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


831-ubc_1994-893885.pdf [ 2.01MB ]
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

Full Text

ESTIMATION OFGENETIC VARIATION INTHAILAND ROSEWOOD (DALBERGIA COCHINCHINENSIS PIERRE)byPrachote SoonhuaeB.Sc. (Forestry), Kasetsart University, 1972M.Sc. (Forestry), Kasetsart University, 1979A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Faculty of Forestry)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIADecember 1993c Prachote Soonhuae, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of__________________The University of British ColumbiaVancouver, CanadaDate 7 / y3DE-6 (2/88)11ABSTRACTGenetic diversity conservation of tropical forest trees, particularly economically valuable andendangered species, is urgently needed for maintaining ecological balances and sources of materials fordirect, indirect and optional uses and their existing values. Thailand rosewood (Dalbergiacochinchinensis Pierre) is an economically important tree species of Southeast Asia. Undisturbedpopulations in natural forest reserves were suspected to have different genetic variability from that ofdisturbed populations in farm and public areas. The objectives of this study were to evaluate geneticvariation and the mating system of Thailand rosewood in order to consider conservation strategies forthe species. The variation was studied using quantitative and isozyme characteristics.Using open-pollinated families, quantitative variation was assessed in seed and 9 months oldseedlings from eight populations, and in four years old seedlings grown in a progeny trial from sevenpopulations. The differences among families were significant for most traits in all of the materialtested. However, significant differences among populations were relatively high for most traits. Thevariation 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 isozymestudies. Twelve enzyme systems coding eighteen loci were analyzed. The overall mean of expectedheterozygosity 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 disturbedpopulations. 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 andisozymes, but the among population variability was relatively high for quantitative traits and moderatefor isozymes. The species was highly outcrossed, with low inbreeding. Due to the fact that remnantpopulations or subpopulations of the species are naturally small and there is difference betweeninpopulations from central and northeastern regions, about one thousand individuals collected from severalpopulations would assure sufficient genetic viability for the conservation in each region. The conservedpopulations should also be distributed throughout the natural range of the species to maintainadaptability of desired quantitative traits.Both in situ and ex situ conservation are needed for this endangered species. Disturbedpopulations should also be maintained as seed sources. Ex situ conservation of about 1,000 individualscollected from subpopulations within each region and planted within the regional area would preservethe broad genetic base and prevent the random drift of the species. In addition, successful conservationrequires strong support from government, involvement from local people, and international cooperationfor technical and financial support.ivTABLE OF CONTENTSPageABSTRACT iiTABLE OF CONTENTS ivLIST OF TABLES viiLIST OF FIGURES ixLIST OF APPENDICES xLIST OF ABBREVIATIONS xiACKNOWLEDGEMENTS xivCHAPTER 1 INTRODUCTION 11. Conservation and Genetic Improvement 31. 1. Morphological variation 51.2. Biochemical variation 61.2.1. DNA variation 61.2.2. Isozyme variation 72. Forest Situation in Thailand 93. Thailand rosewood 124. Study Objectives 155. Study Materials 16CHAPTER 2 VARIATION IN QUANTITATIVE TRAITS 211. Seed Trait Variation 211.1. Study methods 211.2. Data analysis 231.3. Results 24V1.3.1. Variation 241.3.2. Variances and heritabilities 261.3.3. Correlations 261.4. Discussion 272. Seedling Trait Variation 292.1. Study methods 292.2. Data analysis 312.3. Results 322.3.1. Variation 332.3.2. Variances and heritabilities 342.3.3. Correlations 352.4. Discussion 353. Field Progeny Trait Variation 393.1. Study methods 393.2. Data analysis 393.3. Results 393.3. 1. Variation 403.3.2. Variances and heritabilities 423.3.3. Correlations 453.4. Discussion 454. Conclusion 46CHAPTER 3 ISOZYME VARIATION 481. Study Method 482. Data Analysis 492.1. Allelic variation 502.2. Mating system 53vi3. Results 533.1. Allelic variation 543.2. Mating system 734. Discussion 755. Conclusion 78CHAPTER 4 CONSERVATION STRATEGIES 791. Quantitative Trait Consideration 792. Isozyme Consideration 813. Quantitative and Isozyme Correlations 823. Genetic Rationale 843.1. In situ conservation 853.2. Ex situ conservation 874. Socio-economic and Political Considerations 895. Conclusion 91CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS 931. Conclusions 931.1. Quantitative variation 931.2. Isozyme variation 941.3. Conservation strategies 942. Recommendations 95REFERENCES 97APPENDICES 108viiLIST OF TABLESTable Page1 Depletion of forest area in Thailand from 1961-1985 112 List of insect pests of Thailand rosewood 133 Environmental conditions of tree populations in which seed wascollected for studying seed, seedling, and isozyme variation 194 Environmental conditions of tree populations in which seed wascollected for establishing the provenance-progeny plantation 205 Analysis of Variance showing source of variation, degreesof freedom, and expected mean square of seed traits 226 Average seed trait measurements for eight populations 247 Germination, R50 and seed weight variation, variance componentsand family heritahilities for seed traits from eight populations 288 Analysis of Variance for seedling traits 309 Average measurements for seedling traits including height and diameter(HT and DM), total dry weight (TDW), and root/shoot dry weight ratio foreach population 3410 Height and diameter variation, variance components andheritabilities for 9-month nursery seedlings from eight populations 3611 Dry weight variation, variance components and heritabilitiesfor 9-month nursery seedlings from eight populations 3712 Average measurements of progeny traits including height and diameter(HT and DM), persistence of axis (PER), stem straightness (STN), andbranch heaviness (BR) ftr each population 41viii13 Height and diameter variation, variance components andheritabilities for 4-year field progeny from seven populations 4314 Scored trait variation, variance components and heritabilitiesfor 4-year field progeny from seven populations 4415 Variability parameters including mean number of alleles per locus,percentage of loci polymorphic, and mean heterozygosities fordirect-count and Hardy-Weinberg expectation at 18 loci for for thespecies level(data from all populations were combined and treated asone unit for analysis), eaach population, and the means averaged ateach level(SE in parentheses) 5816 Mean allele frequencies for 18 loci across 8 populations 6017 Overall means of unbiased and direct-count heterozygositiesfor each locus based on data combined from eight populations 6318 Chi-square test for deviation from Hardy-Weinberg equilibriumof individual loci in each population 6519 Contingency chi-square analysis for deviation of allele frequenciesfrom those of which were randomly sampled from the total population 6720 Expected and observed heterozygosity and resulting fixationindex for each locus based on the total population 6921 Matrix of genetic distance and for similarity coefficients 7022 F-statistics at all loci from 8 populations 7123 Clustering analysis for Nei’s (1978) unbiased genetic distance usingunweighed pair group method algorithm (UPGMA) 7224 Hierarchical analysis using F-statistics (Wright, 1987) 74ix25 Outcrossing rate (t) and fixation index (f for each population estimatedfrom fifteen loci 7526 Correlation between seed and seedling traits 8127 Correlations estimated from individual family means forheterozygosity, quantitative traits, and mating system 8328 Correlations estimated from population means for heterozygosity,quantitative traits, and mating system 83xLIST OF FIGURESFigure Page1 Location of populations in which seed was collected for the studies 172 Enzyme banding patterns for eighteen loci 553 Electrophoretic gel slices stained with two enzyme systems,PGM and IDH 574 Genetic variahihty including number of alleles per locus, probabilityof polymorphic loci, and observed and expected Hardy-Weinbergheterozygosities from eight loci for each population 685 Clustering tree using Nei’s (1978) unbiased distances 73xiLIST OF APPENDICESAppendix PageA Number of families or individual maternal trees used forthe studies of variation 108B Seed trait means for individual families in eight populations 109C Seedling trait means for height (HT), diameter (DM), totaldry weight (TDW),and root-shoot ratio(RS) 111D Scoring method (Keiding et al., 1984) for stemform 113E Average field progeny trait measures fbr height (HT) anddiameter (DM) at 4 years 114F Average measures of persistence of axis, stem straightness,and branch heaviness 117G Extraction buffer (Liengsiri et al., 1990) 120H Running buffer systems 120I Enzyme staining recipes 121J Allele frequencies and heterozygosity for each locus from eightpopulations 124K Outcrossing rate estimates (t) for individual families from eightpopulations 129L Inferred genotypes for maternal trees from eight populations 130xiiLIST OF ABBREVIATIONSInstituteAOSA Association of Official Seed AnalystsFAO Food and Agriculture Organization of the United NationsITTO International Tropical Timber OrganizationIUCN International Union for the Conservation of Nature and Natural ResourcesNAS National Academy of Sciences of the United States of AmericaNRC National Research Council of the United States of AmericaRFD Royal Forest Department of ThailandOTA Office of Technology Assessment of the United States of AmericaWRI World Resources InstituteTree populationDL Donglan population in Khon-Kaen provinceKH Khaoyai population in Nakomrachasima provinceKK Kangkoi populations (KK1 and KK2) in Saraburi provinceMK Mahasarakam populations (MK1 and MK2) in Mahasarakam provinceML Muakiek populations (ML1 or ML and ML2) in Saraburi provinceMS Same as the population MK1, but different families collectedSK Sisaket populations (SKi and SK2) in the east of Sisaket provinceSKH Sisaket population in the west of Sisaket provinceSM Same as the population SK2, but different families collectedxliiStatisticsCRD Completely Random DesignDF Degree of freedomEMS Expected mean squareP Statistical probabilityRBD Randomized Block DesignSD (or s) Standard deviation from the meanSE Standard error for estimateSV Source of variationX2 Chi-squareParametersBR Branching heaviness of 4-year progenyDM Root collar diameter of a seedling or progenyf Fixation index of an enzyme alleleh2 Narrow sense heritability of a genetic traitHT Total height of a seedling or progeniesPER Persistence of axis for the stern of 4-year progenyR50 Number of days to reach 50% of germinating capacityRDW Root dry weight of a 9-month seedlingRf Migration distance of an enzyme band relative to coloured dye distance on theelectrophoretic gelRS Ratio of root and shoot dry weight of a 9-month seedlingSDW Shoot dry weight of a 9-month seedlingSTN Stem straightness of 4-year progenyxivOutcrossing rate calculated from enzyme dataTDW Total dry weight of a 9-month seedlingEnzymes6PGD 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)xvACKNOWLEDGEMENTSI would like to express my sincere appreciation to my supervisor Dr. Oscar Sziklai for hissupport, advice, continual encouragement, and never-ending energy during my study and living inCanada. The advice, support, assistance, and editorial contribution of Dr. Tim Boyle, of ForestryCanada, 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 hisgenerous help in editing the additional parts.I am grateful to Mr. Pisal Wasuwanich, former Director of ASEAN-Canada Forest Tree SeedCentre, Thailand, for his encouragement to study in Canada. Thanks are also extended to mycolleagues at the ASEAN-Canada Forest Tree Seed Centre for their dedication in field and laboratorywork.Finally, I would like to thank the Canada International Development Agency and PetawawaNational 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.1CHAPTER 1INTRODUCTIONThe world’s tropical forests, which comprise the earth’s most complex and species-richecosystems (Office of Technology Assessment, OTA, 1984), are being destroyed at an unprecedentedrate (National Research Council, NRC, 1991). The estimated net annual loss varies between 14 and20 million hectares (Myers, 1991; World Resources Institute (WRI), 1990). Given the current rate oftropical 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 andsimilarly, the loss of alleles is, effectively, permanent). Conservation of genetic diversity, whichenables a species to adapt to changing environments is of immense benefit to human society. It is vitalthat 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 understandingof the biological dynamics of populations because this information makes it possible to predict trendsin 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 forsampling 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 speciesneeds to be addressed, particularly that of economic species; and basic knowledge about variationamong and within populations is a prerequisite for deciding on sampling strategies (Guries and Ledig,1977).2Genetic improvement of economically important tree species can make a contribution toconservation 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, treeimprovement in Thailand must take place simultaneously with genetic conservation because genetic basepopulations 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 forimprovement from natural populations will contribute directly to genetic conservation and indirectlythrough the future production of improved seed.Diversity conservation of an economically important tree species would provide not onlysources of plants for immediate and future reforestation programmes, but also a wide genetic base forlong-term tree improvement. Because reforestation is expensive, genetically well-adapted materialsare needed to ensure success. Establishment of tree improvement programmes for an economicallyimportant 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 essentialfor continued progress over many generations of improvement (Zobel and Talbert, 1984). In general,tree improvement programmes involve three distinct but related phases: conservation, selection andbreeding, and propagation (Cheliak and Rogers, 1990). The programmes require an integration ofreforestation, silviculture and forest genetics( White, 1987; Zobel and Talbert, 1984). However, theprocedures are complex and vary around the world. It is tempting but not appropriate to considerimprovement 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 speciesin a certain region should be determined to support the implementation of both conservation andimprovement programmes.31. Conservation and Genetic ImprovementThe Food and Agriculture Organization (FAG) (1988) emphasized the need for conservationof a species because of the rapid loss of special valuable genetic materials for use in reforestation andbreeding. Narnkoong (1 984a) recognized genetic conservation as a prerequisite for forest conservationand genetic management; therefore, forest scientists must he responsible for wide population samplingand 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 thebiology of a species is fundamental to adequate conservation and sustained use. Where such knowledgeis lacking or insufficient, research should he carried out in parallel with efforts to conserve and utilizethe species. Roche and Dourojeani (1984) stated that conservation of a species will he determined bythe biology of the species and the degree to which it is known and used by man. Bawa (1976) alsorecommended that the information about genetic variation of natural populations and reproductivebiology 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 thetremendous diversity of tropical forest species is understood, conservation strategies can be made formore sound scientific criteria. Additionally, conservation should be directed towards well adaptedmaterial, with abundant variation in traits to he improved, and should also concern secondarilyimportant traits.Because selectively neutral genes are not functionally different (Kimura and Ohta, l97), theaim of conservation is to collect and preserve adaptive gene complexes (Ledig, l986h; Marshall andBrown, 1975). However. Fryer (1987) has cited that most electrophoretically detected variants areselectively neutral isozymes which are of value primarily as gene markers and are of limited usefulnessin studies of selection in natural populations. Although sampling strategies have focused on singleadaptive 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 loci4which are the real target of genetic conservation. Guries and Ledig (1981) found significantcorrelations 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 stronglyassociated with high levels of polymorphism and heterozygosity, as is in the case of many tropicalspecies.Variability of genetic traits and their modes of inheritance provide a principal insight into thegenetic structure of forest tree populations (Hamrick and Loveless, 1986; El-Kassaby, 1980). Thedistribution of genetic variation within and among breeding populations directs the mating system andgeneflow 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 variationexisting for