G Aug Nachr Univ Gött 21–28 Hattemer HH (2006) On the appropriate size of forest genetic resources. In: Geburek T, Turok J (eds), Conservation and management of forest genetic resources in Europe. Arbora, Zvolen, pp 413–434

281

282

References

Hattemer HH, Gregorius H-R (1996) Bedeutung der biologischen Vielfalt für die Stabilität von Waldökosystemen. In: Müller-Starck G (ed) Biodiversität und nachhaltige Forstwirtschaft. Ecomed, Landsberg, pp 1–10 Hattemer HH, Melchior GH (1993) Genetics and its application to tropical forestry. In: Pancel L (ed) Tropical forestry handbook. Springer, Berlin Heidelberg New York, pp 333–380 Hattemer HH, Müller-Starck G (1990) Evolution von Baumpopulationen als Folge des Waldsterbens? In: Hattemer HH (ed) Erhaltung forstlicher Genressourcen. Sauerländer, Frankfurt, pp 34–45 Hattemer HH, Ziehe M (1997) Genetic control of phenotypic traits with relevance to gene conservation in trees – a survey of methods. In: Mátyás C (ed) Perspectives of forest genetics and tree breeding in a changing world. IUFRO world series, vol 6. IUFRO, Sopron, pp 135–148 Hattemer HH, Gregorius H-R, Ziehe M, Müller-Starck G (1982) Klonanzahl forstlicher Samenplantagen und genetische Vielfalt. Allg Forst-Jagdz 153:183–191 Hattemer HH, Ziehe M, Bergmann F (1993) Einführung in die Genetik für Studierende der Forstwissenschaft. Sauerländer, Frankfurt Hedegart T (1976) Breeding system, variation and genetic improvement of teak (Tectona grandis L.f.). In: Burley J, Styles BT (eds) Tropical trees; variation, breeding and conservation. Academic, London, pp 109–121 Hedegart T, Lauridsen EB, Keiding H (1975) Broadleaved seed orchards. Part D. – Teak. In: Faulkner R (ed) Seed orchards. Her Majesty’s Stationery Office, London, pp 139–142 Hedrick PW (2000) Genetics of populations. Jones and Barlett, Sudbury Heithaus ER, Opler PA, Baker HG (1974) Bat activity and pollination of Bauhinia pauletia: plant-pollinator coevolution. Ecology 55:412–419 Herre EA (1996) An overview of studies on a community of Panamanian figs. J Biogeogr 23:593–607 Herre EA, Machado CA, Bermingham E, Nason JD, Windsor DM, McCaferty SS, Van Houten W, Bachmann K (1996) Molecular phylogenies of figs and their pollinator wasps. J Biogeogr 23:521–530 Hewitt G (2000) The genetic legacy of the Quaternary ice ages. Nature 405:907–913 Hewitt GM (2001) Speciation, hybrid zones and phylogeography – or seeing genes in space and time. Mol Ecol 10:537–549 Hill DS (1967a) Figs (Ficus spp.) of Hong Kong. Hong Kong University Press, Hong Kong Hill DS (1967b) Figs (Ficus spp.) and fig wasps (Chalcidoidea). J Nat Hist 1:413-434 Hill JL, Hill RA (2001) Why are tropical rain forests so species rich? Classifying, reviewing and evaluating theories. Prog Phys Geogr 25:326–354 Hodgson LM (1976a) Some aspects of flowering and reproductive behaviour in Eucalyptus grandis (Hill) Maiden at J.D.M. Keet Forest Research Station. 2. The fruit, seed, seedlings, self fertility, selfing and inbreeding effects. S Afr For J 98:32–43 Hodgson LM (1976b) Some aspects of flowering and reproductive behaviour in Eucalyptus grandis (Hill) Maiden at J.D.M. Keet Forest Research Station. 1. Flowering, controlled pollination method, pollination, receptivity. S Afr For J 97:18–28 Hodgson LM (1977) Methods of seed orchard management for seed production and ease of reaping in Eucalyptus grandis. S Afr For J 100:38–42

References

Hollingsworth PM, Dawson IK, Goodall-Copestake P, Richardson JE, Weber JC, Sotelo Montes C, Pennington RT (2005) Do farmers reduce genetic diversity when they domesticate tropical trees? A case study from Amazonia. Mol Ecol 14:497–501 Holsinger KE (2000) Demography and extinction in small populations. In: Young AG, Clarke GM (eds) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge, pp 55–74 Honadle G (1993) Institutional constraints on sustainable use: lessons from the tropic showing that resource overexploitation is not just an attitude problem and conservation education is not enough. In: Aplet GH, Johnson N, Olson JT, Sample VA (eds), Defining sustainable forestry. Island, Washington , pp 90–126 Hong SO (1975) Vegetative propagation of plant material for seed orchards with special reference to graft-incompatibility problems. In: Faulkner R (ed) Seed orchards. Her Majesty’s Stationery Office, London, pp 38–48 Hosius B (1993) Wird die genetische Struktur eines Fichtenbestandes von Durchforstungseingriffen beeinflußt? Forst Holz 48: 306–308 Howe HF, Smallwood J (1982) Ecology of seed dispersal. Annu Rev Ecol Syst 13: 201–228 Hubbell SP, Foster RB (1992) Short-term dynamics of a neotropical forest: why ecological research matters to tropical conservation and management. Oikos 63:48–61 Hufford KM, Hamrick JL (2003) Viability selection at three early life stages of the tropical tree Platypodium elegans (Fabaceae, Papilionoideae). Evolution 57:518–526 Hughes CE (1989) New opportunities in Leucaena genetic improvement. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 218–226 Hühn M (1986) Theoretical studies on the necessary number of components in mixtures. 3. Number of components and risk considerations. Theor Appl Genet 72:211–218 Huxley J (1938) Clines: an auxiliary taxonomic principle. Nature 143:219–220 Indrioko S, Gailing O, Finkeldey R (2006) Molecular phylogeny of Dipterocarpaceae in Indonesia based on chloroplast DNA. Plant Syst Evol (in press). DOI 10.1007/s00606-006-0435-8 Ishiyama H, Kado T, Iwasaki M, Matsuoka M, Shukor N, Szmidt A, Yamazaki T (2003) Nucleotide variation in the GapC region of four species of Shorea and their putative hybrids. Tropics 13:89–99 Iwata H, Konuma A, Tsumura Y (2000) Development of microsatellite markers in the tropical tree Neobalanocarpus heimii (Dipterocarpaceae). Mol Ecol 9:1661–1686 James T, Vege S, Aldrich P, Hamrick JL (1998) Mating systems of three tropical dry forest tree species. Biotropica 30:587–594 Janson CH, Terborgh J, Emmons LH (1981) Non-flying mammals as pollinating agents in the Amazonian forest. Biotropica 13(Suppl):1–6 Janzen DH (1970) Herbivores and numbers of tree species in tropical forests. Am Nat 104:501–528 Janzen DH (1976) Why bamboos wait so long to flower. Annu Rev Ecol Syst 7:347–391 Janzen DH (1979a) How to be a fig. Annu Rev Ecol Syst 10:13–51 Janzen DH (1979b) How many parents do the wasps from a fig have? Biotropica 11:127–129

283

284

References

Jayasekera NEM, Karunasekera KB, Kearsey MJ (1994) Genetics of reproduction traits in Hevea braliliensis (rubber). I. Changes in genetical control with age. Heredity 73:650–656 Johnsen Ø, Dæhlen OG, Østreng G, Skrøppa T (2005a) Daylength and temperature during seed production interactively affect adaptive performance of Picea abies progenies. New Phytol 168:589–596 Johnsen Ø, Fossdahl CG, Nagy N, Mølmann J, Dæhlen OG (2005b) Climatic adaptation in Picea abies progenies is affected by the temperature during zygotic embryogenesis and seed maturation. Plant Cell Environ 28:1090–1102 Joly HI, Zeh-Nlo M, Danthu P, Aygalent C (1992) Population genetics of an African acacia, Acacia albida. I. Genetic diversity of populations from West Africa. Aust J Bot 40:59–73 Jones N, Burley J (1973) Seed certification, provenance nomenclature and genetic history in forestry. Silvae Genet 22:53–58 Jones N, Burley J (1975) Broadleaved seed orchards. Part B. Notes on tropical and semitropical species other than Eucalyptus species and teak. In: Faulkner R (ed) Seed orchards. Her Majesty’s Stationery Office, London, pp 131–134 Jong K (1976) Cytology of the Dipterocarpaceae. In: Burley J, Styles BT (eds) Tropical trees; variation, breeding and conservation. Academic, London, pp 79–84 Jörgensen J (1990) Conservation of valuable gene resources by cryopreservation in some forest tree species. J Plant Physiol 136: 373–376 Kageyama PY, Kikuti P (1989) Comparison between clones and open-pollinated progenies originating from a population of Eucalyptus grandis (Hill) Maiden in Brazil. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 227–235 Kajita T, Kamiya K, Nakamura K, Tachida H, Wickneswari R, Tsumura Y, Yoshimaru H, Yamazaki T (1998) Molecular phylogeny of Dipterocarpaceae in Southeast Asia based on nucleotide sequences of matK, trnL intron, and trnL-trnF intergenic spacer region of chloroplast DNA. Mol Phylogenet Evol 10:202–209 Kamiya K, Harada K, Ogino K, Kajita T, Yamazaki T, Lee H-S, Ashton PS (1998) Molecular phylogeny of dipterocarp species using nucleotide sequences of two noncoding regions in chloroplast DNA. Tropics 7:195–207 Kamiya K, Harada K, Tachida H, Ashton PS (2005) Phylogeny of PgiC gene in Shorea and its closely related genera (Dipterocarpaceae), the dominant trees in Southeast Asian tropical rain forests. Am J Bot 92:775–788 Kamra SK (1986) Situation pertaining to forestry seed in some developing countries in Asia and measures for improvement. In: Nather J (ed) Proceedings of the international symposium on seed problems under stressful conditions, June 1985. Forstliche Versuchsanstalt, Vienna, pp 228–238 Kang H, Lascoux M, Gullberg U (1998) Systematic tree breeding. In: Mandal AK, Gibson GL (eds) Forest genetics and tree breeding. CBS, New Dehli, pp 38–62 Kanowski PJ, Nikles DG (1989) Family structure in provenance trials, conservation stands, and genetic base populations. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand,

References

November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 68–78 Kaur A, Ha CO, Jong K, Sands VE, Chan HT, Soepadmo E, Ashton PS (1978) Apomixis may be widespread among trees of the climax rain forest. Nature 271:440–442 Keiding H, Boonkird S (1960) Vegetative propagation of teak. Unasylva 14:193–194 Kemp RH, Roche L, Willan RL (1976) Current activities and problems in the exploration and conservation of tropical forest gene resources. In: Burley J, Styles BT (eds) Tropical trees; variation, breeding and conservation. Academic, London, pp 223–233 Kenrick J, Knox RB (1985) Self-incompatibility in the nitrogen-fixing tree, Acacia retinodes: quantitative cytology of pollen tube growth. Theor Appl Genet 69:481–488 Kenta T, Isagi Y, Nakagawa M, Yamashita M, Nakashizuka T (2004) Variation in pollen dispersal between years with different pollination conditions in a tropical emergent tree. Mol Ecol 13:3575–3584 Khurana DK, Khosla PK (1998) Hybrids in forest tree improvement. In: Mandal AK, Gibson GL (eds), Forest genetics and tree breeding. CBS, New Dehli, pp 86–102 Khuspe SS, Nadgauda RS, Jagtap SM, John CK, Mascarenhas AF (1994) Capture genetic resources by in vitro culture: field evaluation and precocious flowering. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 53–60 Kim ZS (1985) Viability selection at an allozyme locus during development in European beech (Fagus sylvatica L.). Silvae Genet 34:181–186 Kim ZS, Yi CH, Lee SW (1994) Genetic variation and sampling strategy for conservation in Pinus species. In: Kim ZS, Hattemer HH (eds) Conservation and manipulation of genetic resources in forestry. Kwang Moon Kag, Seoul, pp 294–321 Kimura M (1983) The neutral theory of natural evolution. Cambridge University Press, Cambridge Kitamura K, Mohamad Yusof AR, Ochiai O, Yoshimaru H (1994) Estimation of outcrossing rate on Dryobalanops aromatica Gaertn. F. in primary and secondary forests in Brunei, Burneo, Southwest Asia. Plant Species Biol 9:37–41 Kjær ED (1996) Estimation of effective population number in a Picea abies (Karst.) seed orchard based on flower assessment. Scand J For Res 11:111–121 Kjær ED, Verapong S (1995) Outcrossing rate of teak (Tectona grandis L.). Silvae Genet 44:145–147 Kjær ED, Lauridsen EB, Wellendorf H (1995) Second evaluation of an international series of teak provenance trials. DANIDA Forest Seed Centre, Humlebæk Konnert M, Schmidt S (1996) Genetische Konsequenzen der Größensortierung in der Baumschule: Erste Ergebnisse. In: Müller-Starck G (ed) Biodiversität und nachhaltige Forstwirtschaft. Ecomed, Landsberg, pp 222–238 Konuma A, Tsumura Y, Lee CT, Lee SL, Okuda T (2000) Estimation of gene flow in the tropical-rainforest tree Neobalanocarpus heimii (Dipterocarpaceae), inferred from paternity analysis. Mol Ecol 9:1843–1852 Koptur S (1984) Outcrossing and pollinator limitation of fruit set: breeding systems of neotropical Inga trees (Fabaceae: Mimosoideae). Evolution 38:1130–1143 Krauss SL, Dixon B, Dixon KW (2002) Rapid genetic decline in a translocated population of the endangered plant Grevillea scapigera. Conserv Biol 16:986–994 Kremer A, Zanetto A, Ducousso A (1997) Multilocus and multitrait measures of differentiation for gene markers and phenotypic traits. Genetics 145:1229–1241

