UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Estimation of genetic variation in Thailand rosewood (Dalbergia cochinchinensis Pierre) Soonhuae, Prachote 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1994-893885.pdf [ 2.01MB ]
Metadata
JSON: 831-1.0099229.json
JSON-LD: 831-1.0099229-ld.json
RDF/XML (Pretty): 831-1.0099229-rdf.xml
RDF/JSON: 831-1.0099229-rdf.json
Turtle: 831-1.0099229-turtle.txt
N-Triples: 831-1.0099229-rdf-ntriples.txt
Original Record: 831-1.0099229-source.json
Full Text
831-1.0099229-fulltext.txt
Citation
831-1.0099229.ris

Full Text

ESTIMATION OF GENETIC VARIATION IN THAILAND ROSEWOOD (DALBERGIA COCHINCHINENSIS PIERRE)  by Prachote Soonhuae B.Sc. (Forestry), Kasetsart University, 1972 M.Sc. (Forestry), Kasetsart University, 1979  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Faculty of Forestry)  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA December 1993 c  Prachote Soonhuae, 1993  In  presenting this thesis  in  partial fulfilment of the requirements  for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  (Signature)  Department of The University of British Columbia Vancouver, Canada Date  DE-6 (2/88)  7  /  y3  11  ABSTRACT  Genetic diversity conservation of tropical forest trees, particularly economically valuable and endangered species, is urgently needed for maintaining ecological balances and sources of materials for direct, indirect and optional uses and their existing values.  Thailand rosewood (Dalbergia  cochinchinensis Pierre) is an economically important tree species of Southeast Asia.  Undisturbed  populations in natural forest reserves were suspected to have different genetic variability from that of disturbed populations in farm and public areas. The objectives of this study were to evaluate genetic variation and the mating system of Thailand rosewood in order to consider conservation strategies for the species. The variation was studied using quantitative and isozyme characteristics. Using open-pollinated families, quantitative variation was assessed in seed and 9 months old seedlings from eight populations, and in four years old seedlings grown in a progeny trial from seven populations. The differences among families were significant for most traits in all of the material tested. However, significant differences among populations were relatively high for most traits. The variation between undisturbed and disturbed populations for all traits was not significant. Root tips of three day old germinants from eight populations provided the material for isozyme studies. Twelve enzyme systems coding eighteen loci were analyzed. The overall mean of expected heterozygosity was relatively high (0.229) and the mean number of alleles per locus was moderate (2.2). F-statistics and Nei’s genetic distances indicated moderate differentiation among the populations (F(ST)  =  0.127).  There was no significant difference between the undisturbed and disturbed  populations. The outerossing rates (t) estimated for the species were close to 1.0 for all populations. The results indicated high genetic variability within populations for both quantitative traits and isozymes, but the among population variability was relatively high for quantitative traits and moderate for isozymes. The species was highly outcrossed, with low inbreeding. Due to the fact that remnant populations or subpopulations of the species are naturally small and there is difference between  in populations from central and northeastern regions, about one thousand individuals collected from several populations would assure sufficient genetic viability for the conservation in each region. The conserved populations should also be distributed throughout the natural range of the species to maintain adaptability of desired quantitative traits. Both in situ and ex situ conservation are needed for this endangered species.  Disturbed  populations should also be maintained as seed sources. Ex situ conservation of about 1,000 individuals collected from subpopulations within each region and planted within the regional area would preserve the broad genetic base and prevent the random drift of the species. In addition, successful conservation requires strong support from government, involvement from local people, and international cooperation for technical and financial support.  iv  TABLE OF CONTENTS Page  ABSTRACT  ii  TABLE OF CONTENTS  iv  LIST OF TABLES  vii  LIST OF FIGURES  ix  LIST OF APPENDICES  x  LIST OF ABBREVIATIONS  xi  ACKNOWLEDGEMENTS  xiv  CHAPTER 1  INTRODUCTION  1  1. Conservation and Genetic Improvement  3  1. 1. Morphological variation  5  1.2. Biochemical variation  6  1.2.1. DNA variation  6  1.2.2. Isozyme variation  7  2. Forest Situation in Thailand  9  3. Thailand rosewood  12  4. Study Objectives  15  5. Study Materials  16  CHAPTER 2  VARIATION IN QUANTITATIVE TRAITS  1. Seed Trait Variation  21 21  1.1. Study methods  21  1.2. Data analysis  23  1.3. Results  24  V  1.3.1. Variation  24  1.3.2. Variances and heritabilities  26  1.3.3. Correlations  26  1.4. Discussion  27  2. Seedling Trait Variation  29  2.1. Study methods  29  2.2. Data analysis  31  2.3. Results  32  2.3.1. Variation  33  2.3.2. Variances and heritabilities  34  2.3.3. Correlations  35  2.4. Discussion 3. Field Progeny Trait Variation  39  3.1. Study methods  39  3.2. Data analysis  39  3.3. Results  39  3.3. 1. Variation  40  3.3.2. Variances and heritabilities  42  3.3.3. Correlations  45  3.4. Discussion 4. Conclusion  CHAPTER 3  35  ISOZYME VARIATION  45  46 48  1. Study Method  48  2. Data Analysis  49  2.1. Allelic variation  50  2.2. Mating system  53  vi 3. Results  53  3.1. Allelic variation  54  3.2. Mating system  73  4. Discussion  75  5. Conclusion  78  CHAPTER 4  CONSERVATION STRATEGIES  79  1. Quantitative Trait Consideration  79  2. Isozyme Consideration  81  3. Quantitative and Isozyme Correlations  82  3. Genetic Rationale  84  3.1. In situ conservation  85  3.2. Ex situ conservation  87  4. Socio-economic and Political Considerations  89  5. Conclusion  91  CHAPTER 5  CONCLUSIONS AND RECOMMENDATIONS  1. Conclusions  93 93  1.1. Quantitative variation  93  1.2. Isozyme variation  94  1.3. Conservation strategies  94  2. Recommendations  95  REFERENCES  97  APPENDICES  108  vii  LIST OF TABLES  Table  Page  1  Depletion of forest area in Thailand from 1961-1985  11  2  List of insect pests of Thailand rosewood  13  3  Environmental conditions of tree populations in which seed was collected for studying seed, seedling, and isozyme variation  4  Environmental conditions of tree populations in which seed was collected for establishing the provenance-progeny plantation  5  19  20  Analysis of Variance showing source of variation, degrees of freedom, and expected mean square of seed traits  22  6  Average seed trait measurements for eight populations  24  7  Germination, R50 and seed weight variation, variance components and family heritahilities for seed traits from eight populations  28  8  Analysis of Variance for seedling traits  30  9  Average measurements for seedling traits including height and diameter (HT and DM), total dry weight (TDW), and root/shoot dry weight ratio for each population  10  Height and diameter variation, variance components and heritabilities for 9-month nursery seedlings from eight populations  11  36  Dry weight variation, variance components and heritabilities for 9-month nursery seedlings from eight populations  12  34  37  Average measurements of progeny traits including height and diameter (HT and DM), persistence of axis (PER), stem straightness (STN), and branch heaviness (BR) ftr each population  41  viii 13  Height and diameter variation, variance components and heritabilities for 4-year field progeny from seven populations  14  Scored trait variation, variance components and heritabilities for 4-year field progeny from seven populations  15  43  44  Variability parameters including mean number of alleles per locus, percentage of loci polymorphic, and mean heterozygosities for direct-count and Hardy-Weinberg expectation at 18 loci for for the species level(data from all populations were combined and treated as one unit for analysis), eaach population, and the means averaged at each level(SE in parentheses)  58  16  Mean allele frequencies for 18 loci across 8 populations  60  17  Overall means of unbiased and direct-count heterozygosities for each locus based on data combined from eight populations  18  Chi-square test for deviation from Hardy-Weinberg equilibrium of individual loci in each population  19  65  Contingency chi-square analysis for deviation of allele frequencies from those of which were randomly sampled from the total population  20  63  67  Expected and observed heterozygosity and resulting fixation index for each locus based on the total population  69  21  Matrix of genetic distance and for similarity coefficients  70  22  F-statistics at all loci from 8 populations  71  23  Clustering analysis for Nei’s (1978) unbiased genetic distance using  24  unweighed pair group method algorithm (UPGMA)  72  Hierarchical analysis using F-statistics (Wright, 1987)  74  ix 25  Outcrossing rate from fifteen  (t)  and fixation index (f for each population estimated  loci  26  Correlation between seed and seedling traits  27  Correlations estimated from individual family means for heterozygosity, quantitative traits, and mating system  28  75 81  83  Correlations estimated from population means for heterozygosity, quantitative traits, and mating system  83  x  LIST OF FIGURES  Figure  Page  1  Location of populations in which seed was collected for the studies  17  2  Enzyme banding patterns for eighteen loci  55  3  Electrophoretic gel slices stained with two enzyme systems, PGM and IDH  4  57  Genetic variahihty including number of alleles per locus, probability of polymorphic loci, and observed and expected Hardy-Weinberg heterozygosities  5  from  eight loci for each population  Clustering tree using Nei’s (1978) unbiased distances  68 73  xi  LIST OF APPENDICES  Page  Appendix A  Number of families or individual maternal trees used for the studies of variation  108  B  Seed trait means for individual families in eight populations  109  C  Seedling trait means for height (HT), diameter (DM), total  D E  dry weight (TDW),and root-shoot ratio(RS)  111  Scoring method (Keiding et al., 1984) for stemform  113  Average field progeny trait measures fbr height (HT) and diameter (DM) at 4 years  F  114  Average measures of persistence of axis, stem straightness, and branch heaviness  117  G  Extraction buffer (Liengsiri et al., 1990)  120  H  Running buffer systems  120  I  Enzyme staining recipes  121  J  Allele frequencies and heterozygosity for each locus from eight populations  K  L  124  Outcrossing rate estimates (t) for individual families from eight populations  129  Inferred genotypes for maternal trees from eight populations  130  xii  LIST OF ABBREVIATIONS  Institute AOSA  Association of Official Seed Analysts  FAO  Food and Agriculture Organization of the United Nations  ITTO  International Tropical Timber Organization  IUCN  International Union for the Conservation of Nature and Natural Resources  NAS  National Academy of Sciences of the United States of America  NRC  National Research Council of the United States of America  RFD  Royal Forest Department of Thailand  OTA  Office of Technology Assessment of the United States of America  WRI  World Resources Institute  Tree population DL  Donglan population in Khon-Kaen province  KH  Khaoyai population in Nakomrachasima province  KK  Kangkoi populations (KK1 and KK2) in Saraburi province  MK  Mahasarakam populations (MK1 and MK2) in Mahasarakam province  ML  Muakiek populations (ML1 or ML and ML2) in Saraburi province  MS  Same as the population MK1, but different families collected  SK  Sisaket populations (SKi and SK2) in the east of Sisaket province  SKH  Sisaket population in the west of Sisaket province  SM  Same as the population SK2, but different families collected  xlii  Statistics CRD  Completely Random Design  DF  Degree of freedom  EMS  Expected mean square  P  Statistical probability  RBD  Randomized Block Design  SD (or s)  Standard deviation from the mean  SE  Standard error for estimate  SV  Source of variation  2 X  Chi-square  Parameters BR  Branching heaviness of 4-year progeny  DM  Root collar diameter of a seedling or progeny  f  Fixation index of an enzyme allele  2 h  Narrow sense heritability of a genetic trait  HT  Total height of a seedling or progenies  PER  Persistence of axis for the stern of 4-year progeny  R50  Number of days to reach 50%  RDW  Root dry weight of a 9-month seedling  Rf  Migration distance of an enzyme band relative to coloured dye distance on the  of germinating capacity  electrophoretic gel RS  Ratio of root and shoot dry weight of a 9-month seedling  SDW  Shoot dry weight of a 9-month seedling  STN  Stem straightness of 4-year progeny  xiv Outcrossing rate calculated from enzyme data TDW  Total dry weight of a 9-month seedling  Enzymes 6PGD  6-phosphogluconate dehydrogenase (6PGD 1 and 6PGD2)  AAT  Aspartate aminotransferase (AAT2)  ALD  Aldolase (ALD3 and ALD4)  DIA  Diaphorase (DIA)  EST  Esterase (EST1, EST2 and EST3)  IDH  Isocitrate dehydrogenase (IDH)  LAP  Leucine-amino peptidase (LAP)  MDH  Malic dehydrogenase (MDH)  ME  Malic enzyme (ME] and ME2)  PGI  Phosphoglucose isomerase (PGII and PGI2)  PGM  Phosphoglucomutase (PGM2)  SDH  Shikiinic acid dehydrogenase (SDH2)  xv  ACKNOWLEDGEMENTS  I would like to express my sincere appreciation to my supervisor Dr. Oscar Sziklai for his support, advice, continual encouragement, and never-ending energy during my study and living in Canada.  The advice, support, assistance, and editorial contribution of Dr. Tim Boyle, of Forestry  Canada, are also appreciated. I recognize my dissertation committee, Drs Robert Guy, Judy Myers, and Christopher Yeatman, for their support and advice. I also acknowledge Mr. Greg O’Neil for his generous help in editing the additional parts. I am grateful to Mr. Pisal Wasuwanich, former Director of ASEAN-Canada Forest Tree Seed Centre, Thailand, for his encouragement to study in Canada.  Thanks are also extended to my  colleagues at the ASEAN-Canada Forest Tree Seed Centre for their dedication in field and laboratory work. Finally, I would like to thank the Canada International Development Agency and Petawawa National Forestry Institute for financial support during the study period. This thesis is dedicated to my family, in particular my wife Suppamas for her understanding, patience, and positive attitude throughout the years of my study.  1  CHAPTER 1  INTRODUCTION  The world’s tropical forests, which comprise the earth’s most complex and species-rich ecosystems (Office of Technology Assessment, OTA, 1984), are being destroyed at an unprecedented rate (National Research Council, NRC, 1991). The estimated net annual loss varies between 14 and 20 million hectares (Myers, 1991; World Resources Institute (WRI), 1990). Given the current rate of tropical forest decline and the instability of these ecosystems, it follows that the genetic diversity (variation) of plant and animal species is greatly threatened. Degraded forestlands and watersheds can be rehabilitated, and denuded hillsides reforested; but total ecological system reestablishment needs a longer period (extinct species do not return and similarly, the loss of alleles is, effectively, permanent).  Conservation of genetic diversity, which  enables a species to adapt to changing environments is of immense benefit to human society. It is vital that plant and animal species retain their potential for adapting to the biological challenges of the future (International Union for the Conservation of Nature and Natural Resources (IUCN), 1980). Conservation of forest genetic diversity and resource management requires an understanding of the biological dynamics of populations because this information makes it possible to predict trends in genetic losses in a given population and develop strategies for preventing such depletion (NRC, 1991).  Information on genetic structure of variation in tree species provides an opportunity for  sampling the diversity of populations efficiently and to ensuring their conservation (NRC, 1991; Cossalter, 1989; Marshall and Brown, 1975). The conservation of the diversity of tropical tree species needs to be addressed, particularly that of economic species; and basic knowledge about variation among and within populations is a prerequisite for deciding on sampling strategies (Guries and Ledig, 1977).  2 Genetic improvement of economically important tree species can make a contribution to conservation of forests in Thailand through more efficient production of wood in productive plantations, thus reducing pressure on natural forests; however, unlike many parts of the temperate zone, tree improvement in Thailand must take place simultaneously with genetic conservation because genetic base populations can be served as both sources of material for improvement and conservation areas. Therefore, genetic improvement  Therefore, a combined study of genetic variation and potential for  improvement from natural populations will contribute directly to genetic conservation and indirectly through the future production of improved seed. Diversity conservation of an economically important tree species would provide not only sources of plants for immediate and future reforestation programmes, but also a wide genetic base for long-term tree improvement. Because reforestation is expensive, genetically well-adapted materials are needed to ensure success.  Establishment of tree improvement programmes for an economically  important tree species is therefore necessary for seed and seedling collection and utilization. Moreover, tree improvement must be concerned with the long-term need to provide the broad genetic base essential for continued progress over many generations of improvement (Zobel and Talbert, 1984). In general, tree improvement programmes involve three distinct but related phases: conservation, selection and breeding, and propagation (Cheliak and Rogers, 1990).  The programmes require an integration of  reforestation, silviculture and forest genetics( White, 1987; Zobel and Talbert, 1984). However, the procedures are complex and vary around the world. It is tempting but not appropriate to consider improvement strategies for tropical trees as mere extensions of those used in boreal and temperate zones (Namkoong, 1984a). Thus genetic variation within and among populations of a particular tree species in a certain region should be determined to support the implementation of both conservation and improvement programmes.  3 1. Conservation and Genetic Improvement  The Food and Agriculture Organization (FAG) (1988) emphasized the need for conservation of a species because of the rapid loss of special valuable genetic materials for use in reforestation and breeding. Narnkoong (1 984a) recognized genetic conservation as a prerequisite for forest conservation and genetic management; therefore, forest scientists must he responsible for wide population sampling and continued development of population diversity as environmental and economic demands change. The FAG (1989) stated that in addition to taxonomic and ecological knowledge, information on the biology of a species is fundamental to adequate conservation and sustained use. Where such knowledge is lacking or insufficient, research should he carried out in parallel with efforts to conserve and utilize the species. Roche and Dourojeani (1984) stated that conservation of a species will he determined by the biology of the species and the degree to which it is known and used by man. Bawa (1976) also recommended that the information about genetic variation of natural populations and reproductive biology is essential for rational genetic conservation and improvement.  O’Malley and Bawa (1987)  emphasized that as much as genetic information explaining the origin and maintenance of the tremendous diversity of tropical forest species is understood, conservation strategies can be made for more sound scientific criteria.  Additionally, conservation should be directed towards well adapted  material, with abundant variation in traits to he  improved, and should also concern secondarily  important traits. Because selectively neutral genes are not functionally different (Kimura and Ohta, l97), the aim of conservation is to collect and preserve adaptive gene complexes (Ledig, l986h; Marshall and Brown, 1975).  However. Fryer (1987) has cited that most electrophoretically detected variants are  selectively neutral isozymes which are of value primarily as gene markers and are of limited usefulness in studies of selection in natural populations.  Although sampling strategies have focused on single  adaptive gene loci, it is impossible to separate the single loci independently of closely linked loci. Thus, while we discuss alleles at single loci, this single loci also holds for other multiple gene loci  4 which are the real target of genetic conservation.  Guries and Ledig (1981) found significant  correlations between the frequencies of a number of allozymes and several climatic variables, particularly temperature. This suggested selection at certain loci considered to have adaptive value. Also Hamrick et al. (1979) discovered that certain life-history characteristics in trees (e.g., long-lived, wind-pollinated, outcrossing, large winged seed, high fecundity, and late successional) were strongly associated with high levels of polymorphism and heterozygosity, as is in the case of many tropical species. Variability of genetic traits and their modes of inheritance provide a principal insight into the genetic structure of forest tree populations (Hamrick and Loveless, 1986; El-Kassaby, 1980). The distribution of genetic variation within and among breeding populations directs the mating system and geneflow of species and subsequently influences the evolution of the populations (Namkoong, 1989). Therefore, Zobel and Talbert (1984) recommended that the amount, cause and nature of variation existing for a trait in a species should be determined for genetic improvement; however, recognition of the value of genetic variation of tropical tree species and its uses is more recent and needs further study. Progressive tree improvement will have to be integrated with the conservation of genetic variability (Cossalter, 1989). Marshall and Brown (1975) concluded that population structure of a species is important for genetic conservation and improvement because the amount of genetic diversity within populations and the range and distribution of this diversity among populations determine the optimum sampling strategy.  Since the sampling strategy depends on population genetic structure  (Guries and Ledig, 1977), an intensive investigation of population structure should therefore preferably precede the initiation of conservation programme. In addition, Bawa (1976) suggested that variation patterns of tropical tree species may differ from those of temperate trees, due to restricted gene flow, low population density, and widely scattered populations. Basically, genetic variation can be studied from quantitative (morphological) traits and biochemical traits. Quantitative trait variation is mainly studied from phenotypic expression of yield,  5 quality, and pest and disease resistance in all possible environments (Brown and Moran, 1979). There are two approaches developed for determination of biochemical variation; they are the study of DNA variation by means of restriction fragment length polymorphisms (RFLPs) (Botstien et at., 1980) and random amplified polymorphic DNAs (RAPDs) (Williams et at., 1990), and the study of isozyme variation by means of polymorphic enzymes (Lewontin and Hubby, 1966).  Study of both  morphological and biochemical traits allows much more information to be obtained. Not only can morphological traits be used to assess the possibility of genetic improvement of adaptive characters, but a comparison of results from both types of traits can improve our understanding of variation patterns and promote more efficient genetic conservation.  1.1. Morphological variation  Morphological characteristics of tree species have been the focus of studies in several different traits and in different developmental stages of species. Seed quality determined through the variation of viability and seed dimensions or mass plays a significant role in producing a large number of healthy seedlings for planting, though these traits are tinder the control of genetic and environmental factors (Tyson, 1989; Bishir and Namkoong, 1987; Thompson,1984). Information on seed quality is therefore necessary for genetic manipulation of a tree species because seed quality traits viability, seed size, germination percentage and energy  -  -  total seed yield,  can carry over to influence the mature  phenotype of the individual (Roach and Wulff, 1987). The phenotypic variation of seedlings in early development is another important factor to study. The large genetic variation in traits of seedlings suggests the potential to increase genetic gain and reduce costs and time of improvement (Lambeth, 1983). Sziklai (1974) has elucidated that information on variation pattern, mode of inheritance and heritability values during developmental stages may be useful in predicting future performance of genetically controlled traits. Numerous studies concerned with early evaluation of seedling traits from either field or nursery growing have been reported for  6 patterns and amounts of variation (Carter et aL, 1990; Rogers et al., 1989; Jiang et al., 1989; Griffin and Cotterill, 1988; Magnussen and Yeatman, 1987; Sweet and Wareing, 1966). In traditional tree improvement, study of progeny variation from a field plantation accounts for much research effort, since a progeny test provides the essential evaluation of parental genetic worth Field performance of progeny is controlled by both genetic and  (Zobel and Talbert, 1984).  environmental factors. Statistical methods (Cochran and Cox, 1957; Steel and Torrie, 1960; Snedecor and  Cochran,1967; Sokal and Rohlf, 1969) provide techniques for partitioning genetic and  environmental variances.  A genetic distribution or family structure will depend on how  progeny-provenance trials, conservation stands, and genetic base populations have been managed at the initiation of genetic manipulation (Kanowski and Nikles, 1989). Thus, the determination of genetic variability within populations is the major concern in selection of trees to form a breeding population (Sniezko, 1986). Moreover, variation evaluated from the progeny of a tree species can be used as fundamental data for determination of the mode of inheritance, heritability and expected genetic gain (Falconer, 1981).  