a trait in a species should be determined for genetic improvement; however, recognitionof the value of genetic variation of tropical tree species and its uses is more recent and needs furtherstudy.Progressive tree improvement will have to be integrated with the conservation of geneticvariability (Cossalter, 1989). Marshall and Brown (1975) concluded that population structure of aspecies is important for genetic conservation and improvement because the amount of genetic diversitywithin populations and the range and distribution of this diversity among populations determine theoptimum sampling strategy. Since the sampling strategy depends on population genetic structure(Guries and Ledig, 1977), an intensive investigation of population structure should therefore preferablyprecede the initiation of conservation programme. In addition, Bawa (1976) suggested that variationpatterns 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 andbiochemical traits. Quantitative trait variation is mainly studied from phenotypic expression of yield,5quality, and pest and disease resistance in all possible environments (Brown and Moran, 1979). Thereare two approaches developed for determination of biochemical variation; they are the study of DNAvariation by means of restriction fragment length polymorphisms (RFLPs) (Botstien et at., 1980) andrandom amplified polymorphic DNAs (RAPDs) (Williams et at., 1990), and the study of isozymevariation by means of polymorphic enzymes (Lewontin and Hubby, 1966). Study of bothmorphological and biochemical traits allows much more information to be obtained. Not only canmorphological traits be used to assess the possibility of genetic improvement of adaptive characters, buta comparison of results from both types of traits can improve our understanding of variation patternsand promote more efficient genetic conservation.1.1. Morphological variationMorphological characteristics of tree species have been the focus of studies in several differenttraits and in different developmental stages of species. Seed quality determined through the variationof viability and seed dimensions or mass plays a significant role in producing a large number of healthyseedlings 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 thereforenecessary for genetic manipulation of a tree species because seed quality traits - total seed yield,viability, seed size, germination percentage and energy - can carry over to influence the maturephenotype 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 andreduce costs and time of improvement (Lambeth, 1983). Sziklai (1974) has elucidated that informationon variation pattern, mode of inheritance and heritability values during developmental stages may beuseful in predicting future performance of genetically controlled traits. Numerous studies concernedwith early evaluation of seedling traits from either field or nursery growing have been reported for6patterns and amounts of variation (Carter et aL, 1990; Rogers et al., 1989; Jiang et al., 1989; Griffinand Cotterill, 1988; Magnussen and Yeatman, 1987; Sweet and Wareing, 1966).In traditional tree improvement, study of progeny variation from a field plantation accountsfor much research effort, since a progeny test provides the essential evaluation of parental genetic worth(Zobel and Talbert, 1984). Field performance of progeny is controlled by both genetic andenvironmental factors. Statistical methods (Cochran and Cox, 1957; Steel and Torrie, 1960; Snedecorand Cochran,1967; Sokal and Rohlf, 1969) provide techniques for partitioning genetic andenvironmental variances. A genetic distribution or family structure will depend on howprogeny-provenance trials, conservation stands, and genetic base populations have been managed atthe initiation of genetic manipulation (Kanowski and Nikles, 1989). Thus, the determination of geneticvariability 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 asfundamental 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 isattributable to genetic differences among individuals. Estimates of heritability, however, are not madewithout error; therefore, the ratios obtained are only a relative indication of genetic control and shouldnot be interpreted as absolute or invariant values (Zobel and Talbert, 1984). The concept of heritabilityis 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 variation1.2.1. DNA variationModern techniques employing the differentiation in molecular DNA and enzymes providepowerful tools for the detection of genetic variability. However, detection by means of restriction7fragment length polymorphisms (RFLPs) or random amplified polymorphic DNAs (RAPDs) requiresequipment and well-trained manpower. The technique, therefore, has been applied only on a limitedbasis in some developed countries, and it has yet to be used in developing tropical regions.1.2.2. Isozyme variationEnzymatic variation remains the available technique for the study of population genetics offorest trees. Many enzymes, a class of specific proteins, have multiple forms and those having similaror 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 widelyaccepted (Feret and Bergmann, 1976). Prakash er al. (1969) introduced the term “allozyme” for thedifferent enzymes produced by different alleles of one or the same gene locus. Allozymes aretherefore allelic isozymes (Cheliak et al., 1987). Isozymes, composed of amino acids, becomepositively charged in acidic environments and negatively charged in basic environments (Feret andBergmann, 1976). The introduction of isozymes into a molecular sieving medium with forced migrationthrough the sieving medium under the influence of a powerful electric current is called“electrophoresis”. The isozymes with different physical characteristics can be separated in solutionsof specific pH values (Feret and Bergmann, 1976).Isozyme electrophoresis is a very effective method for studying genetic variation because thereare small environmental effects, there is codominant expression, and it is possible to study many locifrom each individual (Liengsiri et al., 1990). Brown and Moran (1979) also noted that isozymevariation represents variation close to the DNA level. The present applications of isozymes for thegenetic study in forest trees has been summarised by Liengsiri ci al. (1990) as follows: identificationof species, hybrid, provenance, clonal, cultivar, and seed origin; studying species genetic diversity;allocation of genetic diversity among populations; investigations of genetic organization within apopulation; investigations of mating systems; studying effects of genetic forces; and studying8evolutionary genetics. Several investigators (Brown, 1979 and Hamrick et at., 1979) reviewedelectrophoretic studies of genetic variation among plants. Boyle et at. (1990) have stated that the useof 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 outcrossingmode of mating - are expected to maintain high levels of genetic variation. Mitton (1983) hassuggested that young tissue from seedlings provides good material for enzyme extraction because thehigh level of secondary compounds - e.g. tannins and phenols - in mature tissue may bind andprecipitate proteins, making them inaccessible with simple extraction procedures. Besides theestimation of genetic variation, isozyme analysis can be used for estimation of mating systems of treespecies. The mating system of species is the pattern in which gametes unite to form the nextgeneration (Stern and Roche, 1974). Mating system parameters provide estimates of the degree ofinbreeding in natural stands and indicate the extent of potential inbreeding depression in economicallyimportant traits (Shaw and Allard, 1982).Although mating systems can be estimated from morphological traits (Morgenstern, 1972; Parkand Fowler, 1984), allozyme data are suitable because of the co-dominant expression and the largenumber of loci that can be assayed (Lewontin, 1974). Several statistical models have been developedfor 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 matingsystem in trees has ranged from high degree of inbreeding due to selfing (Fedorov, 1966) to wideoutcrossing (Ashton, 1969). Bawa (1974), Chan (1983), and Bawa et at., (1985) have concluded thatmost tropical tree species are self-incompatible or dioecious. However, the level of inbreeding dueto 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 largenumber of tree species (Moran et at., 1989; Moran and Bell, 1983). In tropical trees, the limited9studies of mating systems suggest that many species may be predominantly outcrossed (O’Malley andBawa 1987). Kimura and Crow (1963) developed the concept of effective population size which is thesize of an equivalent random mating population having the same fixation index as that observed in theactual population. Yasuda (1969) then developed a method for estimating effective population size andthe method is widely used today (Cheliak et al., 1985). However, the effective population size in foresttrees is questionable (NRC, 1991) because the mathematical models can oversimplify more complexbiological realities (Ewens et a!., 1987).Because quantitative trait variation is subject to selection and/or adaptation, conservation oftree 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 populationstructure ( Epperson, 1989; Fumier et a!., 1987) and mating system (Brown, 1989) which are importantfor determining the number, size and distribution of conserved areas. Thus, both quantitative andisozyme information are required for effective conservation of tree species.2. Forest Situation in ThailandThailand covers an area of 51.3 million ha between latitudes 5° 45’ and 200 30’ N andlongitudes 970 30’ and 105° 45’ E. The country has a variety of vegetation types ranging from tropicalevergreen rain forest to dry deciduous forest and savanna forest. The major vegetation types ofThailand, as summarised from FAO (1981) are the following:(a). Evergreen and semi-evergreen forests, including hill evergreen, fresh water swamp, andmangrove forests. This type of forest occurs from sea level to 1000 m elevation where annual rainfallis 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 isbetween 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, generally10either 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 thecountry on sandy or lateritic soils. Bamboo forests, coniferous forests, and scrub formations are themajor sub-types of savanna forests.Due to the rapid rate of population increase, forest land has been encroached upon for severalreasons, The major causes of deforestation are shifting cultivation practices, illegal cutting andprocessing, infrastructure development, settlement, and natural disasters. Factors causing forestdepletion vary considerably across the regions. In the North, the illegal and legal cutting, the extensionof cultivated areas, and shifting cultivation all seriously threaten the forest ecosystems. In theNortheast, the increased extension of agricultural land plays a relatively more important role withreduction of forest land. In the South, rubber cultivation and mining activities appear to play theleading roles. Rates of forest depletion from 1961 to 1985 are presented in Table 1. Total forest landarea covers 15.3 million ha (30%), of which 14.8 million ha (29%) are natural forests and 0.5 millionha (1%) are plantations. The depletion from 1961 to 1985 was 12.6 million ha. The deforestation ratewas 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. Accordingto the current forest policy, at least 40 percent of the country should be kept under forest, 37.5 percentof which should be reserved and 62.5 percent used as productive forest. Thailand is now faced withthe situation where the remaining forests (29%) can no longer provide sufficient wood supplies for thepopulation. Therefore, about 10 percent of the country needs to be reforested, and existing forestsshould 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 beenestablished 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 recognitionTable1DepletionofforestareainThailandfrom1961-1985;thefirstlineineachregionisareadepletionandthesecondlineisforestarealeftintheregionattheperiodspecified.Thenumberinparenthesisispercentageofthetotalregionarea.ForestareaForestareaRegion(mill.ha.)(mill,ha.)196119731976197719821985Northern16.96-0.00(0)1.40(0)0.74(4)0.75(4)0.36(2)11.63(68)11.63(68)10.23(60)9.49(56)8.78(52)8.42(50)Northeastern16.89-2.02(12)0.92(5)1.03(6)0.53(4)0.18(1)7.09(42)5.07(30)4.15(25)3.12(19)2.59(15)2.41(14)Eastern3.65-0.62(17)0.24(6)0.16(4)0.30(8)0.00(0)2.12(57)1.50(40)1.26(34)1.10(30)0.80(22)0.80(22)Central6.75-1.17(17)0.22(3)0.14(2)0.19(3)0.15(3)3.57(53)2.40(36)2.18(33)2.04(31)1.85(28)1.70(25)Southern7.07-0.85(13)0.17(3)0.08(1)0.12(2)0.14(2)2.96(42)2.01(29)1.84(26)1.76(25)1.64(23)1.50(21)Total51.31-5.19(10)3.23(6)1.42(3)1.86(3)0.86(2)27.36(53)22.17(43)18.94(37)17.52(34)15.66(31)14.80(29)*1961Aerialphotos;1973,1976,1977,1982,1985Landsatimagery.Source:RoyalForestDepartment,ForestStatisticsofThailand1986.12that deforestation was the main cause of the worst natural disaster in decades, when flooding andlandslides killed more than 450 people (Rubeli, 1989).3. Thailand rosewoodRosewoods are tree species of the genus Dalbergia belonging to subfamily Papilionaceae,family Leguminosae. Some 250 Dalbergia species are known and most are tropical shrubs andclimbers (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 dozenor so species of the genus produce the rosewood timbers so famous for their rich colours, beautifulgrain, pleasant fragrance, and superlative technical quality. Today, rosewoods are among the mostesteemed 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. cultrataGrah, D. nigrescens Kurz, D. bariensis Pierre, D. dongnaiensis Pierre and D. sissoo Roxb.. The firsttwo species named produce attractive, high quality, and very valuable wood products. Both species arenow very rare, and D. oliveri Gamble is threatened to the extent that almost no populations remain toallow for genetic studies.Dalbergia cochinchinensis Pierre (Thailand rosewood) is an economically important treespecies distributed in eastern and northeastern Thailand, Burma, Cambodia, Laos, and Vietnam inmixed deciduous forests and dry evergreen formations (Keating and Bolza, 1982). The Royal ForestDepartment (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 aterminate leaflet. The leaflet is ovate, 3-4 cm wide, 4-7 cm long, and has an acute tip. The speciesflowers during May and July, producing small, white flowers in small panicles. The seed matures13during October to December in pods that are 1-2 cm wide, 4-6 cm long, thin, flat, oblong, rathernarrow and indehiscent, containing 1-4 flat seeds. The seed is kidney shaped, brown coloured, 4 mmwide, and 7 mm long. Seeds are generally collected from tree crowns and half are damaged by pestsand diseases. However, information about seed pests and diseases is scarce. Insect pests for seedlingsand 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 injuryAntrocephalus sp. Chalcididae Seed boringApoderus sp. Curculionidae Leaf rollingAristobia sp. Cerambycidae Stem boringColasposoma sp. Chrysomelidae Leaf eatingHypomeces squamosus F. Curculionidae Leaf eatingPlecoptera feflexa (Walker) Coctuidae Leaf eatingPsilogramma rnenephron (Cramer) Sphingidae Leaf eatingSphenoptera sp. Buprestidae Stem boringStriglina scitaria Walker Thyridae Leaf rollingThrenetica lacrymans Thomson Cerambycidae Stem boringTrichochrysea severini (Jacoby) Chrysomelidae Leaf eatingUnidentified chrysomelid Chrysomelidae Leaf eatingAlthough this tree species has been planted in some areas for over 30 years, silviculturalinformation is obscure and mainly based on individual observation and experience. Natural seedproduction is highly variable from year to year, though some trees start to produce seed at a very youngage (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 producinguniform seedlings. Piewluang and Liengsiri (1989) used various pretreatments to break the dormancyand found that seedcoat scarification by gently rubbing both flat sides of the seed with sand paperprovided maximum germination (86%).14Soaking seed in concentrated sulphuric acid for one minute and soaking in cool water fortwenty-four hours also provided higher germination percentage (85 % and 80%, respectively) than thecontrol or germinating seed without pretreatment (70% germination). Moreover, soaking seed inconcentrated sulphuric acid pretreatment gave more uniform germination than soaking in water. Thereis no record of seed germination percentage in seed beds. However, from experience, it is found thatthe 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 gmper m2 in the seed bed. Rice ash or saw dust is then used for a covering of about 1 cm thick on theseed bed. The seed bed is watered every day. Seedlings, about 2 cm height, are transplanted to plasticbags during the first week of germination. Five to six month old seedlings from a nursery are usuallyused for outplanting. Although planted seedlings at the Khao-yai Field Station were attacked by insectpests, the rate of survival one year after planting was over 90%.Wood properties of this tree species were described by Keating and Bolza (1982). Thesapwood is up to 40-50 mm wide, well defined and almost white. The heartwood is from light rose-purple to burgundy with darker brown or black streaks, producing attractive patterns. The wooddarkens 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 andlustrous. The timber is hard, very strong, occasionally brittle. Despite its density, sawing can be donewith little difficulty by using saws with shorter, stiffer teeth. The wood machines well, finishes andpolishes to a lovely surface. The best results from seasoning are obtained in log form, but good resultsare reported in kiln-drying if seasoned slowly (Keating and Bolza, 1982). The timber produces anattractive decorative veneer with low moisture movement and is durable in exposed situations. Thetimber is mainly used for special purposes; for example, high-quality furniture, parts of musicalinstruments, and cabinet work. In addition, the wood has been used for making high quality charcoalin local areas.As for most rosewoods, Thailand rosewood is now nowhere abundant; all the accessible stands15have long been logged and destroyed. The species has become scarce and endangered. The specieshas been used in planting programmes in many parts of Thailand. Large amounts of seed of this treespecies are required for such programmes. The rapid loss of forest areas has caused a loss ofgermplasm of the species and some sources of the species are already extinct. Thus, geneticconservation of the species is an urgent task (RFD, 1989a). The species was therefore included in theThai Royal Forest Department /Danish International Development Agency (RFD/DANIDA) hardwoodconservation programme (Sa-ardavut et al., 1989). However, genetic information about variation orpopulation structure as well as the mating system of the species is lacking and needed to ensureeffective management of the conservation programme. Therefore, genetic information is immediatelyrequired for the efficient conservation and seed production of this precious genetic resource (RFD,1989b).4. Study ObjectivesSince 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 tohuman activities than those in national parks and forest reserves. Thus, the populations within publicand farm areas are considered disturbed and generally smaller in size than those in natural forests whichare undisturbed. When populations are small, limited number of trees involve in mating system andtrees will mate between close relatives. This might causes inbreeding depression which could reducevariation 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 thatdiversity within disturbed tree populations was more adversely affected than that within undisturbedones. 