285

286

References

Kremer A, Le Corre V, Mariette S (1999) Population differentiation at adaptive traits and their underlying loci in forest trees: theoretical predictions and experimental results. In: Mátyás C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 59–74 Kremer A, Caron H, Cavers S, Colpaert N, Gheysen G, Gribel R, Lemes M, Lowe AJ, Margis R, Navarro C, Salgueiro F (2005) Monitoring genetic diversity in tropical trees with multilocus dominant markers. Heredity 95:274–280 Krutovsky KV, Neale DB (2005) Nucleotide diversity and linkage disequilibrium in cold-hardiness- and quality-related candidate genes in Douglas fir. Genetics 171: 2029–2041 Kundu SK (1999) The mating system and genetic significance of polycarpy in the neem tree (Azadirachta indica). Theor Appl Genet 99:1216–1220 Kundu SK, Tigerstedt PMA (1998) Variation in net photosynthesis, stomatal characteristics, leaf area and whole.plant phytomass production among ten provenance of neem (Azadirachta indica). Tree Physiol 19:47–52 Ladipo DO, Leakey RRB, Grace J (1991) Clonal variation in apical dominance of Triplochiton scleroxylon K. Schum. in response to decapitation. Silvae Genet 40: 135–140 Ladrach WE (1998) Provenance research: the concept, application and achievement. In: Mandal AK, Gibson GL (eds), Forest genetics and tree breeding. CBS, New Dehli, pp 16–37 Lal P (1993) Economics of mass clonal multiplication of forest trees. In: Davidson J (ed) Proceedings of the regional symposium on recent advances in mass clonal multiplication of forest trees for plantation programmes. RAS/91/004 field document no 4. UNDP/FAO Forest Tree Improvement Project, Los Baños, pp 327–352 Lamb D, Parrotta J, Keenan R, Tucker N (1997) Rejoining habitat remnants: restoring degraded rainforest lands. In: Laurance WF, Bierregaard RO (eds) Tropical forest remnants: ecology, management, and conservation of fragmented communities. University of Chicago Press, Chicago, pp 366–385 Lamprecht H (1986) Waldbau in den Tropen. Parey, Hamburg Lande R (1988) Genetics and demography in biological conservation. Science 241:1455–1460 Langlet O (1936) Studier över tallens fysiologiska variabilitet och dess samband med klimatet. Medd Statl Skogsförsöksanst 29:219–470 Langlet O (1963) Patterns and terms of intraspecific variability. Nature 200:347–348 Langner W (1966) Populationsgenetische Unterschiede und Bestockungsdichte bei der Bestandsbegründung. Allg Forstz 21:643–644 Lara-Gomez G, Gailing O, Finkeldey R (2005) Genetic variation in isolated Mexican populations of the endemic marple Acer skutchii Rehd. Allg Forst-Jagdz 176: 97–103 Larsen JB (1986) Das Tannensterben: Eine neue Hypothese zur Klärung des Hintergrundes dieser rätselhaften Komplexkrankheit der Weißtanne (Abies alba Mill.). Forstwiss Centralbl 105:381–396 Larsen JB, Meki´c F (1991) The geographic variation in European silver fir (Abies alba Mill.). Gass exchange and needle cast in relation to needle age, growth rate, dry matter partitioning and wood density by 15 different provenances at age 6. Silvae Genet 40:188–198

References

Laurance WF, Bierregaard RO (1997a) A crisis in the making. In: Laurance WF, Bierregaard RO (eds) Tropical forest remnants: ecology, management, and conservation of fragmented communities. University of Chicago Press, Chicago, pp XI–XV Laurance WF, Bierregaard RO (eds) (1997b) Tropical forest remnants: ecology, management, and conservation of fragmented communities. University of Chicago Press, Chicago Laurance WF, Lovejoy TE, Vasconcelos HL, Bruna EM, Didham RK, Stouffer PC, Gascon C, Bierregaard RO, Laurance SG, Sampaio E (2002) Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conserv Biol 16:605–618 Law BS, Lean M (1999) Common blossom bats (Syconycteris australis) as pollinators in fragmented Australian tropical rainforest. Biol Conserv 91:201–212 Lawson GJ (1994) Indigenous trees in West African forest plantation: the need for domestication by clonal techniques. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 112–123 Leakey RRB, Newton AC (eds) (1994a) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London Leakey RRB, Newton AC (1994b) Domestication of ‘Cinderella’ species as the start of a woody-plant revolution. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 3–6 Leakey RRB, Last FT, Longman KA (1982) Domestication of tropical trees: an approach securing future productivity and diversity in managed ecosystems. Commonw For Rev 61:33–42 Leakey RRB, Newton AC, McDick J (1994) Capture of genetic variation by vegetative propagation: processes determining success. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 72–83 Le Corre V, Kremer A (2003) Genetic variability at neutral markers, quantitative trait loci and trait in a subdivided population under selection. Genetics 164:1205–1219 Ledig FT (1973) An analysis of methods for the selection of trees from wild stands. For Sci 20:2–16 Ledig FT (1986a) Conservation strategies for forest gene resources. For Ecol Manage 14:77–90 Ledig FT (1986b) Heterozygosity, heterosis, and fitness in outbreeding plants. In: Soulé ME (ed) Conservation biology. The science of scarcity and diversity. Sinauer, Sunderland, pp 77–104 Ledig FT (1992) Human impacts on genetic diversity in forest ecosystems. Oikos 63:87–108 Ledig FT, Kitzmiller JH (1992) Genetic strategies for reforestration in the face of global change. For Ecol Manage 50:153–169 Lee SL (2000) Mating system parameters of Dryobalanops aromatica Gaertn. f. (Dipterocarpaceae) in three different forest types and a seed orchard. Heredity 85:338–345 Lee SL, Wickneswari R, Mahani MC, Zakri AH (2000) Genetic diversity of a tropical tree species, Shorea leprosula Miq. (Dipterocarpaceae), in Malaysia: implications for conservation of genetic resources and tree improvement. Biotropica 32:213–224

287

288

References

Lee SS (ed) (1995) A guide book to Pasoh. FRIM technical information handbook no 3. Forest Research Institute Malaysia, Kepong Lefèvre W (1984) Die Entstehung der biologischen Evolutionstheorie. Ullstein, Frankfurt Leighton M, Leighton DR (1983) Vertebrate responses to fruiting seasonality within a Bornean rain forest. In: Sutton SL, Whitmore TC, Chadwick AC (eds) Tropical rain forest: ecology and management. Blackwell, Oxford, pp 181–196 Leinemann L, Hattemer HH (2006) Genetic variation and mating pattern in a stand of yew (Taxus baccata L.). Allg Forst-Jagdz (in press) Lemes MR, Gribel R, Proctor J, Grattapaglia D (2003) Population genetic structure of mahogany (Swietenia macrophylla King, Meliaceae) across the Brazilian Amazon, based on variation at microsatellite loci: implications for conservation. Mol Ecol 12:2875–2883 Lesica P, Allendorf FW (1995) When are peripheral populations valuable for conservation? Conserv Biol 9:753–760 Lessa EP, Cook JA, Patton JL (2003) Genetic footprints of demographic expansion in North America, but not Amazonia, during the late Quaternary. Proc Natl Acad Sci USA 100:10331–10334 Levin DA, Kerster HW (1974) Gene flow in seed plants. Evol Biol 7:139–220 Lewontin RC (1984) Detecting population differences in quantitative character as opposed to gene frequencies. Am Nat 123:115–124 Li Q, Xu Z, He T (2002) Ex situ genetic conservation of endangered Vatica guangxiensis (Dipterocarpaceae) in China. Biol Conserv 106:151–156 Libby WJ (1973) Domestication strategies for forest trees. Can J For Res 3: 265–276 Libby WJ (1982) What is a safe number of clones per plantation? In: Heybroek HM, Stephan BR, von Weissenberg K (eds) Resistance to diseases and pests in forest trees. Pudoc, Wageningen, pp 342–360 Liengsiri C, Yeh FC, Boyle TJB (1995) Isozyme analysis of a tropical forest tree, Pterocarpus macrocarpus Kurz. in Thailand. For Ecol Manage 74:13–22 Liengsiri C, Boyle TJB, Yeh FC (1998) Mating system in Pterocarpus indicus Kurz in Thailand. J Hered 89:216–221 Lindgren D, Lindgren K (1996) Long distance pollen transfer may make gene conservation difficult. In: Kurm M, Tamm Y (eds) Conservation of forest genetic resources. Nordic group of forest genetics and tree breeding meeting, Estonia, June 1996. Estonian Agricultural University, Tartu, pp 51–62 Lindgren D, Paule L, Shen X, Yazdani R, Segerström U, Wallin J-E, Lejdebro ML (1995) Can viable pollen carry Scots pine genes over long distances? Grana 34:64–69 Lindquist B (1946) Den skogliga rasforskningen och praktiken. Svenska Skogsvårdsföreningen, Stockholm Lira CF, Cardoso SRS, Ferreira PCG, Cardoso MA, Provan J (2003) Long-term population isolation in the endangered tropical tree species Caesalpinia echinata Lam. revealed by chloroplast microsatellites. Mol Ecol 12:3219–3225 Listabarth C (1996) Pollination of Bactris by Phyllotrox and Epurea. Implications of the palm breeding beetles on pollination at the community level. Biotropica 28:69–81 Lobo JA, Quesada M, Stoner KE (2005) Effects of pollination by bats on the mating system of Ceiba pentandra (Bombacaceae) populations in two tropical life zones in Costa Rica. Am J Bot 92:370–376