Heritability values express the proportion of variation in a population that is  attributable to genetic differences among individuals. Estimates of heritability, however, are not made without error; therefore, the ratios obtained are only a relative indication of genetic control and should not be interpreted as absolute or invariant values (Zobel and Talbert, 1984). The concept of heritability is mainly applied for genetic gain evaluation and selection in tree breeding. These genetic parameters, therefore, must be estimated in order to serve for future effective tree improvement.  1.2. Biochemical variation  1.2.1. DNA variation  Modern techniques employing the differentiation in molecular DNA and enzymes provide powerful tools for the detection of genetic variability. However, detection by means of restriction  7 fragment length polymorphisms (RFLPs) or random amplified polymorphic DNAs (RAPDs) requires equipment and well-trained manpower. The technique, therefore, has been applied only on a limited basis in some developed countries, and it has yet to be used in developing tropical regions.  1.2.2. Isozyme variation  Enzymatic variation remains the available technique for the study of population genetics of forest trees. Many enzymes, a class of specific proteins, have multiple forms and those having similar or identical catalytic activities are called isozymes (Feret and Bergrnann, 1976). Markert and Moller (1959) first introduced the term  “isozyme’, hut the alternative term “isoenzyme” is also widely  accepted (Feret and Bergmann, 1976). Prakash er al. (1969) introduced the term “allozyme” for the different enzymes produced by different alleles of one or the same therefore allelic isozymes  (Cheliak et al., 1987).  gene locus.  Allozymes are  Isozymes, composed of amino acids, become  positively charged in acidic environments and negatively charged in basic environments (Feret and Bergmann, 1976). The introduction of isozymes into a molecular sieving medium with forced migration through the  sieving medium under the influence of a powerful electric current is called  “electrophoresis”. The isozymes with different physical characteristics can be separated in solutions of specific pH values (Feret and Bergmann, 1976). Isozyme electrophoresis is a very effective method for studying genetic variation because there are small environmental effects, there is codominant expression, and it is possible to study many loci from each individual (Liengsiri et al., 1990).  Brown and Moran (1979) also noted that isozyme  variation represents variation close to the DNA level. The present applications of isozymes for the genetic study in forest trees has been summarised by Liengsiri ci al. (1990) as follows: identification of species, hybrid, provenance, clonal, cultivar, and seed origin; studying species genetic diversity; allocation of genetic diversity among populations; investigations of genetic organization within a population; investigations of mating systems; studying effects of genetic forces; and studying  8 evolutionary genetics.  Several investigators (Brown, 1979 and Hamrick et at., 1979) reviewed  electrophoretic studies of genetic variation among plants. Boyle et at. (1990) have stated that the use of isozyme analysis has greatly facilitated the investigation of genetic variation among populations. Forest trees having long lived generations, large ranges, high fecundity, and generally an outcrossing -  mode of mating  -  are expected to maintain high levels of genetic variation.  Mitton (1983) has  suggested that young tissue from seedlings provides good material for enzyme extraction because the high level of secondary compounds precipitate proteins, making them  -  e.g. tannins and phenols  -  in mature  tissue may bind and  inaccessible with simple extraction procedures.  Besides the  estimation of genetic variation, isozyme analysis can be used for estimation of mating systems of tree species.  The mating system of species is the pattern  in which gametes unite to form the next  generation (Stern and Roche, 1974). Mating system parameters provide estimates of the degree of inbreeding in natural stands and indicate the extent of potential inbreeding depression in economically important traits (Shaw and Allard, 1982). Although mating systems can be estimated from morphological traits (Morgenstern, 1972; Park and Fowler, 1984), allozyme data are suitable because of the co-dominant expression and the large number of loci that can be assayed (Lewontin, 1974). Several statistical models have been developed for estimation of mating system parameters.  They include either single locus models (Brown et at.,  1975; Clegg et at., 1978; Ritland and El-Kassaby, 1985) or multilocus models (Shaw et at., 1981; Ritland and  Jam, 1981; Ritland and El-Kassaby, 1985).  Iii tropical forests with low density of reproductive individuals, speculation about the mating system in trees has ranged from high degree of inbreeding due to selfing (Fedorov, 1966) to wide outcrossing (Ashton, 1969). Bawa (1974), Chan (1983), and Bawa et at., (1985) have concluded that most tropical tree species are self-incompatible or dioecious. However, the level of inbreeding due to limited gene flow remains unknown (O’Malley and Bawa, 1987).  A mixed mating system,  predominantly outcrossing but with significant inbreeding, has been shown to be common to a large number of tree species (Moran et at., 1989; Moran and Bell, 1983). In tropical trees, the limited  9 studies of mating systems suggest that many species may be predominantly outcrossed (O’Malley and Bawa 1987). Kimura and Crow (1963) developed the concept of effective population size which is the size of an equivalent random mating population having the same fixation index as that observed in the actual population. Yasuda (1969) then developed a method for estimating effective population size and the method is widely used today (Cheliak et al., 1985). However, the effective population size in forest trees is questionable (NRC, 1991) because the mathematical models can oversimplify more complex biological realities (Ewens et a!., 1987). Because quantitative trait variation is subject to selection and/or adaptation, conservation of tree species which is focused on gathering most genes existing should be based on this information (Fumier et al., 1991).  Isozymes, however, provide complementary information on the population  structure ( Epperson, 1989; Fumier et a!., 1987) and mating system (Brown, 1989) which are important for determining the number, size and distribution of conserved areas.  Thus, both quantitative and  isozyme information are required for effective conservation of tree species.  2. Forest Situation in Thailand  Thailand covers an area of 51.3 million ha between latitudes 5° 45’ and longitudes  970  200  30’ N and  30’ and 105° 45’ E. The country has a variety of vegetation types ranging from tropical  evergreen rain forest to dry deciduous forest and savanna forest.  The major vegetation types of  Thailand, as summarised from FAO (1981) are the following: (a). Evergreen and semi-evergreen forests, including hill evergreen, fresh water swamp, and mangrove forests. This type of forest occurs from sea level to 1000 m elevation where annual rainfall is at least 2000 mm and fairly evenly distributed throughout the year. (b). Deciduous forests occur from the plains up to 1000 m elevation where annual rainfall is between 1250 and 2000 mm with well pronounced dry and wet seasons. (c). Dry dipterocarp forests are found on the plains and ridges where soils are poor, generally  10 either sandy or gravelly and subjected to extreme leaching and erosion. (d). Savanna forests can be found in small patches of different types scattered all over the country on sandy or lateritic soils. Bamboo forests, coniferous forests, and scrub formations are the major sub-types of savanna forests. Due to the rapid rate of population increase, forest land has been encroached upon for several reasons,  The major causes of deforestation are shifting cultivation practices, illegal cutting and  processing, infrastructure development, settlement, and natural disasters.  Factors causing forest  depletion vary considerably across the regions. In the North, the illegal and legal cutting, the extension of cultivated areas, and shifting cultivation all seriously threaten the forest ecosystems.  In the  Northeast, the increased extension of agricultural land plays a relatively more important role with reduction of forest land.  In the South, rubber cultivation and mining activities appear to play the  leading roles. Rates of forest depletion from 1961 to 1985 are presented in Table 1. Total forest land area covers 15.3 million ha (30%), of which 14.8 million ha (29%) are natural forests and 0.5 million ha (1%) are plantations. The depletion from 1961 to 1985 was 12.6 million ha. The deforestation rate was about 600,000 ha annually or about 1 % of the country’s land area (FAO, 1987). The depletion of the forest resources in Thailand will continue if no action is taken. According to the current forest policy, at least 40 percent of the country should be kept under forest, 37.5 percent of which should be reserved and 62.5 percent used as productive forest. Thailand is now faced with the situation where the remaining forests (29%) can no longer provide sufficient wood supplies for the population. Therefore, about 10 percent of the country needs to be reforested, and existing forests should be put under a proper management system. Reforestation and conservation of forest land and species are immediate requirements. Concurrent with forest conservation concepts, a number of forest conservation areas have been established as forest reserves, national parks, wildlife sanctuaries, and watershed protection areas. Moreover, all existing logging admissions in the country were terminated in 1988 after the recognition  16.89  3.65  6.75  7.07  51.31  Northeastern  Eastern  Central  Southern  Total 27.36(53)  -  2.96(42)  -  3.57(53)  -  2.12(57)  -  7.09(42)  -  11.63(68)  -  1961  0.17(3) 1.84(26)  3.23(6) 18.94(37)  5.19(10) 22.17(43)  0.22(3) 2.18(33)  Source: Royal Forest Department, Forest Statistics of Thailand 1986.  1.86(3) 15.66(31)  0.12(2) 1.64(23)  0.08(1) 1.76(25)  1.42(3) 17.52(34)  0.19(3) 1.85(28)  0.30(8) 0.80(22)  0.53(4) 2.59(15)  0.75(4) 8.78(52)  1982  0.14(2) 2.04(31)  0.16(4) 1.10(30)  0.24(6) 1.26(34)  0.85(13) 2.01(29)  1.17(17) 2.40(36)  0.62(17) 1.50(40)  1.03(6) 3.12(19)  0.74(4) 9.49(56)  1977  0.92(5) 4.15(25)  1.40(0) 10.23(60)  0.00(0) 11.63(68) 2.02(12) 5.07(30)  1976  1973  Forest area (mill, ha.)  1961 Aerial photos; 1973, 1976, 1977, 1982, 1985 Landsat imagery.  16.96  Northern  *  Forest area (mill. ha.)  Region  in the region at the period specified. The number in parenthesis is percentage of the total region area.  0.86(2) 14.80(29)  0.14(2) 1.50(21)  0.15(3) 1.70(25)  0.00(0) 0.80(22)  0.18(1) 2.41(14)  0.36(2) 8.42(50)  1985  Table 1 Depletion of forest area in Thailand from 1961-1985; the first line in each region is area depletion and the second line is forest area left  12 that deforestation was the main cause of the worst natural disaster in decades, when flooding and landslides killed more than 450 people (Rubeli, 1989).  3. Thailand rosewood  Rosewoods are tree species of the genus Dalbergia belonging to subfamily Papilionaceae, family Leguminosae.  Some 250 Dalbergia species are known and most are tropical shrubs and  climbers (National Academy of Sciences, NAS, 1979). About 120 species occur throughout the tropics, mostly Asiatic (about 90 species), with a very few in Australia and Polynesia (Corner, 1988). A dozen or so species of the genus produce the rosewood timbers so famous for their rich colours, beautiful grain, pleasant fragrance, and superlative technical quality. Today, rosewoods are among the most esteemed and costly timbers in the world (NAS, 1979). In Thailand, among 30 known species of Dalbergia, 22 are trees (Smittinand, 1980). Examples of some of the tree species include: D. cochinchinensis Pierre, D. oliveri Gamble, D. cultrata Grah, D. nigrescens Kurz, D. bariensis Pierre, D. dongnaiensis Pierre and D. sissoo Roxb.. The first two species named produce attractive, high quality, and very valuable wood products. Both species are now very rare, and D. oliveri Gamble is threatened to the extent that almost no populations remain to allow for genetic studies. Dalbergia cochinchinensis Pierre (Thailand rosewood) is an economically important tree species distributed in eastern and northeastern Thailand, Burma, Cambodia, Laos, and Vietnam in mixed deciduous forests and dry evergreen formations (Keating and Bolza, 1982). The Royal Forest Department (RFD, 1983) provided the following taxonomic description. Trees are medium to large sized; the bole is 10-15 m long, sometimes with a fluted base, occasionally slightly bent. The compound leaves are simply pinnate, with 7-9 alternate leaflets and a terminate leaflet. The leaflet is ovate, 3-4 cm wide, 4-7 cm long, and has an acute tip. The species flowers during May and July, producing small, white flowers in small panicles. The seed matures  13 during October to December in pods that are 1-2 cm wide, 4-6 cm long, thin, flat, oblong, rather narrow and indehiscent, containing 1-4 flat seeds. The seed is kidney shaped, brown coloured, 4 mm wide, and 7 mm long. Seeds are generally collected from tree crowns and half are damaged by pests and diseases. However, information about seed pests and diseases is scarce. Insect pests for seedlings and trees of this species (Table 2) have been observed, but methods to control are yet to be developed.  Table 2 List of insect pests of Thailand rosewood (Hutacharern et al, 1988).  Insect pest  Family  Type of injury  Antrocephalus sp. Apoderus sp. Aristobia sp. Colasposoma sp. Hypomeces squamosus F. Plecoptera feflexa (Walker) Psilogramma rnenephron (Cramer) Sphenoptera sp. Striglina scitaria Walker Threnetica lacrymans Thomson Trichochrysea severini (Jacoby) Unidentified chrysomelid  Chalcididae Curculionidae Cerambycidae Chrysomelidae Curculionidae Coctuidae Sphingidae Buprestidae Thyridae Cerambycidae Chrysomelidae Chrysomelidae  Seed boring Leaf rolling Stem boring Leaf eating Leaf eating Leaf eating Leaf eating Stem boring Leaf rolling Stem boring Leaf eating Leaf eating  Although this tree species has been planted in some areas for over 30 years, silvicultural information is obscure and mainly based on individual observation and experience.  Natural seed  production is highly variable from year to year, though some trees start to produce seed at a very young age (3-4 year old). Severe losses of seed viability occur after a short period of storage.  As a result,  the cost of seedling production is high. Due to the hard seedcoat, pretreatments are required for breaking dormancy and producing uniform seedlings. Piewluang and Liengsiri (1989) used various pretreatments to break the dormancy and found that seedcoat scarification by gently rubbing both flat sides of the seed with sand paper provided maximum germination (86%).  14 Soaking seed in concentrated sulphuric acid for one minute and soaking in cool water for twenty-four hours also provided higher germination percentage (85 % and 80%, respectively) than the control or germinating seed without pretreatment (70% germination).  Moreover, soaking seed in  concentrated sulphuric acid pretreatment gave more uniform germination than soaking in water. There is no record of seed germination percentage in seed beds. However, from experience, it is found that the germination is sometimes quite low and varies with seed lots and seed bed preparation. In general, seed bed soil is sprayed with a fungicide a few days before seed sowing. Seed is sown at about 1 gm per m 2 in the seed bed. Rice ash or saw dust is then used for a covering of about 1 cm thick on the seed bed. The seed bed is watered every day. Seedlings, about 2 cm height, are transplanted to plastic bags during the first week of germination. Five to six month old seedlings from a nursery are usually used for outplanting. Although planted seedlings at the Khao-yai Field Station were attacked by insect pests, the rate of survival one year after planting was over 90%. Wood properties of this tree species were described by Keating and Bolza (1982).  The  sapwood is up to 40-50 mm wide, well defined and almost white. The heartwood is from light rosepurple to burgundy with darker brown or black streaks, producing attractive patterns.  The wood  darkens with age, texture is fine to medium, and the grain is narrowly interlocked, often irregular. The cells contain a red, shiny gumniy substance.  Irregular growth rings are clearly visible and  lustrous. The timber is hard, very strong, occasionally brittle. Despite its density, sawing can be done with little difficulty by using saws with shorter, stiffer teeth. The wood machines well, finishes and polishes to a lovely surface. The best results from seasoning are obtained in log form, but good results are reported in kiln-drying if seasoned slowly (Keating and Bolza, 1982). The timber produces an attractive decorative veneer with low moisture movement and is durable in exposed situations. The timber is mainly used for special purposes; for example, high-quality furniture, parts of musical instruments, and cabinet work. In addition, the wood has been used for making high quality charcoal in local areas. As for most rosewoods, Thailand rosewood is now nowhere abundant; all the accessible stands  15 have long been logged and destroyed. The species has become scarce and endangered. The species has been used in planting programmes in many parts of Thailand. Large amounts of seed of this tree species are required for such programmes. The rapid loss of forest areas has caused a loss of germplasm of the species and some sources of the species are already extinct.  Thus, genetic  conservation of the species is an urgent task (RFD, 1989a). The species was therefore included in the Thai Royal Forest Department /Danish International Development Agency (RFD/DANIDA) hardwood conservation programme (Sa-ardavut et al., 1989). However, genetic information about variation or population structure as well as the mating system of the species is lacking and needed to ensure effective management of the conservation programme. Therefore, genetic information is immediately required for the efficient conservation and seed production of this precious genetic resource (RFD, 1989b).  4. Study Objectives  Since Thailand rosewood has been subject to selective logging and other types of deforestation, remnant populations naturally distributed in public and farm areas and are more seriously exposed to human activities than those in national parks and forest reserves. Thus, the populations within public and farm areas are considered disturbed and generally smaller in size than those in natural forests which are undisturbed. When populations are small, limited number of trees involve in mating system and trees will mate between close relatives. This might causes inbreeding depression which could reduce variation within populations. To provide needed information on the species’ variability for the conservation programme, morphological (quantitative) and biochemical (enzyme) variation were studied under the hypothesis that diversity within disturbed tree populations was more adversely affected than that within undisturbed ones. The study objectives were to determine variation in: (a). seed traits;  16 (b). nursery seedling traits; (c). progeny trial; (d). isozyme characteristics and the mating system. All trait variations were considered together in the context of conservation strategies for the species.  5. Study Materials  Seed used in this study came from the two following collections: (a). Seed for the study of variation in traits of seeds, seedlings, and isozymes was collected during November and December 1990 from open-pollinated maternal trees (families) distributed in eight populations in Thailand. Figure 1 presents the distribution area of the tree species and locations of tree populations used in the seed collection. Two populations, Muaklek 1 (ML1) and Muaklek 2 (ML2), were in natural forests and considered undisturbed; the others also natural populations, Kangkoi 1 (KK1), Kangkoi 2 (KK2), Mahasarakam I (MK1), Mahasarakam 2 (MK2), Srisaket 1 (SKi), and Srisaket 2 (5K2), were in public or farm areas and considered disturbed.  Because of the limited  amounts of seed derived from some families and their losses of germination at different stages of the studies, numbers of families used for the studies of seed, seedling, and isozyme variation were unequal (Appendix A).  The variation of seed, seedlings, and isozymes was studied from 69, 56, and 57  families, respectively.  The numbers of families in the individual populations varied from four to  eighteen, three to fourteen, and three to seventeen for the seed, seedling, and isozyme studies, respectively. Geographic and macroclimatic conditions in all populations are quite similar (Table 3). The altitude range is 200-350 m above mean sea level respectively. Average annual rainfall range is 14001650 mm, and average annual temperature is 27-28° C. Soil texture varies from sandy-loam to clay loam with pH between 5 and 6. After collection, seeds were extracted from pods, damaged seeds  17  1000  106°  104°  102°  200  20°  4 18  18°  16°  16°  140  140  100 km  0  0  C)  =  Undisturbed  =  Disturbed 12  12°  1020  100° Figure  1  Location of populations in  (A)  8 2  (B)  populations for  =  KK2, 3  =  which  studying seed,  MLI, 4  =  KH  seed  was  collected for  nursery seddlings,  ML2, 5  7 populations for progeny trial:  SKIT, and 11  106°  104°  3  =  MK1, 6  ML, 5  =  =  and  the  1  isozyme variation:  MK2, 7  MS, 7  studies:  =  =  SKI,  SM, 8  and  8  SK, 9  =  =  KKI, SK2;  DL, 10  18 separated out, and good seeds stored at -2° C. (b). Seed used for planting at the field station was collected in 1987 from 85 open-pollinated families distributed in seven populations in Thailand (Figure 1). Populations Donglan (DL), Muaklek (ML), and Srisaket Khunharn (SKH) were in natural forests and considered undisturbed; populations Khaoyai (KH), Mahasarakam (MS), Srisaket Muang (SM), and Srisaket Khukhan (SK) were in farm or public areas and considered disturbed (Table 4). Numbers of families in the individual populations were different, and ranged from five to twenty (Appendix A). Gographic, macroclimatic, and edaphic conditions in all populations were similar to those of the seed collection for studying seed, seedling, and isozyme variation.  After processing, in late  December 1987, seeds were germinated in seed beds for about 1 week, and seedlings were transplanted into 6 x 4 in. plastic bags and raised in a lath-house for 5 months. The seedlings were then outplanted at the beginning of June 1988, in Khaoyai Field Station. The station is located in Nakornrachasima province, Northeastern Thailand.  lat.  14° 30’N  ML2(4)  14° 35’N  14° 40’N  16° 12’N  16° iN  15° 10N  14° 33’N  KKI(1)  KK2(2)  MK1(5)  MK2(6)  SK1(7)  SK2(8)  Disturbed Population  14° 34’N  MLI(3)  Undisturbed population  Pop(No)  280 300  104° 28’E  280  280  250  250  350  320  elev. m  104° 23’E  103° lO’E  103° 21’E  101° 7’E  101° 2’E  101° 17’E  101° liE  long.  1400  1400  1500  1500  1600  1600  1650  1650  rain mm  28  28  28  28  28  28  27  27  temp. °C  Yearly Ave.  Loamy sand  Sandy loam  Sandy loam  Sandy loam  Clay  Clay  Loam  Loam  soil texture  Medium to large-sized trees along road sides and in school area Medium-sized trees in disturbed-marginal area of natural forest  Large trees patchily distributed in farmland and temple  Large trees patchily distributed in school area  Medium-sized trees widely scattered in scrub formation  Medium-sized trees widely scatter in farm land  Medium-sized trees patchily distributed in marginal area of natural forest Small to medium-sized trees patchily distributed in natural forest  status  Table 3 Environmental conditions of tree populations in which seed was collected for studying seed, seedling, and isozyme variation  lat.  14° 34’N  14° 33’N  ML(3)  SKH(10)  14° 24N  16° l2’N  15° TN  14° 38’N  KH(1 1)  MS(5)  SM(7)  SK(8)  Disturbed population  16° 46’N  DL(9)  Undisturbed population  Pop(No)  240  230  104° 18’E 104° 12E  280  330  103° 20’E  101° 33’E  300  28  28  1400 1400  28  28  28  27  28  temp. °C  1500  1600  1400  1650  320  101° liE 104° 28E  1400  rain mm  230  elev. m  102° 8E  long.  Yearly Ave.  Loamy sand  Medium-sized trees widely scattered in farm land  Medium-sized trees along road-sides  Large-size trees patchily distributed in school area  Sandy loam Sandy loam  Medium-sized trees in private land or along road-sides  Medium-sized trees patchily distributed in marginal area of natural forest Medium-sized trees in marginal forest  Medium-sized trees patchily distributed in forest land  status  Clay  Loamy sand  Loam  Loam  soil texture  Table 4 Environmental conditions of tree populations in which seed was collected for establishing the provenance-progeny plantation  1’) C  21  CHAPTER 2 VARIATION 1N QUANTITATIVE TRAITS  Quantitative traits are regarded as effects of several genes (polygenic) and variation in these traits is generally confounded with environmental effects.  Statistical methods provide a number of  designs for partitioning variances of environmental and genetic effects from each other.  The  quantitative traits investigated in this study involved seed, seedlings, and progeny.  1. Seed Trait Variation  1.1. Study methods  Variation of seed was studied on samples obtained from 69 open-pollinated families of eight populations (Table 3).  Seed quality determines the potential production of healthy seedlings for  planting. Germination capacity is used for evaluation of seed quality. A germination test in controlled environments is a routine practice for studies of seed variation. Results from the test establish the maximum plant-producing potential of families and correlate quite well with emergence under favourable field conditions (St. Clair and Adams, 1991). The germination test, therefore remains the principal and accepted criterion of seed viability (Association of Official Seed Analysts, AOSA, 1983). Additionally, the germination test provides information on seed vigour that high vigour seed will germinate with fast and high percent germination. In this study, the germination test was conducted to measure germination percent and R50 (number of days needed to reach 50% of the germinating capacity). Seed weight is generally affected by genetic factors, but also very much by environmental  22 factors (e.g., nutritional status of parent).  Differences in seed weight can contribute to family  differences in early growth of forest tree progenies (St. Clair and Adams, 1991). A Nested Random Design was used for studying seed trait variation, in four replications. The following statistical model was used.  