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 thespecies.5. Study MaterialsSeed 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 collectedduring November and December 1990 from open-pollinated maternal trees (families) distributed in eightpopulations in Thailand. Figure 1 presents the distribution area of the tree species and locations of treepopulations 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), andSrisaket 2 (5K2), were in public or farm areas and considered disturbed. Because of the limitedamounts of seed derived from some families and their losses of germination at different stages of thestudies, 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 57families, respectively. The numbers of families in the individual populations varied from four toeighteen, 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). Thealtitude range is 200-350 m above mean sea level respectively. Average annual rainfall range is 1400-1650 mm, and average annual temperature is 27-28° C. Soil texture varies from sandy-loam to clayloam with pH between 5 and 6. After collection, seeds were extracted from pods, damaged seeds171000 102° 104° 106°200 20°18°1816° 16°140 14012°____________________________________________________________________________________12100° 1020 104° 106°Figure 1 Location of populations in which seed was collected for the studies:(A) 8 populations for studying seed, nursery seddlings, and isozyme variation: 1 = KKI,2 = KK2, 3 = MLI, 4 ML2, 5 = MK1, 6 = MK2, 7 = SKI, and 8 = SK2;(B) 7 populations for progeny trial: 3 ML, 5 = MS, 7 = SM, 8 SK, 9 DL, 10SKIT, and 11 = KH40 0 100 km= UndisturbedC) = Disturbed18separated out, and good seeds stored at -2° C.(b). Seed used for planting at the field station was collected in 1987 from 85 open-pollinatedfamilies distributed in seven populations in Thailand (Figure 1). Populations Donglan (DL), Muaklek(ML), and Srisaket Khunharn (SKH) were in natural forests and considered undisturbed; populationsKhaoyai (KH), Mahasarakam (MS), Srisaket Muang (SM), and Srisaket Khukhan (SK) were in farmor public areas and considered disturbed (Table 4). Numbers of families in the individualpopulations were different, and ranged from five to twenty (Appendix A).Gographic, macroclimatic, and edaphic conditions in all populations were similar to those ofthe seed collection for studying seed, seedling, and isozyme variation. After processing, in lateDecember 1987, seeds were germinated in seed beds for about 1 week, and seedlings were transplantedinto 6 x 4 in. plastic bags and raised in a lath-house for 5 months. The seedlings were then outplantedat the beginning of June 1988, in Khaoyai Field Station. The station is located in Nakornrachasimaprovince, Northeastern Thailand.Table3Environmentalconditionsoftreepopulationsinwhichseedwascollectedforstudyingseed,seedling,andisozymevariationYearlyAve.Pop(No)lat.long.elev.raintemp.soilstatusmmm°CtextureUndisturbedpopulationMLI(3)14°34’N101°liE320165027LoamMedium-sizedtreespatchilydistributedinmarginalareaofnaturalforestML2(4)14°30’N101°17’E350165027LoamSmall tomedium-sizedtreespatchilydistributedinnaturalforestDisturbedPopulationKKI(1)14°35’N101°2’E250160028ClayMedium-sizedtreeswidelyscatterinfarmlandKK2(2)14°40’N101°7’E250160028ClayMedium-sizedtreeswidelyscatteredinscrubformationMK1(5)16°12’N103°21’E280150028SandyloamLargetreespatchilydistributedinschoolareaMK2(6)16°iN103°lO’E280150028SandyloamLargetreespatchilydistributedinfarmlandandtempleSK1(7)15°10N104°23’E280140028SandyloamMediumtolarge-sizedtreesalongroadsidesandinschoolareaSK2(8)14°33’N104°28’E300140028LoamysandMedium-sizedtreesindisturbed-marginalareaofnaturalforestTable4Environmental conditionsoftreepopulationsinwhichseedwascollectedforestablishingtheprovenance-progenyplantationYearlyAve.Pop(No)lat.long.elev.raintemp.soilstatusmmm°CtextureUndisturbedpopulationDL(9)16°46’NML(3)14°34’NSKH(10)14°33’NDisturbedKH(11)MS(5)SM(7)SK(8)population14°24N16°l2’N15°TN14°38’N102°8E101°liE104°28E101°33’E103°20’E104°18’E104°12E230320300330280230240140016501400160015001400140028 27 28 28 28 28 28LoamLoamLoamysandClaySandyloamSandyloamLoamysandMedium-sizedtreespatchilydistributedinforestlandMedium-sizedtreespatchilydistributedinmarginal areaofnaturalforestMedium-sizedtreesinmarginalforestMedium-sizedtreesinprivatelandoralongroad-sidesLarge-sizetreespatchilydistributedinschoolareaMedium-sizedtreesalongroad-sidesMedium-sizedtreeswidelyscatteredinfarmland1’)C21CHAPTER 2VARIATION 1N QUANTITATIVE TRAITSQuantitative traits are regarded as effects of several genes (polygenic) and variation in thesetraits is generally confounded with environmental effects. Statistical methods provide a number ofdesigns for partitioning variances of environmental and genetic effects from each other. Thequantitative traits investigated in this study involved seed, seedlings, and progeny.1. Seed Trait Variation1.1. Study methodsVariation of seed was studied on samples obtained from 69 open-pollinated families of eightpopulations (Table 3). Seed quality determines the potential production of healthy seedlings forplanting. Germination capacity is used for evaluation of seed quality. A germination test in controlledenvironments is a routine practice for studies of seed variation. Results from the test establish themaximum plant-producing potential of families and correlate quite well with emergence underfavourable field conditions (St. Clair and Adams, 1991). The germination test, therefore remains theprincipal 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 willgerminate with fast and high percent germination. In this study, the germination test was conductedto measure germination percent and R50 (number of days needed to reach 50% of the germinatingcapacity).Seed weight is generally affected by genetic factors, but also very much by environmental22factors (e.g., nutritional status of parent). Differences in seed weight can contribute to familydifferences 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. Thefollowing statistical model was used.Yijk = ,i + P1 + F() + eUi th ti-iWhere YjJk = observed value of the replication in theJ family within the 1population= population meanP1 = population effectFJ() family effecteJ = 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 meansquare of seed traitsSource DF EMSPop p-l VE + rVF(P) + rfVFam (Pop) p(f-1) V + rVF(P)Error pf(r-1) VEWhere V = population varianceVF(p) = family (within population) variance23V = error variancep = number of populations (= 8)f = average number of families per population (= 69/8)r number of replications (= 4)The germination test was performed at 30°C. Fifty seeds from each family were germinatedin one replication. Germinated seeds (germinants) were counted every other day. Seeds wereconsidered to have germinated when their radicles had emerged to twice the seed length. Thegerminants were then removed from the germination boxes. The test was conducted over a period of3 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 measuredusing a different set of seeds from that used in the germination test.1.2. Data analysisData 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) wascalculated using the methods described by Steel and Torrie (1960).The Expected Mean Square (EMS) (Table 5) from each source or factor was partitioned intovariance components (Wright, 1976). The replicated plot means were used for the analysis of seedvariances, thus family heritahilities were calculated from the following formula:V F(PH211, =Vf. + VF(p241.3. ResultsAppendix B presents averaged values of seed trait measures, germination percentage, R50,and seed weight, of each population and family.1.3. 1. Variation(a). GerminationMean germination percentage was 80.59 and ranged from 71.43 (SKi) to 88.50 (MK1) (Table6).Table 6 Average seed trait measurements for eight populationsPopulation %Germination R50 (day) Seed weight (mg)UndisturbedML1 88.42 10.66 24.37ML2 78.96 8.77 23.69Ave. 84.45 9.76 24.03DisturbedKK1 75.07 11.17 24.62KK2 79.75 9.64 23.96MK1 88.50 7.19 23.01MK2 80.75 7.22 27.60SKi 71.43 8.20 25.45SK2 73.75 7.89 20.28Ave. 77.43 8.55 24.15Overall Ave. 80.59 9.25 24.1825Means 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 disturbedpopulations, 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 populationsfor all seed traits studied. Variation among populations for germination percentage was not statisticallysignificant (F = 1.56). On the other hand, variation among families within populations was highlysignificant (confidence level = 0.01, F = 30.3).(b). R50Another criterion analyzed from the germination test data was R50 (number of days neededto reach 50 % of the germination capacity). Means for each population and each family are given inAppendix 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 andundisturbed populations was nonsignificant (F = 1. 16). Nonetheless, among populations, variation washighly significant (confidence level = 0.01, F = 3.86) (Table 7). Overall variation among familieswithin populations was highly significant (F = 27.01).(c). Seed weightSeed weight was the last seed trait studied. Means of single seed weight for populations, eachpopulation, and each family are also presented in Appendix A. The overall mean was 24.18 mg, andmeans of populations ranged from 20.28 (SK2) to 27.60 (MK2) (Table 6). Analysis of varianceindicated that there was no significant difference between undisturbed and disturbed populations (F =0.02). Variation among all populations was nonsignificant(F = 1.50), but among families within allpopulations was highly significant (F = 39.21) (Table 7).261.3.2. Variances and heritahilitiesPartitioned variances for populations, families within all populations,and error for each seed trait are also presented in Table 7. The variances of seed germination were6 %, 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 withinfamilies, respectively. In a similar trend to the variance of germination and R50, seed weight varianceswere 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 atfamily level were high. This is could he the preconditioning maternal environmental effects confoundedwithin the variance components at the families. However, these heritahilities at least revealed thehighest possible parameters for genetic inheritance of the seed traits. If preconditioning maternalenvironmental effects can be controlled or made far uniformity, seed trait heritabilities will bedecreased.1.3.3. CorrelationThe general formula given by Becker (1984) was used for calculating correlation betweentraits. The formula is:CovXYrV’VarX V’VarYWhere r = phenotypic or genetic correlationCovXY = Covariance of phenotypic measures or genetic estimates oftrait X and YVarX and VarY = Variances of phenotypic measures or genetic estimates oftrait X and Y27Coefficients of phenotypic and genetic correlations between seed traits were weak for bothphenotypic and genetic values.1.4. DiscussionSeed quality, as characterized by seed germination percent, R50 and seed weight was foundto 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, thevariations 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 partiallygenetically controlled (Bishir and Namkoong, 1987; Thompson, 1984), but maternal environmentaleffects can have a strong influence (St.Clair and Adams, 1991; Brarnlet et at., 1983). However, thisvariation is useful information in nursery management for producing healthy seedlings that are uniformin size for outplanting. Moreover, seed trait variation can carry over and influence early growth inforest tree progenies (Thompson, 1984; Wilcox, 1983). Therefore, the analysis of seed traitsinvestigated in this study is useful for preliminary identification of seed sources for forestationprogrammes 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 ensureadaptation to local environment. Selection for fast and uniform germinating families should becautiously considered to maintain broad breeding populations. Spatial and temporal variation shouldalso be considered and subject of further study.Table7Germination,R50andseedweightvariation,variancecomponentandfamilyheritabilitiesforseedfromeightpopulationsSourceDFMSF-RatioVar.Comp%Var.CompheritabilityGerminationPopulation71414.831.5615.256.00-Fam(Pop)61909.3830.30**219.8483.000.88Error20730.02-30.0011.00-R50Population775.523.86**1.6924.00-Farn(Pop)6119.5427.0l**4.7066.000.87Error2070.72-0.7210.00-SeedweightPopulation787.231.500.885.000.91Fam(Pop)6158.0039.2l**14.1386.000.90Error2071.48-1.489.00-*Significantdifferenceatthe0.05confidencelevel**Significantdifferenceatthe0.01confidencelevel0029Seed variation provides general genetic information of the species, though environmental preconditioning 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 probablybiased (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-sibfamilies 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. Clairand Adams (1991) found that seed weight was weakly related to emergence percentage and rate. Theabsence of a significant correlation between seed weight and rate of emergence has also been reportedin other conifer species (Mikola, 1984; Perry and Hafley, 1981).2. Seedling Trait Variation2.1. Study methodsFifty-six open-pollinated families of seeds, collected from eight populations, the same as thoseused for the seed trait study, were grown in the nursery to estimate seedling trait variation. Theassumption is made that the pattern of growth of nursery seedlings is significantly related to observedfamily variation in the field. If nursery seedling performance is reliably related to field growth, theimpact on genetic management strategies would he significant (Carter et at., 1990), because nurserystudies are more convenient to control, environments are simpler, they require a shorter period of timethan 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 modelwas as follows.30Yikl = + B + P + Fk) + PB + FBk(J) + ek1ith . th •th •thWhere Ykl = observed value of i seedling of family within J population, in 1block= overall meanB block effectP = population effectFk(j) = family effectPB = interaction between population and blockFBjj) = interaction between family and blockekl = experimental error.The structure of the Analysis of Variance table including expected mean squares (EMS) isshown in Table 8.Table 8 Analysis of Variance for seedling traitsSOURCE DF EMSBlock (b-i) VE + flVFB + nfVPB + nfpV8Population (p-i) V1 + flVFB + nbVF + nfV + nbfVPop*Block (p-l)(b-l) VE + nVFB + nbVF + nfVFam (Pop) p(f-1) VE + nVFB + nbVFFam*Block p(f-l)(b-l) V + nVFBError bpf(n-1) VEWhere VB = block varianceVp = population variance31VF = family variance= variance of population and block interactionVFB = variance of family and block interactionVE = error varianceb = number of blocks (= 4)p = number of populations (= 8)f = average number of families per population (= 56/8)n = number of seedlings per plot or experimental unit (= 6).Seeds were germinated in soil seed beds for one week, until about 2 cm high, and thentransplanted 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 conductedevery day except on rainy days.Seedling height and diameter were measured at 9 months of age. After the 9 monthmeasurement, one seedling from each family in each block was randomly selected for dry weightdetermination. Since only one seedling from each block was used, block variance were then used asexperimental error for F-test. Statistical design for the dry weight variation study was same as thatused for study seed traits variation. Number of families (f), however, was changed to 56/8.2.2. Data analysisThe SAS Computer Program version 6.04 (SAS, 1991) was used for the analysis of varianceof 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 senseheritabilities (Falconer, 1981) at the family level were calculated as follows:32(a). IndividualV + VFB + Vf.