References

Loiselle BA, Ribbens E, Vargas O (1996) Spatial and temporal variation of seed rain in a tropical lowland wet forest. Biotropica 28:82–95 Longman KA (1976) Conservation and utilization of gene resources by vegetative multiplication of tropical trees. In: Burley J, Styles BT (eds) Tropical trees; variation, breeding and conservation. Academic, London, pp 19–24 Loveless MD (1992) Isozyme variation in tropical trees: patterns of genetic organization. New For 6:67–94 Lowenfeld R, Klekowski EJ Jr (1992) Mangrove genetics. I. Mating system and mutation rates of Rhizophora mangle in Florida and San Salvador Island, Bahamas. Int J Plant Sci 153:394–399 Loyn RH, MacAlpine C (2001) Spatial patterns and fragmentation: indicators for conserving biodiversity in forest landscapes. In: Raison RJ, Brown AG, Flinn DW (eds) Criteria and indicators for sustainable forest management. CABI, Wallington, pp 391–422 Lu Z, Johnson WE, Menotti-Raymond M, Yuhki N, Martenson JS, Mainka S, Shi-Qiang H, Zhihe Z, Li G, Pan W, Mao X, O’Brien SJ (2001) Patterns of genetic diversity in remaining giant panda populations. Conserv Biol 15:1596–1607 Lugo AE (1988) Estimating reductions in the diversity of tropical forest species. In: Wilson EO, Peter FM (eds) Biodiversity. National Academy Press, Washington, pp 58–70 Luu HT (2005) Genetic variation and the reproductive system of Dipterocarpus cf. condorensis Pierre in Vietnam. Cuvillier, Göttingen Machado CA, Herre EA, McCaferty SS, Bermingham E (1996) Molecular phylogenies of fig pollinating and non-pollinating wasps and the implications for the origin and evolution of the fig-fig wasp mutualism. J Biogeogr 23:531–542 Mandal AK, Gibson GL (ed) (1998). Forest genetics and tree breeding. CBS, New Delhi Maghembe JA, Kwesiga F, Ngulube M, Prins H, Malaya FM (1994) Domestication potential of indigenous fruit trees of the miombo woodlands of southern Africa. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 220–229 Malcolm JR, Ray JC (2000) Influence of timber extraction routes on Central African small-mammal communities, forest structure, and tree diversity. Conserv Biol 14:1623–1638 Manel S, Schwartz MK, Luikart G, Taberlet P (2003) Landscape genetics: combining landscape ecology and population genetics. Trends Ecol Evol 18:189–197 Margraf J, Milan PP (2004) Rainforestation farming. A farmer’s guide to biodiversity management for the Philippines. Institute of Tropical Ecology, Leyte State University, Baybay Marsh EK (1970) Selecting adapted races of introduced species. In: Proceedings of the 2nd world conference on forest genetics and forest tree breeding. FAO, Rome, pp 1249–1265 Marshall DR, Brown AHD (1975) Optimum sampling strategies in genetic conservation. In: Frankel OH, Hawkes JG (eds) Crop genetic resources for today and tomorrow. Cambridge University Press, Cambridge, pp 53–80 Mátyás C, Yeatman CW (1992) Effect of geographical transfer on growth and survival of jack pine (Pinus banksiana Lamb.) populations. Silvae Genet 41:370–375 Maynard Smith J (1998) Evolutionary genetics. Oxford University Press, Oxford McCauley DE (1991) Genetic consequences of local population extinction and recolonization. Trends Ecol Evol 6:5–8

289

290

References

McCauley DE (1995) The use of chloroplast DNA polymorphism in studies of gene flow in plants. Trends Ecol Evol 10:198–202 McClenaghan LR, Beauchamp AC (1986) Low genetic differentiation among isolated populations of the Californian fan palm (Washingtonia filifera). Evolution 40:315–322 McKay JK, Latta RG (2002) Adaptive population divergence: markers, QTL and traits. Trends Ecol Evol 17:285–291 McKinley CR, van Buijtenen JP (1998) Principles and techniques of selection in tree breeding. In: Mandal AK, Gibson GL (eds) Forest genetics and tree breeding. CBS, New Dehli, pp 1–15 Mehra PN, Bawa KS (1969) Chromosomal evolution in tropical hardwoods. Evolution 23:466–481 Melchior GH (1965) Über die Vegetativvermehrung von Bombacopsis quinata (Jaq.) Dugant. Silvae Genet 14:148–154 Melchior GH (1986a) Genetic differences in ability of aspen families to sustain longterm storage of seeds. In: Nather J (ed) Proceedings of the international symposium on seed problems under stressful conditions. Forstliche Bundesversuchsanstalt, Vienna, pp 58–70 Melchior GH (1986b) Probleme der Saatgutversorgung in Aufforstungsprojekten in Südamerika. In: Nather J (ed) Proceedings of the international symposium on seed problems under stressful conditions. Forstliche Bundesversuchsanstalt, Vienna, pp 203–218 Mendel G (1866) Versuche über Pflanzenhybriden. Verh Naturforsch Ver Bruenn 4:3–47 Menges ES (1991) The application of minimum viable population theory to plants. In: Falk DA, Holsinger KE (eds) Genetics and conservation of rare plants. Oxford University Press, Oxford, pp 45–61 Mesén JF, Boshier DH, Cornelius JP (1994) Genetic improvement of trees in Central America, with particular reference to Costa Rica. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 249–255 Midgley J (1989) Pollen dispersal distances for a conifer canopy species in Knysna forest. S Afr J Bot 55:662–663 Midgley S (1996) Seed collection strategies in a changing world. In: Olesen K (ed) Innovations in tropical tree seed technology. IUFRO symposium of project group P2.04.00 ‘seed problems’. Danish Forest Tree Seed Centre, Humblebæk, pp 175–187 Midgley SJ, Eldridge KG, Doran JC (1989) Genetic resources of Eucalyptus camaldulensis. Commonw For Rev 68:295–308 Milimo PB, Dick JMP, Munro RC (1994) Domestication of trees in semi-arid East Africa: the current situation. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 210–219 Mitton JB (1998) Allozymes in tree breeding research. In: Mandal AK, Gibson GL (eds) Forest genetics and tree breeding. CBS, New Dehli, pp 239–251 Molbo D, Machados CA, Herre EA, Keller L (2004) Inbreeding and population structure in two pairs of cryptic fig wasp species. Mol Ecol 13:1613–1623 Monaco LC (1977) Consequences of the introduction of coffee rust into Brazil. Ann N Y Acad Sci 287:57–71

References

Moncur MV, Mitchell A, Fripp Y, Kleinschmidt GJ (1995) The role of honeybees (Apis mellifica) in eucalypt and acacia seed-production areas. Commonw For Rev 64:350-354 Montagnini F, Jordan CF (2005) Tropical forest ecology. The basis for conservation and management. Springer, Berlin Heidelberg New York Moran GF, Brown AHD (1980) Temporal heterogeneity of outcrossing rates in Alpine ash (Eucalyptus delegatensis R.T. Bak.). Theor Appl Genet 57:101–105 Moran GF, Bell JC (1983) Eucalyptus. In: Tanksley SD, Orton TJ (eds) Isozymes in plant genetics and breeding, part B. Elsevier, Amsterdam, pp 423–441 Moran GF, Hopper SD (1983) Genetic diversity and the insular population structure of the rare granite rock species, Eucalyptus caesia Benth. Aust J Bot 31:161–172 Moran GF, Muona O, Bell JC (1989a) Acacia mangium: a tropical forest tree of the coastal lowlands with low genetic diversity. Evolution 43:231–235 Moran GF, Bell JC, Griffin AR (1989b) Reduction in levels of inbreeding in a seed orchard of Eucalyptus regnans F. Muell. compared with natural populations. Silvae Genet 38:32–36 Morgenstern EK (1996) Geographic variation in forest trees. UBC Press, Vancouver Morin PA, Luikart G, Wayne RK, SNP workshop group (2004) SNPs in ecology, evolution and conservation. Trends Ecol Evol 19:208–216 Moritz A, Lüdders P (1993) Pollenkeimung, Pollenschlauchwachstum und Befruchtungsverhalten verschiedener Klone der Paranuß (Bertholetia excelsa Humb. & Bonpl.). Angew Bot 67:107–112 Mottura MC, Finkeldey R, Verga AR, Gailing O (2005) Development and characterization of microsatellite markers for Prosopis chilensis and Prosopis flexuosa and crossspecies amplification. Mol Ecol Notes 5:487–489 Müller G (1976) A simple method of estimating rates of self-fertilization by analysing isozymes in tree seeds. Silvae Genet 25:15–17 Müller G (1977) Cross-fertilization in a conifer stand inferred from enzyme genemarkers in seeds. Silvae Genet 26:223–226 Müller-Starck G (1982) Reproductive systems in conifer seed orchards. I. Mating probabilities in a seed orchard of Pinus sylvestris L. Silvae Genet 31:188–197 Müller-Starck G (1985) Genetic differences between ‘tolerant’ and ‘sensitive’ beeches (Fagus sylvatica L.) in an environmentally stressed adult forest stand. Silvae Genet 34:241–247 Müller-Starck G (1991) Genetic processes in seed orchards. In: Giertych M, Mátyás C (eds) Genetics of Scots pine. Akadémiai Kiadó, Budapest, pp 147–162 Müller-Starck G (1996) Beiträge der Forstgenetik zur nachhaltigen Waldbewirtschaftung. In: Müller-Starck G (ed) Biodiversität und Nachhaltige Forstwirtschaft. Ecomed, Landsberg, pp 259–283 Müller-Starck G, Ziehe M (1984) Reproductive systems in conifer seed orchards. 3. Female and male fitnesses of individual clones realized in seeds of Pinus sylvestris L. Theor Appl Genet 69:173–177 Muluvi GM, Sprent JI, Soranzo N, Provan J, Odee D, Folkard G, McNicol JW, Powell W (1999) Amplified fragment length polymorphism (AFLP) analysis of genetic variation in Moringa oleifera Lam. Mol Ecol 8:463–470 Muona O (1989) Population genetics in forest tree improvement. In: Brown AHD, Clegg MT, Kahler AL, Weir BS (eds) Plant population genetics, breeding, and genetic resources. Sinauer, Sunderland, pp 282–298

291

292

References

Muona O, Harju A (1989) Effective population size, genetic variability, and mating system in natural stands and seed orchards of Pinus sylvestris. Silvae Genet 38:221–228 Murawski DA, Bawa KS (1994) Genetic structure and mating system of Stemonoporus oblongifolius (Dipterocarpaceae) in Sri Lanka. Am J Bot 81:155–160 Murawski DA, Hamrick JL (1991) The effect of the density of flowering individuals on the mating systems of nine tropical tree species. Heredity 67:167–174 Murawski DA, Hamrick JL (1992) The mating system of Cavanillesia platanifolia under extremes of flowering-tree density: a test of predictions. Biotropica 24:99–101 Murawski DA, Dayanandan B, Bawa KS (1994a) Outcrossing rates of two endemic Shorea species from Sri Lankan tropical rain forests. Biotropica 26:23–29 Murawski DA, Gunatilleke IAUN, Bawa KS (1994b) The effects of selective logging on inbreeding in Shorea megistophylla (Dipterocarpaceae) from Sri Lanka. Conserv Biol 8: 997–1002 Murillo Gamboa O (1997) Genetische Untersuchungen an natürlichen Populationen von Alnus acuminata spp arguta (Schlectendal) Furlow in Costa Rica und Panamá. Cuvillier, Göttingen Myers N (1986) Tropical deforestation and a mega-extinction spasm. In: Soulé ME (ed) Conservation biology. The science of scarcity and diversity. Sinauer, Sunderland, pp 394–409 Nagamitsu T, Inoue T (1997) Cockroach pollination and breeding system of Uvaria elmeri (Annonaceae) in a lowland mixed-dipterocarp forest in Sarawak. Am J Bot 84:208–213 Namkoong G (1969) Non-optimality of the local races. In: Proceedings of the 10th southern conference on forest tree improvement, Raleigh, Southern Forest Tree Improvement Committee, pp 149–153 Namkoong G (1981) Introduction to quantitative genetics in forestry. Castle House, Tunbridge Wells Namkoong G (1984) A control concept of gene conservation. Silvae Genet 33:160–163 Namkoong G (1985) The population genetic basis of breeding theory. In: Gregorius H R (ed) Population genetics in forestry. Springer, Berlin Heidelberg New York, pp 2–15 Namkoong G (1986) Genetics and the forests of the future. Unasylva 38:2–18 Namkoong G (1989) Systems of gene management. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 1–8 Namkoong G (1991) Maintaining genetic diversity in breeding for resistance in forest trees. Annu Rev Phytopathol 29: 325–342 Namkoong G (1999) Plantations vs. naturally regenerated forests: risks and stability in using bred, cloned, or biotech products. In: Mátyás C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 147–154 Namkoong G, Barnes RD, Burley J (1980) A philosophy of breeding strategy for tropical forest trees. Commonwealth Forestry Institute, Oxford Namkoong G, Kang HC, Brouard JS (1988) Tree breeding: principles and strategies. Springer, Berlin Heidelberg New York Namkoong G, Koshy MP, Aitken S (2000) Selection. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford,, pp 101–111