Where  + P 1 + F() +  Yijk  =  ,i  YjJk  =  observed value of the  Ui  e replication in theJ  th  family within the 1  ti-i  population  1 P  =  population mean  =  population effect  FJ()  eJ  family effect  =  experimental error.  The structure of the Analysis of Variance is shown in Table 5.  Table 5 Analysis of Variance showing source of variation, degrees  of freedom, and expected mean  square of seed traits  Source  DF  Pop Fam (Pop) Error  p-l p(f-1) pf(r-1)  Where  EMS  VE + rVF(P) + rfV V + rVF(P) VE  V  =  population variance  VF(p)  =  family (within population) variance  23 V  =  error variance  p  =  number of populations (= 8)  f  =  average number of families per population (= 69/8) number of replications (= 4)  r  The germination test was performed at 30°C. Fifty seeds from each family were germinated in one replication.  Germinated seeds (germinants) were counted every other day.  Seeds were  considered to have germinated when their radicles had emerged to twice the seed length.  The  germinants were then removed from the germination boxes. The test was conducted over a period of 3 weeks. Gemination percentage and R50 were calculated at the end of the test period. Individual seed weight was averaged from 25 seeds per replication. Seed weight was measured using a different set of seeds from that used in the germination test.  1.2. Data analysis  Data on seed trait measurements were analyzed using SAS Programme version 6.04 (SAS, 1991).  The F-ratio, a test of significant variation, for each factor (population and family) was  calculated using the methods described by Steel and Torrie (1960). The Expected Mean Square (EMS) (Table 5) from each source or factor was partitioned into variance components (Wright, 1976). The replicated plot means were used for the analysis of seed variances, thus family heritahilities were calculated from the following formula:  V F(P  11 2 H ,  = Vf.  + VF(p  24  1.3. Results  Appendix B presents averaged values of seed trait measures, germination percentage, R50, and seed weight, of each population and family.  1.3. 1. Variation  (a). Germination  Mean germination percentage was 80.59 and ranged from 71.43 (SKi) to 88.50 (MK1) (Table  6).  Table 6 Average seed trait measurements for eight populations  Population  %Germination  R50 (day)  Seed weight (mg)  Undisturbed ML1 ML2  88.42 78.96  10.66 8.77  24.37 23.69  Ave.  84.45  9.76  24.03  Disturbed KK1 KK2 MK1 MK2 SKi SK2  75.07 79.75 88.50 80.75 71.43 73.75  11.17 9.64 7.19 7.22 8.20 7.89  24.62 23.96 23.01 27.60 25.45 20.28  Ave.  77.43  8.55  24.15  Overall Ave.  80.59  9.25  24.18  25 Means for undisturbed and disturbed populations were 84.45 and 77.43, respectively. Although the germination percentage mean of undisturbed populations was larger than that of disturbed populations, the F-ratio comparing undisturbed and disturbed populations was 3.09 and nonsignificant. Table 7 provides the results of variance analysis among populations and families within all populations for all seed traits studied. Variation among populations for germination percentage was not statistically significant (F  =  1.56). On the other hand, variation among families within populations was highly  significant (confidence level  =  0.01, F  =  30.3).  (b). R50  Another criterion analyzed from the germination test data was R50 (number of days needed to reach 50 % of the germination capacity). Means for each population and each family are given in Appendix A. The overall mean was 9.25 days and the means for populations ranged from 7.19 (MK1) to 11.17 (KK1) days (Table 6). As for germination percentage, R50 variation between disturbed and undisturbed populations was nonsignificant (F highly significant (confidence level  =  =  0.01, F  within populations was highly significant (F  =  1. 16). Nonetheless, among populations, variation was =  3.86) (Table 7). Overall variation among families  27.01).  (c). Seed weight  Seed weight was the last seed trait studied. Means of single seed weight for populations, each population, and each family are also presented in Appendix A. The overall mean was 24.18 mg, and means of populations ranged from 20.28 (SK2) to 27.60 (MK2) (Table 6).  Analysis of variance  indicated that there was no significant difference between undisturbed and disturbed populations (F 0.02). Variation among all populations was nonsignificant(F populations was highly significant (F  =  39.21) (Table 7).  =  =  1.50), but among families within all  26 1.3.2. Variances and heritahilities  Partitioned variances for populations, families within all populations, and error for each seed trait are also presented in Table 7. The variances of seed germination were 6 %, 83 %, and 11 % for among populations, among families, and within families (error), respectively. The R50 variances were 24 %, 66 %, and 10 % for among populations, among families, and within families, respectively. In a similar trend to the variance of germination and R50, seed weight variances were 5 %, 86 %, and 9 % for among populations, among families, and within families, respectively. Family heritahilities of seed traits (Table 7) were similar for all traits.  All heritabilities at  family level were high. This is could he the preconditioning maternal environmental effects confounded within the variance components at the families.  However, these heritahilities at least revealed the  highest possible parameters for genetic inheritance of the seed traits.  If preconditioning maternal  environmental effects can be controlled or made far uniformity, seed trait heritabilities will be decreased.  1.3.3. Correlation  The general formula given by Becker (1984) was used for calculating correlation between traits. The formula is: CovXY r V’VarX V’VarY  Where  r  =  phenotypic or genetic correlation  CovXY  =  Covariance of phenotypic measures or genetic estimates of trait X and Y  VarX and VarY  =  Variances of phenotypic measures or genetic estimates of trait X and Y  27 Coefficients of phenotypic and genetic correlations between seed traits were weak for both phenotypic and genetic values.  1.4. Discussion  Seed quality, as characterized by seed germination percent, R50 and seed weight was found to be similar in trends of variation to most tree species. Variation exists mainly at the family level, more so than at the population level (St.Clair and Adam, 1991; and Jiang et at., 1989). However, the variations were large at population (provenance) level for lodgepole pine (Pinus contorta) (Ying et aL, 1985; Wheeler and Critchfield, 1985; Lindgren, 1983; Wheeler and Guries, 1982; Critchfield, 1980). Only the R50 trait was significantly different among populations.  Seed trait variation is partially  genetically controlled (Bishir and Namkoong, 1987; Thompson, 1984), but maternal environmental effects can have a strong influence (St.Clair and Adams, 1991; Brarnlet et at., 1983). However, this variation is useful information in nursery management for producing healthy seedlings that are uniform in size for outplanting. Moreover, seed trait variation can carry over and influence early growth in forest tree progenies (Thompson, 1984; Wilcox, 1983).  Therefore, the analysis of seed traits  investigated in this study is useful for preliminary identification of seed sources for forestation programmes and future genetic management of the species. Although the variation in seed traits was mainly not significantly different among populations, except the R50 trait, local populations should be used in tree improvement programmes to ensure adaptation to local environment.  Selection for fast and uniform germinating families should be  cautiously considered to maintain broad breeding populations. Spatial and temporal variation should also be considered and subject of further study.  **  *  7 61 207  7 61 207  7 61 207  DF  87.23 58.00 1.48  75.52 19.54 0.72  1414.83 909.38 30.02  MS  Significant difference at the 0.05 confidence level Significant difference at the 0.01 confidence level  Population Fam (Pop) Error  Seed weight  Population Farn (Pop) Error  R50  Population Fam (Pop) Error  Germination  Source  -  1.50 39.2l**  -  3.86** 27.0l**  -  1.56 30.30**  F-Ratio  0.88 14.13 1.48  1.69 4.70 0.72  15.25 219.84 30.00  Var.Comp  5.00 86.00 9.00  24.00 66.00 10.00  6.00 83.00 11.00  % Var.Comp  -  0.91 0.90  -  0.87  -  -  0.88  -  heritability  Table 7 Germination, R50 and seed weight variation, variance component and family heritabilities for seed from eight populations  00  29 Seed variation provides general genetic information of the species, though environmental pre conditioning effects play a substantial role in seed traits (El-Kassaby et al, 1992). The broad sense heritabilities for all seed traits were high; however, these were probably biased (St. Clair and Adams 1991), because seed traits most likely reflect nonheritable maternal effects. The population sizes in this study were quite small (mean  =  9 families); the potential for half-sib  families may not exist in some populations (Surles er al., 1990). The genetic correlation between germination and seed weight was weak. El-Kassaby ci al (1992) found no relationship between germination and seed weight. In the case of Douglas-fir, St. Clair and Adams (1991) found that seed weight was weakly related to emergence percentage and rate. The absence of a significant correlation between seed weight and rate of emergence has also been reported in other conifer species (Mikola, 1984; Perry and Hafley, 1981).  2. Seedling Trait Variation  2.1. Study methods  Fifty-six open-pollinated families of seeds, collected from eight populations, the same as those used for the seed trait study, were grown in the nursery to estimate seedling trait variation.  The  assumption is made that the pattern of growth of nursery seedlings is significantly related to observed family variation in the field. If nursery seedling performance is reliably related to field growth, the impact on genetic management strategies would he significant (Carter et at., 1990), because nursery studies are more convenient to control, environments are simpler, they require a shorter period of time than those carried out in the field. Variation of seedling growth, was then evaluated in terms of height, diameter, and dry weight. A Nested Block Design was employed with 4 blocks. The statistical model was as follows.  30  Where  Yikl  =  Ykl  =  +  B  +  P  +  Fk)  +  PB  FBk(J)  +  ith observed value of i seedling of  .  th  + ek1  family within  •th  J  •th  population, in 1  block =  B  overall mean  block effect  P  =  population effect  Fk(j)  =  family effect  PB  =  interaction between population and block  FBjj)  =  interaction between family and block  =  experimental error.  ekl  The structure of the Analysis of Variance table including expected mean squares (EMS) is shown in Table 8.  Table 8 Analysis of Variance for seedling traits  SOURCE  (b-i) (p-i) (p-l)(b-l) p(f-1) p(f-l)(b-l) bpf(n-1)  Block Population Pop*Block Fam (Pop) Fam*Block Error  Where  EMS  DF  VB  =  block variance  Vp  =  population variance  VE 1 V VE VE V VE  nfpV nfVPB + 8 nbVF + nfV + nbfV + nbVF + nfV  + flVFB + + flVFB + + nVFB + +  nVFB nVFB  +  nbVF  31  =  family variance  =  variance of population and block interaction  VFB  =  variance of family and block interaction  VE  =  error variance  b  =  number of blocks  p  =  number of populations  f  =  average number of families per population  n  =  number of seedlings per plot or experimental unit  VF  (= 4) (= 8) (=  56/8)  (=  6).  Seeds were germinated in soil seed beds for one week, until about 2 cm high, and then transplanted into 6 by 10 inch plastic bags containing a 1:3 ratio of burned rice hull and forest top soil. The seedlings were raised in a lath-house nursery at the Khao-yai field station. Watering was conducted every day except on rainy days. Seedling height and diameter were measured at 9 months of age.  After the 9 month  measurement, one seedling from each family in each block was randomly selected for dry weight determination. Since only one seedling from each block was used, block variance were then used as experimental error for F-test. Statistical design for the dry weight variation study was same as that used for study seed traits variation. Number of families (f), however, was changed to 56/8.  2.2. Data analysis  The SAS Computer Program version 6.04 (SAS, 1991) was used for the analysis of variance of seedling traits. The EMS from each source or factor was partitioned for variance components (Wright, 1976). Because individual maternal difference was used for the analysis of variances, narrow sense heritabilities (Falconer, 1981) at the family level were calculated as follows:  32  (a). Individual V  + VFB +  Vf.(p)  V (b). Family . 1 V  +  14.96  r 1 V  +  Vjr(p)  4  2.3. Results  Seedling traits studied included height (HT), diameter (DM)), total dry weight (TDW) and root-shoot ratio (RS). Appendix C presents means of the seedling traits at population and family levels. Table 10 presents the average values of the seedling traits for each population. In regard to undisturbed and disturbed populations, means of most seedling traits were highest in undisturbed MLI, except mean of RS.  However, seedling trait variation between undisturbed and  disturbed populations was not significant. The initial analyses of all data for seedling height and diameter produced large variation within plots or error terms and the cause for the large error terms was not exactly known, perhaps due to seedlings within families were from different paternal pollen or inconsistent environmental condition within blocks, but they would mask and hamper the detection of variance components at the family and population levels. Schlichting (1986) indicated that variance estimates would he more precise if this amount of within-plot genetic variation was controlled (minimized with clones). Thus, the data which had values departing  from  their family means by more than I standard deviation (is) were deleted  before the variance analysis to decrease within plot variation. The variance components at family and population levels were obviously detected by the latter method. The deletion of one standard deviation  33 criterion was then used for the seedling variation study. The averaged numbers of seedlings within plots were decreased to 3.74.  2.3.1. Variation  (a). Height  MLI population provided the highest mean for HT lowest mean for HT  =  =  63.69 cm; SKI population had the  38.34 cm (Table 9). Table 10 presents the resulting analysis of variance for  height (HT) and diameter (DM) of seedlings.  HT was significantly different both among the  populations and among the families within all populations (F-ratios 17.79 and 3.76, respectively).  (h), Diameter  Diameter mean was largest in MLI  =  0.64 cm and smallest in SKI  =  0.56 cm (Table 9).  The analysis of variance indicated significant differences among families within all populations (F 4.00) but not among populations (F  =  1.17) (Table 10).  (c). Dry weight and root-shoot ratio  Total dry weight (TDW) mean was highest in MLI  =  17.07 gm and lowest in SKi  =  11.94  gm (Table 9). The last parameter estimated, root-shoot ratio RS, provided the highest ratio in SKi 1.48 and lowest ratio in ML2  =  =  0.75. The analysis of variance result for seedling dry weight is shown  in Table 11. TDW was significantly different among families within all populations but not among populations. On the other hand, RS was significantly different both among populations and among families within populations.  34 Average measurements flr seedling traits including height and diameter (HT, and  Table 9  DM), total dry weight (TDW), and root/shoot dry weight ratio (RS)  for each  population  HT (cm)  DM (cm)  TDW (grn)  RS  ML1 ML2  63.69 59.28  0.64 0.62  17.07 16.47  0.78 0.75  Ave.  61.49  0.63  16.77  0.77  KKI KK2 MKI MK2 SKI 5K2  49.31 52.1 I 46.08 40.58 38.34 50.69  0.60 0.62 059 0.56 0.57 0.60  15.65 15.45 16.55 13.63 11.94 15.01  0.82 0.95 1.31 1.19 1.48 1.00  Ave.  46.19  0.59  14.71  1.13  Overall Ave.  53.31  0.61  15.62  0.96  Population  Undisturbed  Disturbed  2.3.2. Variances and heritahilities  The variances of f-IT for  among  populations and among families within all populations were  72.02 (29.45%) and 43.80 (17.91 %), respectively (Table 10). The DM variance among populations was negative , -0.0001 (-0.96%), but presumed zero. This was because sum of other variance components which were calculated from bottom was higher than EMS of population. The variance among ftimilies within populations was 0.0018 (13.80%) (Table 10). The DM variance among populations was negative hut very small, -0.0001 (-0.96%), this was unrealistic and presumed no variance.  The variance among families within populations was 0.0018  35 (13.80%) (Table 10). The dry weight variances for among populations and among families within all populations were: 1.42 (6%) and 6.45 (29%) for TDW; 0.07 (46%) and 0.03 (21%) for RS, respectively. Heritahilities for I-IT and DM are presented in Table 10 and for dry weight traits in Table 11. The individual heritability for HT was over I, theoretically impossible results.  2.3.3. Correlations  There were a relatively high phenotypic correlation between height and diameter (0.63). The TDW was negatively correlated with RS (-0.15).  2.4. Discussion  Seedling growth in a nursery or controlled environment has become an effective means for investigating variation of a tree species and nursery growth often shows significant correlations with growth in a field trial (Jiang  et  a!.. 1989). Seedling performance at an early age in a nursery can he  used as a criterion for increasing the efficiency of selection and breeding. In conifers, inherently rapid growth capacity can be identified at  quite an early age,  3 to 12  months (Pharis et a!., 1991). In this study, height growth variation was significant both among families and populations. Diameter growth variation for seedlings at both ages was significant at the family level. However, variation in diameter growth of nursery seedlings was not different (p  >  0.05)  at the  population level. Unlike height variation, TDW variation existed mainly among families. On the other hand, the RS variation for among populations was double that for among families. The dry weight variation was ambiguous and  must  he carefully interpreted. However, Lambeth  weight as an early testing trait in Douglas-fir.  el  al. (1983) proposed total dry  **  *  3 7 21 48 141 617  3 7 21 48 141 617  DF  0.07 0.07 0.06 0.04 0.01 0.01  1291.36 8336.27 468.60 895.66 238.46 72.18  MS  -  1.57* 1.17 1.50 4.00** 1.00  -  3.09** 17.79** 0.52 3.76** 3.30**  F-Ratio  -0.0000 -0.000 1 0.0016 0.0018 0.0018 0.0077  3.47 72.02 8.67 43.80 44.49 72.18  Var.Comp.  -0.21 -0.96 12.86 13.80 14.28 60.23  1.42 29.45 3.53 17.91 18.19 29.51  % Var.Comp.  -  -  0.63  -  -  -  -  -  1.09  -  -  -  -  -  -  -  0.64  -  -  -  -  0.73  -  -  Heritability (S.E.) Individual Family  Height and diameter variation, variance component and heritability for 9-month nursery seedlings from eight populations  Significant difference at the 0.05 contidence level Significant difference at the 0.01 confidence level  Block Population Block*Pop Fam (Pop) Block*Fam (Pop) Error  Diameter  Block Population Block*Pop Fam (Pop) Block*Farn (Pop) Error  Height  Source  Table 10  C’  **  *  7 48 166  7 48 166  DF  1.92 0.16 0.05  78.02 40.07 14.52  MS  -  l1.69** 3.51**  -  1.95 2.67**  F-Ratio  0.07 0.03 0.05  1.42 6.45 14.52  Var.Comp.  46.00 21.00 33.00  6.00 29.00 65.00  % Var.Comp.  -  0.72  -  -  0.64  -  -  -  0.72  -  -  0.64  Heritability (S.E.) Individual Family  Dry weight variation, variance component and heritability for 9-month nursery seedlings from eight populations  Significant difference at the 0.05 confidence level Significant difference at the 0.01 confidence level  Population Fam(Pop) Error  Root—shoot ratio  Population Fam (Pop) Error  Total dry weight  Source  Table Ii  38 Although selection based on early variation is less accurate, it may be more efficient in terms of selection for gain per unit of time than later selection based on mature measurements (Fins er at. 1990). Moreover, mature outplanted trees are subject to inconsistent environments and damage which can markedly decrease reliability of their measurements (Rehfèldt, 1983), and nursery trials can be used to rank large numbers of families in the juvenile stage (Lambeth, 1983; Carter and Greenwood, 1987; Ritters and Perry, 1987). The average measurements for most seedling traits, especially height and diameter growth, were not significantly different between disturbed and disturbed populations, when variation was analyzed by nesting the populations within their disturbance levels.  Another reason for the larger  average measurements for the undisturbed populations (ML] and ML2) may be that these populations were located closest to the nursery site, and the seedlings were likely to be more suited to the growing environment. For urgent reforestation, seedlings we know more  about  from  local populations are always recommended, until  genetics and ecological interaction of the species.  The heritability for height growth was high both at the family and population levels. Population and family selection can he the best means for genetic improvement of the species. Although an individual heritability of more than one is conceptually impossible, this phenomenon is sometimes encountered in statistical analysis of juvenile trees (Rink, 1984; Hicks and Joly, 1977).  ci  at.., 1977; Adams  This also results from closer relatedness than the half-sib relationship assumed  (Hanover er at., 1991). The heritability for diameter growth was small at the population level, but quite high at the family le’vel. differences in the trait.  Early selection using diameter may not he appropriate due to small  The individual and family heritabilities for dry weight were consistent and  relatively high at the family level, and these might he promising parameters for genetic studies of the species. Phenotypic and genetic correlations between traits were mainly parallel, so phenotypic selection for genetic improvement of traits is thereftre advantageous.  39 3. Field Progeny Trait Variation  3.1. Study methods  Seed of 85 open-pollinated families was planted at Khao-Yai Field Station as described earlier in the section on study materials. A nested block design was used for the progeny test utilizing four blocks. The statistical model, source of variation (SV), degrees of freedom (DF), and expected mean square (EMS) were the same as those used tor studying the seedling trait variation (Table 9), but numbers of populations and families were changed to seven and 85/7, respectively.  Twenty-five  seedlings of each flimily were planted in each block. Height and diameter growth were measured from 10 saplings at 4-years old. The persistence of axis, stem straightness, and branch heaviness were also scored using the applied method of Keicling et at. (1984) (Appendix D).  3.2. Data analysis  The SAS Computer Program version 6.04 (SAS, 1991) was used for variance analysis of field progeny traits. Variance Components for each progeny trait were calculated from EMS (Wright, 1976) and used for estimating narrow sense heritability (Falconer,1981).  3.3 Results  Average measurements for flimily and oulation level of field progeny traits at four years height (NT) and diameter (DM) are presented in Appendix E.  Also, the scored trait means are  presented in Appendix F. Table 12 presents the population means for all progeny traits. To reduce within plot variation, observations in the original data set that were different from family means by more than one standard deviation were deleted to derive more precise results of the  40 variance components at the family  and  population levels (Schlichtin, 1986). The averaged numbers  of trees per lo was then 6.84. Table 14 and 15 present the results for the analysis of variance and genetic parameters for I—IT and DM, and for the scored traits of the progeny, respectively.  3.3. 1 Variation .  (a). Height  Overall mean for NT was 3.48 m (Table 12).  The HT was largest in the undisturbed  population SKH (3.71 m) and smallest in the disturbed population MS (3. 13 m). Although the undisturbed populations tended to have larger height means than those of the disturbed populations, analyses of variance thr height indicated nonsignificant differences between the undisturbed and disturbed populations. The HT (Table 13) was significantly different at both levels, hut the differences  among  confidence level (F-ratio  families within all populations were small and significant at only the 0.05  =  1.30). There were also highly significant differences for the interaction  between blocks and families within all populations for height. This means height traits from different families responded differently to environmental conditions in blocks.  (b). Diameter  Overall mean for DM was 5.95 cm (Table 12). Mean DM was largest in the undisturbed population DL (6.39 cm) and smallest in the undisturbed population ML (5.72 cm). Like height growth, there was no significant difference in diameter between undisturbed and disturbed populations. The variance analysis results for DM are also shown in Table 13. Whereas, the differences among populations for diameter were highly significant, there was no significant difference among families.  41 Average measurements of progeny traits including height and diameter (HT and DM),  Table 12  persistence of axis (PER), stern straightness (STN), and branch heaviness (BR) for each population  1—IT (cm)  DM (cm)  PER  STN  BR  Undisturbed DL ML SKH Ave.  369.20 350.40 370.45 363.35  6.39 5.72 5.99 6.03  2.32 2.39 2.49 2.40  2.54 2.57 2.60 2.57  2.79 2.86 2.95 2.87  Disturbed KR MS SM SK Ave.  350.93 312.88 346.53 355.56 341.48  5.73 5.77 6.11 5.85 5.87  2.44 2.25 2.34 2.42 2.36  2.52 2.39 2.46 2.51 2.47  2.87 2.76 2.88 2.89 2.85  348.23  5.95  2.35  2.50  2.83  Population  Overall Ave.  (c). Scored Traits  Overall means for Persistence of axis (PER), stem straihtness (STN), and branch heaviness (BR) were 2.35, 2.50, and 2.83, respectively (Table 12). PER mean was largest in the undisturbed population SKH (2.49) and smallest in the disturbed population MS (2.25). However, the difference between the undisturbed and disturbed populations was nonsignificant.  There was a highly significant difference among all populations (F  nonsignificant difference among families within all populations (F  =  =  6.55), but  1.13) (Table 14).  STN means for each population also varied only slightly from each other.  Population SKH  produced the largest mean (2.59) for STN and the smallest (2.38) from the population MS (Table 12).  