(p)V(b). FamilyV1. + V1r + Vjr(p)14.96 42.3. ResultsSeedling traits studied included height (HT), diameter (DM)), total dry weight (TDW) androot-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 highestin undisturbed MLI, except mean of RS. However, seedling trait variation between undisturbed anddisturbed populations was not significant.The initial analyses of all data for seedling height and diameter produced large variation withinplots or error terms and the cause for the large error terms was not exactly known, perhaps due toseedlings within families were from different paternal pollen or inconsistent environmental conditionwithin blocks, but they would mask and hamper the detection of variance components at the family andpopulation levels. Schlichting (1986) indicated that variance estimates would he more precise if thisamount of within-plot genetic variation was controlled (minimized with clones). Thus, the data whichhad values departing from their family means by more than I standard deviation (is) were deletedbefore the variance analysis to decrease within plot variation. The variance components at family andpopulation levels were obviously detected by the latter method. The deletion of one standard deviation33criterion was then used for the seedling variation study. The averaged numbers of seedlings withinplots were decreased to Variation(a). HeightMLI population provided the highest mean for HT = 63.69 cm; SKI population had thelowest mean for HT = 38.34 cm (Table 9). Table 10 presents the resulting analysis of variance forheight (HT) and diameter (DM) of seedlings. HT was significantly different both among thepopulations and among the families within all populations (F-ratios 17.79 and 3.76, respectively).(h), DiameterDiameter 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 (F4.00) but not among populations (F = 1.17) (Table 10).(c). Dry weight and root-shoot ratioTotal dry weight (TDW) mean was highest in MLI = 17.07 gm and lowest in SKi = 11.94gm (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 shownin Table 11. TDW was significantly different among families within all populations but not amongpopulations. On the other hand, RS was significantly different both among populations and amongfamilies within populations.34Table 9 Average measurements flr seedling traits including height and diameter (HT, andDM), total dry weight (TDW), and root/shoot dry weight ratio (RS) for eachpopulationPopulation HT DM TDW RS(cm) (cm) (grn)UndisturbedML1 63.69 0.64 17.07 0.78ML2 59.28 0.62 16.47 0.75Ave. 61.49 0.63 16.77 0.77DisturbedKKI 49.31 0.60 15.65 0.82KK2 52.1 I 0.62 15.45 0.95MKI 46.08 059 16.55 1.31MK2 40.58 0.56 13.63 1.19SKI 38.34 0.57 11.94 1.485K2 50.69 0.60 15.01 1.00Ave. 46.19 0.59 14.71 1.13Overall Ave. 53.31 0.61 15.62 0.962.3.2. Variances and heritahilitiesThe variances of f-IT for among populations and among families within all populations were72.02 (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 higherthan EMS of population. The variance among ftimilies within populations was 0.0018 (13.80%) (Table10). The DM variance among populations was negative hut very small, -0.0001 (-0.96%), this wasunrealistic and presumed no variance. The variance among families within populations was 0.001835(13.80%) (Table 10).The dry weight variances for among populations and among families within all populationswere: 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. CorrelationsThere were a relatively high phenotypic correlation between height and diameter (0.63). TheTDW was negatively correlated with RS (-0.15).2.4. DiscussionSeedling growth in a nursery or controlled environment has become an effective means forinvestigating variation of a tree species and nursery growth often shows significant correlations withgrowth in a field trial (Jiang et a!.. 1989). Seedling performance at an early age in a nursery can heused 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 12months (Pharis et a!., 1991). In this study, height growth variation was significant both among familiesand populations. Diameter growth variation for seedlings at both ages was significant at the familylevel. However, variation in diameter growth of nursery seedlings was not different (p > 0.05) at thepopulation 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 variationwas ambiguous and must he carefully interpreted. However, Lambeth el al. (1983) proposed total dryweight as an early testing trait in Douglas-fir.Table10Heightanddiametervariation,variancecomponentandheritabilityfor9-monthnurseryseedlingsfromeightpopulationsSourceDFMSF-RatioVar.Comp.%Var.Comp.Heritability(S.E.)IndividualFamilyHeightBlock31291.363.09**3.471.42--Population78336.2717.79**72.0229.45--Block*Pop21468.600.528.673.53--Fam(Pop)48895.663.76**43.8017.911.090.73Block*Farn(Pop)141238.463.30**44.4918.19--Error61772.18-72.1829.51--DiameterBlock30.071.57*-0.0000-0.21--Population70.071.17-0.0001-0.96--Block*Pop210.061.500.001612.86--Fam(Pop)480.044.00**0.001813.800.630.64Block*Fam(Pop)1410.011.000.001814.28--Error6170.01-0.007760.23--*Significantdifferenceatthe0.05contidencelevel**Significantdifferenceatthe0.01confidencelevelC’TableIiDryweight variation,variancecomponentandheritabilityfor9-monthnurseryseedlingsfromeightpopulationsSourceDFMSF-RatioVar.Comp.%Var.Comp.Heritability(S.E.)IndividualFamilyTotaldryweightPopulation778.021.951.426.00--Fam(Pop)4840.072.67**6.4529.000.640.64Error16614.52-14.5265.00--Root—shootratioPopulation71.92l1.69**0.0746.00--Fam(Pop)480.163.51**0.0321.000.720.72Error1660.05-0.0533.00--*Significantdifferenceatthe0.05confidencelevel**Significantdifferenceatthe0.01confidencelevel38Although selection based on early variation is less accurate, it may be more efficient in termsof 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 whichcan markedly decrease reliability of their measurements (Rehfèldt, 1983), and nursery trials can be usedto 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 wasanalyzed by nesting the populations within their disturbance levels. Another reason for the largeraverage measurements for the undisturbed populations (ML] and ML2) may be that these populationswere located closest to the nursery site, and the seedlings were likely to be more suited to the growingenvironment. For urgent reforestation, seedlings from local populations are always recommended, untilwe know more about 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 issometimes encountered in statistical analysis of juvenile trees (Rink, 1984; Hicks ci at.., 1977; Adamsand Joly, 1977). 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, butquite high at the family le’vel. Early selection using diameter may not he appropriate due to smalldifferences in the trait. The individual and family heritabilities for dry weight were consistent andrelatively high at the family level, and these might he promising parameters for genetic studies of thespecies.Phenotypic and genetic correlations between traits were mainly parallel, so phenotypic selectionfor genetic improvement of traits is thereftre advantageous.393. Field Progeny Trait Variation3.1. Study methodsSeed of 85 open-pollinated families was planted at Khao-Yai Field Station as described earlierin the section on study materials. A nested block design was used for the progeny test utilizing fourblocks. The statistical model, source of variation (SV), degrees of freedom (DF), and expected meansquare (EMS) were the same as those used tor studying the seedling trait variation (Table 9), butnumbers of populations and families were changed to seven and 85/7, respectively. Twenty-fiveseedlings of each flimily were planted in each block. Height and diameter growth were measured from10 saplings at 4-years old. The persistence of axis, stem straightness, and branch heaviness were alsoscored using the applied method of Keicling et at. (1984) (Appendix D).3.2. Data analysisThe SAS Computer Program version 6.04 (SAS, 1991) was used for variance analysis of fieldprogeny traits. Variance Components for each progeny trait were calculated from EMS (Wright, 1976)and used for estimating narrow sense heritability (Falconer,1981).3.3 ResultsAverage measurements for flimily and oulation level of field progeny traits at four yearsheight (NT) and diameter (DM) are presented in Appendix E. Also, the scored trait means arepresented 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 fromfamily means by more than one standard deviation were deleted to derive more precise results of the40variance components at the family and population levels (Schlichtin, 1986). The averaged numbersof trees per lo was then 6.84. Table 14 and 15 present the results for the analysis of variance andgenetic parameters for I—IT and DM, and for the scored traits of the progeny, respectively.3.3. 1 . Variation(a). HeightOverall mean for NT was 3.48 m (Table 12). The HT was largest in the undisturbedpopulation 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 thedisturbed populations, analyses of variance thr height indicated nonsignificant differences between theundisturbed and disturbed populations. The HT (Table 13) was significantly different at both levels,hut the differences among families within all populations were small and significant at only the 0.05confidence level (F-ratio = 1.30). There were also highly significant differences for the interactionbetween blocks and families within all populations for height. This means height traits from differentfamilies responded differently to environmental conditions in blocks.(b). DiameterOverall mean for DM was 5.95 cm (Table 12). Mean DM was largest in the undisturbedpopulation 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 anddisturbed 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 significantdifference among families.41Table 12 Average measurements of progeny traits including height and diameter (HT and DM),persistence of axis (PER), stern straightness (STN), and branch heaviness (BR) foreach populationPopulation 1—IT DM PER STN BR(cm) (cm)UndisturbedDL 369.20 6.39 2.32 2.54 2.79ML 350.40 5.72 2.39 2.57 2.86SKH 370.45 5.99 2.49 2.60 2.95Ave. 363.35 6.03 2.40 2.57 2.87DisturbedKR 350.93 5.73 2.44 2.52 2.87MS 312.88 5.77 2.25 2.39 2.76SM 346.53 6.11 2.34 2.46 2.88SK 355.56 5.85 2.42 2.51 2.89Ave. 341.48 5.87 2.36 2.47 2.85Overall Ave. 348.23 5.95 2.35 2.50 2.83(c). Scored TraitsOverall 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 disturbedpopulation MS (2.25). However, the difference between the undisturbed and disturbed populations wasnonsignificant. There was a highly significant difference among all populations (F = 6.55), butnonsignificant difference among families within all populations (F = 1.13) (Table 14).STN means for each population also varied only slightly from each other. Population SKHproduced the largest mean (2.59) for STN and the smallest (2.38) from the population MS (Table 12).42There was no significant difference between the undisturbed and disturbed populations. Varianceanalysis (Table 14) indicated a significant difference among all studied populations, but no significantdifference 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). Therewas no difference between the undisturbed and disturbed populations. Variation was highly significantamong populations, hut nonsignificant among families within all populations.3.3.2. Variances and heritahilitiesThe 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 relativelyhigh for height at all ages.The diameter variances varied substantially; however, among population variances played asignificant role. Whereas, 3.83% of the total variance was accounted for by the difference among thepopulations, only 0.63% was accounted for by the difference among families.For the scored traits, variances were very small both among populations and among familieswithin all populations. The variances among trees within plots were considerably larger.The height individual heritability was small (19%). The height family heritability was alsosmall (21 %). For diameter, both individual and ftimily heritahilities were very close to zero (0.03 and0.05, respectively). Family heritahilities for the scored traits were moderate for PERST and BR, butsmall for STN. All individual heritahilities for scored traits were small.Table13Heightanddiametervariation,variancecomponentandheritabilityfor4-yearfieldprogenyfromsevenpopulationsSourceDFMSF-RatioVar. Comp.%Var.Comp.Heritability(S.E.)IndividualFamilyHeightBlock36.97302.50**0.00852.02--Population616.54328.43**0.044110.50--Block*Pop181.96170.890.00250.60--Fam(Pop)782.23401.30*0.01724.100.190.21Block*Fam(Pop)2341.722315.59**0.237356.46--Error19870.1105-0.110526.30--DiameterBlock331.49372.67**0.04242.05--Population632.65154•73**0.07933.83--Block*Pop186,90820.89-0.0061-0.30--Fam(Pop)787.77151.070.01310.630.030.05Block*Fam(Pop)2347.25617.08**0.917744.32--Error19871.0243-1.024349.47--*Significantdifferenceatthe0.05confidencelevel**Significantdifferenceatthe0.01confidencelevelTable14Scoredtraitvariation,variancecomponentandheritabilityfor4-yearfieldprogenyfromsevenpopulationsSourceDFMSF-RatioVar.Comp.%Var.Comp.Heritability(S.E.)IndividualFamilyPersistenceofaxisBlock32.48313•39**0.00351.12--Population62.93976.55**0.00762.45--Block*Pop180.44890.930.00020.07--Fam(Pop)780.48421.130.00200.630.030.11Block*Fam(Pop)2340.4268l.55**0.02227.12--Error19870.2761-0.276188.61--StemstraightnessBlock31.3552l.590.00100.37-Population61.99312.83*0.00391.44--Block*Pop180.70311.39*0.00291.05--Pam(Pop)780.50621.080.00110.420.020.06Block*Fam(Pop)2340.46892.04**0.035312.89--Error19870.2294-0.229483.84--BranchingcharacteristicBlock312.037318.39**0.020011.18-Population61.3151355*0.00271.49--Block*Pop180.37000.930.00070.41--Pam(Pop)780.39931.30*0.00321.800.080.22Block*Fam(Pop)2340.30642.43**0.026614.83--Error19870.1260-0.126070.29--*Significantdifferenceatthe0.05confidencelevel**Significantdifferenceatthe0.01confidencelevel453.3.3. CorrelationThe 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, thegenetic correlations were quite high. The phenotypic and genotypic correlations between DM andscored traits were inconsistent. The phenotypic correlations between scored traits were moderate tosmall. There was no genetic correlation between PER and stem straightness (STN), moderate betweenPER and hranchin heaviness (BR), and a neative correlation between STN and BR.3.4. DiscussionField progeny height was the only trait that varied significantly both among populations andfamilies. The diameter trait was not significantly different among families. For the other traits, therewere very small differences at the levels of population and family. The majority of progeny traitvariation was accounted for by differences within families and by environmental differences. Theability of an individual family to alter its morphological response to changes in environmental conditionsis 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 populationsthan those for disturbed populations. On the other hand, the average diameter growth was larger fordisturbed populations than for undisturbed populations. The local populations (undisturbed ML anddisturbed KR) were not consistently different from non-local populations. This was probably becauseof no pronounced difference in environment between the two regions sampled. As a result of thesignificant differences in progeny traits for among populations, conservation should aim at includinghigh levels of variation of these quantitative traits which will provide an opportunity to species to46adaptation to changing environments.Interactions between family and block were significant for all progeny traits. This phenotypicplasticity of the traits should be subject to future study.Individual heritability for height was 0.43. Individual heritability for diameter growth was0.16. 