References

Namkoong G, Boyle T, Gregorius H-R, Joly H, Savolainen O, Wickneswari R, Young AG (1996) Testing criteria and indicators for assessing the sustainability of forest management: genetic criteria and indicators. Center for International Forestry Research, Bogor Namkoong G, Boyle T, El-Kassaby YA, Palmberg-Lerche C, Eriksson G, Gregorius H-R, Joly H, Kremer A, Savolainen O, Wickneswari R, Young AG, Zeh-Nlo M, Prabhu R (2002) Criteria and indicators for sustainable forest management: assessment and monitoring of genetic variation. Working paper FGR37E. FAO, Rome Nanson A (2001) The new OECD scheme for the certification of forest reproductive materials. Silvae Genet 50:181–187 Nason JD, Hamrick JL (1997) Reproductive and genetic consequences of forest fragmentation: two case studies of neotropical canopy trees. J Hered 88:264–276 Nason JD, Herre EA, Hamrick JL (1996) Paternity analysis of the breeding structure of strangler fig populations: evidence for substainal long-distance wasp dispersal. J Biogeogr 23:501–512 Nason JD, Aldrich PR, Hamrick JL (1997) Dispersal and the dynamics of genetic structure in fragmented tropical tree populations. In: Laurence WF, Bierregaard RO (eds). Tropical forest remnants. The University of Chicago Press, Chicago, pp 304–320 Nason JD, Herre EA, Hamrick JL (1998) The breeding structure of a tropical keystone plant resource. Nature 391:685–687 Navarro C, Cavers S, Pappinen A, Tigerstedt P, Lowe A, Merilä J (2005) Contrasting quantitative traits and neutral gene markers for genetic resource assessment of Mesoamerican Cedrela orodrata. Silvae Genet 54:281–292 Nei M (1972) Genetic distance between populations. Am Nat 106:283–292 Nei M (1973) Analysis of gene diversity in subdivided populations. Proc Natl Acad Sci USA 70:3321–3323 Nei M (1987) Molecular evolutionary genetics. Columbia University Press, New York Nei M, Maruyama T, Chakraborty R (1975) The bottleneck effect and genetic variability in populations. Evolution 29:1–10 Neil PE (1990) Conservation and management possibilities for Agathis macrophylla in Vanuatu. For Ecol Manage 35:239–248 Nester MR (1994) Modulo tile constructions for systematic seed orchard designs. Silvae Genet 43:312–321 Newbury HJ, Ford-Lloyd BV (1993) The use of RAPD for assessing variation in plants. Plant Growth Regul 12:43–51 Newton AC, Leakey RRB, Powell W, Chalmers K, Waugh R, Tchoundieu Z, Mathias PJ, Alderson PG, Mesén JF, Baker P, Ramnarine S (1994) Domestication of mahoganies. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 256–266 Ng KKS, Lee SL, Koh CL (2004) Spatial structure and genetic diversity of two tropical tree species with contrasting breeding systems and different ploidy levels. Mol Ecol 13:657–669 Nienstaedt H, Riemenschneider DE (1985) Changes in heritability estimates with age and site in white spruce, Picea glauca (Moench) Voss. Silvae Genet 34:34–41 Nikles DG (1989) The development of Pinus hybrids for operational use in Queensland. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry.

293

294

References

Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 272–282 Nikles DG (1993) Breeding methods for production of interspecific hybrids in clonal selection and mass propagationprogrammes in the tropics and sub-tropics. In: Davidson J (ed) Proceedings of the regional symposium on recent advances in mass clonal multiplication of forest trees for plantation programmes. RAS/91/004 field document no 4. UNDP/FAO Forest Tree Improvement Project, Los Baños, pp 218–252 Nikles DG (1995) Hybrids of the slash -Caribbean- central American pine complex: characteristics, bases of superiority and potential utility in South China and elsewhere. In: Shen X (ed) Forest tree improvement in the Asia-Pacific region. China Forestry Publishing House, Beijing, pp 168–186 Nikles DG, Newton RS (1984) Provenance Resource Stands (PRSs) of Pinus caribaea var. hondurensis in Queensland. In: Barnes RD, Gibson GL (eds) Provenance and genetic improvement strategies in tropical forest genetics. Commonwealth Forestry Institute, Oxford, and Forest Research Centre, Harare, pp 582–583 Noss RF (1993) Sustainable forestry or sustainable forests? In: Aplet GH, Johnson N, Olson JT, Sample VA (eds) Defining sustainable forestry. Island, Washington, pp 17–43 Novick RR, Dick CW, Lemes MR, Navarros C, Caccone A, Bermingham E (2003) Genetic structure of Mesoamerican populations of big-leaf mahogany (Swietenia macrophylla) inferred from microsatellite analysis. Mol Ecol 12:2885–2893 Nybom H (2004) Comparison of different nuclear DNA markers for estimating intraspecific genetic diversity in plants. Mol Ecol 13:1143–1155 Obayashi K, Tsumura Y, Ihara-Ujino T, Niiyama K., Tanouchi H, Suyama Y, Washitani I, Lee C-T, Lee SL, Muhammad N (2002) Genetic diversity and outcrossing rate between undisturbed and selectively logged forests of Shorea curtisii (Dipterocarpaceae) using microsatellite DNA analysis. Int J Plant Sci 163:151–158 Oddou-Muratorio S, Guesnet D, Ozdemir E, Petit RJ, Demesure B (2001) Patterns of seed dispersal in a scattered forest tree species (Sorbus torminalis) based on multiscale investigation of population genetic structure for chloroplast DNA. In: MüllerStarck G, Schubert R (eds) Genetic response of forest systems to changing environmental conditions, Kluwer, Dordrecht, pp 271–280 Okafor JC, Lamb A (1994) Fruit trees: diversity and conservation strategies. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 34–41 O’Malley DM, Buckley DP, Prance GT, Bawa KS (1988) Genetics of Brazil nut (Bertholletia excelsa Humb. & Bonpl.: Lecythidaceae). Theor Appl Genet 76:929–932 Opler PA, Bawa KS (1978) Sex ratios in tropical forest trees. Evolution 32:812–821 Opler PA, Baker HG, Frankie GW (1975) Reproductive biology of some Costa Rican Cordia species (Boraginaceae). Biotropica 7:234–247 Otegbeye GO (1985) Provenance productivity in Eucalyptus camaldulensis Dehnh. and its implications to genetic improvement in the Savanna region of Nigeria. Silvae Genet 34:121–126 Otegbeye GO (1998) Forestry mating design and progeny testing: principles, methods and application. In: Mandal AK, Gibson GL (eds) Forest genetics and tree breeding. CBS, New Dehli, pp 63–86 Otero JT, Ackerman JD, Bayman P (2004) Differences in mycorrhizal preferences between two tropical orchids. Mol Ecol 13:2393–2404

References

Ouédraogo G (1997) Genetische Variation und Reproduktionssystem in Populationen von Acacia macrostachya Reichenbach x Benth (Mimosaceae) in Burkina Faso. Institute of Forest Genetics and Forest Tree Breeding, Göttingen Owino F, Oduol PA, Esegu F (1994) Domestication of Sesbania sesban for agroforestry system in eastern and southern Africa. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 205–209 Pacheco LF, Simonetti JA (2000) Genetic structure of a mimosoid tree deprived of its seed disperser, the spider monkey. Conserv Biol 14:1766–1775 Pakkanen A, Nikkanen T, Pulkkinen P (2000) Annual variation in pollen contamination and outcrossing in a Picea abies seed orchard. Scand J For Res 15:399-404 Palmberg C, Esquinas-Alcazar JT (1990) The role of the United Nations agencies and other international organizations in the conservation of plant genetic resources. For Ecol Manage 35:171–197 Palmer JR (1994) Designing commercially promising tropical timber species. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 16–24 Pandey M, Gailing O, Leinemann L, Finkeldey R (2004) Molecular markers provide evidence for long-distance planting material transfer during plantation establishment of Dalbergia sissoo Roxb. in Nepal. Ann For Sci 61:603–606 Park YK, Son SH (1996) Forest genetic conservation based on in vitro culture systems. In: Park YK, Sakamoto S (eds) Biodiversity and conservation of plant genetic resources in Asia. Japan Scientific Society Press, Tokyo, pp 153–173 Park YK, Bonga JM, Mullin TJ (1998) Clonal forestry. In: Mandal AK, Gibson GL (eds), Forest genetics and tree breeding, CBS, New Dehli, pp 143–167 Pascarella JB (1997) The mating system of the tropical understory shrub Ardisia escallonioides (Myrsinaceae). Am J Bot 84:456–460 Pastorino MJ, Gallo LA, Hattemer HH (2004) Genetic variation in natural populations of Austrocedrus chilensis, a cypress of the Andean-Patagonian forest. Biochem Syst Ecol 32: 993–1008 Paton DC (2000) Disruption of bird-plant pollination systems in Southern Australia. Conserv Biol 14:1232–1234 Paule L, Lindgren D, Yazdani R (1993) Allozyme frequencies, outcrossing rate and pollen contamination in Picea abies seed orchards. Scand J For Res 8:8–17 Petit R, El Mousadik A, Pons O (1998) Identifying populations for conservation on the basis of genetic markers. Conserv Biol 12:844–855 Petit R, Brewer S, Bordács S, Burg K, Cheddadi R, Coart E, Cottrell J, Csaikl U, van Dam B, Deans D, Espinel S, Fineschi S, Finkeldey R, Glaz I, Goicoechea PG, Jensen JS, König AO, Lowe AJ, Madsen SF, Mátyás G, Munro RC, Popescu F, Slade D, Tabbener H, de Vries SGM, Ziegenhagen B, de Beaulieu J-L, Kremer A (2002) Identification of refugia and post-glacial colonisation routes of European white oaks based on chloroplast DNA and fossil pollen evidence. For Ecol Manage 156:49–74 Phillips MA, Brown AHD (1977) Mating system and hybridity in Eucalyptus pauciflora. Aust J Biol Sci 30:337-344 Pinyopusarerk K, Williams ER (2000) Range-wide variation in growth and morphological characteristics of Casuarina equisetifolia grown in Northern Australia. For Ecol Manage 134:219–232

295

296

References

Poke FS, Vaillancourt RE, Potts BM, Reid JB (2005) Genomic research in Eucalyptus. Genetica 125:79–101 Prabhu R, Ruitenbeek HJ, Boyle TJB, Colfer CJP (2001) Between voodoo science and adaptive management: the role and research needs for indicators of sustainable forest management. In: Raison RJ, Brown AG, Flinn DW (eds) Criteria and indicators of sustainable forest management. CABI, Wallington, pp 39–66 Prance GT (1985) The pollination of Amazonian plants. In: Prance GT, Lovejoy TE (eds) Key environments. Amazonia. Pergamon, Oxford, pp 166–191 Prance GT (1994) The resource of usful tree species: identification of priorities for domestication. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 7–15 Primack R, Corlett R (2005) Tropical rain forests. Blackwell, Oxford Prober SM, Brown AHD (1994) Conservation of grassy white box woodlands: population genetics and fragmentation of Eucalyptus albens. Conserv Biol 8:1003–1013 Proctor M, Yeo P, Lack A (1996) The natural history of pollination. HarperCollins, London Pryor LD, Byrne GR (1969) Variation and taxonomy in Eucalyptus camaldulensis. Silvae Genet 18:64–71 Putz FE, Romero C (2001) Biologists and timber certification. Conserv Biol 15:313–314 Putz FE, Blate GM, Redford KH, Fimble R, Robinson J (2001) Tropical forest management and conservation of biodiversity: an overview. Conserv Biol 15:7–20 Quaile DR, Mullin LJ (1984) Provenance and progeny testing in Eucalyptus camaldulensis in Zimbabwe. In: Barnes RD, Gibson GL (eds) Provenance and genetic improvement strategies in tropical forest genetics. Commonwealth Forestry Institute, Oxford, and Forest Research Centre, Harare, pp 438–450 Rabouam C, Comes AM, Bretagnolle V, Humbert J-F, Periquets G, Bigot Y (1999) Features of DNA fragments obtained by random amplified polymorphic DNA (RAPD) assays. Mol Ecol 8:493–503 Raison RJ, Flinn DW, Brown AC (2001) Application of criteria and indicators to support sustainable forest management: some key issues. In: Raison RJ, Brown AG, Flinn DW (eds) Criteria and indicators for sustainable forest management. CABI, Wallington, pp 5–18 Ramirez N, Sobrevila C, de Enrech NX, Ruiz-Zapata T (1984) Floral biology and breeding system of Bauhinia benthamiana Taub. (Leguminosae), a bat-pollinated tree in Venezuelan ‘Llanos’. Am J Bot 71:273–280 Rautner M (2001) Designer trees. Biotechnol Dev Moni 44:2–7 Raven PH (1988) Our diminishing tropical forests. In: Wilson EO, Peter FM (eds) Biodiversity. National Academy Press, Washington, pp 119–122 Raymond CA, Lindgren D (1990) Genetic flexibility – a model for determining the range of suitable environments for a seed source. Silvae Genet 39:112–120 Reddy KV, Rockwood DL (1989) Breeding strategies for coppice production in a Eucalyptus grandis base population with four generations of selection. Silvae Genet 38: 148–151 Reed DH, Frankham R (2001) How closely correlated are molecular and quantitative measures of genetic variation? A meta-analysis. Evolution 55:1095–1101 Reed DH, Frankham R (2003) Correlation between fitness and genetic diversity. Conserv Biol 17:230–237