42 There was no significant difference between the undisturbed and disturbed populations.  Variance  analysis (Table 14) indicated a significant difference among all studied populations, but no significant difference among families within all populations. In a similar fashion to PER and STN means, BR means for each population varied very little. The largest mean was also in population SKH (2.95) and the smallest in population MS (2.75). There was no difference between the undisturbed and disturbed populations. Variation was highly significant among populations, hut nonsignificant among families within all populations.  3.3.2. Variances and heritahilities  The HT variance among populations (10.50%) was over twice larger than that among families (4.10%) The variances for interaction between block and families within all populations were relatively high for height at all ages. The diameter variances varied substantially; however, among population variances played a significant role. Whereas, 3.83% of the total variance was accounted for by the difference among the populations, only 0.63% was accounted for by the difference among families. For the scored traits, variances were very small both among populations and among families within all populations. The variances among trees within plots were considerably larger. The height individual heritability was small (19%).  The height family heritability was also  small (21 %). For diameter, both individual and ftimily heritahilities were very close to zero (0.03 and  0.05, respectively). Family heritahilities for the scored traits were moderate for PERST and BR, but small for STN. All individual heritahilities for scored traits were small.  **  *  3 6 18 78 234 1987  3 6 18 78 234 1987  DF  31.4937 32.6515 6,9082 7.77 15 7.2561 1.0243  6.9730 16.5432 1.9617 2.2340 1.7223 0.1105  MS  -  0.89 1.07 7.08**  2.67** 4•73**  -  2.50** 8.43** 0.89 1.30* 15.59**  F-Ratio  0.0424 0.0793 -0.0061 0.0131 0.9 177 1.0243  0.0085 0.0441 0.0025 0.0172 0.2373 0.1105  Var. Comp.  2.05 3.83 -0.30 0.63 44.32 49.47  2.02 10.50 0.60 4.10 56.46 26.30  % Var.Comp.  -  -  -  0.03  -  -  0.05  -  -  -  -  -  -  -  -  -  -  0.21  -  -  0.19  -  -  Heritability (S.E.) Individual Family  Height and diameter variation, variance component and heritability for 4-year field progeny from seven populations  Significant difference at the 0.05 confidence level Significant difference at the 0.01 confidence level  Block Population Block*Pop Fam (Pop) Block*Fam (Pop) Error  Diameter  Block Population Block*Pop Fam (Pop) Block*Fam (Pop) Error  Height  Source  Table 13  *  3 6 18 78 234 1987  3 6 18 78 234 1987  3 6 18 78 234 1987  12.0373 1.3151 0.3700 0.3993 0.3064 0. 1260  1.3552 1.9931 0.7031 0.5062 0.4689 0.2294  2.4831 2.9397 0.4489 0.4842 0.4268 0.2761  MS  Significant difference at the 0.05 confidence level  Block Population Block*Pop Pam (Pop) Block*Fam (Pop) Error  Branching characteristic  Block Population Block*Pop Pam (Pop) Block*Fam (Pop) Error  Stem straightness  Block Population Block*Pop Fam (Pop) Block*Fam (Pop) Error  DF  **  -  0.0200 0.0027 0.0007 0.0032 0.0266 0. 1260  0.0010 0.0039 0.0029 0.0011 0.0353 0.2294  0.0035 0.0076 0.0002 0.0020 0.0222 0.2761  Var.Comp.  11.18 1.49 0.41 1.80 14.83 70.29  0.37 1.44 1.05 0.42 12.89 83.84  1.12 2.45 0.07 0.63 7.12 88.61  % Var.Comp.  -  -  -  -  -  0.22 -  0.08  -  -  -  -  -  -  -  -  0.06  -  0.02  -  -  -  -  -  -  -  -  0.11  -  -  0.03  -  -  Heritability (S.E.) Individual Family  Significant difference at the 0.01 confidence level  0.93 1.30* 2.43**  18.39** 355*  -  l.59 2.83* 1.39* 1.08 2.04**  -  3•39** 6.55** 0.93 1.13 l.55**  F-Ratio  Scored trait variation, variance component and heritability for 4-year field progeny from seven populations  Persistence of axis  Source  Table 14  45 3.3.3. Correlation  The phenotypic and genetic correlations between HT and DM were 0.71 and 0.64, respectively. Whereas, the phenotypic correlations between I-IT and scored traits were very small, the genetic correlations were quite high.  The phenotypic and genotypic correlations between DM and  scored traits were inconsistent. The phenotypic correlations between scored traits were moderate to small. There was no genetic correlation between PER and stem straightness (STN), moderate between PER and hranchin heaviness (BR), and a neative correlation between STN and BR.  3.4. Discussion  Field progeny height was the only trait that varied significantly both among populations and families. The diameter trait was not significantly different among families. For the other traits, there were very small differences at the levels of population and family.  The majority of progeny trait  variation was accounted for by differences within families and by environmental differences.  The  ability of an individual family to alter its morphological response to changes in environmental conditions is termed phenotypic plasticity, and is considered adaptive if plant fitness is enhanced (Schlichting, 1986). Interaction effects between ftimnily and block were significant for all progeny traits. Bradshaw (1965) indicated that the plasticity of a character is a character in itself and is under genetic control. The average height growth and scored traits were slightly larger for undisturbed populations than those for disturbed populations. On the other hand, the average diameter growth was larger for disturbed populations than for undisturbed populations. The local populations (undisturbed ML and disturbed KR) were not consistently different from non-local populations. This was probably because of no pronounced difference in environment between the two regions sampled.  As a result of the  significant differences in progeny traits for among populations, conservation should aim at including high levels of variation of these quantitative traits which will provide an opportunity to species to  46 adaptation to changing environments. Interactions between family and block were significant for all progeny traits. This phenotypic plasticity of the traits should be subject to future study. Individual heritability for height was 0.43. 0.16.  Individual heritability for diameter growth was  Franklin (1979) suggested that heritability of growth traits may be high and then decrease,  possibly to zero, at about the time of crown closure and then increase again as the stand reaches maturity.  Due to the very small individual heritability estimates for all scored traits, individual  selection based on these traits may not appropriate for tree improvement of this tree species. However, Bastien and Roman-Arnat (1990) proposed that very early evaluation is not possible for all traits. The phenotypic and genetic correlations between traits were not parallel; this is due to the traits responded differently to environments.  The phenotypic correlation between DM and STN was  moderate, but the genetic correlation was zero. Conversely, the phenotypic correlation between HT and each scored trait was very weak, hut the genetic correlation was moderate to high.  Therefore  selection from phenotypic traits of progeny may not he appropriate for genetic improvement.  4. Conclusion  The study of variation in quantitative traits revealed different patterns of variation among traits. Seed variation mainly existed among families.  differences among populations. differences both  among  R50 was the only seed trait which had significance  Nursery seedling and field progeny traits produced significant  populations and among flimilies, but  no  significant (Iifference was observed  among populations for total dry weight. The significant effects of interactions between family and block would be caused by phenotypic plasticity of the progeny traits. Variation in quantitative traits was not significant between undisturbed and disturbed populations. Family heritability estimates for seed traits were high. This may have been confounded by  47 environmental pre-conditioning effects. The heritability estimates for progeny traits were quite low. The correlations between seed R50 and seedling height and diameter were significant. R50, seedling and progeny traits showed comparable patterns of genetic variation. therefore, be considered in a sampling strategy for conservation.  These adaptive traits should,  48  CHAPTER 3  ISOZYME VARIATION  The isozyme technique is a quick and inexpensive means for the investigation of genetic variation of forest tree species. The study of both isozyme and other quantitative traits provides useful information for genetic management as well as conservation strategies. However, the study of isozyme variation of tropical forest tree species has only a short history. In Thailand, isozyme studies on forest tree species have just been carried out fhr the last few years by the Royal Forest Department and the ASEAN-Canada Forest Tree Seed Centre.  The studies are mainly in the infancy stage and need to  be extended to derive valuable information on genetic structure of forest tree populations.  1. Study Method  Fifty-seven of the open-pollinated families were used for the isozyme study.  Seed was  germinated in petri dishes for three days. The root tips or radicles, which had emerged to about a half of seed length were used for enzyme extraction. The isozyme procedures followed were those described by Liengsiri er al. (1990).  In  summary, embryonic tissue from the radicle was separated from the germinating seed and ground in 0.04 ml of extraction buffer (Appendix G) using a rotating teflon grinding head.  The resulting  homogenate was absorbed onto filter paper (Whatman No. 3) wicks, approximately 14 x 1 mm in dimension. These were then introduced into a 12.5 % starch (Connaught Laboratory, Ontario, Canada) gel by making a vertical cut along the axis of the gel, about 1 cm from one edge, placing the wicks vertically along the resulting edge, and then pushing the two slabs together.  49 The loaded gel was placed on a wooden stand equipped with two tanks of electrode buffer. Bridge wicks (J-cloth or sponge paper) saturated with electrode buffer from the two buffer tanks were then placed on the gel surface about 1.5 cm from both electrode sides. The stand was then placed in a refrigerator for electrophoresis. Two buffer systems, H and B buffer systems (Appendix H), were used for gel preparation and electrode finning. The applied electric currents for the buffer systems H and B were 90 amperes and 30 amperes, respectively.  Initially 50 % of the final applied current, 45 amperes for the H buffer  system and 15 amperes for the B buffer system, was used for 10 to 30 minutes until a coloured dye marker migrated about 0.5 cm beyond the origin. At this stage all sample wicks were removed from the gel. The electric current was then re-applied to the system at the final wnning level. Once the enzyme had migrated a sufficient distance through the gel, as indicated by the progress of the coloured dye (5 cm), the gel was sliced into 1 mm thick slices, which were then stained for enzymes shown in Appendix I. Gel slices from buffer system H were used for enzyme systems aldolase (ALD), isocitrate dehydrogenase (IDH), malic dehydrogenase (MDH), malic enzyme (ME), phosphoglucomutase (PGM), shikimic acid dehydrogenase (SDH) and 6-phosphogluconate dehydrogenase (6PGD). Enzyme systems aspartate aminotransferase (AAT), diaphorase (DIA), esterase (EST), leucine-amino peptidase (LAP) and phosphoglucose isomerase (PGT) were stained from gel slices run with buffer system B. Twenty arrays or radicles from each family were scored for analysis of isozyme parameters. Some enzyme systems were scored for more than one locus. A numeric system was applied to alleles at each locus.  2. Data Analysis  Genetic variability and mating system parameters of the tree species were estimated as follows.  50  2.1. Allelic variation  The Biosys-1 Computer Program version 1.7 (Swofford and Selander, 1989) was used for analysis of the following parameters.  (a). Allele frequencies  The frequencies of alleles occurring at each locus are estimated for individual populations and loci.  (b). Heterozygosity  Observed heterozygosity (Ho) is the proportion of all genotypes that are heterozygotes. Expected heterozygosity (He) Where  p  =  frequency of  jth  =  ) 2 1 (1-Ep  allele  (c). Number of alleles per locus  This parameter was the numbers of alleles recorded at each locus.  The average number of  alleles for the overall population were also calculated.  (d). Percentage of loci polymorphic  Percentages of loci possessing more than one allele in individual populations were calculated to represent another aspect of variation.  51 (e). Deviation from Hardy-Weinberg equilibrium  A chi-square test for deviation of allelic frequencies from Hardy-Weinberg expectation will test the null hypothesis that there are no biological phenomena or sampling biases with a net effect sufficient to cause significant differences between observed and expected proportions. Contingency chi-square tests also provide information as to whether differences in genotype proportions among eight populations were no greater than among ten random samples drawn from the total population.  (f). Genetic similarity and distance coefficients  Genetic distance provides a measure of the relationship among individual populations. Various measures of genetic distance have been derived and the most frequently quoted measure is one derived by Nei (1972): J  =  JxyIV’JxJy  I D  1-He  =  TnT  =  expected heterozygosity  J  =  genetic identity  Jxy  =  covariance of J for the population x and y  Ix  =  variance of I for population x  Jy  =  variance of J for population y  I  =  normalized genetic identity for populations x and y  D  =  genetic distance.  Where He  Nei’s (1978) unbiased genetic distance and Nei’s (1978) unbiased identity were calculated in this study.  52 (g). Genetic differentiation of populations by F-statistics (Wright, 1965, 1978)  F-statistics (Wright, 1965) is a set of statistics which indicates the relative distribution of genetic variation within and among populations. The formulae are:  Where  15 F  =  1-Ho/He  F  =  Vp/pq  1 F  =  ) F 15 15 + (1-F F  15 F  =  the mean deviation of genetic proportions from Hardy  -  Weinberg  expectation for each population. F  =  the correlation between random gametes within a population with gametes in the total of all populations  FIT  =  the correlation among gametes for the total of all populations  Vp  =  variance of allele frequencies in a population from frequencies over all populations.  p & q  =  allele frequencies.  G-statistics (Nei, 1973) is an extension of the concept of F-statistics for a locus possessing more than two alleles:  Where  HT  =  D 59 + 5 H -  5 G  =  DST/HT  HT  =  gene diversity (expected heterozygosity) for total of all populations.  H  =  average of the individual population gene diversities.  =  average gene diversity among populations  =  59 genetic difference among populations, equivalent to Wright’s F  53  (h). Cluster analysis  Unbiased Nei’s genetic distances were used to conduct a cluster analysis using the unweighed pair groups method algorithm (UPGMA) (Sneath and Sokal, 1973).  2.2. Mating system  The Generalized Multilocus Estimation Programme or MLTF (Ritland, 1986) was applied for the mating system analysis of the species. Only fifteen loci with high variability were selected for the mating system analysis. The programme provided information as follows.  (a). Outcrossing rate estimate (t) (b). Fixation index estimate (f)  The mating systems of all populations, and comparisons of undisturbed populations and disturbed populations were estimated.  3. Results  Reproducible banding patterns were observed from twelve enzyme systems coding for eighteen loci. Seven enzyme systems produced more than one locus. The loci included 6-phosphogluconate dehydrogenase I and 2 (6PGDI and 6PGD2), aspartate arninotransferase 2 (AAT2), aldolase 3 and 4 (ALD3 and ALD4), diaphorase (DIA), esterase 1,2 and 3 (EST1, EST2 and EST3), isocitrate dehydrogenase (IDH), leucine-arnino peptidase (LAP), malic dehydrogenase (MDH), malic enzyme 1 and 2 (MEl and ME2), phosphoglucose isomerase I and 2 (PGJ1 and PGI2), phosphoglucomutase 2 (PGM2), and shikimic acid dehydrogenase 2 (SDH2).  54 All enzyme banding patterns are displayed in Figure 2. slices presents in Figure 3.  An example of electrophoretic gel  Allele variability parameters for individual loci and tree populations  provided valuable details on the genetic population structure of the studied species.  3.1. Allelic variation  The genetic variability parameters are shown in Table 15. The mean zygotic sample size per  locus for each population ranged from 60.0 (population SK2) to 324.2 (population ML1). The overall mean zygotic sample size averaged across all eight populations was 135.8, and the overall mean sample sizes averaged across two undisturbed populations and six disturbed populations were 268.8 and 91.5, respectively.  (a). Allelic frequencies  The allelic frequencies of each locus for individual populations are presented in Appendix J. The across population means for eighteen loci are presented in Table 16. The 6PGD enzyme of individual seedlings displayed two loci, 6PGD1 and 6PGD2. Although five alleles were scored at locus 6PGD1, only four frequency means were derived because the fifth allele was observed in just the population MK2 with very low frequency (0.007) and its frequency mean was presumed zero. The frequency means for alleles one and two were 0.891 and 0.087, respectively. The frequency means of alleles three and four were also low (0.008 and 0.004, respectively) and the alleles were found only in a few populations investigated.  55 Rf  6PGD  Rf  AID  6PGD1 .385  ALD3 .423  .372 5 .333  .410  3  .321 4 1 .295  1 .398 3  2  2 6PGD2 .288  .287  ALD4  5  2 .282 .192 .231  3 .166 2  1  .218  .154 1  4 Rf  3  IDH  Rf  MDH  .379 .364  I  .372  2 5  2.346 .333  .318  I I I  3 .288 .273  1 3  4 Rf  I I  I I I f I Figure 2  ME .394 MEl .364 2 1 ME2 .166 .152 2 1 .136  1  I I Rf  PGM PGM2 .368  I .338  3  1 .324 2  3  Enzyme banding patterns for eighteen loci. Band  (  ) indicates allele(s), three  figures over a band (e g .345) indicate relative migrating distance (Rf) of the band or allele, and a figure under a band indicates the allele number in that locus.  56 Figure 2 (continued)  Rf  SDH  Rf  D1A  .225  .683  .200 2  .667  3 .650  .188 1  1 3  Rf  2  [ST .727  ESTI .712  2 .667  3 .652  1 4  .470 EST2 .364  2  1  Rf  AAT AAT2  I .333  .122 1 .273  EST3 .108 3  2 .182  .095 2  3 Rf  1  PGI  Rf  LAP .500  .424  PGI1 2 .480  1 .409 3 2  .400  I—. .288  1  .386  .273 3 1 .243 2  PGI2  4  57  Figure 3  Electrophoretic gel slices stained with two enzyme systems, PGM and IDH. A). Above gel slice was stained with PGM enzyme and scored 1 for one locus (PGM2) which was homozygous (two alleles were the same). B). Below gel slice was stained with IDH enzyme and scored for one locus (IDH) which was either homozygous or heterozygous (two alleles were different)  Table 15  MEAN  2. ML2  1. ML1  UNDISTURBED POPULATION  OVERALL POPULATION (species)  Population  94.4  2.3 (0.2)  268.8 (4.9)  86.1  78.8  97  88.9  Percentage of loci polymorphic*  (0.2) 2.4 (0.1)  3.4 (0.2)  Mean no. of alleles per locus  324.2 (5.2) 213.3 (4.6)  1085.9 (17.8)  Mean sample size per Locus  0.333 (0.069)  0.320 (0.075) 0.346 (0.063)  0.289 (0.056)  Directcount  Mean heterozygosity  0.272 (0.046)  0.244 (0.046) 0.300 (0.045)  0.266 (0.042)  HdyWbg expt**  treated as one unit for analysis), each population, and the means averaged at each level (SE in parentheses)  for direct-count and Hardy-Weinberg expectation at 18 loci for the species level (data from all populations were combined and  Variability parameters including mean number of alleles per locus, percentage of loci polymorphic, and mean heterozygosities  00  **  *  105.3 (4.2) 92.2 (5.4) 74.4 (3.5) 79.3 (0.6) 137.7 (2.2) 60.0 (0.0) 91.5 (2.7) 135.8 (3.2)  Mean sample size per Locus  1.8 (0.1) 1.7 (0.2) 2.2 (0.3) 2.2 (0.3) 2.4 (0.2) 2.5 (0.2) 2.1 (0.2) 2.2 (0.2)  Mean no. of alleles per locus  72.9  68.5  88.9  72.2  55.6  55.6  61.1  77.8  Percentage of loci polymorphic*  0.289 (0.061) 0.203 (0.063) 0.200 (0.066) 0.257 (0.079) 0.240 (0.068) 0.33 1 (0.066) 0.253 0.067) 0.273 (0.068)  Directcount  Mean heterozygosity  A locus is considered polymorphic if the frequency of the most common allele does not exceed 0.99 Unbiased estimate (see Nei, 1978)  OVERALL MEAN  MEAN  8. SK2  7. SKI  6. MK2  5. MKI  4. KK2  3. KKI  DISTURBED POPULATION  Population  Table 15 (continued)  0.221 (0.042) 0.157 (0.042) 0.172 (0.051) 0.207 (0.055) 0.230 (0.052) 0.299 (0.050) 0.214 (0.049) 0.229 (0.048)  HdyWbg expt**  60 Five alleles were recorded at locus 6PGD2. The frequency means for alleles one, two, and three were 0.703, 0.263, and 0.035, respectively. The frequency means of alleles four and five were quite low (0.005 and 0.001, respectively) and the alleles were observed with low frequencies in two and one populations, respectively. AAT displayed at least two loci but only locus AAT2 was consistent and scorable.  Three  alleles were recorded at locus AAT2. The frequency means for alleles one, two ,and three were 0.526, 0.294, and 0. 179, respectively. ALD codes for at least four loci. Two of these (ALD3 and ALD4) were scorable. There were three alleles observed at locus ALD3. The frequency means for alleles one, two, and three were 0.785, 0.164, and 0.051, respectively.  Mean allele frequencies for 18 loci across 8 populations  Table 16  Locus  Allele  No 1  6PGD1 6PGD2 AAT2 ALD3 ALD4 DIA EST1 EST2 EST3 IDH LAP MDH MEl ME2 PGI1 PGI2 PGM2 SDH2 * **  141 142 105 125 128 143 137 135 135 143 142 133 143 143 133 135 143 142  0.89 0.70 0.53 0.79 0.98 0.98 0.63 0.76 0.81 0.74 0.72 0.90 0.98 0.91 0.99 0.94 0.91 0.82  2 0.09 0.26 0.29 0.16 0.01 0.02 0.34 0.20 0.16 0.15 0.26 0.09 0.02 0.08 0.01 0.00 0.06 0.12  3 0.00 0.04 0.18 0.05 0.01 -  0.02 0.04 0.03 0.03 0.02 0.01  4 -  -  -  -  -  0.01 -* -  -  -  0.01 -  -  -  -  0.03 0.06  5  -  IDH, allele 4 existed in some populations but the overall mean was very close to zero PGM2, allele 3 existed in some populations but the overall mean was very close to zero  0.08  61 Locus ALD4 also had three alleles. The frequency mean for allele one was high (0.98), but quite low for those of alleles two and three (both were 0.01). DIA enzyme displayed mainly a single and thick band of one locus, though three alleles were recorded. The frequency means for allele one and two were 0.98 and 0.02. The frequency mean for allele three approached zero (0.001) EST was scored for three loci (EST1, EST2, and EST3). Four alleles were recorded at the first locus, EST1.  The frequency means  for alleles one, two, three, and four were 0.626, 0.339,  0.024, and 0.011, respectively. Three alleles were recorded at the second locus, EST2. The frequency means for alleles one, two, and three were 0.757, 0.206, and 0. 37, respectively. Three alleles were recorded at the third locus, EST3. The frequency means for alleles one, two, and three were 0.807, 0.160, and 0.026, respectively. IDH displayed clearly distinct hands for one locus which appeared to he dimeric. Locus IDH had five alleles. The frequency mean was low for allele four (0.004). The frequency means for alleles one, two, three, and five were 0.738, 0.153, 0.033, and 0.076, respectively. LAP displayed only one clear banding locus with four alleles. The frequency means for alleles one, two, and three were 0.718, 0.261, and 0.019, respectively. The frequency mean for allele four was very low, approaching zero (0.001). MDH displayed one scorable locus with quite clear bands. The locus possessed three alleles. The frequency means across all populations were 0.898, 0.088, and 0.014 for alleles one, two, and three, respectively. ME displayed two loci (MEl and ME2). Locus MEl provided two alleles with the frequency means for alleles one and two of 0.98 and 0.02, respectively. Three alleles were recorded at locus ME2 with frequency means for alleles one and two of 0.92 and 0.076, respectively. The frequency mean for allele three was low (0.006). PGI had two loci scorable (PGII and P012). Two alleles were recorded at the first locus,  62 PGI1. The frequency mean was high for allele one (0.992) and very low for allele two (0.003). PGM yielded two loci, but only one clear locus, PGM2, was scored with four alleles being recorded. Alleles one, two, and four gave frequency means of 0.911, 0.056, and 0.03, respectively, whereas allele three had a very low frequency mean of 0.003. SDH was also scored for one clear locus, SDH2, which produced four alleles. The frequency means for alleles one, two, and three were 0.824, 0.116, and 0.061, respectively. Three alleles were recorded at the last locus, P012. The frequency means for alleles one and two were 0.94 and 0.06, respectively. Because allele three occurred with very low frequency (0.002) in only the population ML1, the frequency mean across all populations was approximately zero.  (b). Heterozygosity  Mean heterozygosities for all studied levels, including species, undisturbed populations, disturbed populations, and individual populations, are presented in Table 15.  The mean  heterozygosities for direct-count were higher than those for unbiased estimates at all levels. The mean heterozygosities at the species level or total population (eight populations were treated as one unit for the analysis) were 0.287 and 0.266 for direct-count and unbiased estimate (Nei, 1978), respectively. The mean heterozygosities (averaged from the mean heterozygosities of eight populations) were 0.273 and 0.229 for direct-count and unbiased estimate, respectively. Mean heterozygosities for undisturbed and disturbed populations were calculated by averaging from the mean heterozygosities of populations within each disturbance level correspondingly. The mean heterozygosities for individual loci at each population and their averages across eight populations are presented in Appendix I.  Mean heterozygosities based on combined data of eight  populations for each enzyme locus are also given in Table 18. The EST1 locus had the largest directcount and unbiased heterozygosities (0.714 and 0.479, respectively). The PGI1 locus gave the smallest heterozygosities (0.015 for both unbiased and direct-count). Loci 6PGDI, ALD4, 1DB, MEl, PGM2,  63 and DIA had lower mean heterozygosities for direct-count than those for unbiased estimates.  Table 17  Overall means of unbiased and direct-count heterozygosities for each locus based on data combined from eight populations  Locus Direct-count  6PGDI 6PGD2 AAT2 ALD3 ALD4 DIA EST1 EST2 EST3 IDH LAP MDH MEl ME2 PGI1 PGI2 PGM2 SDH2 *  Heterozygosity HdyWbg expt  0.156’ 0.565 0.634 0.423 0.033* 0.028* 0.714 0.486 0.307 0.297* 0.429 0.194 0.030* 0.155 0.015 0.122 0.079* 0.288  0.189 0.420 0.440 0.329 0.039 0.034 0.479 0.354 0.277 0.377 0.321 0.164 0.038 0.145 0.015 0.111 0.140 0.246  Direct-count smaller than unbiased expectation  (c). Number of alleles per locus  Number of alleles per locus are also displayed in Table 15. number of alleles per locus was 3.4.  