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 reachesmaturity. Due to the very small individual heritability estimates for all scored traits, individualselection 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 alltraits.The phenotypic and genetic correlations between traits were not parallel; this is due to the traitsresponded differently to environments. The phenotypic correlation between DM and STN wasmoderate, but the genetic correlation was zero. Conversely, the phenotypic correlation between HTand each scored trait was very weak, hut the genetic correlation was moderate to high. Thereforeselection from phenotypic traits of progeny may not he appropriate for genetic improvement.4. ConclusionThe study of variation in quantitative traits revealed different patterns of variation among traits.Seed variation mainly existed among families. R50 was the only seed trait which had significancedifferences among populations. Nursery seedling and field progeny traits produced significantdifferences both among populations and among flimilies, but no significant (Iifference was observedamong populations for total dry weight. The significant effects of interactions between family and blockwould be caused by phenotypic plasticity of the progeny traits. Variation in quantitative traits was notsignificant between undisturbed and disturbed populations.Family heritability estimates for seed traits were high. This may have been confounded by47environmental 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, seedlingand progeny traits showed comparable patterns of genetic variation. These adaptive traits should,therefore, be considered in a sampling strategy for conservation.48CHAPTER 3ISOZYME VARIATIONThe isozyme technique is a quick and inexpensive means for the investigation of geneticvariation of forest tree species. The study of both isozyme and other quantitative traits provides usefulinformation for genetic management as well as conservation strategies. However, the study of isozymevariation of tropical forest tree species has only a short history. In Thailand, isozyme studies on foresttree species have just been carried out fhr the last few years by the Royal Forest Department and theASEAN-Canada Forest Tree Seed Centre. The studies are mainly in the infancy stage and need tobe extended to derive valuable information on genetic structure of forest tree populations.1. Study MethodFifty-seven of the open-pollinated families were used for the isozyme study. Seed wasgerminated in petri dishes for three days. The root tips or radicles, which had emerged to about a halfof seed length were used for enzyme extraction.The isozyme procedures followed were those described by Liengsiri er al. (1990). Insummary, embryonic tissue from the radicle was separated from the germinating seed and ground in0.04 ml of extraction buffer (Appendix G) using a rotating teflon grinding head. The resultinghomogenate was absorbed onto filter paper (Whatman No. 3) wicks, approximately 14 x 1 mm indimension. 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 wicksvertically along the resulting edge, and then pushing the two slabs together.49The 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 werethen placed on the gel surface about 1.5 cm from both electrode sides. The stand was then placed ina refrigerator for electrophoresis.Two buffer systems, H and B buffer systems (Appendix H), were used for gel preparation andelectrode finning. The applied electric currents for the buffer systems H and B were 90 amperes and30 amperes, respectively. Initially 50 % of the final applied current, 45 amperes for the H buffersystem and 15 amperes for the B buffer system, was used for 10 to 30 minutes until a coloured dyemarker migrated about 0.5 cm beyond the origin. At this stage all sample wicks were removed fromthe gel. The electric current was then re-applied to the system at the final wnning level. Once theenzyme had migrated a sufficient distance through the gel, as indicated by the progress of the coloureddye (5 cm), the gel was sliced into 1 mm thick slices, which were then stained for enzymes shown inAppendix I.Gel slices from buffer system H were used for enzyme systems aldolase (ALD), isocitratedehydrogenase (IDH), malic dehydrogenase (MDH), malic enzyme (ME), phosphoglucomutase (PGM),shikimic acid dehydrogenase (SDH) and 6-phosphogluconate dehydrogenase (6PGD). Enzyme systemsaspartate 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 allelesat each locus.2. Data AnalysisGenetic variability and mating system parameters of the tree species were estimated as follows.502.1. Allelic variationThe Biosys-1 Computer Program version 1.7 (Swofford and Selander, 1989) was used foranalysis of the following parameters.(a). Allele frequenciesThe frequencies of alleles occurring at each locus are estimated for individual populations andloci.(b). HeterozygosityObserved heterozygosity (Ho) is the proportion of all genotypes that are heterozygotes.Expected heterozygosity (He) = (1-Ep2)Where p = frequency of jth allele(c). Number of alleles per locusThis parameter was the numbers of alleles recorded at each locus. The average number ofalleles for the overall population were also calculated.(d). Percentage of loci polymorphicPercentages of loci possessing more than one allele in individual populations were calculatedto represent another aspect of variation.51(e). Deviation from Hardy-Weinberg equilibriumA chi-square test for deviation of allelic frequencies from Hardy-Weinberg expectation will testthe null hypothesis that there are no biological phenomena or sampling biases with a net effect sufficientto cause significant differences between observed and expected proportions. Contingency chi-squaretests also provide information as to whether differences in genotype proportions among eightpopulations were no greater than among ten random samples drawn from the total population.(f). Genetic similarity and distance coefficientsGenetic distance provides a measure of the relationship among individual populations. Variousmeasures of genetic distance have been derived and the most frequently quoted measure is one derivedby Nei (1972):J = 1-HeI JxyIV’JxJyD = TnTWhere He = expected heterozygosityJ = genetic identityJxy = covariance of J for the population x and yIx = variance of I for population xJy = variance of J for population yI = normalized genetic identity for populations x and yD = genetic distance.Nei’s (1978) unbiased genetic distance and Nei’s (1978) unbiased identity were calculated inthis study.52(g). Genetic differentiation of populations byF-statistics (Wright, 1965, 1978)F-statistics (Wright, 1965) is a set of statistics which indicates the relative distribution ofgenetic variation within and among populations. The formulae are:F15 = 1-Ho/HeF = Vp/pqF1- = F15 + (1-F15)FWhere F15 = the mean deviation of genetic proportions from Hardy - Weinbergexpectation for each population.F = the correlation between random gametes within a population withgametes in the total of all populationsFIT = the correlation among gametes for the total of all populationsVp = variance of allele frequencies in a population from frequencies overall populations.p & q = allele frequencies.G-statistics (Nei, 1973) is an extension of the concept of F-statistics for a locus possessingmore than two alleles:HT = H5+D9-G5 = DST/HTWhere HT = gene diversity (expected heterozygosity) for total of all populations.H = average of the individual population gene diversities.= average gene diversity among populations= genetic difference among populations, equivalent to Wright’s F5953(h). Cluster analysisUnbiased Nei’s genetic distances were used to conduct a cluster analysis using the unweighedpair groups method algorithm (UPGMA) (Sneath and Sokal, 1973).2.2. Mating systemThe Generalized Multilocus Estimation Programme or MLTF (Ritland, 1986) was applied forthe mating system analysis of the species. Only fifteen loci with high variability were selected for themating 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 anddisturbed populations were estimated.3. ResultsReproducible banding patterns were observed from twelve enzyme systems coding for eighteenloci. Seven enzyme systems produced more than one locus. The loci included 6-phosphogluconatedehydrogenase 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), isocitratedehydrogenase (IDH), leucine-arnino peptidase (LAP), malic dehydrogenase (MDH), malic enzyme 1and 2 (MEl and ME2), phosphoglucose isomerase I and 2 (PGJ1 and PGI2), phosphoglucomutase 2(PGM2), and shikimic acid dehydrogenase 2 (SDH2).54All enzyme banding patterns are displayed in Figure 2. An example of electrophoretic gelslices presents in Figure 3. Allele variability parameters for individual loci and tree populationsprovided valuable details on the genetic population structure of the studied species.3.1. Allelic variationThe genetic variability parameters are shown in Table 15. The mean zygotic sample size perlocus for each population ranged from 60.0 (population SK2) to 324.2 (population ML1). The overallmean zygotic sample size averaged across all eight populations was 135.8, and the overall mean samplesizes averaged across two undisturbed populations and six disturbed populations were 268.8 and 91.5,respectively.(a). Allelic frequenciesThe 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. Althoughfive alleles were scored at locus 6PGD1, only four frequency means were derived because the fifthallele was observed in just the population MK2 with very low frequency (0.007) and its frequency meanwas 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 thealleles were found only in a few populations investigated.Rf 6PGD Rf AIDRf IDH Rf MDH.379.364 .372I 25 2.346.333.318 3I .288 1I 1 .273I 3 I4 IRf ME Rf PGMI .394 MEl PGM2I .364 .36821 IME2 .338 3I .166I .152 1 .324I 2f 1 .136 2I 3556PGD1.385.372 5.333.321 41 .29532ALD3.423.410 31 .39826PGD2 .288.287 52 .282.231 31 .2184ALD4.192.166 2.15413Figure 2 Enzyme banding patterns for eighteen loci. Band ( ) indicates allele(s), threefigures over a band (e g .345) indicate relative migrating distance (Rf) of the bandor allele, and a figure under a band indicates the allele number in that locus.Figure 2 (continued)56.225.200 2.18813[ST.727 ESTI.7122.667 31.652.683.667 3.65012Rf AATI AAT21 .273 EST32.1823.122.108 3.095 21.424 PGI11 .409Rf.5002 .4803Rf SDH Rf D1ARf4.470EST2.364 21.333PGI LAP2RfI—..288.2733 PGI21.2432.4001 .386457Figure 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 weredifferent)Table15Variabilityparametersincludingmeannumberofallelesperlocus,percentageoflocipolymorphic,andmeanheterozygositiesfordirect-countandHardy-Weinbergexpectationat18lociforthespecieslevel(datafromallpopulationswerecombinedandtreatedasoneunitforanalysis),eachpopulation,andthemeansaveragedateachlevel(SEinparentheses)MeanheterozygosityMeansampleMeanno.PercentagesizeperofallelesoflociDirect-HdyWbgPopulationLocusperlocuspolymorphic*countexpt**OVERALLPOPULATION(species)UNDISTURBEDPOPULATION1.ML12.ML2MEAN1085.9(17.8)324.2(5.2)213.3(4.6)268.8(4.9)3.4(0.2)97(0.2)2.4(0.1)2.3(0.2)88.978.894.486.10.289(0.056)0.320(0.075)0.346(0.063)0.333(0.069)0.266(0.042)0.244(0.046)0.300(0.045)0.272(0.046)00Table15(continued)MeanheterozygosityMeansampleMeanno.PercentagesizeperofallelesoflociDirect-HdyWbgPopulationLocusperlocuspolymorphic*countexpt**DISTURBEDPOPULATION3.KKI105.31.877.80.2890.221(4.2)(0.1)(0.061)(0.042)4.KK292.21.761.10.2030.157(5.4)(0.2)(0.063)(0.042)5.MKI74.*Alocusisconsideredpolymorphicifthefrequencyofthemostcommonalleledoesnotexceed0.99**Unbiasedestimate(seeNei,1978)60Five alleles were recorded at locus 6PGD2. The frequency means for alleles one, two, andthree were 0.703, 0.263, and 0.035, respectively. The frequency means of alleles four and five werequite low (0.005 and 0.001, respectively) and the alleles were observed with low frequencies in twoand one populations, respectively.AAT displayed at least two loci but only locus AAT2 was consistent and scorable. Threealleles 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 werethree alleles observed at locus ALD3. The frequency means for alleles one, two, and three were 0.785,0.164, and 0.051, respectively.Table 16 Mean allele frequencies for 18 loci across 8 populationsLocus No Allele1 2 36PGD1 141 0.89 0.09 0.00 -6PGD2 142 0.70 0.26 0.04 -AAT2 105 0.53 0.29 0.18 -ALD3 125 0.79 0.16 0.05 -ALD4 128 0.98 0.01 0.01 -DIA 143 0.98 0.02 - -EST1 137 0.63 0.34 0.02 0.01EST2 135 0.76 0.20 0.04 -EST3 135 0.81 0.16 0.03 -IDH 143 0.74 0.15 0.03 -*LAP 142 0.72 0.26 0.02 -MDH 133 0.90 0.09 0.01 -MEl 143 0.98 0.02 -ME2 143 0.91 0.08 0.01 -PGI1 133 0.99 0.01 - -PGI2 135 0.94 0.00 - -PGM2 143 0.91 0.06 0.03SDH2 142 0.82 0.12 0.06 -* 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 zero4 50.0861Locus ALD4 also had three alleles. The frequency mean for allele one was high (0.98), butquite 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 wererecorded. The frequency means for allele one and two were 0.98 and 0.02. The frequency mean forallele three approached zero (0.001)EST was scored for three loci (EST1, EST2, and EST3). Four alleles were recorded at thefirst 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 IDHhad five alleles. The frequency mean was low for allele four (0.004). The frequency means for allelesone, 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 allelesone, two, and three were 0.718, 0.261, and 0.019, respectively. The frequency mean for allele fourwas 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, andthree, respectively.ME displayed two loci (MEl and ME2). Locus MEl provided two alleles with the frequencymeans 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 of0.92 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,62PGI1. 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 beingrecorded. 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 frequencymeans 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 andtwo 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). HeterozygosityMean heterozygosities for all studied levels, including species, undisturbed populations,disturbed populations, and individual populations, are presented in Table 15. The meanheterozygosities for direct-count were higher than those for unbiased estimates at all levels. The meanheterozygosities at the species level or total population (eight populations were treated as one unit forthe 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 0.273and 0.229 for direct-count and unbiased estimate, respectively. Mean heterozygosities for undisturbedand disturbed populations were calculated by averaging from the mean heterozygosities of populationswithin each disturbance level correspondingly.The mean heterozygosities for individual loci at each population and their averages across eightpopulations are presented in Appendix I. Mean heterozygosities based on combined data of eightpopulations for each enzyme locus are also given in Table 18. The EST1 locus had the largest direct-count and unbiased heterozygosities (0.714 and 0.479, respectively). The PGI1 locus gave the smallestheterozygosities (0.015 for both unbiased and direct-count). Loci 6PGDI, ALD4, 1DB, MEl, PGM2,63and 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 ondata combined from eight populationsLocus HeterozygosityDirect-count HdyWbg expt6PGDI 0.156’ 0.1896PGD2 0.565 0.420AAT2 0.634 0.440ALD3 0.423 0.329ALD4 0.033* 0.039DIA 0.028* 0.034EST1 0.714 0.479EST2 0.486 0.354EST3 0.307 0.277IDH 0.297* 0.377LAP 0.429 0.321MDH 0.194 0.164MEl 0.030* 0.038ME2 0.155 0.145PGI1 0.015 0.015PGI2 0.122 0.111PGM2 0.079* 0.140SDH2 0.288 0.246* Direct-count smaller than unbiased expectation(c). Number of alleles per locusNumber of alleles per locus are also displayed in Table 15. At the species level, meannumber of alleles per locus was 3.4. Mean number of alleles per locus was largest in the SK2population (2.5) and smallest in the population KK2 (1.7). The averaged number of alleles per locusacross 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 disturbedpopulations (averaged from six disturbed populations) was 2.1.64(d). Percentage of loci polymorphicA locus was considered to be polymorphic if more than one allele with frequency 0.01 or morewas detected. The percentage of polymorphic loci for each population ranged from 55.6 (populationMK1 and MK2) to 94.4 (population ML2) (Table 15). The percentages of polymorphic loci averagedacross 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 thanthose of disturbed populations. The percentage of polymorphic loci for the species level was 88.9.(e). Deviation from Hardy-Weinberg equilibriumChi-square tests for Hardy-Weinberg equilibrium provided different results for each locus inindividual populations. Table 18 presents the chi-square (X2) value and significance probability for eachlocus in the individual populations. The test was valid at a particular locus in a certain populationbecause the test will be performed only when every progeny (radicles) from every family in thatpopulation had been scored. The smallest population in the study consisted 3 families; the progeny usedfor scoring then 60 (20 from each families); therefore, every locus which was tested for deviation fromHardy-Weinberg equilibrium was at least scored from 60 progeny.However, at the species level (eight populations treated as one total population), only sevenloci 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 offreedom, which was highly significant.65Table 18 Chi-square test for deviation from Hardy-Weinberg equilibrium of individual loci ineach population.LOCI KK1 KK2 MLI ML2 MKI MK2 SKi SK26PGD1X2AP6PGD2X2APAAT2X2APALD3X2APALD4X2APDIAX2APEST 1X2APEST2X2APEST3X2APIDHX2AP0.96 1.17 0.002 2 20.33 0.28 0.9716.35 33.96 65.242 2 20.00 0.00 0.00-- 59.92-- 7-- 0.0019.80 0.09 45.813 2 20.00 0.77 0.00-- 23.99-- 7-- 0.0021.4920.0010.5120.003.7720.0536.3270.0072.60 82.354 40.00 0.0015.03 21.214 50.02 0.0264.35 88.543 30.00 0.001.06 8.743 30.79 0.030.00 -7 -1.00 -16.7030.004.4570.0431.8840.000.8130.8510.6630.0159.0030.002.8730.411.8570.172.3670.1221.0530.008.3330.045.4620.0222.6330.003.04 0.002 20.08 1.0023.64 0.253 30.00 0.97101.56 6.703 30.00 0.08100.90 5.633 30.00 0.133.25 0.003 20.36 1.00-- 0.00 --- 2 --- 0.96 -36.76 287.26 79.18 6.262 3 3 40.00 0.00 0.00 0.4032.52 218.87 7.75 0.253 3 2 20.00 0.00 0.01 0.621.31 12.74 0.93 0.032 3 2 20.25 0.01 0.33 0.866.93 0.05 4.99 27.212 2 2 50.01 0.84 0.03 0.0014.9120.006.1530.101.4220.2347.8930.0066Table 18 (continued)LAPX2 2.03 - 36.81A 2 - 2P 0.15 - 0.00MDHX2 14.33 0.01 6.60A 2 2 2P 0.00 0.94 0.01MEl - - 8.28A - - 2P - - 0.00ME2X2 4.12 0.81 0.82A 2 2 3P 0.04 0.37 0.85PGI 1X2 0.11 -A 2 - -P 0.74 - -PGI2X2 4.66 1.04 16.96A 2 2 3P 0.03 0.31 0.00PGM2X2 0.07 - -A 2 - -P 0.80 - -SDH2X2 13.67 2.34 37.06A 2 3 3P 0.00 0.51 0.00X2 Chi-squareA = Number of alleles at a locusP = significance probability0.0220.880.0090.95-- 0.21-- 2-- 0.65- 0.07 134.24- 2 3- 0.8 0.00-- 0.00-- 9-- 0.95Finally, coefficients for heterozygote deficiency or excess, including the fixation index of eachgene 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 individualloci. 