References

Regal PJ (1982) Pollination by wind and animals: ecology of geographic patterns. Annu Rev Ecol Syst 13:497–524 Remington DL, Purugganan MD (2003) Candidate genes, quantitative trait loci, and functional trait evolution in plants. Int J Plant Sci 164:7–20 Richards C (2000) Genetic and demographic influences on population persistence: gene flow and genetic rescue in Silene alba. In: Young AG, Clarke GM (eds) Genetics, demography and viability of populations. Cambridge University Press, Cambridge, pp 271–291 Richards PW (1981) The tropical rain forest. Cambridge University Press, Cambridge Richardson JE, Pennington RT, Pennington TD, Hollingsworth PM (2001) Rapid diversification of a species-rich genus of neotropical rain forest trees. Science 293:2242–2245 Ritland K (1996a) Inferring the genetic basis of inbreeding depression in plants. Genome 39:1–8 Ritland K (1996b) A marker-based method for inferences about quantitative inheritance in natural populations. Evolution 50:1062–1073 Ritland K, El-Kassaby YA (1985) The nature of inbreeding in a seed orchard of Douglas fir as shown by an efficient multilocus model. Theor Appl Genet 71:375–384 Ritland K, Jain S (1981) A model for the estimation of outcrossing rate and gene frequencies using n independent loci. Heredity 47:35–52 Roberds JH, Namkoong G, Skrøppa T (1990) Genetic analysis of risk in clonal populations of forest trees. Theor Appl Genet 79:841–848 Roche LR (ed) (1975) The methodology of conservation of forest genetic resources. Report on a pilot study. FAO, Rome Rossetto M, McNally J, Henry RJ, Hunter J, Matthes M (2000) Conservation genetics of an endangered rainforest tree (Fontainea oraria – Euphorbiaceae) and implications for closely related species. Conserv Genet 1: 217–229 Rotach P (2000) Zur Gefährdung seltener Baumarten in der Schweiz: Grundsätzliche Überlegungen, Situationsanalyse und zwei Fallbeispiele. For Snow Landscape Res 75:267–284 Rothe GM (1994) Electrophoresis of enzymes. Laboratory methods. Springer, Berlin Heidelberg New York Russel JR, Weber JC, Booth A, Powell W, Sotelo-Montes C, Dawson IK (1999) Genetic variation of Calycophyllum spruceanum in the Peruvian Amazon Basin, revealed by amplified fragment length polymorphism (AFLP) analysis. Mol Ecol 8:199–204 Sa-Ardavut P, Chairuangsirikui T, Chanpaisaeng S (1989) In situ gene conservation in northeast Thailand. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 291–297 Saatchi S, Agosti D, Alger K, Delabie J, Musinsky J (2001) Examining fragmentation and loss of primary forest in the southern Bahian Atlantic Forest of Brarazil with radar imagery. Conserv Biol 15:867–875 Sakai KI (1985) Studies on breeding structure in two tropical tree species. In: Gregorius H-R (ed) Population genetics in forestry. Springer, Berlin Heidelberg New York, pp 212–225 Salwasser H, MacCleery DW, Snellgrove TA (1993) An ecosystem perspective on sustainable forestry and new directions for the US National Forest system. In: Aplet GH,

297

298

References

Johnson N, Olson JT, Sample VA (eds) Defining sustainable forestry. Island, Washington, pp 44–89 Sampson JF, Hopper SD, James SH (1989) The mating system and population genetic structure in a bird-pollinated mallee, Eucalyptus rhodantha. Heredity 63:383–393 Sanchez-Azofeifa GA, Quesada-Mateo C, Dayanandan S, Bawa K (1998) Protected areas and conservation of biodiversity in the tropics. Conserv Biol 13:407–411 Savolainen O, Kärkkäinen K, Kuittinen H (1992) Estimating numbers of embryonic lethals in conifers. Heredity 69:308–314 Sayer JA, Whitmore TC (1990) Tropical moist forests: destruction and species extinction. Biol Conserv 55:199–213 Schaal BA, Hayworth DA, Olsen KM, Rauscher JT, Smith WA (1998) Phylogenetic studies in plants: problems and prospects. Mol Ecol 7:465–474 Schlichting CD (1986) The evolution of phenotypic plasticity in plants. Annu Rev Ecol Syst 17:667–693 Schlichting CD (2002) Phenotypic plasticity in plants. Plant Species Biol 17:85–88 Schmidt L (1993a) Selection of plus-trees. Guidelines on selection practices in tree improvement. Field manual no 2 RAS/91/004. UNDP/FAO Forest Tree Improvement Project, Los Baños Schmidt L (1993b) Seed orchards. Guidelines on Establishment and Management Practices. Field manual no 4 RAS/91/004. UNDP/FAO Forest Tree Improvement Project, Los Baños Schmidt L (1993c) Seed stands. Guidelines on Establishment and Management Practices. Field manual no 3 RAS/91/004. UNDP/FAO Forest Tree Improvement Project, Los Baños Schmidt L (2000) Guide to handling of tropical and subtropical forest seed. Danida Forest Seed Centre, Humlebæk Schmidt-Adam G, Young A, Murray BGM (2000) Low outcrossing rates and shift in pollinators in New Zealand pohutukawa (Metrosideros excelsa; Myrtaceae). Am J Bot 87:1265–1271 Schonewald-Cox CM, Chambers SM, MacBryde B, Thomas L (eds) (1983) Genetics and conservation. A reference for managing wild animal and plant populations. Benjamin/Cummings, Menlo Park Schwartz MW, Hermann SM, van Mantgem PJ (2000) Population persistence in Florida torreya: comparing modeled projections of a declining coniferous tree. Conserv Biol 14:1023–1033 Schwartzman S, Moreira A, Nepstad D (2000) Rethinking tropical forest conservation: perils in parks. Conserv Biol 14:1351–1357 Sedgley M, Harbard J, Smith R-M, Wickneswari R, Griffin AR (1992) Reproductive biology and interspecific hybridization of Acacia mangium and Acacia auriculiformis A. Cunn. ex Benth. (Leguminosae: Mimosoideae). Aust J Bot 40:37–48 Shaffer ML (1981) Mimimum population sizes for species conservation. Bioscience 31:131–134 Shapcott A (1999) Comparison of the population genetics and densities of five Pinanga palm species at Kuala Belalong, Brunei. Mol Ecol 8: 1641–1654 Shelbourne CJA (1992) Genetic gains from different kinds of breeding and propagating populations. S Afr For J 160:49–65

References

Sherwin WB, Moritz C (2000) Managing and monitoring genetic erosion. In: Young AG, Clarke GM (eds) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge, pp 9–34 Sim BL (1984) The genetic base of Acacia mangium Willd. in Sabah. In: Barnes RD, Gibson, GL (eds) Provenance and genetic improvement strategies in tropical forest genetics. Commonwealth Forestry Institute, Oxford, and Forest Research Centre, Harare, pp 597–603 Simons AJ, MacQueen, Stewart JL (1994) Strategic concepts in the domestication of non-industrial trees. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 91–102 Simpson D (1998) Selection of superior trees. In: Puri S (ed) Tree improvement. Applied research and technology transfer. Science, Enfield, pp 107–124 Sinclair FL, Verinumbe I, Hall JB (1994) The role of tree domestication in agroforestry. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 124–136 Siregar IZ, Hattemer HH (2001) Gene flow and mating system in a seedling seed orchard and a natural stand of Pinus merkusii Jung. et de Vriese in Indonesia. In: Müller-Starck G, Schubert R (eds) Genetic response of forest systems to changing environmental conditions. Kluwer, Dordrecht, pp 281–292 Skrøppa T, Johnsen Ø (1999) Patterns of adaptive genetic variation in forest tree species; the reproductive environment as an evolutionary force in Picea abies. In: Mátyás C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 49–58 Smith NJH, Williams JT, Plucknett DL, Talbot JP (1992) Tropical forests and their crops. Cornell University Press, New York Smits WTM (1993) Mass propagation of dipterocarps by vegetative methods in Indonesia. In: Davidson J (ed) Proceedings of the regional symposium on recent advances in mass clonal multiplication of forest trees for plantation programmes. RAS/91/004 field document no 4. UNDP/FAO Forest Tree Improvement Project, Los Baños, pp 38–45 Smits WTM, de Fraiture AC, Yasman I (1994) Production of dipterocap planting stock by cuttings in Indonesia. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 267–272 Smouse PE, Dyer RJ, Westfall RD, Sork VL (2001) Two-generation analysis of pollen flow across a landscape. I. Male gamete heterogeneity among females. Evolution 55:260–271 Sneath PHA, Sokal RR (1973) Numerical taxonomy. Freeman, San Francisco Sokal RR, Rohlf FJ (1998) Biometry, 3rd edn. Freeman, New York Soonhuae P, Boyle T, Yeh F (1995) The population genetics of Dalbergia cochinenesis Pierre and implications for genetic conservation. In: Baradat P, Adams WT, MüllerStarck G (eds) Population genetics and genetic conservation of forest trees. SPB, Amsterdam, pp 371–385 Sork VL, Campbell D, Dyer R, Fernandez J, Nason J, Petit R, Smouse P, Steinberg E (1998) Proceedings from a workshop on gene flow in fragmented, managed, and continuous populations. Research paper no 3. National Center for Ecological Analysis and

299

300

References

Synthesis, Santa Barbara. http://www.nceas.ucsb.edu/nceas-web/projects/2057/nceaspaper3/ Sotelo Montes C, Vidaurre H, Weber J (2003) Variation in stem-growth and branchwood traits among provenances of Calycophyllum spruceanum Benth. from the Preruvian Amazon. New For 26:1–16 Soulé ME (ed) (1986) Conservation biology. The science of scarcity and diversity. Sinauer, Sunderland Soulé ME (1987) Introduction. In: Soulé ME (ed) Viable populations for conservation. Cambridge University Press, Cambridge, pp 1–10 Soulé ME (1991) Conservation tactics for a constant crisis. Science 253:744–750 Sousa VA, Hattemer HH (2003) Pollen dispersal and gene flow by pollen in Araucaria angustifolia. Aust J Bot 51:309–317 Sousa VA, Sebbenn AM, Hattemer HH, Ziehe M (2005) Correlated mating of a dioecious Brazilian conifer, Araucaria angustifolia (Bert.) O. Ktze. For Genet 12:107–119 Spichiger R, Palese R, Chauems A, Ramella L (1995) Origin, affinities and diversity hot spots of the Paraguayan dendrofloras. Candollea 50:515–537 Squillace AE (1974) Average genetic correlations among offsping from open-pollinated forest trees. Silvae Genet 23:149–153 Squillace AE, Schreuder HT, van Vredenburch CLH (1975) Genetic variation and breeding of Scots pine in the Netherlands. For Sci 21:341–352 Squirrell J, Hollingsworth PM, Woodhead M, Russell J, Lowe AJ, Gibby M, Powell W (2003) How much effort is required to isolate nuclear microsatellites from plants? Mol Ecol 12:1339–1348 Stacy EA, Hamrick JL, Nason JD, Hubbell SP, Foster RB, Condit R (1996) Pollen dispersal in low-density populations of three neotropical tree species. Am Nat 148:275–298 Starke R (1996) Die Reproduktion der Buche (Fagus sylvatica L.) unter verschiedenen waldbaulichen Gegebenheiten. In: Müller-Starck G (ed) Biodiversität und Nachhaltige Forstwirtschaft. Ecomed, Landsberg, pp 135–159 Steininger MK, Tucker CJ, Ersts P, Killeen TJ, Villegas Z, Hecht SB (2001) Clearance and fragmentation of tropical deciduous forest in the Tierras Bajas, Santa Cruz, Bolivia. Conserv Biol 15:856–866 Stern K (1963) Herkunftsversuche für Zwecke der Forstplanzenzüchtung, erläutert am Beispiel zweier Modellversuche. Zuechter 34:181–219 Stern K, Gregorius HR (1972) Schätzungen der effektiven Populationsgröae bei Pinus silvestris. Theor Appl Genet 42:107–110 Stern K, Roche L (1974) Genetics of forest ecosystems. Springer, Berlin Heidelberg New York Strabala TJ (2004) Expressed sequence tag databases from forest tree species. In: Kumar S, Fladung M (ed) Molecular genetics and breeding of forest trees. Haworth, Binghamton, pp 19–51 Stuart AN, Marshall AG (1976) Nectarivorous bats as pollinators of trees in West Malaysia. In: Burley J, Styles BT (eds) Tropical trees; variation, breeding and conservation. Academic, London, pp 141–150 Styles BT (1972) Flower biology of the Meliaceae and its bearing on tree breeding. Silvae Genet 21:175–182 Szmidt AE, Wang XR (2000) Genetic markers in forest genetics–the tunnel remains dark. In: Mátyás C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 31–48