At the species level,  mean  Mean number of alleles per locus was largest in the SK2  population (2.5) and smallest in the population KK2 (1.7). The averaged number of alleles per locus across eight populations was 2.2. Mean number of alleles per locus for the undisturbed populations (averaged across two populations) was 2.3.  Mean number of alleles per locus for the disturbed  populations (averaged from six disturbed populations) was 2.1.  64 (d). Percentage of loci polymorphic  A locus was considered to be polymorphic if more than one allele with frequency 0.01 or more was detected. The percentage of polymorphic loci for each population ranged from 55.6 (population MK1 and MK2) to 94.4 (population ML2) (Table 15). The percentages of polymorphic loci averaged across all populations, undisturbed populations, and disturbed populations were 72.9, 86.1, and 68.5, respectively. The percentages of polymorphic loci for undisturbed populations were slightly higher than those of disturbed populations. The percentage of polymorphic loci for the species level was 88.9.  (e). Deviation from Hardy-Weinberg equilibrium  Chi-square tests for Hardy-Weinberg equilibrium provided different results for each locus in individual populations. Table 18 presents the chi-square (X ) value and significance probability for each 2 locus in the individual populations. The test was valid at a particular locus in a certain population because the test will be performed only when every progeny (radicles) from every family in that population had been scored. The smallest population in the study consisted 3 families; the progeny used for scoring then 60 (20 from each families); therefore, every locus which was tested for deviation from Hardy-Weinberg equilibrium was at least scored from 60 progeny. However, at the species level (eight populations treated as one total population), only seven loci which were tested for the deviation in all eight populations were valid. These loci were 6PGD1, 6PGD2, ALD3, EST1, EST2, IDH, and MDH. Table 19 gives contingency chi-square for eighteen loci was 6725.837 with 301 degrees of freedom, which was highly significant.  65  Table 18  Chi-square test for deviation from Hardy-Weinberg equilibrium of individual loci in each population.  LOCI  6PGD1 X2 A P 6PGD2 X2 A P AAT2 X2 A P ALD3 X2 A P ALD4 X2 A P DIA X2 A P EST 1 X2 A P EST2 X2 A P EST3 X2 A P IDH X2 A P  KK1  KK2  MLI  ML2  MKI  MK2  SKi  0.96 2 0.33  1.17 2 0.28  0.00 2 0.97  3.04 2 0.08  0.00 2 1.00  72.60 4 0.00  82.35 4 0.00  0.81 3 0.85  16.35 2 0.00  33.96 2 0.00  65.24 2 0.00  23.64 3 0.00  0.25 3 0.97  15.03 4 0.02  21.21 5 0.02  10.66 3 0.01  59.92 0.00  101.56 3 0.00  6.70 3 0.08  64.35 3 0.00  88.54 3 0.00  59.00 3 0.00  45.81 2 0.00  100.90 3 0.00  5.63 3 0.13  1.06 3 0.79  8.74 3 0.03  2.87 3 0.41  23.99  3.25 3 0.36  0.00 2 1.00  0.00  -  -  -  -  19.80 3 0.00  0.09 2 0.77  -  -  -  -  7  7  0.00  -  -  -  -  0.00 2 0.96  7  1.00  -  -  -  SK2  1.85 7  0.17 2.36  -  7  -  0.12  -  21.49 2 0.00  36.76 2 0.00  287.26 79.18 3 3 0.00 0.00  6.26 4 0.40  16.70 3 0.00  14.91 2 0.00  21.05 3 0.00  10.51 2 0.00  32.52 3 0.00  218.87 3 0.00  7.75 2 0.01  0.25 2 0.62  4.45  6.15 3 0.10  8.33 3 0.04  3.77 2 0.05  1.31 2 0.25  12.74 3 0.01  0.93 2 0.33  0.03 2 0.86  1.42 2 0.23  5.46 2 0.02  36.32  6.93 2 0.01  0.05 2 0.84  4.99 2 0.03  27.21 5 0.00  47.89 3 0.00  22.63 3 0.00  7  0.00  7  0.04  31.88 4 0.00  66 Table 18 (continued)  LAP X2 A P MDH X2 A P MEl A P ME2 X2 A P PGI 1 X2 A P PGI2 X2 A P PGM2 X2 A P SDH2 X2 A P  2.03 2 0.15 14.33 2 0.00  -  -  -  0.01 2 0.94  -  -  -  -  -  -  4.12 2 0.04 0.11 2 0.74 4.66 2 0.03 0.07 2 0.80 13.67 2 0.00  0.81 2 0.37  36.81 2 0.00  0.01 1.15 3 0.77 13.95  0.82 3 0.85  1.22 3 0.75  64.74 3 0.00  79.00  0.19 2 0.66  0.39 3 0.94  9  0.00  -  -  -  16.96 3 0.00  -  -  -  -  -  -  37.06 3 0.00  16.84 3 0.00  0.87 4 0.99  0.02  0.01  9  9  0.88  0.93  0.02  1.14  9  0.00 2.09 3 0.55  2  0.88  9  7  0.82 1.09 2 0.30 1.01 9  0.32 20.57 3 0.00  7  0.29  0.00  0.05  -  2.34 3 0.51  9  6.60 2 0.01 8.28 2 0.00  -  1.04 2 0.31  8.00  0.95 -  -  -  -  -  -  0.07 2 0.8  -  -  -  -  0.21 2 0.65  0.20 2 0.66  134.24 32.79 3 4 0.00 0.00 0.00 9  0.95  4.26 9  0.04  X2 Chi-square A = Number of alleles at a locus P = significance probability  Finally, coefficients for heterozygote deficiency or excess, including the fixation index of each gene locus, were analyzed from the pooled population (treating eight populations as one pooled unit), under the Hardy-Weinberg principle. Table 21 gives the coefficients and fixation indices of individual loci. The fixation index was least for locus EST1 (-0.537) and greatest for locus PGM2 (0.46 1). Figure 4 shows the genetic variability parameters from the isozymes analysis for individual  67 populations.  Table 19  Contingency chi-square analysis for deviation of allele frequencies from those of which were randomly sampled from the total population  Locus  6PGD1 6PGD2 AAT2 ALD3 ALD4 DIA ESTI EST2 EST3 IDH LAP MDH MEl ME2 PGII PGI2 PGM2 SDH2  No.of alleles  5 5 3 3 3 3 4 3 3 5 4 3 2 3 2 3 4 3  (Totals)  Chi-square  D.F.  P  314.032 321.491 700.223 236. 124 154.322 353.214 350.495 375.276 536.853 805.559 720.349 291.266 103.005 77.647 34.850 68.058 721.486 561.588  28 28 14 14 14 14 21 14 14 28 21 14 7 14 7 14 21 14  .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00001 .00000 .00000 .00000  6725.837  301  .00000  (f). Genetic similarity and distance coefficients  Genetic similarity and distance coefficients provide the information about magnitude or extent of differences among each pair of populations. Different models have been developed for estimation of genetic similarity and distance. Nei’s (1978) unbiased genetic distances and Nei’s (1978) unbiased genetic identities are shown in Table 21.  Figure 4  0  0.5  1  1.5  2  2.5  3  MKi  MK2  Ski  SK2  Hardy-Weinberg heterozygosities from eighteen loci for each population  Genetic variability including number of alleles per locus, probability of polymorphic loci, and observed and expected  Exp. heterozygosity  ML2  Obs. heterozygosity  ML1  Prob.polymorioci  KK2  No alleles/locus  KK1  00  69  Table 20  Expected and observed heterozygosity, and resulting fixation index for each locus, based on the population.  A negative fixation index represents a heterozygote  deficiency  Locus 6PGD1 6PGD2 AAT2 ALD3 ALD4 DIA EST1 EST2 EST3 IDH LAP MDH MEl ME2 PGI1 PGI2 PGM2 SDH2  Heterozygotes Observed  Expected  182.533 508.543 466.070 406.785 50.921 20.832 551.381 447.594 383.802 417.319 419.034 225.010 90. 161 176.564 15.887 134.520 140.998 488.527  152 662 536 518 40 15 847 598 353 323 375 235 70 181 16 143 76 410  Fixation index (F) .167 -.302 -.151 -.274 .214 .280 -.537 -.337 .080 .226 .105 -.045 .223 -.026 -.008 -.064 .461 .160  Nei’s genetic distance coefficients were largest between the populations KKI and MKI 0.082 and smallest populations MK2 and SKi  =  0.009.  70 Matrix of genetic distance and/or similarity coefficients  Table 21  A. Nei (1978) unbiased genetic distance  Population 1 2 3 4 5 6 7 8  DISTURBED-KK1 DISTURBED-KK2 UNDISTURBED-ML1 UNDISTURBED-ML2 DISTURBED-MK1 DISTURBED-MK2 DISTURBED-SKi DISTURBED-SK2  KK1  KK2  .014 .021 .035 .082 .073 .083 .057  ***  .018 .052 .077 .057 .075 .042  ML1  ML2  MK1  MK2  SKi  SK2  *****  .024 .067 .052 .067 .050  ***  .060 .058 .057 .046  *****  .019 .014 .046  *****  .009 .037  .038  *****  MLI  ML2  MKI  MK2  SKI  SK2  .991 .964  *****  B. Nei (1978) unbiased genetic identity  Population 1 2 3 4 5 6 7 8  DISTURBED-KKI DISTURBED-KK2 UNDISTURBED-ML1 UNDISTURBED-ML2 DISTURBED-MK1 DISTURBED-MK2 DISTURBED-SKi DISTURBED-SK2  KKI  KK2  .986 .979 .965 .921 .930 .921 .945  **  .982 .950 .926 .944 .928 .959  *‘  .977 .936 .950 .936 .952  *****  .941 .944 .944 .955  **  .981 .986 .955  .963  *****  (g). Genetic differentiation of populations by F-statistics (Wright, 1965, 1978)  F-statistics describing the structure of variation within and among populations were calculated for individual alleles in each locus. Means of F(IS), F(IT), and F(ST) of each locus are presented in Table 22. The overall locus mean of F(IS) was -0.200. F(IS) was largest for locus PGM2 (0.423) and smallest for locus EST1 (-0.498). The overall locus mean of F(IT) was -0.048.  Similar to the F(IS) mean, F(IT) mean was  71 largest for locus PGM2 (0.524) and smallest for locus ESTI (-0.450). The overall locus mean of F(ST) was 0.127. F(ST) mean was largest for locus AAT2 (0.278) and smallest for locus PGII (0.019).  F-statistics at all loci from 8 populations (S.E. in parenthesis): F(IS) and F(IT) are  Table 22  the fixation indices within populations and within total population, and F(ST) is the genetic differentiation among populations  Locus  F(IS)  F(IT)  F(ST)  6PGD1 6PGD2 AAT2 ALD3 ALD4 DIA EST1 EST2 EST3 IDH LAP MDH MEl ME2 PGI1 PGI2 PGM2 SDH2  0.166 -0.351 -0.453 -0.293 0.166 0.181 -0.498 -0.381 -0.109 0.208 -0.343 -0.190 0.196 -0.071 -0.027 -0.111 0.432 -0.022  0.214 -0.274 -0.049 -0.194 0.197 0.293 -0.450 -0.269 0.042 0.296 -0.033 -0.044 0.247 -0.036 -0.008 -0.062 0.524 0.179  0.057 0.057 0.278 0.077 0.038 0.137 0.032 0.081 0.137 0.111 0.231 0.122 0.064 0.033 0.019 0.044 0.162 0.196  Mean  -0.200  -0.048  0.127(0.135)*  *  Standard error for F(ST) mean averaged from 1 8 loci  (h). Cluster Analysis  Nei’s (1978) unbiased genetic distance was used for cluster analysis and construction of a cluster tree. The results of the cluster analysis using the unweighed pair group method are shown in Table 23. The analysis provided seven clustering levels with 4 cycles of pairing. The cluster analysis  72 tree is shown in Figure 5.  Table 23  Cluster analysis for Nei ‘5 (1978) unbiased genetic distance using unweighed pair group method algorithm (UPGMA)  Population or cluster numbers joined KK1 MK2 KKI MKI KK1 MK1 KK1  KK2 SKI ML1 MK2 ML2 SKi MK1  Cycle  Clustering level .01394 .00893 .01986 .01634 .03687 .04053 .06272  2 2 3 3 4  Subsequently, based on Wright (1978), hierarchical analysis was carried out by nesting the populations KK1, KK2, MLI, and ML2 to the central region and nesting the populations MK1, MK2, SKi, and SK2 to the northeast region. Table 24 presents the F-statistics and variance components for among populations within regions (F(SR)), among populations within total or across populations (F(ST)), and among regions within total or across populations (F(RT)) of each locus and combined across loci. The F(SR), F(ST), and F(RT) for combined across loci were 0.086, 0. 122, and 0.040, respectively.  The difference among populations F(ST)  from  this calculation was comparable to the  F(ST) in Table 28; they were 0.122 and 0. 127, respectively. The difference among populations was significant, but among regions was nonsignificant. Negative variance components resulted from the orthogonal partition of total variance; when preceding variance components exceeded the total variance, the remaining variance components took negative values (Swofford and Selander, 1989). In summary, the variability parameters of this tree species were relatively high for all studied levels.  Variability parameters differed among the populations.  The parameter estimates averaged  across undisturbed populations were somewhat larger than those of disturbed populations. The number  73 of alleles per locus and percentage of polymorphic loci also varied among populations. The chi-square test for deviation from Hardy  -  Weinberg equilibrium was significant for half of the loci analyzed. The  cluster analysis divided populations into two main geographical groups, one from central and the other from northeastern Thailand.  .10  .08  .07  Distance .05 .03  .02  .00  +----+----+----+----+----+----+----+----+----+----+----+----+  DISTURBED- KK1 DISTURBED-KK2 UNDISTURBED-ML1 UNDISTURBED-ML2 DISTURBED -MK1 DISTURBED-MK2 DISTURBED-SKi DISTURBED - SK2  .10  Figure 5  .08  .07  .05  .03  .02  .00  Clustering tree using Nei’s (1978) unbiased distances  3.2. Mating system  The fifteen loci selected for the mating system analysis included 6PGD1, 6PGD2, AAT2, ALD3, ALD4, EST2, EST3, IDH, MDH, ME], ME2, PGTI, PGT2, PGM2, and SDH2. The outcrossing rate and fixation index for each population are shown in Table 25. Appendix K presents the outcrossing rates for individual families. Due to the poor convergence of estimates in most families, the family outcrossing rates were overestimated (values more than 1.0) and standard error estimates were invalid. The overall mean for population outcrossing rate was 0.971. The means of outcrossing rates for undisturbed and disturbed populations were slightly and non-significantly different, 0.980 and 0.968, respectively. The mean fixation index over all populations was 0.05 1. The mean fixation index of undisturbed populations was 0.005; disturbed populations had a larger fixation  74 index mean, 0.051. Outcrossing rates for most individual families were high and mainly greater than 1.0. This problem always occurs with mutiple loci analysis of mating system when population size quite small. Only ten of fifty-seven families had outcrossing rates of I .0 or lower. For instance, the outcrossing rate for one family in the population KK1 was only 0.2, and 0.69 for another family in the population MK1. Inferred maternal genotypes for each family are presented in Appendix L.  Hierarchical analysis  Table 24  using  F-statistics (Wright, 1978).  F-statistics for among  populations within regions (F(SR)), among populations within total populations (F(ST)), and among regions within total populations (F(RT)); variance component is in parenthesis (S.E. in parenthesis)  Locus  F(SR)  F(ST)  F(RT)  6PGDI 6PGD2 AAT2 ALD3 ALD4 DIA EST1 EST2 EST3 IDH LAP MDH MEl ME2 PGI1 PGI2 PGM2 SDH2  0.063 0.050 0.157 0.051 0.039 0.147 0.020 0.059 0.163 0.085 0.116 0.088 0.056 0.026 0.016 0.024 0.176 0.098  0.052 0.053 0.272 0.072 0.033 0.130 0.027 0.077 0.133 0.106 0.227 0.119 0.062 0.029 0.015 0.040 0.157 0.194  -0.011 0.003 0.136 0.021 -0.007 -0.020 0.007 0.019 -0.037 0.023 0.126 0.033 0.007 0.002 -0.001 0.017 -0.023 0.107  ALL LOCI  0.086  0.122 (0.135)*  *  Standard error for F(ST) and F(RT) means averaged from 18 loci  0.040(0.002)*  75  Table 25  Outcrossing rate (t), and fixation index for each population estimated from fifteen loci (S.E. in parenthesis)  Population  f  t  Undisturbed ML1 ML2  0.984(0.020) 0.976 (0.020)  0.004(0.100) 0.005 (0.132)  Ave.  0.980 (0.020)  0.005 (0.116)  KK1 KK2 MK1 MK2 SKi SK2  0.972 0.998 0.928 0.965 0.956 0.987  0.293 0.087 0.009 0.002 0.001 0.000  AVE.  0.968 (0.085)  0.065 (0.739)  0.971 (0.069)  0.051 (0.583)  Disturbed  Overall Ave.  (0.212) (0.142) (0.000) (0.000) (0.027) (0.127)  (0.167) (0.266) (1.000) (1.000) (1.000) (1.000)  4. Discussion  A high amount of genetic variability exists at the species level of Dalbergia cochinchinensis  Pierre. Compared with average values for gymnosperms (Harnrick and Godt, 1989), mean number of alleles per locus, percentage of loci polymorphic, and mean of expected heterozygosity of this tree species are higher.  Most loci at the species level were polymorphic; indeed, some of the initially  surveyed loci were excluded before the analysis if their band patterns were not clear or segregated. The expected mean heterozygosity in this species was 0.229 (Table 15), compared with 0.173 for conifers, 0.177 for long-lived woody perennials and 0. 137 for narrow ranging species (Hamrick and  76 Godt, 1989). The expected heterozygosity is comparable to other tropical tree species, which range from 0.106 to 0.374 (Ledig, 1986a). However, the heterozygosity was higher than that averaged for twenty-nine taxa of tropical woody angiosperms, which was 0.111 (Hamrick and Loveless, 1986). Mean heterozygosity for sixteen taxa of common (more than four individuals per hectare) tropical trees was 0.211 (Hamrick and Loveless, 1989).  Hamrick and Murawski (1991) also found mean  heterozygosity for ten taxa of uncommon (less than five individuals per hectare) tropical trees to be  0.142. In comparison with Dalbergia cochinchinensis Pierre, which is considered uncommon or rare, the mean heterozygosity for other uncommon tropical trees is, therefore, only about 50% of the value calculated for this species. At the population level, genetic variability parameters of this tree species were relatively high. The small population sizes and difference in number of families in the populations (from 3-17 families) might contribute to the high variability among the populations. The direct-counted heterozygosity is higher than the expected heterozygosity in every population.  The possible reasons for excess of  heterozygotes are heterozygous advantage, negative assortative mating and differences in allele frequencies between male and female gamete pools (Schnabel and Hamrick, 1990; Surles et al., 1990). The first reason (that is selection favouring heterozygous genes) may be the most likely for this species.  The second and third reasons are not likely because the species has perfect flowers and no variation in flower phenology. However, half of the loci showed heterozygote deficiency when analyzed across populations (or at the species level). Means of all genetic variability parameters of undisturbed and disturbed populations were not different, because some disturbed populations had variability parameter more or less than undisturbed populations. The disturbed populations might be degraded in terms of genetic variability in some instances; however, variability in the disturbed populations remained high and valuable for future management. F-statistics indicated that mean deviation of genotype frequencies from Hardy-Weinberg expectations for overall populations, F(IT), and within individual populations, F(IS), were -0.048 and -0.200, respectively. The proportion of genetic variability among populations was 0.127. This means  77 that almost 13 % of total allozyrne variability is accounted for by differences in allele frequencies among populations. The proportions of genetic variability for eight tropical tree species have been reported by Hamrick and Murawski (1991) to vary from 0.064 and 0.163. Ledig and Conkle (1983) reported three to twelve percent F(ST) values for conifers. This could be a result of limited gene flow and the development of spatial genetic structure (Schnabel et at, 1991). The cluster analysis using Nei’s (1978) genetic unbiased distance divided the studied populations into two main geographic groups. The central area group included the KK1, KK2, ML1, and ML2 populations and the northeast area group included MK1, MK2, SKi, and SK2.  However, the hierarchical analysis using Wright’s (1978) statistics  indicated no significant difference among the regions (F(RT)  =  0.040). There were two interesting  anomalies. First, the MLI and ML2 populations, which were very close in geographic distance, did not link together in the first cycle, but in the third. Second, the SKi and SK2 populations were also geographically close but the two linked together in the third cycle instead of the first. The reason for these results probably the populations had been decreased and lost of genes (genetic drift), consequently, different genes remained at different population. The mating system of the species appears to be highly outcrossed with a mean outcrossing rate (t) of 97.1% (S.E. 0.069). Similar outcrossing rates have been reported for  Pithecellobium pedicellare,  a tropical forest leguminous tree, with t averaging nearly 1.0 (O’Malley and Bawa, 1987). The possible reasons for the high outcrossing rates estimated from this study are negative assortative mating which may be a consequence of self-incompatibility, heterotic selection against inbreeding offspring, and effective pollen-mediated gene flow (O’Malley and Bawa, 1987). However, the most likely reason has not been studied. High outcrossing rates are characteristic of populations with low densities because flight distances of pollinators, e.g. insects, are maximized. Consequently, the frequency of mating among non-related individuals is also maximized (Ellstrand et at., 1987). Furthermore, sampling error is likely when outcrossing rate is near 1 .0 and sample sizes are small (Adams and Bikes, 1988). Another tropical species, Berthoiletia excelsa, has been reported to have an 85% outcrossing rate. Although mating system in a plant species varies in space and time (Hamrick,  78 1982), the results from this study indicated no significant differences among the tree populations.  5. Conclusion  From twelve enzyme systems producing eighteen loci, isozyrne analysis revealed high genetic variability of Thailand rosewood (Daibergia cochinchinensis Pierre).  The individual populations  showed different degrees of variability. However, the differences between undisturbed and disturbed populations were not significant. Half of the loci had genotype frequencies deviating from Hardy Weinberg expectation. Nevertheless, numbers of loci varied significantly among populations (from 2 loci in population MK1 to 12 loci in population ML 1). Nei’s (1978) unbiased genetic distance and Fstatistics (Wright, 1978) indicated moderate levels of differences among populations. Cluster analysis divided populations into two geographic regions, central and northeast, but the regional difference was not significant. High outcrossing was found in all the populations with the outcrossing rate very close to 1.0. The fixation indices for most populations were correspondingly very low.  79  CHAPTER 4  CONSERVATION STRATEGIES  A conservation programme for this tree species has been established by the Thai Royal Forest Department and Danish International Development Agency (Sa-ardavut et al., 1989). However, the numbers of trees and locations have not been based on genetic information about the variation and dynamics of population structure. The genetic variability and mating system information derived from this study was therefore applied in support of the existing conservation programme, to improve its scientific rationale. Due to the inconsistent amounts and patterns of variation among all traits studied, resolution of these differences has to be drawn step by step for considering efficient conservation strategies. Subsequently, the genetic information was considered together with the socio-economic situation for considering strategies to revise the conservation programme for future implementation.  1. Quantitative Trait Considerations  Quantitative traits are complex and under the control of environmental conditions. Variation in quantitative traits is of great importance in gene conservation because the conservation strategy is to determine efficient sampling to preserve genes adapted to the environment throughout a species range (Fumier et al., 1991). The variation in seed germination, R50 and seed weight was mainly accounted for by differences among families within populations. The variation in R50 among populations was, however, significant.  The proportions of variation among populations to among families for seed  germination, R50, and weight were 6.74, 26.67, and 5.49%, respectively. The general explanation for  80 high variation among families was that only one measure was used for each replication and this measure was a mean or treated as a mean of the replication; therefore, this replicated mean was confounded with variation within a family which generally was assigned to error term when more than one measures from each experimental plot were used for analysis; the replicated mean confounded with variation within a family caused decreasing the error term and increasing variation among families and consequently, high estimated F-ratio for variation among families. The variance analysis for seed traits was simply conducted in a well controlled laboratory environment and consequently decreased unknown errors.  The small error terms were mainly the results of variation among replications and were  required for variation study. Moreover, using replicated means withdrew most of genetic variation within families from the error terms. Seed traits influence survival and reproductive success of species and the information is desirable for nursery implementation. The variation in most seedling and progeny traits, on the other hand, existed both among populations and within populations. The proportions of variation among populations to among families for seedling height, diameter, and dry weight were 62.18, 0.00, and 42.89%, respectively.  The  proportions of variation among populations to among families for progeny height, diameter, and stem form were 71.90, 85.87, and 67.41, respectively. Generally, the variation among populations for quantitative traits in this study was moderate for R50 (F  =  3.86), relatively high for seedlings height and root-shoot ratio  ( F  =  17.79 and 11.69,  respectively), and moderate to relatively high for progeny height, diameter, persistence of axis, stem straightness, and branching characteristic (F  =  8.43, 4.73, 6.55, 2.83, and 3.55, respectively). This  reflected the adaptation and/or selection of polygenic traits under different environments. However, the variation of seed germination, seed weight, which preconditioning maternal environmental effects of individual families were likely to dominate the variation, were not different among populations. The variation among population for seedling diameter and dry weight was also not different, probably these traits carried over preconditioning maternal environmental effects longer than other seedling traits. The only seed trait correlated with seedling traits was R50. (Table 26). Although statistical  81 correlations between seed and progeny traits and between seedling and progeny traits were not possible due to different families (Appendix A) being used for seed and seedling studies and the progeny test, the amounts and patterns of variation were compared to those for seed and seedling traits. R50 was the only seed trait producing a comparable amount and pattern of variation to those of seedling and progeny traits. Consequently, the conclusion for quantitative traits was that the genetic variation was relatively high within and among the populations of this tree species. The seed trait variation provided more reliable results than those for seedling and progeny variation because the experimental errors for seed variation estimates were reduced under the controlled environmental conditions. On the other hand, seedling and progeny variation were subject to environmental differences of individual seedlings and unknown within family variation.  