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 individual79.0090.000.3930.9416.8430.000.0290.8864.7430.000.1920.662.0930.558.0090.011.1530.7713.9590.001.2230.750.0570.821.0920.301.0190.3220.5730.000.8740.990.0190.931.1470.290.2020.6632.7940.004.2690.0467populations.Table 19 Contingency chi-square analysis for deviation of allele frequencies from those of whichwere randomly sampled from the total populationNo.ofLocus alleles Chi-square D.F. P6PGD1 5 314.032 28 .000006PGD2 5 321.491 28 .00000AAT2 3 700.223 14 .00000ALD3 3 236. 124 14 .00000ALD4 3 154.322 14 .00000DIA 3 353.214 14 .00000ESTI 4 350.495 21 .00000EST2 3 375.276 14 .00000EST3 3 536.853 14 .00000IDH 5 805.559 28 .00000LAP 4 720.349 21 .00000MDH 3 291.266 14 .00000MEl 2 103.005 7 .00000ME2 3 77.647 14 .00000PGII 2 34.850 7 .00001PGI2 3 68.058 14 .00000PGM2 4 721.486 21 .00000SDH2 3 561.588 14 .00000(Totals) 6725.837 301 .00000(f). Genetic similarity and distance coefficientsGenetic similarity and distance coefficients provide the information about magnitude or extentof differences among each pair of populations. Different models have been developed for estimationof genetic similarity and distance. Nei’s (1978) unbiased genetic distances and Nei’s (1978) unbiasedgenetic identities are shown in Table 21.32.5 21.5 10.5 0Noalleles/locusProb.polymorioci___I Obs.heterozygosityExp.heterozygosityFigure4Geneticvariabilityincludingnumberofallelesperlocus,probabilityofpolymorphicloci,andobservedandexpectedHardy-WeinbergheterozygositiesfromeighteenlociforeachpopulationKK1KK2ML1ML2MKiMK2SkiSK20069Table 20 Expected and observed heterozygosity, and resulting fixation index for each locus,based on the population. A negative fixation index represents a heterozygotedeficiencyHeterozygotes FixationLocus Expected Observed index (F)6PGD1 152 182.533 .1676PGD2 662 508.543 -.302AAT2 536 466.070 -.151ALD3 518 406.785 -.274ALD4 40 50.921 .214DIA 15 20.832 .280EST1 847 551.381 -.537EST2 598 447.594 -.337EST3 353 383.802 .080IDH 323 417.319 .226LAP 375 419.034 .105MDH 235 225.010 -.045MEl 70 90. 161 .223ME2 181 176.564 -.026PGI1 16 15.887 -.008PGI2 143 134.520 -.064PGM2 76 140.998 .461SDH2 410 488.527 .160Nei’s genetic distance coefficients were largest between the populations KKI and MKI0.082 and smallest populations MK2 and SKi = 0.009.70Table 21 Matrix of genetic distance and/or similarity coefficientsA. Nei (1978) unbiased genetic distancePopulation KK1 KK2 ML1 ML2 MK1 MK2 SKi SK21 DISTURBED-KK12 DISTURBED-KK2 .014 ***3 UNDISTURBED-ML1 .021 .018 *****4 UNDISTURBED-ML2 .035 .052 .024 ***5 DISTURBED-MK1 .082 .077 .067 .060 *****6 DISTURBED-MK2 .073 .057 .052 .058 .019 *****7 DISTURBED-SKi .083 .075 .067 .057 .014 .0098 DISTURBED-SK2 .057 .042 .050 .046 .046 .037 .038 *****B. Nei (1978) unbiased genetic identityPopulation KKI KK2 MLI ML2 MKI MK2 SKI SK21 DISTURBED-KKI2 DISTURBED-KK2 .986 **3 UNDISTURBED-ML1 .979 .982 *‘4 UNDISTURBED-ML2 .965 .950 .977 *****5 DISTURBED-MK1 .921 .926 .936 .941 **6 DISTURBED-MK2 .930 .944 .950 .944 .9817 DISTURBED-SKi .921 .928 .936 .944 .986 .991 *****8 DISTURBED-SK2 .945 .959 .952 .955 .955 .964 .963 *****(g). Genetic differentiation of populations by F-statistics (Wright, 1965, 1978)F-statistics describing the structure of variation within and among populations were calculatedfor individual alleles in each locus. Means of F(IS), F(IT), and F(ST) of each locus are presented inTable 22.The overall locus mean of F(IS) was -0.200. F(IS) was largest for locus PGM2 (0.423) andsmallest for locus EST1 (-0.498).The overall locus mean of F(IT) was -0.048. Similar to the F(IS) mean, F(IT) mean was71largest 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).Table 22 F-statistics at all loci from 8 populations (S.E. in parenthesis): F(IS) and F(IT) arethe fixation indices within populations and within total population, and F(ST) is thegenetic differentiation among populationsLocus F(IS) F(IT) F(ST)6PGD1 0.166 0.214 0.0576PGD2 -0.351 -0.274 0.057AAT2 -0.453 -0.049 0.278ALD3 -0.293 -0.194 0.077ALD4 0.166 0.197 0.038DIA 0.181 0.293 0.137EST1 -0.498 -0.450 0.032EST2 -0.381 -0.269 0.081EST3 -0.109 0.042 0.137IDH 0.208 0.296 0.111LAP -0.343 -0.033 0.231MDH -0.190 -0.044 0.122MEl 0.196 0.247 0.064ME2 -0.071 -0.036 0.033PGI1 -0.027 -0.008 0.019PGI2 -0.111 -0.062 0.044PGM2 0.432 0.524 0.162SDH2 -0.022 0.179 0.196Mean -0.200 -0.048 0.127(0.135)** Standard error for F(ST) mean averaged from 1 8 loci(h). Cluster AnalysisNei’s (1978) unbiased genetic distance was used for cluster analysis and construction of acluster tree. The results of the cluster analysis using the unweighed pair group method are shown inTable 23. The analysis provided seven clustering levels with 4 cycles of pairing. The cluster analysis72tree is shown in Figure 5.Table 23 Cluster analysis for Nei ‘5 (1978) unbiased genetic distance using unweighed pair groupmethod algorithm (UPGMA)Population or cluster Clustering Cyclenumbers joined levelKK1 KK2 .01394MK2 SKI .00893KKI ML1 .01986 2MKI MK2 .01634 2KK1 ML2 .03687 3MK1 SKi .04053 3KK1 MK1 .06272 4Subsequently, based on Wright (1978), hierarchical analysis was carried out by nesting thepopulations 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 foramong 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 combinedacross 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 theF(ST) in Table 28; they were 0.122 and 0. 127, respectively. The difference among populations wassignificant, but among regions was nonsignificant. Negative variance components resulted from theorthogonal 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 studiedlevels. Variability parameters differed among the populations. The parameter estimates averagedacross undisturbed populations were somewhat larger than those of disturbed populations. The number73of alleles per locus and percentage of polymorphic loci also varied among populations. The chi-squaretest for deviation from Hardy - Weinberg equilibrium was significant for half of the loci analyzed. Thecluster analysis divided populations into two main geographical groups, one from central and the otherfrom northeastern Thailand.Distance.10 .08 .07 .05 .03 .02 .00+----+----+----+----+----+----+----+----+----+----+----+----+DISTURBED- KK1DISTURBED-KK2UNDISTURBED-ML1UNDISTURBED-ML2DISTURBED -MK1DISTURBED-MK2DISTURBED-SKiDISTURBED - SK2.10 .08 .07 .05 .03 .02 .00Figure 5 Clustering tree using Nei’s (1978) unbiased distances3.2. Mating systemThe 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. AppendixK presents the outcrossing rates for individual families. Due to the poor convergence of estimates inmost families, the family outcrossing rates were overestimated (values more than 1.0) and standarderror estimates were invalid. The overall mean for population outcrossing rate was 0.971. The meansof outcrossing rates for undisturbed and disturbed populations were slightly and non-significantlydifferent, 0.980 and 0.968, respectively. The mean fixation index over all populations was 0.05 1. Themean fixation index of undisturbed populations was 0.005; disturbed populations had a larger fixation74index mean, 0.051.Outcrossing rates for most individual families were high and mainly greater than 1.0. Thisproblem 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 outcrossingrate for one family in the population KK1 was only 0.2, and 0.69 for another family in the populationMK1. Inferred maternal genotypes for each family are presented in Appendix L.Table 24 Hierarchical analysis using F-statistics (Wright, 1978). F-statistics for amongpopulations within regions (F(SR)), among populations within total populations(F(ST)), and among regions within total populations (F(RT)); variance component isin parenthesis (S.E. in parenthesis)Locus F(SR) F(ST) F(RT)6PGDI 0.063 0.052 -0.0116PGD2 0.050 0.053 0.003AAT2 0.157 0.272 0.136ALD3 0.051 0.072 0.021ALD4 0.039 0.033 -0.007DIA 0.147 0.130 -0.020EST1 0.020 0.027 0.007EST2 0.059 0.077 0.019EST3 0.163 0.133 -0.037IDH 0.085 0.106 0.023LAP 0.116 0.227 0.126MDH 0.088 0.119 0.033MEl 0.056 0.062 0.007ME2 0.026 0.029 0.002PGI1 0.016 0.015 -0.001PGI2 0.024 0.040 0.017PGM2 0.176 0.157 -0.023SDH2 0.098 0.194 0.107ALL LOCI 0.086 0.122 (0.135)* 0.040(0.002)** Standard error for F(ST) and F(RT) means averaged from 18 loci75Table 25 Outcrossing rate (t), and fixation index for each population estimated from fifteen loci(S.E. in parenthesis)Population t fUndisturbedML1 0.984(0.020) 0.004(0.100)ML2 0.976 (0.020) 0.005 (0.132)Ave. 0.980 (0.020) 0.005 (0.116)DisturbedKK1 0.972 (0.212) 0.293 (0.167)KK2 0.998 (0.142) 0.087 (0.266)MK1 0.928 (0.000) 0.009 (1.000)MK2 0.965 (0.000) 0.002 (1.000)SKi 0.956 (0.027) 0.001 (1.000)SK2 0.987 (0.127) 0.000 (1.000)AVE. 0.968 (0.085) 0.065 (0.739)Overall Ave. 0.971 (0.069) 0.051 (0.583)4. DiscussionA high amount of genetic variability exists at the species level of Dalbergia cochinchinensisPierre. Compared with average values for gymnosperms (Harnrick and Godt, 1989), mean number ofalleles per locus, percentage of loci polymorphic, and mean of expected heterozygosity of this treespecies are higher. Most loci at the species level were polymorphic; indeed, some of the initiallysurveyed 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 forconifers, 0.177 for long-lived woody perennials and 0. 137 for narrow ranging species (Hamrick and76Godt, 1989). The expected heterozygosity is comparable to other tropical tree species, which rangefrom 0.106 to 0.374 (Ledig, 1986a). However, the heterozygosity was higher than that averaged fortwenty-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 treeswas 0.211 (Hamrick and Loveless, 1989). Hamrick and Murawski (1991) also found meanheterozygosity for ten taxa of uncommon (less than five individuals per hectare) tropical trees to be0.142. 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 valuecalculated 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 ishigher than the expected heterozygosity in every population. The possible reasons for excess ofheterozygotes are heterozygous advantage, negative assortative mating and differences in allelefrequencies 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 inflower phenology. However, half of the loci showed heterozygote deficiency when analyzed acrosspopulations (or at the species level). Means of all genetic variability parameters of undisturbed anddisturbed populations were not different, because some disturbed populations had variability parametermore or less than undisturbed populations. The disturbed populations might be degraded in terms ofgenetic variability in some instances; however, variability in the disturbed populations remained highand valuable for future management.F-statistics indicated that mean deviation of genotype frequencies from Hardy-Weinbergexpectations 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 means77that almost 13 % of total allozyrne variability is accounted for by differences in allele frequencies amongpopulations. The proportions of genetic variability for eight tropical tree species have been reportedby Hamrick and Murawski (1991) to vary from 0.064 and 0.163. Ledig and Conkle (1983) reportedthree to twelve percent F(ST) values for conifers. This could be a result of limited gene flow and thedevelopment 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 centralarea group included the KK1, KK2, ML1, and ML2 populations and the northeast area group includedMK1, MK2, SKi, and SK2. However, the hierarchical analysis using Wright’s (1978) statisticsindicated no significant difference among the regions (F(RT) = 0.040). There were two interestinganomalies. First, the MLI and ML2 populations, which were very close in geographic distance, didnot link together in the first cycle, but in the third. Second, the SKi and SK2 populations were alsogeographically close but the two linked together in the third cycle instead of the first. The reason forthese 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 negativeassortative mating which may be a consequence of self-incompatibility, heterotic selection againstinbreeding 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 populationswith 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 havean 85% outcrossing rate. Although mating system in a plant species varies in space and time (Hamrick,781982), the results from this study indicated no significant differences among the tree populations.5. ConclusionFrom twelve enzyme systems producing eighteen loci, isozyrne analysis revealed high geneticvariability of Thailand rosewood (Daibergia cochinchinensis Pierre). The individual populationsshowed different degrees of variability. However, the differences between undisturbed and disturbedpopulations were not significant. Half of the loci had genotype frequencies deviating from HardyWeinberg expectation. Nevertheless, numbers of loci varied significantly among populations (from 2loci in population MK1 to 12 loci in population ML 1). Nei’s (1978) unbiased genetic distance and F-statistics (Wright, 1978) indicated moderate levels of differences among populations. Cluster analysisdivided populations into two geographic regions, central and northeast, but the regional difference wasnot significant. High outcrossing was found in all the populations with the outcrossing rate very closeto 1.0. The fixation indices for most populations were correspondingly very low.79CHAPTER 4CONSERVATION STRATEGIESA conservation programme for this tree species has been established by the Thai Royal ForestDepartment and Danish International Development Agency (Sa-ardavut et al., 1989). However, thenumbers of trees and locations have not been based on genetic information about the variation anddynamics of population structure. The genetic variability and mating system information derived fromthis study was therefore applied in support of the existing conservation programme, to improve itsscientific rationale.Due to the inconsistent amounts and patterns of variation among all traits studied, resolutionof 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 forconsidering strategies to revise the conservation programme for future implementation.1. Quantitative Trait ConsiderationsQuantitative traits are complex and under the control of environmental conditions. Variationin quantitative traits is of great importance in gene conservation because the conservation strategy isto 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 accountedfor 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 seedgermination, R50, and weight were 6.74, 26.67, and 5.49%, respectively. The general explanation for80high variation among families was that only one measure was used for each replication and this measurewas a mean or treated as a mean of the replication; therefore, this replicated mean was confounded withvariation within a family which generally was assigned to error term when more than one measuresfrom each experimental plot were used for analysis; the replicated mean confounded with variationwithin a family caused decreasing the error term and increasing variation among families andconsequently, high estimated F-ratio for variation among families. The variance analysis for seed traitswas simply conducted in a well controlled laboratory environment and consequently decreased unknownerrors. The small error terms were mainly the results of variation among replications and wererequired for variation study. Moreover, using replicated means withdrew most of genetic variationwithin families from the error terms. Seed traits influence survival and reproductive success of speciesand the information is desirable for nursery implementation.The variation in most seedling and progeny traits, on the other hand, existed both amongpopulations and within populations. The proportions of variation among populations to among familiesfor seedling height, diameter, and dry weight were 62.18, 0.00, and 42.89%, respectively. Theproportions of variation among populations to among families for progeny height, diameter, and stemform were 71.90, 85.87, and 67.41, respectively.Generally, the variation among populations for quantitative traits in this study was moderatefor 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, stemstraightness, and branching characteristic (F = 8.43, 4.73, 6.55, 2.83, and 3.55, respectively). Thisreflected the adaptation and/or selection of polygenic traits under different environments. However,the variation of seed germination, seed weight, which preconditioning maternal environmental effectsof individual families were likely to dominate the variation, were not different among populations. Thevariation among population for seedling diameter and dry weight was also not different, probably thesetraits carried over preconditioning maternal environmental effects longer than other seedling traits.The only seed trait correlated with seedling traits was R50. (Table 26). Although statistical81correlations between seed and progeny traits and between seedling and progeny traits were not possibledue 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 wasthe only seed trait producing a comparable amount and pattern of variation to those of seedling andprogeny traits. Consequently, the conclusion for quantitative traits was that the genetic variation wasrelatively high within and among the populations of this tree species. The seed trait variation providedmore reliable results than those for seedling and progeny variation because the experimental errors forseed variation estimates were reduced under the controlled environmental conditions. On the otherhand, seedling and progeny variation were subject to environmental differences of individual seedlingsand unknown within family variation. However, all