References

Tangmitcharoen S, Owens JN (1997) Floral biology, pollination, pistil receptivity and pollen tube growth of teak (Tectona grandis Linn f.). Ann Bot 79:227–241 Telles MPC, Coelho ASG, Chaves LJ, Diniz-Filho JAF, D’Ayala Valva F (2003) Genetic diversity and population structure of Eugenia dysenterica DC. (‘cagaiteira’Myrtaceae) in Central Brazil: Spatial analysis and implications for conservation and management. Conserv Genet 4:685–695 Terborgh J (1986) Keystone plant resources in the tropical forest. In: Soulé, ME (ed) Conservation biology. The science of scarcity and diversity. Sinauer, Sunderland, pp 330–344 Tewari DN (1992) A Monograph on Teak (Tectona grandis Linn. f.). International Book Distributors, Dehra Dun Thamarus KA, Groom K, Murrel J, Byrne M, Moran GF (2002) A genetic linkage map for Eucalyptus globulus with candidate loci for wood, fibre, and floral traits. Theor Appl Genet 104:379–387 The Angiosperm Phylogeny Group (2003) An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Bot J Linn Soc 141:399–436 The Arabidopsis Genome Initiative (2000) Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408:796–815 Thomas WW (1999) Conservation and monographic research on the flora of tropical America. Biodivers Conserv 8:1007–1015 Thomson JD, Herre EA, Hamrick JL, Stone JL (1991) Genetic mosaics in strangler fig trees: implications for tropical conservation. Science 254:1214–1216 Thrall PH, Burdon JJ, Murray BR (2000) The metapopulation paradigm: a fragmented view of conservation biology. In: Young AG, Clarke GM (eds) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge, pp 75–95 Tompsett PB (1994) Capture of genetic resources by collection and storage of seed: a physiological approach. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the rebuilding of forest resources. Her Majesty’s Stationery Office, London, pp 61–71 Trapnell DW, Hamrick JL, Nason JD (2004) Thee-dimensional fine-scale genetic structure of the neotropical epiphytic orchid, Laelia rubescens. Mol Ecol 13:1111–1118 Tsumura Y, Kawahara T, Wickneswari R, Yoshimura K (1996) Molecular phylogeny of Dipterocarpaceae in Southeast Asia using RFLP of PCR-amplified chloroplast genes. Theor Appl Genet 93:22–29 Turesson G (1922) The genotypical response of the plant species to the habitat. Hereditas 3:211–350 Ujino T, Kawahara T, Tsumura Y, Nagamitsu T, Yoshimaru H, Wickneswari R (1998) Development and polymorphism of simple sequence repeat DNA markers for Shorea curtisii and other Dipterocarpaceae species. Heredity 81:422–428 Ulrich B (1981) Zur Stabilität von Waldökosystemen. Forstarchiv 52:165–170 van Buijtenen JP (1975) Advanced generation breeding. In: Proceedings of the 13th southern forest tree improvement conference, Raleigh, Southern Forest Tree Improvement Committee, pp 63–72 van Buijtenen JP (1976) Mating designs.In: Proceedings of the IUFRO joint meeting of genetic working parties on advanced generation breeding. INRA, Bordeaux, pp 11–29

301

302

References

van Wyk G (1977) Pollen handling, controlled pollination and grafting of Eucalyptus grandis. S Afr For J 101:47–53 van Wyk G (1981) Inbreeding effects in Eucalyptus grandis in relation to degree of relatedness. S Afr For J 116:60–63 van Wyk G, Schönau APG, Schön PP (1989) Growth potential and adaptability of young eucalypt hybrids in South Africa. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 325–333 Vavilov NI (1992) Origin and geography of cultivated plants. Cambridge University Press, Cambridge Vavilov NI (1996) Five continents. IPGRI, Rome Vellend M (2005) Species diversity and genetic diversity: parallel processes and correlated patterns. Am Nat 166:199–215 Vendramin GG, Lelli L, Rossi P, Morgante M (1996) A set of primers for the amplification of 20 choloroplast microsatellites in Pinaceae. Mol Ecol 5:585–598 Verkerke W (1989) Structure and function of the fig. Experientia 45:612–622 Viana VM, Tabanez AAJ, Batista LF (1997) Dynamics and restoration of forest fragments in the Brazilian Atlantic moist forest. In: Laurance WF, Bierregaard RO (eds) Tropical forest remnants: ecology, management, and conservation of fragmented communities. University of Chicago Press, Chicago, pp 351–365 Viera MF, de Carvalho-Okano RM (1996) Pollination biology of Mabea fistulifera (Euphorbiaceae) in Southeastern Brazil. Biotropica 28:61–68 Villaincourt RE, Potts BM, Watson M, Volker PW, Hodge GR, Reid JB, West AK (1995) Detection and prediction of heterosis in Eucalyptus globulus. For Genet 2:11–19 von Gadow K (2005) Forsteinrichtung. Analyse und Entwurf der Waldentwicklung. Universitätsverlag Göttingen, Göttingen von Gadow K, Kleinn C (2005) Managing forest ecosystems. In: Peterson CE, Maguire DA (eds) Balancing ecosystem values – innovative experiments for sustainable forestry. USDA Forest Service general technical report PNW-GTR-635, pp 15–24 von Humboldt A (1808) Ansichten der Natur. Neuausgabe 2004. Eichborn, Frankfurt Vornam B, Kuchma O, Kuchma N, Arkhipov A, Finkeldey R (2004) SSR markers as tools to reveal mutation events in Scots pine (Pinus sylvestris L.) from Chernobyl. Eur J For Res 123:245–248 Vornam B, Blödner C, Schubert R, Baumann R, Müller-Starck G, Hattemer HH, Polle A (2003) Preliminary results of expression profiling of spruce (Picea abies (L.) Karst.) in response to drought stress. In: Galris D (ed) Free radicals and oxidative stress: chemistry, biochemistry and pathophysiological implications. Proceedings of the meeting of the society for free radical research – European section, Ioannina, Greece, 26–29 June 2003. Medimont, Bologna, pp 109–113 Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Hornes M, Frijters A, Pot J, Peleman J, Kuiper M, Zabeau M (1995) AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res 23:4407–4414 Wang K (2001) Gene flow and mating system in European beech (Fagus sylvatica L.). Cuvillier, Göttingen Warren JM, Kalai, Misir S (1995) An unnatural breeding system polymorphism in cacao (Theobroma cacao, Sterculiaceae) in Trinidad. Am J Bot 82:1126–1130

References

Webb CJ, Bawa KS (1983) Pollen dispersal by hummingbirds and butterflies: a comparative study of two lowland tropical forest plants. Evolution 37:1258–1270 Weising K, Gardner RC (1999) A set of conserved PCR primers for the analysis of simple sequence repeat polymorphisms in chloroplast genomes of dicotyledonous angiosperms. Genome 42:9–19 Weising K, Nybom H, Wolff K, Kahl G (2005) DNA fingerprinting in plants. Principles, methods, and applications. Taylor and Francis, Boca Raton Wellendorf H, Kaosa-Ard A (1988) Teak improvement strategy in Thailand. DSR, Copenhagen Wells OO, Wakeley PC (1966) Geographic variation in survival, growth and fusiformrust infection of planted loblolly pine. For Sci Monogr 11:1–40 Wells OO, Wakeley PC (1970) Variation in longleaf pine from several geographic sources. For Sci 16: 28–42 Westfall RD (1992) Developing seed transfer zones. In: Fins L, Friedman ST, Brotschol JV (eds) Handbook of quantitative forest genetics. Kluwer, Dordrecht, pp 313–398 Wheeler NC, Jech KS (1992) The use of electrophoretic markers in seed orchard research. New For 6:311–328 Whelan RJ, Ayre DJ, England PR, Llorens T, Beynon F (2000) Ecology and genetics of Grevillea (Proteaceae): implications for conservation of fragmented populations. In: Young AG, Clarke GM (eds) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge, pp 253–269 White G, Powell W (1997) Isolation and characterization of microsatellite loci in Swietenia humilis (Meliaceae): an endangered tropical hardwood species. Mol Ecol 6:851–860 White GM, Boshier DH (2000) Fragmentation in Central American dry forests: genetic impacts on Swietenia humilis (Meliaceae). In: Young AG, Clarke GM (eds) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge, pp 293–311 White GM, Boshier DH, Powell W (1999) Genetic variation within a fragmented population of Swietenia humilis Zucc. Mol Ecol 8:1899–1909 White TL (1987) A conceptual framework for tree improvement programs. New For 4: 325–342 Whitmore TC (1975) Tropical rain forest of the far east. Clarendon, Oxford Whitmore TC (1998) An introduction to tropical rain forests. Oxford University Press, Oxford Wickneswari R, Boyle TJ (2000) Effects of logging and other forms of harvesting on genetic diversity in humid tropical forests. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford, pp 115–122 Wickneswari R, Norwati M (1993) Genetic diversity of natural populations of Acacia auriculiformis. Aust J Bot 41:65–77 Wickneswari R, Norwati M (1995) Spatial heterogeneity of outcrossing rates in Acacia auriculiformis A. Cunn. ex Benth. in Australia and Papua New Guinea. In: Baradat P, Adams WT, Müller-Starck G (eds) Population genetics and genetic conservation of forest trees. SPB, Amsterdam, pp 329–337 Wickneswari R, Lee SL, Din M (1995) Utility of RAPD markers in detecting genetic variation in multiple seedlings of Hopea odorata Roxb. (Dipterocarpaceae). Malay For 58:42–50

303

304

References

Wickneswari R, Lee CT, Muhammad N, Boyle TJB (1999) Impact of logging on genetic diversity in humid tropical forests. In: Mátyás C (ed) Forest genetics and sustainablility. Kluwer, Dordrecht, pp 171–181 Wiebes JT (1979) Co-evolution of figs and their insect pollinators. Annu Rev Ecol Syst 10:1–12 Wilcox BA, Murphy DD (1985) Conservation strategy: the effect of fragmentation on extinction. Am Nat 125:879–887 Williams CG, Savolainen O (1996) Inbreeding depression in conifers: implications for breeding strategy. For Sci 42:102–117 Williams ER, Matheson AC, Harwood CE (2002) Experimental design and analysis for tree improvement. CSIRO, Melbourne Wilson EO (1988) The current state of biological diversity. In: Wilson EO, Peter FM (eds) Biodiversity. National Academy Press, Washington, pp 3–18 World Commission of the Environment and Development (1998) http://www.sustainablemeasures. com/Training/Indicators/index.html Wright JW (1976) Introduction to forest genetics. Academic, London Wright S (1931) Evolution in Mendelian populations. Genetics 16:97–159 Yang RC, Yeh FC, Yanchuk AD (1996) A comparison of isozyme and quantitative genetic variation in Pinus contorta ssp. latifolia by FST. Genetics 142:1045–1052 Yates CJ, Broadhurst LM (2002) Assessing limitations on population growth in two critically endangered Acacia taxa. Biol Conserv 108:13–26 Yeh F (1989) Isozyme analysis for revealing population structure for use in breeding strategies. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and genetic improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO Conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford, and Winrock International, Arlington, pp 119–131 Yeh F, Brune A, Cheliak WM, Chipman DC (1983) Mating system of Eucalyptus citriodora in a seed-production area. Can J For Res 13:1051–1055 Young A, Boshier D, Boyle T (eds) (2000) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford Young AG, Boyle TJ (2000) Forest fragmentation. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford,, pp 123–134 Young AG, Clarke GM (eds) (2000) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge Young AG, Merriam GH, Warwick SI (1993) The effects of forest fragmentation on genetic variation in Acer saccharum Marsh (sugar maple) populations. Heredity 71:277–289 Young AG, Boyle T, Brown T (1996) The population genetic consequences of habitat fragmentation for plants. Trends Ecol Evol 11:413–418 Young AG, Brown AHD, Zich FA (1999) Genetic structure of fragmented populations of the endangered daisy Rutidosis leptorrynchoides. Conserv Biol 13:256–265 Young AG, Brown AHD, Murray BG, Thrall PH, Miller CH (2000) Genetic erosion, restricted mating and reduced viability in fragmented populations of the endangered grassland herb Rutidosis leptorrhynchoides. In: Young AG, Clarke GM (eds) Genetics, demography and viability of fragmented populations. Cambridge University Press, Cambridge, pp 335–359