However, all quantitative traits from this study are mainly  parallel in patterns of the variation and should he combined for the effective conservation of adaptive genes. The amount and pattern of variation will be compared with those of isozymes to make the final decision on conservation strategies.  Table 26  Correlation between seed and seedling traits  TRAIT  R50  GER  -.0615  GER R50 WT HT DM TDW  -.0615 .2046* .1369 .0912 -.0575  N of cases:  221  .0592 .3240** .1963* .1832*  1-tailed Signif:  * -  WT  HT  DM  .2046* .0592  .1369 .3240** .0944  .0912 .1963* .0796 .7478**  .0944 .0796 -.0936  .01  ** -  .7478** .3015**  .2535**  .001  2. Isozyme Consideration  The isozyme analysis indicated that genetic variability at 18 loci was relatively high with mean  82 unbiased heterozygosity averaged from eight populations averaging 0.229, polymorphic loci 72.9 %, and number of alleles per locus 2.2.  The genetic differences derived from Nei’s (1978) unbiased  genetic distance ranged from 0.009 to 0.083. There was also an indication of relatively high variation among populations from Wright’s F-statistics (12.7%).  This was smaller than the mean variation  among the populations for quantitative traits (31,6%) and only one-fifth of the seedling height (63.72%). Because high variation for isozymes appears to exist within populations, judging from both quantitative and isozyme analyses, the conservation programme of this tree species must involve a large number of families.  However, only sound genetic base would not enough to consider conservation  strategies in tropical regions due to the growing human population surviving on forest land and products. Therefore, to be effective, the conservation requires the integration of socio-economic and political situations properly.  3. Quantitative and Isozyme Correlations  The comparison of quantitative and biochemical variation was based on forty-eight common individual trees (families). Quantitative trait variation for seed germination, R50, seed weight, nursery  seedling height, diameter, total dry weight, and root and shoot dry weight ratio were compared to isozyme variation (expected heterozygosity calculated for each common family by using formula page  51) and the mating system (Appendix K).  The comparison was conducted at both the family and  population levels by using their corresponding means. The correlation coefficients at the family and population levels are presented in Table 27 and 28, respectively. When family means were compared, only the seedling height was significantly correlated with heterozygosity.  However, there was no  correlation between quantitative trait and isozyme variation at the population level.  83 Correlations estimated from individual family means for heterozygosity, quantitative  Table 27  traits and mating system  TRAIT  HETEROZYGOSITY  HETEROZYGOSITY GERMINATION R50 SEED WEIGHT HEIGHT DIAMETER TOTAL DRY WEIGHT ROOT-SHOOT RATIO N of cases: 48  0.1126 0.0206 -0.0692 0.0370 0.0090 -0.1242 0.0550 0.0228  1.0000 0.0748 -0.0067 -0.1265 0.4258* 0.2051 0.1410 -0.2885 *  1-tailed Signif:  -  **  .01  MATING  -  .001  Correlations estimated from population means for heterozygosity, quantitative traits  Table 28  and mating system  HETEROZYGOSITY  TRAIT HETEROZYGOSITY GERMINATION R50 SEED WEIGHT HEIGHT DIAMETER TOTAL DRY WEIGHT ROOT-SHOOT RATIO N of cases:  8  1-tailed Signif:  0.3119 -0.2437 0.2168 0.0600 -0.0504 0.5246 -0.0754 -0.6411  1.0000 0.0447 -0.3156 0. 1097 0.0707 0.2403 0.1718 -0.34 12 * -  .01  **  MATING  -  .001  There was an agreement on high variation among families for quantitative and isozyme traits. The variation among populations was from moderate to high for quantitative traits which were expected to be under control of selection. The variation among populations for isozymes, on the other hand, was just at the moderate level. Isozymes may be selectively neutral or nearly neutral (Kimura, 1983) and the low isozyme differentiation among populations may reflect sufficient levels of gene flow for neutral alleles (Lewontin, 1974).  However, these levels of gene flow may be insufficient to prevent  84 differentiation at loci under selection (Muona, 1989). Therefore, isozyme variation alone is not enough to predict the distribution of variation for selectively important traits for designing sampling strategies for gene conservation programmes (Muona, 1989; Falkenhagen, 1985; Hamrick, 1983). The effective gene conservation programmes continue to rely on data from quantitative trait variation (Fumier et al., 1991).  Nevertheless, isozymes remain useful for studies of genetic structure within populations  (Epperson, 1989; Fumier et al., 1987) and plant mating systems (Brown, 1989).  These genetic  estimates also provide the necessary information for determining the number, size and distribution of conservation areas.  Thus the most effective gene conservation should be implemented on the  complementary bases of both quantitative and isozyine variation.  4. Genetic Rationale  Genetic resources are the foundation of biological diversity which forms the essential link between evolution in the past and future adaptation to environmental change (Yeatman, 1987). Conservation of tropical forests is one of the world’s most pressing concerns today because the losses of genetic resources are rapid and cause changes in the world’s ecosystem (Noss, 1991). Conservation of genetic resources requires that selected, representative populations be regenerated from generation to generation (Yeatman, 1987). The principles of genetic conservation are the same for all living organisms. However, the strategies and methodology vary according to the distribution and biological nature of the species (FAO, 1989). The biological diversity is so complex and intangible that its conservation cannot be approached without focusing on specific levels including genes, populations, species, communities, ecosystems, and biomes (Salwasser, 1990). Conservation must he carried out at all levels (Namkoong, 1990).  However, current conservation efforts seem to emphasize the community habitat level and  higher, and rarely operate at the species and lower level (Falk, 1990). Conservation efforts for rare and endangered plant species, many of which are narrowly endemic and have restricted ranges which  85 cause serious threats to their survival, require a sound basis in scientific knowledge. Since biological diversity is the principal goal of conservation, integrating endangered species conservation into ecosystem management should be the best strategy (Salwasser, 1991). This strategy, however, requires empirical and theoretical knowledge in ecology.  Lande (1988) has addressed  population demographic basis important for biological conservation, and advocated incorporation of population demography and genetics into future conservation plans. Dalbergia cochinchinensis Pierre is an endangered species and most of its populations have  been disturbed. These populations might be difficult to conserve under the increasing land requirements for agriculture and industrial infrastructure in Thailand. Yeatman (1987) indicated that indigenous tree species and populations are most effectively maintained in situ by natural regeneration and/or by planting trees of local origin and wide parentage.  Therefore, both in situ and ex situ conservation  strategies are needed for retention of the genetic variability of this endangered species.  4.1. In situ conservation  In situ conservation is an ideal strategy which protects and maintains species and populations with the ecosystem where they occur. This will conserve an ecosystem and maintain the variation of  the species and the dynamics of its interactions with the environment and other species (FAO, 1989). Theoretically, in situ conservation of threatened or endangered species requires the estimation and design of minimum viable population sizes. The minimum viable population size is a threshold for the minimum number of individuals which enable a population in a given habitat to persist and regenerate indefinitely (Gilpin and Soulá, 1986).  Three broad approaches to estimating minimum viable  population size have been taken so far (NRC, 1991). The first two approaches, based on effective population size (Franklin, 1980; Soulé, 1980), might he inappropriate to Daihergia cochinchinensis Pierre due to the problems of estimating effective population size for forest species. The third approach  is based on population size which will minimize the sampling loss of low-frequency alleles (NRC,  86 1991).  Namkoong (1984a) estimated that in species with low levels of inbreeding and population  structure or high variation among population, a sample size of 1,000 individuals will minimize the loss of an low frequency allele below the probability of 0.01. Namkoong (1984b) has also suggested that to minimize loss of useful alleles, a population size of 10 to 20 should be maintained in each of 20 to 50 or more subpopulations. Within each suhpopulation, simple recurrent selection or hybridization may simply generate new subpopulations for genetic improvement.  Moreover, because outcrossing is the  mating system of this tree species, with this number of individuals in each subpopulation, conservation should be rational because their genetic constitutions are distinct. Due to the high amount of genetic variability among the families for Thailand rosewood and the relatively high amount of differentiation among populations, the genetic conservation strategies should include a large number of families to cover a substantial proportion of variability.  Thus  Namkoong’s (1984a) criteria would be appropriately applied to this species due to its naturally small population size and scattered distribution. Both quantitative and isozyme results indicated two separate populational regions, central and northeastern regions. Therefore, about 1,000 individuals gathered from small populations or subpopulations in each region would he adequate. The conserved populations should also be distributed throughout the natural range of the species in each region. In many areas, since natural regeneration is rarely successful, in situ conservation should be promoted by collecting seed from many trees to sample (genetically) a given population adequately for planting the next generation at the population site. The success of enrichment planting is dependent on the potential of plants derived from seed. Therefore, designation, protection, and management of seed sources for production of required quantities are prerequisites (Yeatman, 1987).  The current  conservation programme of this tree species, however, rarely emphasizes in situ conservation due to the challenges and high requirements of natural forest land. The natural populations of this tree species largely abide in the national parks, forest reserves, wildlife sanctuaries, and public areas, but the actual sites or locations, sizes, and circumstances of most populations are unknown. Therefore, an extensive survey must be carried out to obtain information and to determine a sampling strategy.  87  In situ conservation can also be applied to disturbed forest areas, including public areas and private areas, in order to sample the genetic variability of disturbed populations to ensure the most viable genetic base for future management. Furthermore, the outcrossing rate of the species is quite high both in undisturbed and disturbed populations. This implies that high quality seed can be obtained from either type of population. However, in situ conservation for disturbed populations is even more challenging than for undisturbed populations. An additional means to ensure this future availability of most genetic variability is, thus, to establish cx situ conservation to complement the in situ strategy.  In situ conservation is not always possible at all areas where populations occur. Also, human intervention is possible for refining and strengthening the programme. While some forest reserves will be devoted to ecosystem conservation, others may he dedicated to conserve intraspecies variation of the target species or subspecific populations (Yeatrnan, 1987; Kemp, 1992).  4.2. Ex situ conservation  Ex situ conservation is a strategy which protects and maintains genetic resources outside their environment as production stands, breeding material, conservation/seed stands, seed, pollen and/or tissue (Bonner, 1985).  Due to the population pressure on destruction of tropical forests, in situ  conservation might not enable the tree species to he retained in all desirable locations and environments. Conservation as cx situ may be required to ensure the availability of particular genetic materials for future management (FAO, 1989). However, cx situ conservation should he done to complement in situ conservation only; it can never replace in situ conservation fully. So far, cx situ approaches have formed the main emphasis of the conservation programme of this species. Five seed sources, consisting of thirty families each, were proposed for planting at five locations, 10 hectares each, in 1991 (RFD, l989a).  However, in the sampling strategy, several  geographically proximate populations were included in the same seed source.  The results from this  study indicate that genetically distinct populations may have been inadvertently mixed. Genetic integrity  88 of different populations would be lost by random mating with among populations, such an approach to conservation is not recommended. Due to the high genetic variability among populations, more populations should be added to the programme.  At least one ex situ plantation regenerated from 1,000 individuals of small and  remnant populations or subpopulations in each region should, if possible, be located in the selected area of the region. This will ensure the maintenance of a broad genetic base and prevention of the random genetic drift caused by inbreeding within the natural small population size of the species. The number of ex situ plantations can be more than one in each region. This depends on socio-economic situations which must be considered in making the programme practical. Whereas the success of ex situ conservation is dependent on functional and structural properties of an ecosystem, information on both population demography and environmental randomness are required. However, interaction with species of plants and animals are rarely understood. Some species might positively interact with the conserved species. Therefore, an ex situ plantation would be better established at the original sites of populations. However, when forest or land areas are changed by human intervention, the environmental conditions become poor and may no longer he appropriate for ex situ conservation.  Brune (1990) gave a good example of relationships among tree species,  pollinators, and environments in Brazil-nut plantations.  In the past, the Brazil-nut trees were left  standing, and other species were cleared from natural forest. The land around the standing trees was burned, effectively eliminating the natural habitat. The Brazil-nut trees, isolated from native vegetation, did not bear fruit becuase of the inappropriate habitats for the insect pollinators.  Only by careful  consideration and integration of genetic, ecological and other biological aspects of the target species can conservation success be ensured. Climatic, geographic, and edaphic factors, although demonstrating little variation, have to be matched with the species requirements. More information on these aspects is, however, required to increase the efficiency of the ex situ conservation. Although seed of Thailand rosewood can he stored at cool temperature, seed viability decreases very quickly. Therefore, a seed bank conservation method has to be developed.  Research on this  89 aspect should he emphasized for future application. However, today, plantations still serve as the best means of ex situ conservation. Furthermore, to support the conservation success, tree improvement for this tree species should be initiated using information on quantitative variation derived from this study. The improved seed or planted material then will increase the production of this tree species and decrease the pressure on harvesting trees from natural forests and conservation areas.  S. Socio-economic and Political Considerations  A genetic conservation programme cannot he segregated functionally and must be integrated consciously with socio-econornic conditions. Genetic conservation should be an integral part of national policies; the formulation of national strategies should be carried out by the Royal Forest Department, but needs the support and cooperation of other government sections (Kemp, 1992).  As for many  tropical countries, human population growth and increased social demands for land and income support  in Thailand challenge the protection of the full range of genetic variability.  To accomplish the  conservation of endangered species, a goal of land and resource management must be to sustain biological diversity, and goals for conservation must then he merged with other goals for land use (Salwasser, 1991). Complementary management of other forest land areas is necessary to ensure the conservation of a species from the local to the national level (International Tropical Timber Organization, ITTO, 1991).  Therefore, national forest policy must integrate the conservation  programme into all forest activities. The people responsible for these activities must be well educated in the necessity of forest resource conservation. Of greatest concern are local people or villagers who mainly make their living from forest contributions.  The conservation programme then should  emphasize the participation of these people. Basically, villagers, who live closest to the forest area to be conserved, are the poor or poorest in most tropical countries. A conservation programme can be effective if local or rural people are involved and benefit from it.  incorporation of conservation  90 objectives with rural development programmes should be a powerful means of success. Since direct and indirect values of conservation are ambiguous to them, obvious benefits from conservation should be demonstrated and extended to them: for example, economic values such as seed, wood, watershed protection, and income from recreation in conserved areas. Conservation and utilization therefore must be made compatible and beneficial. Moreover, conservation ethics should be strongly emphasized at all levels of education. The economic returns of poor rural people who rely on forest resources will be affected, in the short term, by a change from over-exploited forests to conservation forests.  Common cause  between conservationists and the local people has to he made before the programme is implemented. The land requirement for growing cash  crops  is the general problem for local people. An agroforestry  system which integrates tree planting with agricultural practices can be one option. Agroforestry can be located in marginal, low productivity forests, which will provide added benefit as buffer zones for designated forest conservation areas.  To conserve genetic diversity of an economically important  species, in situ conservation is needed with the complement of ex situ conservation. If local people’s demands on forest products are excessive, consideration is needed for sustainable management to provide an appropriate means of harvesting forest products. A large number of new jobs concerned with conservation can replace the old ones. These include forest guards, seed collectors, tree planters, tourist guides, biological research helpers, food, beverage and souvenir dealers, and handicraft work. A good example comes from Khaoyai National Park in northeast Thailand.  Problems with forest encroachment by local villagers seeking economic  benefits from poaching of timber and animals were alleviated by recruitment of forest rangers and guides. Many who were life-long poachers are now playing a significant role in conservation, deriving a consistent and legal livelihood in the process. This kind of activity must be developed either by the government or by concerned non-governmental agencies. The key to success is the coordination of people from all levels, especially local people. The programme then must be promoted extensively to the public; information must be delivered.  91 Furthermore, for the sake of conservation, the designated forests must be treated by the people as their own property and these people have to be responsible for the retention of the forests. The problems of forest destruction created and/or supported by powerful people are usual in a developing country. Thus, mass media, public awareness should be generated against them and their activities. Punishment by all legal and social means might he the best avenue to control these illegal activities. On the other hand, people who dedicate themselves, their abilities, and/or their properties to conservation should be rewarded with great public honour and financial benefits. Because the conservation programme of this tree species was established on a bi-government basis and several tree species were included in the programme, limited work has been carried out so far. Funds and personnel are constrained and solutions are needed to this problem. The programme should be revised and extended to include other countries in which the tree species or genus is distributed. Only with the success of conservation efforts for this valuable species can present and future benefits from its sustainable use he ensured.  6. Conclusion  Due to variability of the species being high within and among populations among populations, in situ conservation should include more populations and families than in the currently existing programme which has operated with only small number of populations and families.  Disturbed  populations could also be conserved to cover most genetic variability and provide materials for research and reforestation. Ex situ conservation should be carried out to complement in situ measures to ensure available genetic materials for future management. Ex situ plantations should be established in the same locality as where the seed was collected.  Finally, participation of local people in the conservation  programme has to he developed. The government, public and other agencies must play a substantial role in conservation of all levels from community or ecosystem to species and gene. The programme should be in cooperation with other neighbouring countries within which the species is distributed. This  92 will make it easier to acquire starting funding from government and/or nongovernment agencies.  93  CHAPTER 5  CONCLUSIONS AND RECOMMENDATIONS  1. Conclusions  1.1. Quantitative variation  The study of genetic variation of Thailand rosewood both through quantitative traits and isozyme characteristics provided useful information for genetic management and conservation of the species.  Quantitative traits, including seed, nursery seedling, and field progeny traits revealed high  variability exists among families within populations and relatively high among populations. There were no significant differences in these traits between undisturbed and disturbed populations. The observed quantitative variation may be an adaptation to different environments or due to regional isolation and should form the basis of conservation strategies. However, the heritability estimates for most traits may be insufficient to justify a genetic improvement programme. The genetic correlations among traits are weak. Nevertheless, the variation in seed and seedling traits remains valuable fur nursery management and reforestation because both quantity and quality of seed and seedlings are required. The progeny variation was significant among families for only some traits, except diameter, persistence of axis, and  stem straightness. The variation of interactions between families and blocks was significant for all progeny traits.  94 1.2. Isozyme variation  Twelve enzyme systems provided eighteen scorable loci which were used for analyses of genetic variability and the mating system. The genetic variability parameters, number of alleles per locus and mean heterozygosity were high within populations and moderate and differed significantly among populations, but not between undisturbed and disturbed populations.  Nei’s (1978) genetic  distance and F-statistics (Wright, 1965) indicated moderate genetic differentiation among populations. Cluster analysis using Nei’s genetic distance divided populations into two geographic regions: central and northeastern Thailand. However, the genetic distance between regions was not significant. Outcrossing rate  (t)  was high, almost I .0.  The reasons for this are probably the self-  incompatibility, negative assortative mating, and selection to minimize inbreeding of the species. Also small population sizes may have contributed to the sampling error.  1.3 Conservation strategies  Conservation of Thailand rosewood was initiated in 1989.  However, a genetic basis for  rational conservation has rarely been available and applied. The information derived from this study will therefore be used to support and strengthen the existing programme. The high quantitative and isozyme variability among populations necessitates the conservation of several populations populations. However, remnant populations of this tree species are naturally small.  The appropriate number of individuals to retain adequate variability should be considered.  About one thousand individuals for each geographic region gathered from several small populations or subpopulations distributed throughout the region should be enough for maintaining most and even low frequency alleles (Namkoong, 1984a). Both in situ and ex situ conservation should be established for this endangered tree species.  Disturbed populations should also he conserved as sources of genetic  material for genetic studies and reforestation. Ex situ conservation should be established at least one  95 location in each region by collecting about I ,000 individuals from populations or subpopulations within the region to maintain the broad genetic base and prevent random genetic drift of the species. Tree improvement of this tree species should also he simultaneously carried out to increase production of the species and decrease the need for harvesting trees from natural forests and conservation areas. the traits.  High variation in both quantitative traits and isozyme made possible to improve  However, heritabilities of the traits were not very consistent, so selection based on high  heritability traits will make more progress on tree improvement.  2. Recommendations  Conservation should be nniltidisciplinary; integration of population genetic and ecological information with socio-economic and political aspects of the country should be considered.  