quantitative traits from this study are mainlyparallel in patterns of the variation and should he combined for the effective conservation of adaptivegenes. The amount and pattern of variation will be compared with those of isozymes to make the finaldecision on conservation strategies.Table 26 Correlation between seed and seedling traitsTRAIT GER R50 WT HT DMGER -.0615 .2046* .1369 .0912R50 -.0615 .0592 .3240** .1963*WT .2046* .0592 .0944 .0796HT .1369 .3240** .0944 .7478**DM .0912 .1963* .0796 .7478**TDW -.0575 .1832* -.0936 .3015** .2535**N of cases: 221 1-tailed Signif: * - .01 ** - .0012. Isozyme ConsiderationThe isozyme analysis indicated that genetic variability at 18 loci was relatively high with mean82unbiased 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) unbiasedgenetic distance ranged from 0.009 to 0.083. There was also an indication of relatively high variationamong populations from Wright’s F-statistics (12.7%). This was smaller than the mean variationamong 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 bothquantitative and isozyme analyses, the conservation programme of this tree species must involve a largenumber of families. However, only sound genetic base would not enough to consider conservationstrategies in tropical regions due to the growing human population surviving on forest land andproducts. Therefore, to be effective, the conservation requires the integration of socio-economic andpolitical situations properly.3. Quantitative and Isozyme CorrelationsThe comparison of quantitative and biochemical variation was based on forty-eight commonindividual trees (families). Quantitative trait variation for seed germination, R50, seed weight, nurseryseedling height, diameter, total dry weight, and root and shoot dry weight ratio were compared toisozyme variation (expected heterozygosity calculated for each common family by using formula page51) and the mating system (Appendix K). The comparison was conducted at both the family andpopulation levels by using their corresponding means. The correlation coefficients at the family andpopulation 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 nocorrelation between quantitative trait and isozyme variation at the population level.83Table 27 Correlations estimated from individual family means for heterozygosity, quantitativetraits and mating systemTRAIT HETEROZYGOSITY MATINGHETEROZYGOSITY 1.0000 0.1126GERMINATION 0.0748 0.0206R50 -0.0067 -0.0692SEED WEIGHT -0.1265 0.0370HEIGHT 0.4258* 0.0090DIAMETER 0.2051 -0.1242TOTAL DRY WEIGHT 0.1410 0.0550ROOT-SHOOT RATIO -0.2885 0.0228N of cases: 48 1-tailed Signif: * - .01 ** - .001Table 28 Correlations estimated from population means for heterozygosity, quantitative traitsand mating systemTRAIT HETEROZYGOSITY MATINGHETEROZYGOSITY 1.0000 0.3119GERMINATION 0.0447 -0.2437R50 -0.3156 0.2168SEED WEIGHT 0. 1097 0.0600HEIGHT 0.0707 -0.0504DIAMETER 0.2403 0.5246TOTAL DRY WEIGHT 0.1718 -0.0754ROOT-SHOOT RATIO -0.34 12 -0.6411N of cases: 8 1-tailed Signif: * - .01 ** - .001There 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 expectedto be under control of selection. The variation among populations for isozymes, on the other hand, wasjust at the moderate level. Isozymes may be selectively neutral or nearly neutral (Kimura, 1983) andthe low isozyme differentiation among populations may reflect sufficient levels of gene flow for neutralalleles (Lewontin, 1974). However, these levels of gene flow may be insufficient to prevent84differentiation at loci under selection (Muona, 1989). Therefore, isozyme variation alone is not enoughto predict the distribution of variation for selectively important traits for designing sampling strategiesfor gene conservation programmes (Muona, 1989; Falkenhagen, 1985; Hamrick, 1983). The effectivegene 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 geneticestimates also provide the necessary information for determining the number, size and distribution ofconservation areas. Thus the most effective gene conservation should be implemented on thecomplementary bases of both quantitative and isozyine variation.4. Genetic RationaleGenetic resources are the foundation of biological diversity which forms the essential linkbetween 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 lossesof genetic resources are rapid and cause changes in the world’s ecosystem (Noss, 1991). Conservationof genetic resources requires that selected, representative populations be regenerated from generationto generation (Yeatman, 1987). The principles of genetic conservation are the same for all livingorganisms. However, the strategies and methodology vary according to the distribution and biologicalnature of the species (FAO, 1989).The biological diversity is so complex and intangible that its conservation cannot beapproached 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 andhigher, and rarely operate at the species and lower level (Falk, 1990). Conservation efforts for rareand endangered plant species, many of which are narrowly endemic and have restricted ranges which85cause serious threats to their survival, require a sound basis in scientific knowledge.Since biological diversity is the principal goal of conservation, integrating endangered speciesconservation into ecosystem management should be the best strategy (Salwasser, 1991). This strategy,however, requires empirical and theoretical knowledge in ecology. Lande (1988) has addressedpopulation demographic basis important for biological conservation, and advocated incorporation ofpopulation demography and genetics into future conservation plans.Dalbergia cochinchinensis Pierre is an endangered species and most of its populations havebeen disturbed. These populations might be difficult to conserve under the increasing land requirementsfor agriculture and industrial infrastructure in Thailand. Yeatman (1987) indicated that indigenous treespecies and populations are most effectively maintained in situ by natural regeneration and/or byplanting trees of local origin and wide parentage. Therefore, both in situ and ex situ conservationstrategies are needed for retention of the genetic variability of this endangered species.4.1. In situ conservationIn situ conservation is an ideal strategy which protects and maintains species and populationswith the ecosystem where they occur. This will conserve an ecosystem and maintain the variation ofthe 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 anddesign of minimum viable population sizes. The minimum viable population size is a threshold for theminimum number of individuals which enable a population in a given habitat to persist and regenerateindefinitely (Gilpin and Soulá, 1986). Three broad approaches to estimating minimum viablepopulation size have been taken so far (NRC, 1991). The first two approaches, based on effectivepopulation size (Franklin, 1980; Soulé, 1980), might he inappropriate to Daihergia cochinchinensisPierre due to the problems of estimating effective population size for forest species. The third approachis based on population size which will minimize the sampling loss of low-frequency alleles (NRC,861991). Namkoong (1984a) estimated that in species with low levels of inbreeding and populationstructure or high variation among population, a sample size of 1,000 individuals will minimize the lossof an low frequency allele below the probability of 0.01. Namkoong (1984b) has also suggested thatto minimize loss of useful alleles, a population size of 10 to 20 should be maintained in each of 20 to50 or more subpopulations. Within each suhpopulation, simple recurrent selection or hybridization maysimply generate new subpopulations for genetic improvement. Moreover, because outcrossing is themating system of this tree species, with this number of individuals in each subpopulation, conservationshould be rational because their genetic constitutions are distinct.Due to the high amount of genetic variability among the families for Thailand rosewood andthe relatively high amount of differentiation among populations, the genetic conservation strategiesshould include a large number of families to cover a substantial proportion of variability. ThusNamkoong’s (1984a) criteria would be appropriately applied to this species due to its naturally smallpopulation size and scattered distribution. Both quantitative and isozyme results indicated two separatepopulational regions, central and northeastern regions. Therefore, about 1,000 individuals gatheredfrom small populations or subpopulations in each region would he adequate. The conserved populationsshould 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 bepromoted by collecting seed from many trees to sample (genetically) a given population adequately forplanting the next generation at the population site. The success of enrichment planting is dependent onthe potential of plants derived from seed. Therefore, designation, protection, and management of seedsources for production of required quantities are prerequisites (Yeatman, 1987). The currentconservation programme of this tree species, however, rarely emphasizes in situ conservation due tothe challenges and high requirements of natural forest land. The natural populations of this tree specieslargely abide in the national parks, forest reserves, wildlife sanctuaries, and public areas, but the actualsites or locations, sizes, and circumstances of most populations are unknown. Therefore, an extensivesurvey must be carried out to obtain information and to determine a sampling strategy.87In situ conservation can also be applied to disturbed forest areas, including public areas andprivate areas, in order to sample the genetic variability of disturbed populations to ensure the mostviable genetic base for future management. Furthermore, the outcrossing rate of the species is quitehigh both in undisturbed and disturbed populations. This implies that high quality seed can be obtainedfrom either type of population. However, in situ conservation for disturbed populations is even morechallenging than for undisturbed populations. An additional means to ensure this future availability ofmost 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, humanintervention is possible for refining and strengthening the programme. While some forest reserves willbe devoted to ecosystem conservation, others may he dedicated to conserve intraspecies variation of thetarget species or subspecific populations (Yeatrnan, 1987; Kemp, 1992).4.2. Ex situ conservationEx situ conservation is a strategy which protects and maintains genetic resources outside theirenvironment as production stands, breeding material, conservation/seed stands, seed, pollen and/ortissue (Bonner, 1985). Due to the population pressure on destruction of tropical forests, in situconservation 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 forfuture management (FAO, 1989). However, cx situ conservation should he done to complement in situconservation only; it can never replace in situ conservation fully.So far, cx situ approaches have formed the main emphasis of the conservation programme ofthis species. Five seed sources, consisting of thirty families each, were proposed for planting at fivelocations, 10 hectares each, in 1991 (RFD, l989a). However, in the sampling strategy, severalgeographically proximate populations were included in the same seed source. The results from thisstudy indicate that genetically distinct populations may have been inadvertently mixed. Genetic integrity88of different populations would be lost by random mating with among populations, such an approach toconservation is not recommended.Due to the high genetic variability among populations, more populations should be added tothe programme. At least one ex situ plantation regenerated from 1,000 individuals of small andremnant populations or subpopulations in each region should, if possible, be located in the selected areaof the region. This will ensure the maintenance of a broad genetic base and prevention of the randomgenetic drift caused by inbreeding within the natural small population size of the species. The numberof ex situ plantations can be more than one in each region. This depends on socio-economic situationswhich must be considered in making the programme practical.Whereas the success of ex situ conservation is dependent on functional and structural propertiesof an ecosystem, information on both population demography and environmental randomness arerequired. However, interaction with species of plants and animals are rarely understood. Some speciesmight positively interact with the conserved species. Therefore, an ex situ plantation would be betterestablished at the original sites of populations. However, when forest or land areas are changed byhuman intervention, the environmental conditions become poor and may no longer he appropriate forex 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 leftstanding, and other species were cleared from natural forest. The land around the standing trees wasburned, 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 carefulconsideration and integration of genetic, ecological and other biological aspects of the target species canconservation success be ensured. Climatic, geographic, and edaphic factors, although demonstratinglittle variation, have to be matched with the species requirements. More information on these aspectsis, however, required to increase the efficiency of the ex situ conservation.Although seed of Thailand rosewood can he stored at cool temperature, seed viability decreasesvery quickly. Therefore, a seed bank conservation method has to be developed. Research on this89aspect should he emphasized for future application. However, today, plantations still serve as the bestmeans of ex situ conservation.Furthermore, to support the conservation success, tree improvement for this tree species shouldbe initiated using information on quantitative variation derived from this study. The improved seed orplanted material then will increase the production of this tree species and decrease the pressure onharvesting trees from natural forests and conservation areas.S. Socio-economic and Political ConsiderationsA genetic conservation programme cannot he segregated functionally and must be integratedconsciously with socio-econornic conditions. Genetic conservation should be an integral part of nationalpolicies; 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 manytropical countries, human population growth and increased social demands for land and income supportin Thailand challenge the protection of the full range of genetic variability. To accomplish theconservation of endangered species, a goal of land and resource management must be to sustainbiological 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 theconservation of a species from the local to the national level (International Tropical TimberOrganization, ITTO, 1991). Therefore, national forest policy must integrate the conservationprogramme into all forest activities. The people responsible for these activities must be well educatedin the necessity of forest resource conservation. Of greatest concern are local people or villagers whomainly make their living from forest contributions. The conservation programme then shouldemphasize the participation of these people. Basically, villagers, who live closest to the forest area tobe conserved, are the poor or poorest in most tropical countries. A conservation programme can beeffective if local or rural people are involved and benefit from it. incorporation of conservation90objectives with rural development programmes should be a powerful means of success. Since directand indirect values of conservation are ambiguous to them, obvious benefits from conservation shouldbe demonstrated and extended to them: for example, economic values such as seed, wood, watershedprotection, and income from recreation in conserved areas. Conservation and utilization therefore mustbe made compatible and beneficial. Moreover, conservation ethics should be strongly emphasized atall levels of education.The economic returns of poor rural people who rely on forest resources will be affected, inthe short term, by a change from over-exploited forests to conservation forests. Common causebetween 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 agroforestrysystem which integrates tree planting with agricultural practices can be one option. Agroforestry canbe located in marginal, low productivity forests, which will provide added benefit as buffer zones fordesignated forest conservation areas. To conserve genetic diversity of an economically importantspecies, 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 forsustainable management to provide an appropriate means of harvesting forest products.A large number of new jobs concerned with conservation can replace the old ones. Theseinclude 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 NationalPark in northeast Thailand. Problems with forest encroachment by local villagers seeking economicbenefits from poaching of timber and animals were alleviated by recruitment of forest rangers andguides. Many who were life-long poachers are now playing a significant role in conservation, derivinga consistent and legal livelihood in the process. This kind of activity must be developed either by thegovernment or by concerned non-governmental agencies.The key to success is the coordination of people from all levels, especially local people. Theprogramme then must be promoted extensively to the public; information must be delivered.91Furthermore, for the sake of conservation, the designated forests must be treated by