References

Zabala NQ (1994) Mass multiplication of clonal planting material and establishment of plantation of dipterocarps in the Phillipines. UNDP/FAO Forest Tree Improvement Project, Los Baños Zheng YQ, Ennos R (1997) Changes in the mating system of populations of Pinus caribaea Morelet var. caribaea under domestication. For Genet 4:209–215 Zheng YQ, Ennos RA (1999) Genetic variability and structure of natural and domesticated populations of Caribbean pine (Pinus caribaea Morelet). Theor Appl Genet 98:765–771 Ziehe M (2006) Genomische Assoziationen durch Selbst-und Fremdbefruchtung und ihre Bedeutung für die Interpretation genetischer Strukturen am Beispiel der Buche (Fagus sylvatica L.). Schr Forstl Fak Univ Gött Niedersächs Forstl Versuchsanst (in press) Ziehe M, Gregorius H-R (1996) Beurteilung der Gefährdung genetischer Ressourcen anhand von Streamerkmalen. In: Müller-Starck G (ed) Biodiversität und Nachhaltige Forstwirtschaft. Ecomed, Landsberg, pp 300–317 Ziehe M, Hattemer HH (1998) The significance of heterozygosity in tree breeding and conservation. For Tree Improv 26:3–25 Ziehe M, Hattemer HH (2002) Target diameter felling and consequences for genetic structures in a beech stand (Fagus sylvatica L.). In: von Gadow K, Nagel J, Saborowski J (eds) Continuous cover forestry. Kluwer, Dordrecht, pp 91–105 Ziehe M, Roberds JH (1989) Inbreeding depression due to overdominance in partially self-fertilizing plant populations. Genetics 121:861–868 Ziehe M, Gregorius H-R, Glock H, Hattemer HH, Herzog S (1989) Gene resources and gene conservation in forest trees: general concepts. In: Scholz F, Gregorius H-R, Rudin D (eds) Genetic effects of air pollutants in forest tree populations. Springer, Berlin Heidelberg New York, pp 173–185 Ziehe M, Starke R, Hattemer HH, Turok J (1998) Genotypische Strukturen in BuchenAltbeständen und ihren Samen. Allg Forst- Jagdz 169:91–99 Ziehe M, Hattemer HH, Müller-Starck G, Müller-Starck R (1999) Genetic structures as indicators for adaptation and adaptational potentials. In: Mátyás C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 75–89 Zobel B (1992) Vegetative propagation in production forestry. J For 90:29–33 Zobel B, Talbert J (1984) Applied forest tree improvement. Wiley, New York Zobel BJ, van Buijtenen JP (1989) Wood variation. Its causes and control. Springer, Berlin Heidelberg New York Zobel BJ, van Wyk G, Stahl P (1987) Growing exotic forests. Wiley, New York

305

Index

A Abies alba 248 Acacia albida 29 auriculiformis 31, 77, 193, 214, 246 crassicarpa 214 mangium 11, 29, 193, 214, 217 mearnsii 202, 215 nilotica 214 retinodes 84 Acer saccharum 118 skutchii 10, 29 Adansonia digitata 56 adaptability 3, 87, 142, 243 adaptation 21, 87, 113, 147, 163 evolutionary 88 physiological 87, 88 adaptedness 3, 20, 87, 155, 206 adaptive genetic marker 248 adaptive potential 23, 87, 247 adaptive trait expression 247ff. AFLP see amplified fragment length polymorphism agamospermy 51 Agathis spp. 131 Albertia edulis 48 allele 15, 17 deleterious 80 dominant 15 recessive 15 allelic richness 25 allelic structure 16, 17, 22 allochthonous 137, 148f. allogamy 45 allopatric 149 speciation 101, 104

Allophyllus occidentalis 48 Altingia excelsa 134 amplified fragment length polymorphism 9, 35, 189 Ancistrocladus korupensis 79 Andira inermis 57 androdioecy 44 anemogamy see pollination by wind Anguria spp. 58 Anisoptera spp. 104 Annonaceae 58 Apis spp. 62, 64 apomixis 51, 52 Araucaria angustifolia 55, 157, 214 araucana 151, 248f. cunninghamii 194, 214 Ardisia escallonioides 73, 76, 77 Aristolochia spp. 58 Astrocaryum mexicanum 77 Austrocedrus chilensis 249 autochthonous 137, 148 autogamy 44 Azadirachta indica 151, 165 B Bambusa vulgaris 50 barochory 66 base population 174, 208 basic material 221 basic rule of conservation 124 bats as fruit dispersers 117 as pollinators 56, 59, 121 Bauhinia spp. 56 pauletia 45 ungulata 45, 46

308

Index

bees as pollinators 57, 59, 62, 64 beetles as pollinators 58, 64 Beilschmiedia pendula 72, 77 Bernardia nicaraguensis 47 Bertholletia excelsa 77, 84 biparental inbreeding 133, 204, 225 birds as pollinators 57, 59, 117 Blastophaga quadraticeps 92 Bombacaceae 66 Bombacopsis quinata 212, 224 bottleneck 20, 122 breeding population 208 products 215, 231ff. progress 184, 186, 207 strategy 195, 207ff. Brosimum alicastrum 72, 77 budding 199 bugs as pollinators 59 bulbil 53 Bursera simarouba 47 tomentosa 47 butterflies as pollinators 58, 59 C Caesalpinia echinata 10, 68, 115 Calamus spp. 45, 48 Calliandra spp. 57 Calophyllum longifolium 63, 77 Calycophyllum spruceanum 10, 29, 151 candidate gene 7, 89 Carapa guianensis 77, 133 procera 77 Carica papaya 44 Caryocar brasiliense 8 Castanea sativa 191 Castanopsis spp. 45 Casuarina spp. 214, 220 equisetifolia 151, 165 nana 51 Castilla elastica 45 cauliflory 56 Cavanillesia platanifolia 72, 76, 77, 85 Cecropia spp. 55 obtusifolia 73 peltata 47

Cedrela odorata 10, 11, 37, 68, 77, 246 Cedrelinga catenaeformis 222 Ceiba pentandra 56, 72, 77 Ceropegia 58 certification 142, 221 Chlorophora tinctoria 47 chloroplast DNA 14, 18, 37, 68 sequences of 103 cicadas as pollinators 59 cistron 6, 7 Citrus spp. 52 clonal archive 234 forestry 204f., 229, 232 mixture 232ff. propagation 204f. seed orchard 55, 60, 74, 196ff., 223f. variation 164, 167 clone 50 cluster analysis 27 Coccoloba caracasana 48 floribunda 48 coevolution 90, 91 Coffea arabica 215f. Coleoptera 58 combining ability 185 conservation 88, 90 dynamic 251ff. ex situ 251, 253ff. of figs 96 in situ 251ff. objectives 241ff. static 251, 255ff. strategy 100 continuous-cover system 137 Cordia alliodora 50, 77, 84, 202 collococca 47, 84 dentata 84 inermis 84 panamensis 47, 84 pringlei 84 correlated matings 133, 138, 156f., 204 Corythophora alta 67, 140 cosexuality 42, 44 Cotylelobium spp. 104 Couroupita guianensis 56

Index

cpDNA see chloroplast DNA criteria of sustainable forest management 143f. Cryptomeria japonica 89, 232 Cunninghamia lanceolata 214, 232 Cupressus lutitanica 214 cutting propagation 224 Cycadaceae 58 D Daemorops verticillaris 131 Dalbergia sissoo 13, 17, 26, 151, 216 Darwin, C. 19, 100, 174f. demographic stochasticity 125 density of population 112, 116, 130f., 198, 223, 230 diallel mating design 186 diaspore 65 differentiation of provenances 153 parameters 247 Dinizia excelsa 64, 214 differentiation see genetic differentiation dioecy (dioecious) 43, 44, 46, 84, 132, 157 functional 94 Diospyros 45 Diptera 58 Dipterocarpaceae 12, 45, 65, 103ff. Dipterocarpus spp. 104 cf. condorensis 77, 79, 85 tempehes 62, 78 disease 90 disturbance 145f. distyly 83 DNA 7 sequencing 7 domestication 174ff. dominance 180f. dominance hypothesis 80 drift see genetic drift drought 151f., 248f. Dryobalanops spp. 104 aromatica 132, 204, 225 Dunalia arborescens 66

E Ebenaceae 46 ecological niche 101 ecotypic variation 164 edge effects 112, 131 endozoochory 66, 93 effective density of population 131, 133 effective number of alleles 25 effective size 116f. endosperm of conifers 14 Enterolobium cyclocarpum 118 environmental matching 166f., 216 epistatic effect 180 equilibrium 157, 181, 196, 221 Erythrina spp. 57 Erythroxylon rotundifolium 47 eucalypt plantations 81 Eucalyptus spp. 45, 80 albens 254 camaldulensis 89, 151, 169f., 214 citriodora 77, 214, 223 considensiana 139 deglupta 214 delegatensis 77, 81, 164, 249f. exserta 214 globulus 167, 192, 214 grandis 77, 81, 163, 192, 194, 202, 207, 214, 229 kitsoniana 77 obliqua 77 pauciflora 77 regnans 14, 77, 81 rhodantha 77 robusta 214 saligna 77, 194, 214 sieberi 139 stellulata 77 stoatei 77 tereticornis 165, 214 urophylla 150, 192, 194, 214 Eugenia dysenterica 33 evolution 19 cumulative 101 evolutionary factor 19 exotic species 148, 215 exozoochory 66

309

310

Index

expected heterozygosity see heterozygosity, expected experimental design 159ff. series 162 site 161ff. F factorial mating design 185 Fagaceae 45 Fagus sylvatica 49, 54, 88 family structure 67, 115, 117, 134, 136 Fertilization 54 Ficus spp. 45, 48, 57, 59, 90, 91ff., 106, 118, 251 flowers and inflorescence of 92 field trial 89, 151ff. fig wasp 57, 90, 92ff. fitness 20, 80 domestic 88, 175 fixation index 33 flies as pollinators 58 flowering phenology 117, 120, 197 flying foxes 114 Fontainea oraria 34 forest management 137 fragmentation 111ff. Fraxinus excelsior 45, 49 G gall flower 91 gamete 41 gametophyte 53 gametophytic self-incompatibility 83 geitonogamy 45 gene bank 231, 255f. gene flow 20, 53, 121, 137, 149, 168 actual 60 potential 60 gene locus 6, 13 gene marker 15, 23, 210 gene pool 24 gene resource 238ff. genetic conservation 238 genetic controlledness 176ff. genetic differentiation 27, 30, 35

genetic distance 27 genetic diversity 24, 25, 100 genetic drift 20, 29, 32, 100, 113, 118, 206, 216 genetic identity 200, 210, 234 genetic incompatibility 156, 242 genetic inventory 23, 245ff. genetic load 153, 168 genetic multiplicity 24 genetic structure 16 genetic variance components 182 genetic variation measurement of 24 of tropical forest trees 30 Genipa caruto 48 genotype 16, 21 genotypic structure 16, 17, 22 geographic variation pattern 163ff., 170 glaciation 67 Gliricidia spp. 10 sepium 8, 29 Gmelina arborea 83, 151, 215 gonosome 42 grafting 197 Grevillea spp. 214, 220 macleyana 114 Gurania spp. 58 gymnosperms 84 gynodioecy 44, 94f. H habitat corridors 126 haplodiploidy 96 haplotype 16, 18, 68 Hardy-Weinberg structure 26, 72, 78, 79, 90 Heliconia spp. 57 Heliconius spp. 58 hemiepiphyte 91 hermaphroditism 44, 46, 49, 82, 84 heterosis 192ff. heterozygosity 124, 192, 237 expected 26, 29, 78 observed 26, 79, 89, 90 heterozygous 15 Hevea brasiliensis 215