The  conservation strategy has to he forimilated by national policies to benefit local people. The participation of local people is the key to success; therefore, development agencies and operations should emphasize the provision of benefits to local communities. Education and job training should he provided to local people to help improve their knowledge and skills. Furthermore, agroforestry, forest community, and forest village projects which were successful in several areas should he promoted and extended throughout the country. This study would provide the principal knowledge of gentic variation strategies for conservation and future genetic management of this tree species, hut information is still inexact.  Therefore, a  number of studies are required to develop the understanding of the genetics and ecology of tropical forests.  This should serve as the principle for conservation which is to be scientifically based and  rational.  Socio-economic and political refinements are also needed to support and strengthen the  effectiveness of conservation efforts. The activities involve: (a). Extend the exploration of tree populations and collection of seed for comparative study of both quantitative and isozyme variation.  96 (b). Repeat studies of quantitative variation in different years and planting sites to further the understanding of the effects of temporal and spatial variation. (c). Repeat studies of isozyrne variation using a larger number of families and populations in different years to obtain a better understanding of the genetic structure of Thailand rosewood and its temporal variation. (d). Incorporate ecological studies (particularly genecology) of Thailand rosewood and its associated species, especially pollinators, for more efficient conservation. (e). Compare distribution, variation, and mating systems of other tree species in the genus Dalbergia to get an insight into evolutionary processes within the genus. (f). Study and improve seed storage technology of species for seed bank conservation. (g). Stress the urgency of the conservation needs of this species and genus to government and nongovernment agencies. (h). Ensure regeneration of the target species and a coherent system of economic incentives for greater attention in genetic conservation.  97  REFERENCES  Adams, W.T. and D.S. Bikes. 1988. Estimating mating patterns in forest tree populations. In: Proc. International Workshop on Plant Biology Oct. 11-13 1988. Porano-Ovieto, Italy. pp. 7584.  Adams, W.T. and R.J. Joly. 1977. Analysis of genetic variation of height growth and survival in openpollinated progenies of eastern white pine. In: Proc. Northeast For. Tree Improv. Conf 25:117-13 1. Ashton, P.S. 1969. Speciation among tropical forest trees: some deductions in light of recent evidence. Biol. J. Linn. Soc. London. 1:155-196. Association of Official Seed Analysts (AOSA). 1983. Seed Vigor Testing Handbook. Seed Vigor Test Committee of the Association of Official Seed Analysts. International Seed Testing Association, Zurich, Zwitzerland. 88 pp. Bastien, J.C. and B. Roman-Arnat. 1990. Predicting Douglas fir (Pseudotsuga menziesii (Mirb) Franco) volume at age 15 with early traits. Silvae Genetica 39:29-35.  Bawa, K.S  .  1974. Breeding systems of tree species of a lowland tropical community. Evolution 28:85-92.  Bawa, K.S. 1976. Breeding of tropical hardwoods: an evaluation of underlying bases, current status and future prospects. In: Burley. J and B.T. Stiles (Eds.), Tropical Trees Variation, breeding and conservation. pp 43-59 Linn. Soc. Symp. Series number 2 Academic Press, London. .  Bawa, K.S.; D.R. Perry; J.H. Beach. 1985. Reproductive biology of tropical lowland rain forest trees. I.Sexual systems and incompatibility mechanisms. Amer. J. Bot. 72:33 1-345. Becker, W.A. 1984. Manual of Quantitative Generics. 4th edition. Academic Enterprises, Pullman, Washington. 190 pp. Bishir, I. and G. Namkoong. 1987. Unsound seeds in conifers: Estimation of Numbers of Lethal Alleles and Magnitude of Effects Associated with the Maternal Parent. Silvae Genetica 36:180-185. Bonner, F.T. 1985. Technologies to Maintain Tree Germplasm Diversii. Assessment, U.S. Congress, Washington, D.C.  Office of Technology  Botstien, D.; R.L. White; M. Skolnick; R.W. Davis. 1980.Constructionofgeneticlinkagemap in man using restriction fragment length polymorphisms. Amer. J. Hum. Genet. 32:314-331.  98 Boyle, T.J.B.; C. Liengsiri; C. Piewluang. 1990. Genetic studies in a tropical pine Pinus kesiya. II. Genetic variation among four populations in northern Thailand. (Submitted to J. Tropical Forest Sci.) -  Bradshaw, A.D. 1965. Evolutionary significance of phenotypic plasticity in plants. Adv. Genet. 13:115155. Bramlet, D.L.; T.R. Dell; W.D. Pepper. 1983. Genetic and maternal influences on Virginia pine seed germination. Silvae Genetica 32: 1-4. Brown, A.H.D. 1979. Enzyme polymorphism in plant populations. Theor. Popul. Biol. 15:1-42. Brown, A.H.D. 1989. Genetic characterization of plant mating systems. In: Brown,A.H.D.; M.T. Clegg; A.L. Kahler; B.S. Weir (Eds.), Plant Population Genetics, Breeding, and Genetic Resources. Sinauer Associates, Sunderland, M.A. pp 145-162. Brown, A.H.D. 1990. The role of isozyme studies in molecular systematics. Atist. Syst. Bot. 3:39-46. Brown, A.H.D. and G.F. Moran 1979. Isozyme and the Genetic Resources of Forest Trees. In: Proc. Symp. on Isozymes of N. Amer. Forest Trees and Forest Insects. July 27, 1979, Berkeley, California. pp. 1-10. Brown, A.H.D.; A.C. Matheson; K.G. Eldridge. 1975. Estimation of the mating system of Eucalyptus obliqua L’Herit. by using allozyme polymorphisms. Aust. J.Bot. 23:931-949. Brune, A. 1990. Reproductive biology and tropical plantation forestry. In: Bawa, K.S.; M. Hadley (Eds.), Reproductive Ecology of Tropical Forest Plants. Man and the Biosphere Series Vol 7. UNESCO, Paris. pp. 349-354. Carter, K.K.; G.W. Adams; M.S. Greenwood; P. Nitschke. 1990. Early family selection in jack pine. ran. J. For. Res. 20:285-291. Carter, K.K. and M.S. Greenwood. 1987. Tree Improvement. In: 1987 Annual Report of the cooperative Forest Research Unit. Univ. of Maine, College of Forest Resources, Orono, Maine. pp. 21-22. Chan, H.T. 1983. Reproductive biology of some Malaysian dipterocarps. Malaysian For. 44:28-341.  III. Breeding systems.  Cheliak, W.M.; B.P. Dancik; K. Morgan; F.C.H. Yeh; C. Strobeck. 1985. Temporal variation of the mating system in a natural population of jack pine. Genetics 109:569-584. Cheliak, W.M.; F.C.H. Yeh; J.A. Pitel. 1987. programs. For. C’hronicle 63:89-96.  Use of electrophoresis in tree improvement  Cheliak, W.M. and D.L. Rogers. 1990. Integrating biotechnology into tree improvement programs. Can. J. For. Res. 20:452-463. Clegg, M.T.; A.L. Kahler; R.W. Allard. 1978. Estimation of life cycle components of selection in an experimental plant population. Genetics 89:765-792.  99 Cochran, W.G. and G.M. Cox. 1957. Experimental De.vign. John Wiley & Sons, Inc., New York. Corner, E.J.H. 1988. Wayside Trees of Malaya. Malaysia. 476 pp.  Vol. 1 United Selangor Press, Kuala Lumpur,  Cossalter, C. 1989. Genetic conservation : A cornerstone of breeding strategies. In: Gibson, G.L., A.R. Griffin and A.C. Matheson (Eds.), Proc. JUFRO Conf on Breeding Tropical Trees: Population Structure and Genetic lmprov. Strategies in Clonal and Seedling Forestry. Pattaya, Thailand Nov. 19-Dec. 3, 1988. pp. 28-3 8. Critchfield, W.B. 1980. Genetics of lodgepole pine. USDA For. Serv. Res. Pap. WO-37, 57 pp. Isozyme patterns of selected Pseudotsuga menziesii (Mirb.) Franco El-Kassaby, Y.A. 1980. population. Ph.D. Thesis. Dept. of For., Univ. of British Columbia. 137 pp. El-Kassaby, Y.A.; D.G.W. Edward; D.W. Taylor. 1992. Genetic control of germination parameters in Douglas-fir and its importance for domestication. Silvae Genetica 4 1:48-54. Ellstrand, N.C.; A.M. Torres; D.A. Levin. 1987. Density and the rate of apparent outcrossing Helianthus annuus (Asteraceae). Svst. Bor. 3:403-407. Epperson, B.K. 1989. Spatial patterns of genetic variation within plant populations. In: Brown, A.H.D.; M.T. Clegg; A.L. Kahler; B.S. Weir (Eds.), Plant Population Genetics, Breeding, and Genetic Resources. Sinauer Associates, Sunderland, M.A. pp 229-253. Ewens, W.J.; P.J. Brockwell; J.M. Gani; S.!. Resnick. 1987. Minimum viable population size in the presence of catastrophes. In: Soule, M .E. (Ed.), Viable Populations for Conservation. Cambridge Univ. Press, New York. pp. 59-68. Falconer, D.S. 1981. Introduction to Forest Genetics. Academic Press, New York. 463 pp. Falk, D.A. 1990. Endangered forest resources in the U.S.: Integrated strategies for conservation of rare species and genetic diversity. Forest Ecology and Management 35:91-117. Falkenhaken, E.R. 1985. Isozymes studies in provenance research of forest trees. Theor. AppI. Genet. 69: 335-347. Federov, A.A. 1966. The structure of the tropical rain forests and speciation in humid tropics. J. Ecol. 54: 1-11. Feret, P.P. and F. Bergmann. 1976. Gel electrophoresis of proteins and enzymes. In: Miksche, J.P. (Ed.), Modern Methods in Forest Genetics. Spring-Verlag, New York. pp. 49-77. Fins, L.; M. Rust; B. Wilson. 1990. Using nursery trials to predict Douglas-fir field performance. In: Proc. Joint meeting WFGA/JUFRO Aug. 20-24, 1990. Olympia, Washington, U.S.A.. Food and Agriculture Organization (FAQ). 1981. Forest Resources of Tropical Asia. FAO/UNEP. Rome, Italy. FAO. 1987. Forest Situations of Tropical Asia. FAQ, Rome, Italy.  100 FAO. 1988. in Situ conservation: conserving Plant Genetic Resources in Situ to Meet Today r and Tomorrow’s Needs (leaflet,). Rome, Italy. FAO. 1989. Plant Genetic Resources: Their conservation in situ for human use. Rome, Italy. Franklin, E.C. 1979. Model relating levels of genetic variance to stand development of four North America conifers. Silvae Genetica 25:207-212. Franklin, I.R. 1980. Evolutionary change in small populations. In: Soul, M.E. and B.A. Wilcox (Eds.), Conservation Biology: An Evolutionary-Ecological Perspective. Si nauer Associates, Sunderland, Mass. pp. 135-149. Fryer, J.H. 1987. Agreement between patterns of morphological variability and isozyme band phenotypes in pitch pine. Silvae Genetica 36:199-206. Fumier, G.R.; P Knowles; M.A. Clyde; B.P. Dancik. 1987. Effect of avian seed dispersal on the genetic structure of whitebark pine populations. Evolution 41:607-6 12. Fumier, G.R.; M. Stine; C.A. Mohn; M.A. Clyde. 1991. Geographic patterns of variation in allozymes and height growth in white spruce. can. J. For. Res. 21:707-712. Gilpin, M.E. and M.E. Soulá. 1986. Minimum viable populations: Processes of species extinction. In: Soulé. M.E. (Ed.), Conservation Biology: The Science of Scarcity and Diversity. Sinauer Associates, Sunderland, Mass. pp. 19-34. Griffin, A.R. and P.P. Cotterill. 1988. Genetic Variation in Growth of Outcrossed, Selfed and Open-Pollinated Progenies of Eucalyptus regnans and some Implications for Breeding Strategy. Silvae Genetica 37:124-130. Guries, R.P. and F.T. Ledig. 1977. Analysis of population structure from allozyme frequencies. In: Proc. 14th 5. For. Tree improv. Conf Gainesville, Fla. pp. 246-253.  Guries, R.P. and F.T. Ledig. 1981. Genetic structure of populations and differentiation in forest trees. In: Proc. isozymes of North America Forest Trees and Forest insects. Tech. Rep. PSW-48. pp. 42-47. Hamrick, J.L. 1982. Plant population genetics and evolution. Amer. J. Bot. 69:1685-1693. Hamrick, J.L. 1983. The distribution of genetic variation within and among natural plant populations. In: Schonewald-Cox, C.M.; S.M. Chambers; B. MacBryde; W.L. Thomas (Eds), Genetics and conservation-a reference for managing wild animal and plant populations. Benjamin/Commings, Menlo Park, CA. pp 335-348. Hamrick, J.L. and M.J.W. Godt. 1989. Allozyme diversity in plant species. In: Brown, A.D.H.; M.T. Clegg; A.L. Kahier; B.S. Weir (Eds.), Plant Population Genetics, Breeding and Genetic Resources. Sinauer Associates, Inc., Sunderland, Mass. pp. 43-63. Hamrick, J.L.; Y.B. Linhart; J.B. Mitton. 1979. Relationships between life history characteristics and electrophoretically detectable genetic variation in plants. Ann. Rev. Ecol. Syst. 10: 173-200.  101 Hamrick, J.L. and M.D. Loveless. 1986. Isozyrne variation in tropical trees: procedures and preliminary results. Biotropica 18:201-207.  Hamrick, J.L. and M.D. Loveless. 1989. The genetic structure of tropical tree populations: Associations with reproductive biology. In: Bock, J.H. and Y.B. Linhart (Eds.), The Evolutionary Ecology ofPlants. Westview Press, Boulder, San Francisco and London. pp. 129-146. Hamrick, J.L. and D.A. Murawski. 1991. Levels of allozyme diversity in populations of uncommon neotropic tree species. J. Tropical Ecol. 7:395-399. Hanover, J.W.; T. Mebrathn; P. Bloese. 1991. Genetic improvement of black locust: a prime agroforestry species. Forestry chronicle 67:227-231. Hicks, R.R.Jr.; G. Rink; B.E. Cutter. 1977. Variation and heritability of seedling height in East Texas sycamore and rich birch. In: Proc. Northea.vt For. Tree Improv. Conf 25:133-137. Hutacharern, C.; S. Choldamrongkul; S. Eungwijarnpanya; A. Choldamrongkul; P. Pholwicha. 1988. Check Lists of Forest Insects in Thailand. Royal Forest Department, Bangkok, Thailand. 72 pp. International Tropical Timber Organization (ITTO). 1991. Report on working group on guidelines for the conservation of biological diversity in tropical production forests. Ninth session, Nov Dec 4, 1991. Yokohama. In: PC’FIX/IO. ITTO and Permanent Committee on 28 Reforestation and Forest Management. -  International Union for the Conservation of Nature and Natural Resources (IUCN). 1980. World conservation Strategy: Resource Conservation For Sustainable Development. IUCN, Grand, Switzerland. Jiang, I.B.; J, A. Jonsson; G. Eriksson. 1989. Within and between population variation in growth of Pinus contorta var. iatfolia: A combined study of growth-chamber and field-trial experiments. Silvae Genetica 38:201-211. Kanowski, P.J. and D.G. Nikies. 1989. Family structure in provenance trials, conservation stands, and genetic base populations. In: Gibson, G.L.; A.R. Griffin; A.C. Matheson (Eds.), Proc. JUFRO conf on Breeding Tropical Trees: Population Structure and Genetic Iniprov. Strategies in Clonal and Seedling Fore.rtrv. Pattaya, Thailand. Nov. 19 Dec. 3, 1988. pp. 236-249. -  South-east Asia, Northern Australia and the PacJic Keating, N.G. and E. Bolza. 1982. characteristics, Properties and Uses of Timhe,. vol. 1. Inkata Press, Melbourne. 126pp. Keiding, H.; E.B. Lauridren; H. Wellendorf. 1984. Evaluation of a series of teak and gmelina provenance trials: selection of traits, their assessment and analysis of observation. In: Proc. Joint Meeting Work Conf on Provenance and Genetic Improv. Strategies in Tropical Forest Trees. Matare, Zimbabwe 9-14 April 1984. . 69 30 pp. Kemp, R.H. 1992. The conservation of genetic resources in managed tropical forests. Unasylva 169 V43:34-40.  102 Kimura, M. 1983. The Neutral Theory of Molecular Evolution. Cambridge University Press, Cambridge.  Kimura, M. andJ.F. Crow. 1963. The measurement of effective population number. Evolution 17:279288. Kimura, M. and T. Ohta. 1971. Protein polymorphism as a phase of molecular evolution. Nature 229:467-469. Lambeth, C.E. 1983. Early testing An overview with emphasis on loblolly pine. In: Proc. 7th S. For. Ti-ce Irnprov. Conf pp. 297-3 11. -  Lambeth, C.E.; J.P. van Buijtenen; S.D. Duke; R.B. Mccullough. 1983. Early Selection is Effective in 20-Year-Old Genetic Tests of Loblolly Pine. Silvae Genetica 32:210-215. Lande, R. 1988. Genetic and demography in biological conservation. Science 241:1455-1460. Ledig, F.T. 1986a. Heterozygosity, heterosis and fitness in outbreeding plants. In: Soulc, M.E. (Ed.), Conservation Biology: The Sc-ience of Scarcity and diversii. Sinauer Associates, Sunderland, Mass. pp. 77-104. Ledig, F.T. 1986b. Conservation strategies for forest gene resources. For. Ecol. Manag. 14:77-90. Ledig, F.T and M.T. Conkle. 1983. Gene diversity and genetic structure in a narrow endemic, Torrey pine (Pinus torrevana Parry cx Carr). Evolution 37:79-85. Lewontin, R.C. 1974. The Genetic BLV1s of Evolutionary c’hange. Columbia Univ. Press, New York. 346 pp. Lewontin, R.C. and J.L. Hubby. 1966. Molecular approach to the study of genetic heterozygosity in natural populations. II. Amount of variation and degree of heterozygosity in natural populations of Drosophila pseudoobscura. Genetics 54:595-609. Liengsiri, C.; C. Piewluang; T.J.B. Boyle. 1990. Gel Electrophoresis of Tropical Trees ASEAN-Canada Forest Tree Seed Centre. Muak Lek, Saraburi,Thailand.  -  A Manual.  Lindgren, K. 1983. Provenance of Pinus contorta in Northern Sweden. Ph.D. Dissertation. Dept. For. Genet. and Plant Physiology. The Swedish Univ. of Agricultural Science. 91 pp. Magnussen, S. and C.W. Yeatman. 1987. Early testing of jack pine. 1. Phenotypic response to spacing. Can. J. For. Rev. 17:453-459. Marshall, D.R. and A.H.D. Brown. 1975. Optimum sampling strategies in genetic conservation. In: Frankel,O.H. and J.G. Hawkes (Eds.), Crop Genetic Resources for Today and Tomorrow. Cambridge Univ. Press. 492 pp. Markert, C.L. and F. MoIler. 1959. Multiple forms of enzymes: tissue, ontogenetic, and species specific patterns. In: Proc. Nat. Acaci. Sc-i. (USA) 45:573-763.  103 Mikola, J. 1984. Relationship between height growth differences of Scots pine full-sib families and variation in seed size, annual growth rhythm and some foliage characteristics. In: Tigerstedt, P.M.A.; P. Puttonen; V. Koski (Eds.), Crop Physiology of Forest Trees. Helsinki Univ. Press, Finland. pp. 233-243. Mitton, J.B. 1983. Conifers. In: isozyrnes iii Plant Genetics and Breeding Part B, Elsevier Science Publishers, Amsterdam. pp. 443 -472. Moran, G.F.; 0. Muona ; J.C. Bell. 1989. Breeding systems and genetic diversity in Acacia auriculjformis and A. crassicarpa. Biotropica 21:250-256. Moran, G.F. and J.C. Bell. 1983. Eucalyptus. In: Tanksley, D. and T.J. Orton (Eds.), Isozymes in Plant Genetics and Breeding Part B, Elsevier Science Publishers, Amsterdam. pp. 423-441. Morgenstern, E.K. 1972. Preliminary estimates of inbreeding in natural population of black spruce Picea marina. Can. J. Genet. ‘ytoi. 14:443-446. Muona, 0. 1989. Population genetics in forest tree improvement. In: Brown, A.H.D.; M.T. Clegg; A.L. Kahler; B.S. Weir (Eds.), Plant population genetics, breeding, and genetic resources. Sinauer Associates, Sunderland, M.A. pp 282-298. Myers, N. 1991. Tropical forests : present status and future outlook. Climatic Change 19:3-32. Namicoong, G. 1984a. Genetic structure of forest tree populations. In: Genetics: New Frontiers Vol. 4, Applied Genetics. Proc. of the 15th International congress of Genetics. Mohan Primlani, Oxford, and TBH Publishers, New Delhi. pp. 35 1-360. Namkoong, G. 1984b. Strategies for gene conservation in forest tree breeding. In: Yeatman, C.W.; D. Kafton; 0. Wilkes (Eds.), Plant Generic Resources: A Conservation Imperative. Westview Press Inc. Boulder, Cob. pp. 79-89. Namkoong, G. 1989. System of gene management. In: Griffin and A.C.Matheson (Eds.), Proc. JUFRO C’onf on Breeding Tropical Trees: Population Structure and Genetic lmprov. Strategies in Clonal and seedling Forestry. Pattaya, Thailand. Nov. 19 Dec. 3, 1988. pp. 1-8. -  Namicoong, G. 1990. Biological Diversity: Its Conservation and Use for Sustainable Agricultural, Forestry and Fisheries Development. Food and Agriculture Organization. 36 pp. National Academy of Sciences (NAS). 1979. Tropical Legumes: Resources for the Future. NAS, Washington, D.C. 332 pp. National Research Council (NRC). 1991. Managing Global Genetic Resources: Forest Trees. National Academy Press, Washington, D.C. 228 pp. Nei, M. 1972. Genetic distance between populations. Amer. Nat. 106:283-292. Nei, M. 1973. Analysis of gene diversity in subdivided populations. In: Proc. Nat. Acad, Sci. (USA) 70:3321-3323.  104 Nei, M. 1975. Molecular population genetics and evolution In: Neuberger, A., E.L. Tatum (Eds.), Frontiers of Biol. vol. 40 Amer. Elsevier, New York. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590. Noss, R.F. 1991. From endangered species to biodiversity. In: Kohm, K.A. (Ed.), Balancing on the Brink of Extinction. Island Press, Washington, D.C. pp. 227-245. Office of Technology Assessment (OTA). 1984. Technology to Sustain Tropical Forest Resources. OTA F214 Washington, D.C. : U.S. Department of the Interior. -  O’Malley, D.M. and K.S. Bawa. 1987. Mating system of tropical rain forest tree species. Amer. J. Bot. 74(8):1149-1987. Park, Y.S. and D.P. Fowler. 1984. Inbreeding in black spruce ( Picea marina [Mill.1 B.S.P.): self-fertility, genetic load and performance. can. J. For. Rev. 14:17-21. Perry, T.O. and W.L. Hafley. 1981. Variation in seedling growth rates: Their genetics and physiological bases. In: Proc. 16th South For. Tree Improv. Conf Blackburg, VA. pp. Pharis, R.P.; F.C. Yeh; B.P. Dancik. 1991. Superior growth potential in trees: What is its bias, and can it be tested for at an early age? C’an. J. For. Res. 2 1:368-374. Piewluang, C. and C. Liengsiri. 1989. Germination pretreatment for Daibergia cochinchinensis Pierre. In: Proc. Silviculture cotif Royal Forest Department. Bangkok, Thailand.( in Thai). pp. 175-183. Prakash, S.; R.C. Lewontin; J.L. Hubby. 1969. A molecular approach to the study of genetic heterozygosity in natural populations of Drosophila pseudoobscura. Genetics 61:841-858. Rehfeldt, G.E. 1983. Veracity of short-term testing. In: Proc. Servicewide Genetics Workshop on Progeny Testing. Charleston, South Carolina, Dec. 5-9, 1983. USDA Forest Service. pp. 118-148. Rink, G. 1984. Trends in genetic control of juvenile black walnut height growth. For. Sci. 32:10491057. Ritland, K. 1986. Generalized Mulrilocus Ritland, Toronto, Canada  Estimation  Program MLTF. Documentation Manual. K.  Ritland, K.R. and S.K. Jam. 1981. A model for the estimation of outcrossing rate and gene frequencies using and independent loci. Heredity 47:35-52. Ritland, K.R. and Y.A. El-Kassaby. 1985. The nature of in breeding in seed orchard of Douglas fir as shown by an efficient multilocus model. Theor. AppI. Genet.71:375-384. Ritters, K.H. and D.A. Perry. 1987. Early genetic evaluation of open-pollinated Douglas-fir families. For. Sci. 33:577-582. Roach, D.A. and R.D. Wulff. 1987. Maternal effects in plants. Ann. Rev. Ecol. Syst. 18:209-215.  105 Roche, L. and M.J. Dourojeani 1984. A Guide to in situ conservation of Genetic Resources of Tropical Woody Species. FAO. 196 pp. Rogers, D.L.; R.F. Stettler; P.E. Heilmann. 1989. Genetic variation and productivity of Populus trichocarpa and its hybrids. 111. Structure and pattern of variation in a 3-year field test. Can. J. For. Res. 19:372-377. Royal Forest Department (RFD). 1983. Economic Trees of Thailand Vol 3. Royal Forest Department, Bangkok, Thailand. (in Thai) 105 pp. RFD. 1986. Forest Statistics of Thailand. Royal Forest Department, Bangkok, Thailand. 78 pp.  RFD. 1989a. Proposal for Hardwood conservation Program. Royal Forest Department, Bangkok, Thailand. 10 pp. RFD. 1989b. Establishment of cx situ gene conservation stands of some dipterocarp and leguminous species Research Project. Royal Forest Department, Bangkok, Thailand. 102 pp. Rubeli, K. 1989. Pride and Protest in Malaysia. New Scientist 124:49-52. Sa-ardavut, P; T. Chairuangsirikul; S. Chanpaisaeng; J.J. Granhof. 1989. In situ gene conservation f on Breeding Tropical Trees: Population 1 in northeast Thailand. In: Proc. JUFRO Cot Structure and Generic Imprm’. Strategies in Clonal and Seedling Forestiy Pattaya, Thailand. Nov 19-Dec. 3/1988. pp. 291-297. Saiwasser, H. 1990. Conserving biological diversity: A perspective on scope and approaches. For. Ecol. Mgr. 35:79-90. Saiwasser, H. 1991. In search of an ecosystem approach to endangered species conservation. In: Kohm, K.A. (Ed.), Balancing on the Brink of Extinction. Island Press, Washington, D.C. pp. 227-245. SAS. 1991. SAS/STAT User guide. SAS Institute Inc., Cary, N.C. Schlichting, C.D. 1986. The evolution of phenotypic plasticity in plants. Ann. Rev. Ecol. Syst. 17:667693. Schnabel, A. and J.L. Hamrick. 1990. Organization of genetic diversity within and among populations of Gleditsia triacanthos (Legurninosae). Amer. J. Bot. 77(8): 1060-1069. Schnabel, A.; R.H. Laushman; iL. Hamrick. 1991. Comparative genetic structure of two co-occurring tree species, Maclura porn ifera (M oraceae) and Gieditsia triacanthos (Leguminosae). Heredity 67:357-364. Shaw, D.V. and R.W. Allard. 1982. Estimation of outcrossing rate in Douglas-fir using isozyme markers. Theor. Appi. Genet. 63: 113-120. Shaw, D.V.; A.L. Kahler; R.W. Allard. 1981. A multilocus estimator of mating system parameters in plant populations. Proc. Nat. Acad. Sci (USA) 78: 129-1302.  106 Sneath, P.H.A. and R.R. Sokal. 1973. Numerical Taxonomy. W.H. Freeman, San Francisco. 644 pp. Sniittinand, T. 1980. Botanical Names for Forest Species of Thailand. Royal Forest Department, Bangkok, Thailand. 205 pp. Snedecor, G.W. and W.G. Cochran. 1967. Statistical Methods. Iowa State Univ. Press. Ames. 503 pp. Sniezko, R.A. 1986. Simultaneous progeny/provenance testing of subtropical and tropical forest tree species: experiences in Zimbabwe. In: Proc. JUFRO Joint Meeting of Working Parties on Breeding Theoty, Progeny Testing and Seed Orchards. Williarnburg, VA. 13-17 October 1986. pp. 310-324. Sokal, R.G.D. and F.J. Rohlf. 1969. Biometrv. W.H. Freeman, San Francisco. 219 pp. Soule, M.E. 1980. Thresholds for survival: Maintaining fitness and evolutionary potential. In: SouI, M E.; B. A. Wilcox (Eds.), Conservation Biology: An Evolutionary-Ecological Perspective. Sinauer Associates, Sunderland, Mass. pp. 153-169. .  St. Clair, J.B. and W.T. Adams. 1991. Effects of seed weight and rate of emergence on early growth of open-pollinated Douglas-fir families. Forest Science 34:984-997. Steel, R.G.D. and J.H. Torrie. 1960. Principles and Procedures of Statistics. York.  McGraw-Hill, New  Stern, K. and L. Roche. 1974. Genetics of Forest Ecosystems. New York, Springer-Verlag. 330 pp. Surles, S.E.; J.L. Hamrick; B.C. Bongarten. 1990. Mating system in open-pollinated families of black locust (Rohinia pseudoacacia). Siivae Genetica 39:35-42. Sweet, G.B. and P.F. Wareing. 1966. The relative growth rates and small seedlings in forest tree species. Forestry (supplement): 110-117. Swofford, D.L. 1981. On the utility of the distance Wagner procedure. In: Funk, V.A. and D.R. Brooks (Eds.), Advances in Cladistics: Proceedings of the First Meeting of the Willies Hennig Society. New York Botanical Garden, Bronx, New York. pp. 25-43. Swofford, D.L. and R.B. Selander. 1989. Biosys-1 A Computer Program for the Analysis of Allelic Variation in Population Genetics and Biochemical Sysrematics. Release 1.7. D.L. Swofford, Illinois History Survey. 43 pp. .  Sziklai, 0. 1974. Juvenile-mature correlation In: Proc. JUFRO Joint Meeting of Work Parties on Population and Ecological Genetics, Breeding Theory and Progeny Testing. Stockholm pp. 217-235.  Thompson, J.N. 1984. Variation among individual seed mass in Lomarium grayi (Urnbelliferae) under controlled conditions: Magnitude and partitioning of variance. Ecology 65:626-63 1. Tyson, H.  1989. Genetic control of seed weight in flax (Linum usitatissimum) and possible implications. Theor. Appi. Genet. 77:260-270.  107 Wheeler, N.C. and W.B. Critchfield. 1985. The distribution and botanical characteristics of lodgepole pine: biogeographical and management implications. In: Baumgartner, D.M.; R.G. Krebill; J.T. Arnott; G.F. Weetman (Eds.), Lodgepole pine: the species and its management Symp. Proc. May 8-10, 1984, Spokane, Washington, U.S.A. and May 14-16, 1984, Vancouver, B.C. Canada. pp. 45-53. Wheeler, N.C. and R.P. Guries. 1982. Population structure, genetic diversity, and morphological variation in Pinus contorta doug!. Can. J. For. Res. 12:595-606. White, T.L. 1987. A conceptual framework for tree improvement programs. New Forests 4:325-342. Wilcox, M.D. 1983. Reciprocal cross effects in Pinus radiata. NZ. J. For. Sd. 13:37-45. Williams, J.G.K.; A.R. Kuheli; K.J. Livak; J.A. Rafaiski; S.V. Tingey. 1990. DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nuc. Acids Res. 1 8:6531 6535 -  World Resources Institute (WRI). 1990. World Resources. A Guide to the Global Environment. The World Resources Institute: Oxford University Press, Oxford and New York. Wright, J.W. 1976. Introduction to Forest Genetics. Academic Press, New York. 463 pp. Wright, S. 1965. The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution 19:395-420. Wright, S. 1978. Evolution and Genetics of Populations, vol 4. Variability Within and Among Natural Populations. University of Chicago Press, Chicago. Yasuda, W. 1969. The estimation of the variance effective population number base on gene frequency. Jpn. J. Hum. Genet. 14:10-15. Yeatman, C.W. 1987. Conservation of genetic resources within managed natural and man-made forests. Malaysian Forester 50:1-8. Yeh, F.C.H. 1989. Isozyme analysis for revealing population structure for use in breeding strategies. In: Proc. ?UFRO Conf on Breeding Tropical Trees: Population Structure and Genetic Improv. Strategies in Clonal and Seedling Forestry. Pattaya, Thailand Nov 19-Dec 3 /1988. pp. 119-131. Ying, C.C.; K. Illingworth; M. Carlson. 1985. Geographic variation in lodgepole pine and its implications for tree improvement in British Columbia. In: Baumgartner, D.M.; R.G. Krebill; J.T. Arnott; G.F. Weetman (Eds.), Lodgepole pine: the species and its management Symp. Proc. May 8-10, 1984, Spokane, Washington, U.S.A. and May 14-16, 1984, Vancouver, B.C. Canada. pp. 45-53. Zobel, B.J. and J.T. Talbert. 1984. Applied Forest Tree improvement John Wiley & Sons, New York. 505 pp.  108  APPENDICES  Appendix A Number of families or individual maternal trees used for the studies of variation: (A). for seed, seedling, and isozyme variation, seed was collected in 1990 from eight populations (B). for progeny trial, seed was collected in 1987 from seven populations Number of families  (A). Seed  Seedling  Isozyme  18 13  14 11  17 11  KKI KK2 MKI MK2 SKi SK2  7 10 4 4 7 6  6 8 4 4 6 3  6 5 4 4 7 3  TOTAL  69  56  57  Population Undisturbed ML1 ML2 Disturbed  Number of families  (B). Population  Progeny trial  Undisturbed DL ML SKH  20 15 5  Disturbed KH MS SM SK  9 19 5 12  TOTAL  85  109  Appendix B Seed trait means for individual families in eight populations  POPULATION  GERMINATION %  OVERALL MEAN  80.5870  R50 (day)  SEED WEIGHT (mg)  9.2517  24.1748  UNDISTURBED POPULATION POPULATION ML1 FAMILY 1 FAMILY 2 FAMILY 3 FAMILY 4 FAMILY 5 FAMILY 6 FAMILY 7 FAMILY 8 FAMILY 9 FAMILY 10 FAMILY 11 FAMILY 12 FAMILY 13 FAMILY 14 FAMILY 15 FAMILY 16 FAMILY 17 FAMILY 18  88.4167 87.5000 79.0000 7 1.5000 95.0000 88.5000 86.5000 90.0000 92.0000 92.5000 94.5000 93.5000 89.0000 87.5000 90.5000 93.0000 91.5000 78.0000 91.5000  10.6628 12.4875 8.2425 7.6775 12.7000 11.0150 8.1575 9.0325 11.1700 13.5050 9.7125 9.0175 10.6075 13.7825 12.7150 8.6350 11.3400 13.5275 8.6050  24.3676 26.4850 21.5450 20.0200 25.0275 27.5250 24.7400 25.3325 27.2575 22.7125 22.7950 16.9200 24.9150 29.3325 23.4600 23.3725 28.0850 26.8600 22.2325  POPULATION ML2 FAMILY 1 FAMILY 2 FAMILY 3 FAMILY 4 FAMILY 5 FAMILY 6 FAMILY 7 FAMILY 8 FAMILY 9 FAMILY 10 FAMILY 11 FAMILY 12 FAMILY 13  78.9615 88.0000 92.5000 93.5000 69.0000 92.5000 77.0000 86.5000 72.5000 97.5000 95.0000 70.5000 52.5000 59.0000  8.7677 12.4075 10.3550 6.3750 6.6925 9.9575 11.2725 10.6700 7.5625 8.8375 8.7525 7.3250 10.8100 9.3500  23.6936 19.0200 25.0850 23.6250 23.8925 24.9025 20.4850 28.6850 22.3550 18.2410 25.2620 25.1520 21.5975 32.0860  11.1564 10.2625 13.4775  24.6226 30.3710 16.4490  DISTURBED POPULATION POPULATION KK1 FAMILY 1 FAMILY 2  75.0714 88.0000 26.0000  110 Appendix B (continued)  81.5000 74.5000 83.0000 82.5000 90.0000  14.6275 6.1100 13.4450 8.8925 11.2800  23.1135 25.8590 30.5540 23.7440 22.2680  POPULATION KK2 FAMILY 1 FAMILY 2 FAMILY 3 FAMILY 4 FAMILY 5 FAMILY 6 FAMILY 7 FAMILY 8 FAMILY 9 FAMILY 10  79.7500 75.5000 87.5000 93.0000 86.0000 68.5000 78.0000 84.0000 79.5000 35.5000 90.5000  9.6355 10.4275 8.5125 5.7925 6.4775 7.3250 10.5475 10.0850 9.7375 6.5575 14.5050  23.9697 21.0120 20.0990 23.6550 21.0370 24.5820 25.67 10 33.0500 16.5720 25.1200 26.5275  POPULATION MKI FAMILY 1 FAMILY 2 FAMILY 3 FAMILY 4  88.5000 87.0000 86.0000 90.0000 91.0000  8.2013 6.8200 9.3850 8.9750 7.6250  25.4536 28.0330 25.4820 21.7225 26.5770  POPULATION MK2 FAMILY 1 FAMILY 2 FAMILY 3 FAMILY 4  80.7500 90.5000 81.5000 76.0000 75.0000  7.8906 7.9975 6.9175 7.9450 8.7025  20.2830 20.8780 18.6670 17.53 10 24.0560  POPULATION SK1 FAMILY 1 FAMILY 2 FAMILY 3 FAMILY4 FAMILY 5 FAMILY 6 FAMILY 7  71.4286 85.0000 27.5000 92.0000 87.5000 44.5000 89.5000 74.0000  7.1871 8.2400 8.3325 5.6225 5.1175 8.8025 6.1900 8.0050  23.0084 24.6700 27.4460 19.7610 27.1670 13.0070 24.4600 24.5480  POPULATION SK2 FAMILY 1 FAMILY 2 FAMILY 3 FAMILY 4 FAMILY 5 FAMILY 6  73.7500 75.0000 78.0000 88.0000 49.0000 65.0000 87.0000  7.2212 4.9375 5.9425 5.4425 13.5000 6.3625 7.1425  27.6025 25.3370 29.0630 24.8480 27.0120 29.4650 29.8900  FAMILY FAMILY FAMILY FAMILY FAMILY  3 4 5 6 7  111 Appendix C Seedling-trait means for height (HT), diameter (DM), total dry weight (TDW), and root-shoot ratio (RS).  Trait  HT (cm)  DM (cm)  TDW (gm)  RS  Overall mean  53.3090  .6099  15.6205  .9577  ML1 AVE. 1 FAM FAM 2 FAM 3 FAM 4 FAM 5 FAM 6 FAM 7 8 FAM 9 FAM FAM 10 FAM 11 FAM 12 FAM 13 FAM 14  63.6896 62.4417 62.3334 59.9667 65.4167 65. 1500 65.8750 64.7500 63.0168 58.9167 71.6875 70.7500 60.6417 57.8333 62.8750  .6386 .6762 .6418 .6269 .6250 .6493 .6229 .5975 .6064 .6059 .6463 .7321 .6214 .6696 .6187  17.0650 20.4225 18.3300 18.2400 14.2950 14.4125 17.0200 17.4400 16.6925 17.5425 15.6675 18.1950 16.3725 15.9400 18.3400  .7809 .6956 .6662 .7107 .6918 .5461 .6581 .6706 .8950 .9210 .8692 .7371 .9763 1.0135 .8821  ML2AVE. FAM 1 FAM 2 FAM 3 FAM 4 5 FAM FAM 6 FAM 7 FAM 8 FAM 9 FAM 10 FAM 11  59.2837 61.2917 63.9250 43.1667 70.8667 49.6667 66.7917 72.8959 45.3167 59.2084 56.4167 62.5750  .6210 .6129 .6395 .5200 .7089 .5954 .6907 .6694 .5321 .6085 .6528 .6011  16.4659 19.4025 16.3900 14.3475 16.4550 16.4250 16.9625 17.4550 12.8675 14.5600 18.9300 17.3300  .7523 .6421 .7942 .8720 .7887 .7793 .7958 .6345 .7395 .8653 .5854 .7783  49.3056 42.5500 50.7500 55.7167 54.5667 45.3333 46.9167  .6013 .5582 .6331 .5962 .5745 .6394 .6066  15.6509 14.1200 22.1467 10.6375 13.9200 17.9875 16.7175  .8199 .9081 .8033 .5951 .7545 .9921 .8621  Undisturbed population  Disturbed population KK1 AVE. 1 FAM FAM 2 FAM 3 FAM 4 FAM 5 FAM 6  112 Appendix C (cont  KK2 AVE. FAM 1 FAM 2 FAM 3 4 FAM 5 FAM 6 FAM 7 FAM 8 FAM  52.1146 54.7083 48.8333 45.9083 58.5000 54.4000 64.1917 36.0000 54.3750  .6237 .6569 .6246 .6000 .6400 .5934 .7429 .5188 .6133  15.4468 12.4875 15.4725 14.3900 18.0667 17. 1275 23.0275 12.0550 11.6025  .9484 1.2542 .8778 .9858 1.0848 .8956 .8925 .9203 .7107  MK1 AVE. 1 FAM FAM 2 3 FAM FAM 4  46.0802 49.5000 52.8708 41.5750 40.3750  .5890 .6394 .5968 .5636 .5563  16.5547 15.6675 12.6300 22.5933 16.8375  1.3141 1.3854 1.0118 1.7439 1.1154  MK2 AVE. FAM 1 2 FAM FAM 3 FAM 4  40.5781 42.8667 37.5417 54.4875 27.4167  .5579 .5552 .5346 .6797 .4621  13.6306 13.6400 19.9125 17.7200 13.2500  1.1854 1.3042 .9055 1.0700 1.4620  SKi AVE. 1 FAM 2 FAM 3 FAM 4 FAM FAM 5 6 FAM  38.3437 27.9375 47.8750 41.2083 44.4083 34.5000 34.1333  .5672 .5119 .6346 .5915 .6032 .4900 .5721  11.9413 10.9700 13.7000 14.1725 9.6625 13.3050 9.8375  1.4791 2.0420 1.3569 1.4624 1.3549 .8566 1.8019  SK2 AVE. 1 FAM FAM 2 FAM 3  50.6944 43.6667 58.4584 49.9583  .5987 .5475 .6496 .5990  15.0133 11.8700 19.1875 13.9825  1.0014 .9979 1.0786 .9277  113  Appendix D Scoring method  ( Keiding et a!, 1984) for stemform score  Class Persistence of axis Double or multiple stems from ground level Axis branches out in 1st quarter of the tree Axis branches out in 2nd quarter of the tree Axis branches out in 3rd quarter of the tree Axis branches out in 4th quarter of the tree Complete persistence  1 2 3 4 5 6  Straightness Crooked more than three serious bends Crooked one or two serious bends Slightly crooked, many bends Slightly crooked, few bends Straight  1 2 3 4 5  Branch heaviness* Very heavy branches from 1/2 to 3/4 of stem Heavy branches about 1/2 of stem Medium branches from 1/4 to 1/2 of stem Light branches about 1/4 of stem Very light branches less than 1/4 of stem  2 3 4 5  -  -  -  -  -  -  -  *  Branch heaviness is a relative measure of branch size in proportion to the stem at the base of the branches.  114 Appendix E Average field progeny trait measures for height (HT) and diameter (DM) at 4 years  TRAIT  HT (m)  DM (cm)  Overall mean  3.4823  5.9480  Undisturbed population DL AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  3.6920 3.8050 3.7225 3.1438 3.6110 3.2688 3.4413 3.8650 3.7760 3.6912 3.7187 3.6662 3.9588 3.5835 4.0175 3.6625 3.7162 3.9037 3.4775 3.9713 3.8400  6.3918 6.7048 5.8952 6.0303 6.4035 5.7410 6.0445 6.4695 6.9902 5.9440 6.4260 6.8827 6.8180 5.8777 6.9828 6.0760 6.1863 6.7200 6.2398 6.5377 6.8650  ML AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15  3.5040 3.6613 3.6800 3.4600 3.0250 3.5263 3.2900 3.1831 3.4038 3.1825 3.9100 3.5050 3.9200 3.8338 3.5525 3.4113  5.7200 6.3618 6.0422 5.4875 5.0188 5.9737 5.8387 4.8785 5.5681 4.6802 6.3468 5.8160 6.2628 6.0752 5.4412 5.9765  115 Appendix E (continued)  SKH AVE. FAM FAM FAM FAM FAM  1 2 3 4 5  3.7045 3.5062 3.9650 3.8775 3.8063 3.3675  5.9882 5.8483 6.2088 6.1260 6.5073 5.2508  Disturbed population KH AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM  1 2 3 4 5 6 7 8 9  3.5093 3.0775 3.8800 2.9625 3.8500 3.6100 3.4938 3.4962 3.6687 3.5450  5.7343 5.2963 6.5940 4.8375 6.2435 6.2110 5.5878 5.6438 5.5375 5.6575  MS AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19  3. 1288 3.4638 3.5512 2.8738 3.7188 3.4013 3.1388 3.2512 2.8225 2.9943 3.3563 3.4662 2.9850 2.3913 2.9538 2.7238 3.2075 2.8450 2.7449 3.5300  5.7650 5.7117 6.0432 5.3323 7.1325 6.2620 5.6053 6.2217 4.9893 5.7915 6. 1063 6.7655 5.1958 4.9705 5.3356 5.2590 5.9940 5.0707 5.4008 6.2710  SM AVE. FAM FAM FAM FAM FAM  1 2 3 4 5  3.4653 3.6063 3.2712 3.6800 3.4500 3.3188  6.1108 6.3267 6.0305 6.2970 5.8898 6.0097  116 Appendix E (continued)  SK AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM  1 2 3 4 5 6 7 8 9 10 11 12  3.5556 3.5487 3.1700 3.4925 3.4562 3.9005 3.7612 3.7975 3.7603 3.5783 3.2400 3.7875 3. 1750  5.8543 6.1018 5.7785 5.5467 6.0670 6.9512 6.0827 6. 1480 5.8307 5.5698 5.3647 5.7330 5.0772  117 Appendix F Average measures of persistence of axis, stern straightness, and branch heaviness  PER  STN  BR  2.3531  2.5024  2.8319  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  2.3 150 2.4750 2.6250 2.2750 2.3500 2.4500 2.3000 2. 1750 2. 1750 2.4000 2.3750 2.1750 2.2750 2.2250 2.2000 2.3250 2.3000 2.2250 2.2750 2.4000 2.3000  2.5363 2.7000 2.7500 2.4750 2.6500 2.4750 2.4500 2.4500 2.3750 2.5750 2.7750 2.3500 2.5250 2.4250 2.5250 2.5500 2.4250 2.7000 2.3750 2.7500 2.4250  2.7925 2.8500 2.8750 2.8500 2.6750 2.8500 2.8000 2.6750 2.9000 2.8500 2.9000 2.8000 2.7500 2.8000 2.8250 2.8500 2.6000 2.8750 2.7750 2.7000 2.6500  ML AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15  2.3859 2.1500 2.5500 2.3750 2.4500 2.2000 2.4000 2.2821 2.4054 2.4750 2.3500 2.5000 2.4250 2.3500 2.5500 2.3250  2.5654 2.3750 2.6000 2.4750 2.4750 2.4500 2.5500 2.6667 2.7297 2.7000 2.6000 2.6500 2.6250 2.4750 2.6750 2.4500  2.8557 2.5500 3.0000 2.8750 2.7750 2.7750 2.8500 2.8974 2.9189 2.9250 2.9250 2.5750 2.9750 2.9500 2.9250 2.9250  SKH AVE. FAM FAM FAM  1 2 3  2.4850 2.3250 2.5750 2.5000  2.5950 2.6500 2.7000 2.5250  2.9500 2.8250 2.9750 3.0000  POPULATION Overall mean Undisturbed population  DL AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM  118  Appendix F (continued)  FAM FAM  4 5  2.3750 2.6500  2.5000 2.6000  2.9250 3.0250  Disturbed population KH AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM  1 2 3 4 5 6 7 8 9  2.4444 2.2250 2.3750 2.3250 2.3000 2.3750 2.4250 2.6250 2.6250 2.7250  2.5194 2.5000 2.4750 2.3500 2.4500 2.5250 2.4750 2.5750 2.4750 2.8500  2.8667 2.8500 2.8500 2.7750 2.8250 2.8500 2.9000 2.8250 2.8750 3.0500  MS AVE. FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM FAM  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19  2.2493 2.3250 2.3750 2.3000 2. 1250 2.3250 2.2750 2. 1250 2.1500 2.1000 2.2750 2.3250 2.2778 2.1750 2.2821 2.3000 2.3000 2.2250 2.1538 2.3250  2.3886 2.5000 2.6250 2.2000 2.2500 2.5000 2.3500 2.4750 2.2750 2.2500 2.4500 2.3250 2.3889 2.3250 2.3590 2.4500 2.4000 2.3750 2.3590 2.5250  2.7573 2.6500 2.8250 2.8750 2.6500 2.9750 2.7750 2.9000 2.5750 2.8250 2.6750 2.7250 2.8889 2.7750 2.7949 2.6000 2.7500 2.6500 2.7949 2.7000  SM AVE. FAM FAM FAM FAM FAM  1 2 3 4 5  2.3400 2.2750 2.3750 2.5000 2.4000 2.1500  2.4550 2.4750 2.5750 2.5500 2.4250 2.2500  2.8800 2.8250 2.9000 2.9250 2.9250 2.8250  SK AVE. FAM FAM FAM  1 2 3  2.4208 2.3500 2.3250 2.3250  2.5146 2.5500 2.5750 2.4750  2.8896 2.9250 3.0500 2.7500  119 Appendix F (continued)  FAM FAM FAM FAM FAM FAM FAM FAM FAM  4 5 6 7 8 9 10 11 12  2.4000 2.3250 2.7000 2.2750 2.4500 2.4750 2.4000 2.5250 2.5000  2.6250 2.5000 2.4500 2.4500 2.6500 2.5000 2.4500 2.5500 2.4000  2.8750 2.9750 2.9500 2.7750 2.8500 2.9750 2.9000 2.8250 2.8250  120  Appendix G Extraction buffer (Liengsiri et al, 1990)  Borate pH 8.0 0.05 M MgC12 Ascorbic acid 0.01 M CaC12 Cysteine-HCL 5.4 mM PEG 20 M Tergitol 0.5 M Sucrose B-mercaptoethanol Tween 80 1.0 %  0.2 0.2 1.0 1.0 0.3  % % % % %  pH adjusted to 8.0 with iN NaOH  Appendix H Running buffer systems  System  Electrod buffer  H Pitel and Cheliak (1984)  Tris 0. 125 M EDTA 1.4 mM pH adjusted to 7.2 with 1M citric acid  B Ridgeway ci at (1970)  Lithiumhydroxide 0.06 M 0.3 M Boric acid pH adjusted to 8.1 with iN NaOH  Gel buffer  Histidine-HCL 0.05 M pH adjusted to 7.2 with 1 M Tris 0.03 M Tris Citric acid 0.005 M Electrode buffer B 1 % pH adjusted to 8.5 with iN NaOH  121 Appendix I Enzyme staining recipes  1. Aldolase (ALD) E.C.4.i.2.3 25.0 ml 0.2 M Tris—HCL, pH 8.0 125.0 mg Fructose-1-6-diphosphase (tetrasodium salt) 38.0 mg Asenic acid (disodium salt) 150 units Glyceradehyde-3-phosphase dehydrogenase 0.5 ml NAD 0.5 ml MTT 0.5 ml PMS Note:Incubated in the dark about 60 mm. at 37 oc until dark blue bands appear. 2. Aspartate aminotransferase (AAT) E.C.2.6. 1.1 4.0 mg Pyridoxal-5-phosphase 100.0 mg Fast blue BB salt 25.0 ml AAT substrate solution Note:Incubated in the dark about 30 mm. at 37oc until blue bands appear on a pink background. 3. Diaphorase (DIA) E.C.1.6.4.3 0.2 M Tris-HCL, pH 8.0 2,6-dichiorophenol-indophenol B-NADH MTT Note:Incubated in the dark about 45 mm. at 37 oc  25.0 ml 0.5 mg 13.0 mg 0.5 ml  4. Esterase/colorimetric (EST) E.C.3. 1.1.1 25.0 ml 0.2 M phosphase buffer, ph 6.4 50.0 mg L-naphthyl acetate 50.0 mg B-naphthyl acetate 100.0 mg Fast blue RR salt Note:L-naphthyl acetate and B-naphthyl acetate were dissolved in 2.5 ml acetone before adding other substrates. :Incubated in the dark about 60 mm. at room temperature until bands appear. 5. Isocitrate dehydrogenase (IDH) E. C. 1. 1. 1.42 25.0 ml 0.2 M Tris-HCL, pH 8.0 200.0 mg DL-isocitric acid (trisodium salt) 0.5 ml 1 % MgCI2 (wlv) 0.5 ml ADP 0.5 ml NBT 0.5 ml PMS Note:Incubated in the dark about 60 mm. at 37 oc until dark bands appear.  122 Appendix I (continued)  6. Leucine-amino peptidase (LAP) E.C.3.4.11.1 50.0 ml (a) 0.5 M boric acid 25.0 ml (b) 0.2 M Tris-HCL, pH 8.0 25.0 ml 0.2 M maleic anhydride Mixed and adjusted pH to 5.3 with 1 N NaOH just prior to staining 0.5 ml (c) 10 % MgC12 50.0 mg L-leucine B-naphthylamide-HCL 50.0 mg Fast black K salt Note:Soaked gel in (a) for 15 mm., emptied solution, and rinsed gel with distilled water. :Mixed (b) and (c) and added to the gel. :Incubated in the dark at 37 oc. 7. Malate dehydrogenase (MDH) E.C.l.l.l.37 0.2 M Tris-HCL, pH 8.0 0.5 M DL-malic acid, pH 7.0 NAD NBT PMS Note: Incubated in the dark about 45 mm. at 37 oc until dark  12.5 ml 12.5 ml 0.5 ml 0.5 ml 0.5 ml blue bands appear.  8. Malic enzyme (ME) E.C.1.1.l.40 12.5 ml Electrode buffer H 12.5 ml 0.5 M DL-malic acid, ph 7.0 0.5 ml 1 % MgC12 0.5 ml NADP 0.5 ml MTT 0.5 ml PMS Note:Incubated in the dark at 37 oc until dark blue bands appear. 9. Phosphoglucose isomerase (PGI) E.C.5.3.l.9 25.0 ml 0.2 M Tris-HCL, pH 8.0 12.5 mg Fructose-6-phosphate (disodiurn salt) 5 units Glucose-6-phosphate dehydrogenase 0.5 ml 1 % MgCI2 0.5 ml NADP 0.5 ml MTT 0.5 ml PMS Note:Incubated in the dark at 37 oc until dark blue bands appear. 10. Phosphoglucomutase (PGM) E.C.2.7.5. I 0.2 M Tris-HCL, pH 8.0 Glucose-i-phosphate (disodium salt) Glucose-i ,6-diphosphate solution Glucose-6-phosphate dehydrogenase 1 % MgC12 NADP MTT PMS Note:Incubated in the dark at 37 oc until bands appear.  25.0 ml 250.0 mg 0.5 ml 25 units 0.5 ml 0.5 ml 0.5 ml 0.5 ml  123 Appendix I (continued)  11. Shikimic acid dehydrogenase (SDH) E.C.l.1.1.25 0.2 M Tris-HCL, pH 8.0 Shikimic acid 1 % MgCI2 NADP NBT PMS Note:Incubated in the dark at 37 oc.  25.0 ml 50.0 mg 1.0 ml 0.5 ml 1.0 ml 0.5 ml  12. 6-phosphogluconate dehydrogenase (6PGD) E. C. 1.1.1.44 5.0 ml 0.2 M Tris-HCL, pH 8.0 15.0 mg 6-phosphogluconic acid (trisodium salt) 1.0 ml 1 % MgCI2 1.0 ml NADP 1.0 ml MTT ml 0.5 PMS Note:Incubated in the dark about 45 mm. at 37 oc until dark blue bands appear.  124  Appendix J Allele frequencies and heterozygosity for each locus in eight populations  Population KKI  KK2  MLI  ML2  MK1  MK2  SKI  5K2  118  100  340  220  80  69  140  60  .915 .085 .000 .000 .000  .900 .100 .000 .000 .000  .997 .003 .000 .000 .000  .864 .136 .000 .000 .000  .994 .006 .000 .000 .000  .717 .196 .080 .000 .007  .846 .071 .054 .029 .000  .892 .100 .008 .000 .000  H(b) H(unb) H(DC)  .155 .156 .169  .180 .181 .200  .006 .006 .006  .236 .236 .264  .012 .013 .013  .441 .441 .203  .275 .276 .179  .195 .196 .217  6PGD2 (N) 1 2 3 4 5  119 .592 .408 .000 .000 .000  100 .630 .370 .000 .000 .000  340 .688 .312 .000 .000 .000  220 .727 .245 .027 .000 .000  80 .944 .013 .044 .000 .000  78 .692 .244 .032 .032 .000  140 .604 .239 .150 .004 .004  60 .700 .275 .025 .000 .000  H(b) H(unb) H(DC)  .483 .485 .664  .466 .469 .740  .429 .430 .618  .410 .411 .527  .107 .108 .112  .459 .462 .615  .556 .558 .643  .434 .437 .600  80 1.000 .000 .000  100 1.000 .000 .000  260 .660 .340 .000  140 .489 .421 .089  20 .150 .375 .475  80 .250 .512 .237  100 .160 .485 .355  60 .500 .225 .275  .000 .000 .000  .000 .000 .000  .449 .450 .665  .575 .577 .836  .611 .627 .650  .618 .622 .950  .613 .616 .970  .624 .629 1.000  300 .718 .282 .000  220 .595 .357 .048  80 .788 .131 .081  80 .894 .038 .069  140 .854 .050 .096  60 .817 .075 .108  Locus  6PGD1 (N) 1 2 3 4 5  AAT2 (N) 1 2 3 H(b) H(unb) H(DC)  ALD3 (N) 1 2 3  100 .690 .305 .005  20 .925 .075 .000  125 Appendix J (continued)  H(b) H(unb) H(DC)  .431 .433 .620  .139 .142 .150  .405 .405 .563  .516 .517 .809  .356 .358 .425  .195 .196 .213  .260 .261 .236  .316 .318 .367  ALD4 (N) 1 2 3  100 1.000 .000 .000  40 1.000 .000 .000  300 .957 .043 .000  220 .966 .032 .002  80 .994 .006 .000  80 .994 .000 .006  140 1.000 .000 .000  60 .925 .000 .075  H(b) H(unb) H(DC)  .000 .000 .000  .000 .000 .000  .083 .083 .060  .066 .066 .059  .012 .013 .013  .012 .013 .013  .000 .000 .000  .139 .140 .117  DIA (N) 1 2 3  120 1.000 .000 .000  100 1.000 .000 .000  340 1.000 .000 .000  220 .995 .000 .005  80 1.000 .000 .000  80 1.000 .000 .000  140 1.000 .000 .000  60 .842 .158 .000  H(b) H(unb) H(DC)  .000 .000 .000  .000 .000 .000  .000 .000 .000  .009 .009 .009  .000 .000 .000  .000 .000 .000  .000 .000 .000  .267 .267 .217  EST1 (N) 1 2 3 4  80 .656 .344 .000 .000  100 .605 .395 .000 .000  340 .501 .476 .022 .000  220 .586 .409 .005 .000  80 .650 .131 .156 .063  80 .688 .306 .006 .000  139 .698 .302 .000 .000  60 .625 .350 .000 .025  H(b) H(unb) H(DC)  .451 .454 .688  .478 .480 .770  .521 .522 .956  .489 .490 .773  .532 .535 .600  .434 .436 .613  .422 .423 .561  .486 .490 .750  EST2 (N) 1 2 3  80 .731 .269 .000  100 .635 .315 .050  320 .547 .444 .009  220 .841 .159 .000  80 .944 .056 .000  80 .806 .194 .000  140 .825 .079 .096  60 .725 .133 .142  H(b) H(unb) H)DC)  .393 .396 .538  .495 .498 .730  .504 .505 .906  .268 .268 .318  .106 .107 .112  .312 .314 .387  .304 .305 .350  .437 .440 .550  126 Appendix J  (continued)  EST3 (N) 1 2 3  80 .819 .181 .000  100 .895 .000 .105  320 .841 .058 .102  220 .530 .470 .000  80 .894 .106 .000  80 1.000 .000 .000  140 .796 .204 .000  60 .683 .317 .000  H(b) H(unb) H(DC)  .297 .299 .363  .188 .189 .210  .280 .280 .300  .498 .499 .532  .190 .191 .188  .000 .000 .000  .324 .325 .293  .433 .436 .567  IDH (N) 1 2 3 4 5  120 .954 .046 .000 .000 .000  100 .840 .160 .000 .000 .000  340 .815 .185 .000 .000 .000  220 .791 .209 .000 .000 .000  80 .762 .013 .131 .025 .069  80 .606 .306 .000 .006 .081  140 .729 .164 .000 .000 .107  60 .408 .242 .000 .000 .350  H(b) H(unb) H(DC)  .087 .088 .042  .269 .270 .200  .302 .302 .306  .331 .331 .282  .396 .398 .325  .532 .535 .438  .431 .432 .300  .652 .658 .483  LAP (N) 1 2 3 4  119 .882 .118 .000 .000  100 1.000 .000 .000 .000  340 .909 .091 .000 .000  220 .832 .168 .000 .000  80 .475 .500 .025 .000  80 .500 .500 .000 .000  140 .332 .554 .114 .000  60 .817 .158 .017 .008  H(b) H(unb) H(DC)  .208 .208 .235  .000 .000 .000  .166 .166 .112  .280 .280 .227  .524 .527 .950  .500 .503 1.000  .570 .572 .607  .308 .310 .300  MDH (N) 1 2  80 .700 .300  100 .990 .010  300 .870 .130  220 .773 .223  80 .950 .000  80 .931 .013  140 .986 .014  60 .983 .017  .000  .000  .000  .005  .050  .056  .000  .000  .420 .423 .600  .020 .020 .020  .226 .227 .260  .353 .354 .373  .095 .096 .100  .129 .130 .138  .028 .028 .029  .033 .033 .033  3 H(b) H(unb) H(DC)  127 Appendix J (continued)  MEl (N) 1 2  120 1.000 .000  100 1.000 .000  340 .910 .090  220 .925 .075  80 1.000 .000  80 1.000 .000  140 1.000 .000  60 1.000 .000  H(b) H(unb) H(DC)  .000 .000 .000  .000 .000 .000  .163 .164 .138  .139 .139 .105  .000 .000 .000  .000 .000 .000  .000 .000 .000  .000 .000 .000  ME2 (N) 1 2 3  120 .842 .158 .000  100 .915 .085 .000  340 .912 .075 .013  220 .893 .084 .023  80 .931 .063 .006  80 1.000 .000 .000  140 .986 .014 .000  60 .875 .125 .000  H(b) H(unb) H(DC)  .267 .268 .317  .156 .156 .170  .163 .163 .165  .195 .195 .195  .129 .130 .112  .000 .000 .000  .028 .028 .029  .219 .221 .250  PGI 1 (N) 1 2  100 .965 .035  100 1.000 .000  320 1.000 .000  200 .983 .018  60 1.000 .000  80 1.000 .000  140 .993 .007  60 1.000 .000  H(b) H(unb) H(DC)  .068 .068 .070  .000 .000 .000  .000 .000 .000  .034 .034 .035  .000 .000 .000  .000 .000 .000  .014 .014 .014  .000 .000 .000  PGI2 (N) 1 2 3  120 .833 .167 .000  100 .905 .095 .000  320 .939 .059 .002  200 .930 .070 .000  60 1.000 .000 .000  80 1.000 .000 .000  140 .961 .039 .000  60 .942 .058 .000  H(b) H(unb) H(DC)  .278 .279 .333  .172 .173 .190  .115 .115 .119  .130 .131 .140  .000 .000 .000  .000 .000 .000  .075 .076 .079  .110 .111 .117  PGM2 (N) 1 2 3 4  120 .975 .025 .000 .000  100 1.000 .000 .000 .000  340 1.000 .000 .000 .000  220 .875 .125 .000 .000  80 1.000 .000 .000 .000  80 .969 .031 .000 .000  140 .879 .118 .004 .000  60 .592 .150 .017 .242  128 Appendix I (continued)  H(b) H(unb) H(DC)  .049 .049 .050  .000 .000 .000  .000 .000 .000  .219 .219 .205  .000 .000 .000  .061 .061 .063  .214 .215 .014  .569 .574 .300  SDH2 (N) 1 2 3  120 .746 .254 .000  100 .865 .105 .030  335 .575 .248 .178  220 .468 .259 .273  80 1.000 .000 .000  80 1.000 .000 .000  140 .993 .000 .007  60 .942 .058 .000  H(b) H(unb) H(DC)  .379 .381 .508  .240 .241 .270  .577 .578 .594  .639 .641 .532  .000 .000 .000  .000 .000 .000  .014 .014 .014  .110 .111 .083  H(b) = Heterozygosity (biased) H(unb) = Heterozygosity (unbiased) H(DC) = Heterozygosity (direct counted)  129 Appendix K Outcrossing rate estimates (t) for individual families from eight populations  Population  t  Population  1.97 1.97 1.97 1.97 1.97 1.97 1.97 1.97 1.97 1.97 1.97 1.97 1.37 1.97  Fam 15 Farn 16 Fam 17 ML2 Fam 1 Fam2 Farn3 Farn 4 Earn 5 Farn 6 Fam 7 Farn 8 Fam 9 Fain 10  Undisturbed population ML1 Fam 1 Fam 2 Fam 3 Fam4 Fam 5 Fam6 Fam7 Fam 8 Fam 9 Fam 10 Fam 11 Fam 12 Fam 13 Fam 14  1.97 1.97 0.90* 0.95* 1.95 0.91* 1.95 1.00* 1.95 1.95 1.95 1.95 1.95  Disturbed population KK1 Fam 1 Fam2 Fam3 Fam 4 FamS Fam 6 KK2 Fam 1 Fam2 Fam 3 Fam4 FamS MK1 Fam 1 Fam 2 Farn 3 Fam4 *  1.01 1.94 1.94 1.94 1.94 0.20* 1.07 2.00 2.00 2.00 2.00 1.11 2.00 2.00 0.69*  Family with outcrossing rate 1.0 or smaller  MK2 Fain 1 Fam2 Fam3 Farn 4 SKi Fam 1 Farn2 Fam 3 Fain 4 Fam 5 Fam6 Farn7 SK2 Fain 1 Fam 2 Fain 3  1.93 1.93 0.93* 0.95* 1.91 1.91 1.91 1.91 1.91 Q79* 0.91* 1.97 1.97 1.97  L.) — .  c. i.)  —  —  —  —  —(-————  —  ‘a Qo -  C’ L .  L’J  —  —  —  (-a———  (—)I-J—I->  —  —  I-.) D—  tJ  -  —  —  —  —  U—  -(->k)-  ——“--I_J  i-)  —  —  —  J—  —  —  I-_)————  —  — —  —  I—)L-JI(J—(-)  —  —  —  —  —  I—)  —  --)t-JL-.)—  C’  —  —  ‘C 00  -3—I)—I-  I-.)  (-J.)  )(—)L.tJ——(-  —  c  —  — I——I-)  —  —  —  -  —  —t)L.  —  —--—I_)I--) — ———I_)I)l-)——(_)t)—  c  lI)(--I)  ---D-)  —1 C’ c,  t-)  I-  ———().—.—  ———(_)  ()—k)—-”--  I-.)  )————  4i(_)l_)—— -Jk)k)I—  ——L.a  —  ———tJ— —L-(.(  —  —  —  ———(‘Jt-_)I_J Itkt-  D—) L-) 1-) tJ I-.)  — —  — —  0 0  C  (0 a_  C  —  —  7’  —  —(-J——t-Jt->  —-—-L)t’-)  —  t-I-JI-  —--)t.,)  () t-.) — — t—)-.) —  Q’  ——L--)——  —  —  -)1-_)  -  —  —  ——to—— ——k)—  .  — —.  0 00 C.-)  U 0 00  0 00  C.,  C)  00 a I’-)  G  I—)  G  U  C,  C,  I-.)  a  U  C  a  a  C..)  a  00 a.  0( 0  CD CD  C  C  CD (10  C  j)  -  CD  -  0•  CD  C  CD  CD C  >  CD  C  — —  —  —  —  —  ———  —  —  —  —  ————  —  —  —  (9  (9  (9  (9  (9  (9  I3I3—l3($($(3  —  —13—  (9——  1313(3  —13—  —13——131313  —(9I3—--($-  (9  )3)——$.)  —  —  1313  1313  13($  (3  il)k)  —  —  —  —  —  13  —  —  —  —  —  —  ——13—  ——-—  (3—1313  ————  —-l3I3  (9  3  13  —1313—  ——t3--  13  t)  —  —  (9  —  —  —L’3l)L’3  —  —  3k)——  —  —  E EEEEEE EEE EE  —  ç_(/  C  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0099229/manifest

Comment

Related Items