Index

homology 103 homoplasy 103 homozygous 15 honey bees 114, 202, 225 Hopea spp. 104 odorata 52 hummingbird 57 as pollinators 117 hybrid seed orchard 198 hybrid vigor 192 hybridization 193 hydrochory 66 Hymenoptera 43, 57 I imprinting 152 inbreeding 20, 73, 78, 96, 216 coefficient 33, 34, 79, 90, 117 depression 51, 78, 80, 89, 115, 134, 192, 216f. -effective size of population 116 incompatibility 43, 58, 75, 80, 82 indicators of sustainable forest management 143ff. indigenous population 148 individual effective pollen cloud 156 Inga spp. 84, 102, 104ff. inheritance 14 biparental 14 maternal 15, 18 paternal 15 study 12, 14 uniparental 14, 18 insect diversity 100 interaction 164, 179 internal transcribed spacer 103 Ionopsis utricularoides 91 Iryanthera macrophylla 44 isoenzyme see isozyme isolation 113, 115, 117, 197, 200, 202 isozyme 12 electrophoresis 12 gene locus 12, 30, 31, 33, 63, 73, 76, 81, 85

J Jacaranda copaia 77 Juglans neotropica 222 regia 51 K keystone species 91 L Laelia rubescens 50 land race 149f., 154, 165, 167, 194, 217, 230 Lecythidaceae 57 legislation 221 Leucaena spp. 213 leucocephala 193, 216, 232 linkage map 189ff. Lovoa trichilioides 219 M Mabea fistulifera 59 Macaranga spp. 103 mangrove 53 mating contact 115, 132 mating design 184ff. mating preference 42, 43, 96 mating, random 69 mating system 20, 22, 69, 132, 138 megagametophyte 14 Meliaceae 46 Mendel, G. 13 metapopulation 125, 139, 254f. Metrosideros excelsa 114 microsatellite 8, 33, 62, 64, 78, 89 migration 20, 53, 65, 118 minimum viable population 124ff. mitochondrial DNA 14 mixed mating system 76, 80, 81, 85 molecular marker 7 monoecy 44, 46, 84, 94 Monotes kerstingii 104 Moringa oleifera 10, 29 moths as pollinators 58, 62 mtDNA see mitochondrial DNA multiple-population breeding 195, 206f., 240f., 253

311

312

Index

mutation 19, 22 mycorrhiza 90 N natural forest 136 nature reserves 244, 252 neighborhood model 64 Neobalanocarpus heimii 8, 62 neophyte 148 neutral genetic marker 245ff. norm of reaction 87 O Ocotea tenera 45 open-pollinated progeny see single-tree progeny orang-utan 68 orchard, seed 224ff. orchid 50 origin 166, 216 outcrossing rate 73, 76, 77, 90, 132f., 203, 223ff. species 80 overdominance (hypothesis) 80, 181, 194 P Pachira quinata 117 oleaginea 51 Pachycereus pringlei 45 pair comparison in the field 177f. Palmae 58 panmixis 69, 72 Parashorea spp. 104 parent-offspring regression 196 Parkia spp. 56 pasture trees 116, 119, 126 paternity analysis 61 exclusion 120 Paulownia taiwaniana 232 PCR see polymerase chain reaction Pentaclethra macroloba 29 persistence of population 123ff., 126 phenology 94 phenotype 21

phenotypic plasticity 87 structure 21 variation 89 phylogenetic tree 103 phylogeny 99 molecular 91, 102ff. Picea abies 54, 71, 72, 138, 152, 210, 228, 230 glauca 224 Pinanga spp. 79 Pinus spp. 55, 89, 131 banksiana 194, 224 caribaea 29, 133, 151, 214, 223f. elliottii 214 kesiya 214 merkusii 14, 29, 76, 77, 163, 214, 225, 259ff. oocarpa 214 palustris 168 patula 214 pinaster 193f. strobus 149 sylvestris, 49, 54, 55, 149, 160, 194, 198, 228 taeda 168 tecunumanii 254 Pisonia macranthocarpa 47 Pithecellobium elegans 8, 123 planting stock 159, 227f. Platypodium elegans 72, 73, 77, 89 plus-tree selection 187f., 225 Podocarpaceae 45 Podocarpus spp. 55 pollen 54 allele frequency 69 cloud 69 contamination 198, 252 dispersal 118, 120f., 133 limitation 114, 204, 225 transport 52, 53, 60 vector 54, 59, 113, 125 pollination 54 by animals 55, 59, 65 by birds 56, 59 by insects 57, 59

Index

by wind 54, 59 controlled 13 trap 58 pollinator see pollen vector polymerase chain reaction 7 polyploidy 122 Pongo pygmaeus 68 population 5 density 28, 85, 100 size 20, 22 size, reproduction effective 29 Populus spp. 89 tremula 158 PPL see proportion of polymorphic loci progeny tests 183ff., 202f. propagation population 208 proportion of polymorphic loci 25, 29 Prosopis spp. 8 juliflora 50 protogyny 93 provenance 148 choice 165ff. hybrid 165 resource stand 223f. Prunus africana 10 Pseudotsuga menziesii 149, 225 Pterocarpus indicus 14, 72, 77, 79 macrocarpus 29, 77, 79 Q qualified reproductive material 222 quantitative trait locus (QTL) 189ff. Quararibea asterolepis 72, 77 Quercus spp. 54 macrocarpa 55 robur 191 R race 149 Rafflesia arnoldii 58 Randia spinosa 48 subcordata 48 random amplified polymorphic DNA 9, 35, 189, 192, 211 ranking of provenances 163

RAPD see random amplified polymorphic DNA recolonization, postglacial 67 recombination rate 189f. refugia 67, 100, 101 regeneration artificial 213ff. natural 136, 230 of forest genetic resources 257ff. region of provenance 148 relatedness 78 of neighbor trees 115, 134, 139 of seeds 117, 119, 138 remote pollen 115, 120, 137 reproduction 41 asexual 49 effective size 115f., 124, 126f. sexual 41 system 20 vegetative 50, 53 reproductive balance 115, 117, 157, 197 material 220ff. restriction fragment length polymorphism 10, 189 RFLP see restriction fragment length polymorphism Rhizophora mangle 45, 73 root sucker 50, 53 rotation forest management 137, 229 Ruprechtia costata 48 S Salicaceae 46 Samanea saman 119 sapromyiophily 58 seed collection 156ff., 226 seed dispersal 31 by bats 117 by birds 66 by mammals 66 by water 66 by wind 65 vectors 65 seedling seed orchard 201ff., 225 seed production area 165, 222f.

313

314

Index

seed source 148 seed storage 158, 219, 226f. seed transfer 167ff., 216f., 227, 229 seed vector 113, 125 selected reproductive material 221 selection 20, 73 artificial 180ff., 186ff., 209f. dysgenic 130, 135f. fertility 20, 22, 42, 43, 88 gametic 48 marker assisted 189ff. mass 135, 187 of clones 189, 195 of families 188 of genetic resources 243ff. viability 20, 22, 42, 43, 48, 84, 88, 179, 227, 229 selective logging 130 thinning 188, 202f., 230 self-fertilization 114ff., 132f., 156, 198, 225 self-incompatibility see incompatibility selfing 20, 45 rate 73, 80 self-pollination 115, 193, 197 self-sterility 82, 84 series of experiments 162, 248 Sesbania spp. 213, 220 sex determination 42 sex ratio 48 sexual function 49 structure 47 system 42, 44, 69 type 42 shelterwood system 138f. Shorea spp. 104, 106 congestiflora 77 curtisii 8 leprosula 29, 35, 67 megistophylla 77 ovalis ssp. sericea 52, 67 parvifolia 11, 35 trapezifolia 77 Simarouba glauca 48

simple sequence repeat see microsatellite single nucleotide polymorphism 7, 89 single-tree seed lots 154, 157, 183 size of forest fragments 112 S-locus 83 SNP see single nucleotide polymorphism solitary tree 115, 120, 126 Sorbus torminalis 67 Sorocea affinis 72, 77 source-identified reproductive material 222 spatial genetic structure 115, 121, 125, 156 species diversity 99ff. -elimination trial 154, 163, 218 interactions 90, 101 -provenance trial 154 richness 21 Spondias mombin 63, 77, 121 nigrescens 47 sporophytic self-incompatibility 83 SSR see microsatellite Stemmadenia donnel-smithii 77 Stemonoporus oblongifolius 52, 77 strangling fig 91 sustainability 141, 231 sustainable forest management 141f., 240 Swietenia spp. 214, 219 humilis 8, 118f., 121, 123 macrophylla 29, 33, 79, 123 syconium 91 symbiosis 90, 91 sympatric 149 Symphonia globulifera 73, 116, 126, 140 T Tachigali versicolor 72, 77 target-diameter felling 130, 135 Tectona grandis 10, 45, 60, 71, 72, 74, 83, 151, 160, 167, 170ff., 198ff., 214 flowers and inflorescence of 75 Terminalia spp. 214, 219 tested reproductive material 222 Theobroma cacao 84

Index

thinning 138 thrips as pollinators 59 Tolumnia variegata 91 tree improvement 88 Trichilia anisopleura 47 cuneata 47 tuberculata 72, 77 Trigona spp. 62 trioecy 44 Triplaris americana 48 Triplochiton scleroxylon 219, 228 triploidy 51 Turpinia occidentalis 63, 77 TwoGener method 64, 69 U unisexual 42 Upuna spp. 104 Uvaria elmeri 59 V variance-effective size 116f. variation parameters 246f.

Vatica spp. 104 vegetative propagation 195, 197, 228f., 232, 256f. verifiers of population status 144f. vertebrate pollinators 56 W Wallace, A. 19, 100 X xenogamy 45 Y Z Zanthoxylum setulosum 48 zoogamy see pollination by animals zygote 41

315

Tropical Forest Genetics (Tropical Forestry)

Tropical Forest Seed (Tropical Forestry)

Tropical Forest Ecology

Tropical Forest Community Ecology

Tropical Forest Seed

Forest Road Operations in the Tropics (Tropical Forestry)

Tropical Agroforestry (Tropical Agriculture)

Tropical Forest Ecology: The Basis for Conservation and Management (Tropical Forestry)

Harvesting Operations in the Tropics (Tropical Forestry)

Living in a Dynamic Tropical Forest Landscape

Tropical Sin

Tropical Terror

Tropical Sin

Tropical Rainforests

Plantation Technology in Tropical Forest Science

Tropical Freeze

Tropical Sin

Tropical Interiors

Tropical Eden

Tropical High

Tropical Forests

Tropical Mariculture

Tropical Getaway

Tropical Getaway

Tropical Sin

Tropical Forest Genetics (Tropical Forestry)

Tropical Forest Seed (Tropical Forestry)

Tropical Forest Ecology

Tropical Forest Community Ecology

Tropical Forest Seed

Forest Road Operations in the Tropics (Tropical Forestry)

Tropical Agroforestry (Tropical Agriculture)

Tropical Forest Ecology: The Basis for Conservation and Management (Tropical Forestry)

Harvesting Operations in the Tropics (Tropical Forestry)

Living in a Dynamic Tropical Forest Landscape

Tropical Sin

Tropical Terror

Tropical Sin

Tropical Rainforests

Plantation Technology in Tropical Forest Science

Tropical Freeze

Tropical Sin

Tropical Interiors

Tropical Eden

Tropical High

Tropical Forests

Tropical Mariculture

Tropical Getaway

Tropical Getaway

Tropical Sin

Recommend Documents