the people as theirown 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 ina developing country. Thus, mass media, public awareness should be generated against them and theiractivities. Punishment by all legal and social means might he the best avenue to control these illegalactivities. On the other hand, people who dedicate themselves, their abilities, and/or their propertiesto conservation should be rewarded with great public honour and financial benefits.Because the conservation programme of this tree species was established on a bi-governmentbasis and several tree species were included in the programme, limited work has been carried out sofar. Funds and personnel are constrained and solutions are needed to this problem. The programmeshould be revised and extended to include other countries in which the tree species or genus isdistributed. Only with the success of conservation efforts for this valuable species can present andfuture benefits from its sustainable use he ensured.6. ConclusionDue 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 existingprogramme which has operated with only small number of populations and families. Disturbedpopulations could also be conserved to cover most genetic variability and provide materials for researchand reforestation. Ex situ conservation should be carried out to complement in situ measures to ensureavailable genetic materials for future management. Ex situ plantations should be established in the samelocality as where the seed was collected. Finally, participation of local people in the conservationprogramme has to he developed. The government, public and other agencies must play a substantialrole in conservation of all levels from community or ecosystem to species and gene. The programmeshould be in cooperation with other neighbouring countries within which the species is distributed. Thiswill make it easier to acquire starting funding from government and/or nongovernment agencies.9293CHAPTER 5CONCLUSIONS AND RECOMMENDATIONS1. Conclusions1.1. Quantitative variationThe study of genetic variation of Thailand rosewood both through quantitative traits andisozyme characteristics provided useful information for genetic management and conservation of thespecies. Quantitative traits, including seed, nursery seedling, and field progeny traits revealed highvariability exists among families within populations and relatively high among populations. There wereno significant differences in these traits between undisturbed and disturbed populations. The observedquantitative variation may be an adaptation to different environments or due to regional isolation andshould form the basis of conservation strategies. However, the heritability estimates for most traits maybe insufficient to justify a genetic improvement programme. The genetic correlations among traits areweak. Nevertheless, the variation in seed and seedling traits remains valuable fur nursery managementand reforestation because both quantity and quality of seed and seedlings are required. The progenyvariation was significant among families for only some traits, except diameter, persistence of axis, andstem straightness. The variation of interactions between families and blocks was significant for allprogeny traits.941.2. Isozyme variationTwelve enzyme systems provided eighteen scorable loci which were used for analyses ofgenetic variability and the mating system. The genetic variability parameters, number of alleles perlocus and mean heterozygosity were high within populations and moderate and differed significantlyamong populations, but not between undisturbed and disturbed populations. Nei’s (1978) geneticdistance and F-statistics (Wright, 1965) indicated moderate genetic differentiation among populations.Cluster analysis using Nei’s genetic distance divided populations into two geographic regions: centraland 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. Alsosmall population sizes may have contributed to the sampling error.1.3 Conservation strategiesConservation of Thailand rosewood was initiated in 1989. However, a genetic basis forrational conservation has rarely been available and applied. The information derived from this studywill therefore be used to support and strengthen the existing programme.The high quantitative and isozyme variability among populations necessitates the conservationof several populations populations. However, remnant populations of this tree species are naturallysmall. 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 orsubpopulations distributed throughout the region should be enough for maintaining most and even lowfrequency alleles (Namkoong, 1984a). Both in situ and ex situ conservation should be established forthis endangered tree species. Disturbed populations should also he conserved as sources of geneticmaterial for genetic studies and reforestation. Ex situ conservation should be established at least one95location in each region by collecting about I ,000 individuals from populations or subpopulations withinthe 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 increaseproduction of the species and decrease the need for harvesting trees from natural forests andconservation areas. High variation in both quantitative traits and isozyme made possible to improvethe traits. However, heritabilities of the traits were not very consistent, so selection based on highheritability traits will make more progress on tree improvement.2. RecommendationsConservation should be nniltidisciplinary; integration of population genetic and ecologicalinformation with socio-economic and political aspects of the country should be considered. Theconservation strategy has to he forimilated by national policies to benefit local people. The participationof local people is the key to success; therefore, development agencies and operations should emphasizethe provision of benefits to local communities. Education and job training should he provided to localpeople to help improve their knowledge and skills. Furthermore, agroforestry, forest community, andforest village projects which were successful in several areas should he promoted and extendedthroughout the country.This study would provide the principal knowledge of gentic variation strategies for conservationand future genetic management of this tree species, hut information is still inexact. Therefore, anumber of studies are required to develop the understanding of the genetics and ecology of tropicalforests. This should serve as the principle for conservation which is to be scientifically based andrational. Socio-economic and political refinements are also needed to support and strengthen theeffectiveness of conservation efforts. The activities involve:(a). Extend the exploration of tree populations and collection of seed for comparative study ofboth quantitative and isozyme variation.96(b). Repeat studies of quantitative variation in different years and planting sites to further theunderstanding of the effects of temporal and spatial variation.(c). Repeat studies of isozyrne variation using a larger number of families and populations indifferent years to obtain a better understanding of the genetic structure of Thailand rosewood and itstemporal variation.(d). Incorporate ecological studies (particularly genecology) of Thailand rosewood and itsassociated species, especially pollinators, for more efficient conservation.(e). Compare distribution, variation, and mating systems of other tree species in the genusDalbergia 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 andnongovernment agencies.(h). Ensure regeneration of the target species and a coherent system of economic incentivesfor greater attention in genetic conservation.97REFERENCESAdams, 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. 75-84.Adams, W.T. and R.J. Joly. 1977. Analysis of genetic variation of height growth and survival in open-pollinated progenies of eastern white pine. In: Proc. Northeast For. Tree Improv. Conf25:117-13 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 TestCommittee of the Association of Official Seed Analysts. International Seed TestingAssociation, 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. Evolution28:85-92.Bawa, K.S. 1976. Breeding of tropical hardwoods: an evaluation of underlying bases, current statusand 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 LethalAlleles and Magnitude of Effects Associated with the Maternal Parent. Silvae Genetica36:180-185.Bonner, F.T. 1985. Technologies to Maintain Tree Germplasm Diversii. Office of TechnologyAssessment, U.S. Congress, Washington, D.C.Botstien, D.; R.L. White; M. Skolnick; R.W. Davis. 1980.Constructionofgeneticlinkagemap in manusing restriction fragment length polymorphisms. Amer. J. Hum. Genet. 32:314-331.98Boyle, 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. 13:115-155.Bramlet, D.L.; T.R. Dell; W.D. Pepper. 1983. Genetic and maternal influences on Virginia pine seedgermination. 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 GeneticResources. 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 Eucalyptusobliqua 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 Vol7. 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 thecooperative 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. III. Breeding systems.Malaysian For. 44:28-341.Cheliak, W.M.; B.P. Dancik; K. Morgan; F.C.H. Yeh; C. Strobeck. 1985. Temporal variation of themating system in a natural population of jack pine. Genetics 109:569-584.Cheliak, W.M.; F.C.H. Yeh; J.A. Pitel. 1987. Use of electrophoresis in tree improvementprograms. For. C’hronicle 63:89-96.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 inan experimental plant population. Genetics 89:765-792.99Cochran, 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. Vol. 1 United Selangor Press, Kuala Lumpur,Malaysia. 476 pp.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.El-Kassaby, Y.A. 1980. Isozyme patterns of selected Pseudotsuga menziesii (Mirb.) Francopopulation. 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 parametersin 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 outcrossingHelianthus 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 thepresence 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 rarespecies 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.100FAO. 1988. in Situ conservation: conserving Plant Genetic Resources in Situ to Meet Today r andTomorrow’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 NorthAmerica 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 bandphenotypes 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 thegenetic 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 inallozymes 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. SinauerAssociates, Sunderland, Mass. pp. 19-34.Griffin, A.R. and P.P. Cotterill. 1988. Genetic Variation in Growth of Outcrossed, Selfed andOpen-Pollinated Progenies of Eucalyptus regnans and some Implications for BreedingStrategy. 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), Geneticsand 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 GeneticResources. Sinauer Associates, Inc., Sunderland, Mass. pp. 43-63.Hamrick, J.L.; Y.B. Linhart; J.B. Mitton. 1979. Relationships between life history characteristics andelectrophoretically detectable genetic variation in plants. Ann. Rev. Ecol. Syst. 10: 173-200.101Hamrick, J.L. and M.D. Loveless. 1986. Isozyrne variation in tropical trees: procedures andpreliminary 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.), TheEvolutionary 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 uncommonneotropic tree species. J. Tropical Ecol. 7:395-399.Hanover, J.W.; T. Mebrathn; P. Bloese. 1991. Genetic improvement of black locust: a primeagroforestry species. Forestry chronicle 67:227-231.Hicks, R.R.Jr.; G. Rink; B.E. Cutter. 1977. Variation and heritability of seedling height in East Texassycamore 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 forthe conservation of biological diversity in tropical production forests. Ninth session, Nov28 - Dec 4, 1991. Yokohama. In: PC’FIX/IO. ITTO and Permanent Committee onReforestation and Forest Management.International Union for the Conservation of Nature and Natural Resources (IUCN). 1980. Worldconservation Strategy: Resource Conservation For Sustainable Development. IUCN,Grand, Switzerland.Jiang, I.B.; J, A. Jonsson; G. Eriksson. 1989. Within and between population variation in growthof Pinus contorta var. iatfolia: A combined study of growth-chamber and field-trialexperiments. Silvae Genetica 38:201-211.Kanowski, P.J. and D.G. Nikies. 1989. Family structure in provenance trials, conservation stands, andgenetic 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.Keating, N.G. and E. Bolza. 1982. South-east Asia, Northern Australia and the PacJiccharacteristics, 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 gmelinaprovenance trials: selection of traits, their assessment and analysis of observation. In:Proc. Joint Meeting Work Conf on Provenance and Genetic Improv. Strategies in TropicalForest Trees. Matare, Zimbabwe 9-14 April 1984. pp.30-69Kemp, R.H. 1992. The conservation of genetic resources in managed tropical forests. Unasylva 169V43:34-40.102Kimura, 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 17:279-288.Kimura, M. and T. Ohta. 1971. Protein polymorphism as a phase of molecular evolution. Nature229:467-469.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 Effectivein 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, Torreypine (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 innatural populations. II. Amount of variation and degree of heterozygosity in naturalpopulations of Drosophila pseudoobscura. Genetics 54:595-609.Liengsiri, C.; C. Piewluang; T.J.B. Boyle. 1990. Gel Electrophoresis of Tropical Trees - A Manual.ASEAN-Canada Forest Tree Seed Centre. Muak Lek, Saraburi,Thailand.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 tospacing. 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 speciesspecific patterns. In: Proc. Nat. Acaci. Sc-i. (USA) 45:573-763.103Mikola, J. 1984. Relationship between height growth differences of Scots pine full-sib families andvariation 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 SciencePublishers, Amsterdam. pp.443-472.Moran, G.F.; 0. Muona ; J.C. Bell. 1989. Breeding systems and genetic diversity in Acaciaauriculjformis 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 inPlant 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 sprucePicea 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 geneticresources. 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. MohanPrimlani, 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. NationalAcademy 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)70:3321-3323.104Nei, 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 ofindividuals. Genetics 89:583-590.Noss, R.F. 1991. From endangered species to biodiversity. In: Kohm, K.A. (Ed.), Balancing on theBrink 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 andphysiological 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, andcan 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 geneticheterozygosity in natural populations of Drosophila pseudoobscura. Genetics 61:841-858.Rehfeldt, G.E. 1983. Veracity of short-term testing. In: Proc. Servicewide Genetics Workshop onProgeny 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. 32:1049-1057.Ritland, K. 1986. Generalized Mulrilocus Estimation Program MLTF. Documentation Manual. K.Ritland, Toronto, CanadaRitland, K.R. and S.K. Jam. 1981. A model for the estimation of outcrossing rate and genefrequencies 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 firas 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. 18:209-215.105Roche, L. and M.J. Dourojeani 1984. A Guide to in situ conservation of Genetic Resources ofTropical Woody Species. FAO. 196 pp.Rogers, D.L.; R.F. Stettler; P.E. Heilmann. 1989. Genetic variation and productivity of Populustrichocarpa 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 leguminousspecies 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 conservationin northeast Thailand. In: Proc. JUFRO Cot1f on Breeding Tropical Trees: PopulationStructure 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. 17:667-693.Schnabel, A. and J.L. Hamrick. 1990. Organization of genetic diversity within and among populationsof Gleditsia triacanthos (Legurninosae). Amer. J. Bot. 77(8): 1060-1069.Schnabel, A.; R.H. Laushman; iL. Hamrick. 1991. Comparative genetic structure of two co-occurringtree 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 isozymemarkers. Theor. Appi. Genet. 63: 113-120.Shaw, D.V.; A.L. Kahler; R.W. Allard. 1981. A multilocus estimator of mating system parametersin plant populations. Proc. Nat. Acad. Sci (USA) 78: 129-1302.106Sneath, 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 treespecies: experiences in Zimbabwe. In: Proc. JUFRO Joint Meeting of Working Parties onBreeding Theoty, Progeny Testing and Seed Orchards. Williarnburg, VA. 13-17 October1986. 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 earlygrowth 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. McGraw-Hill, NewYork.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 ofblack 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 treespecies. 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 WilliesHennig 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 AllelicVariation 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 onPopulation 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) undercontrolled conditions: Magnitude and partitioning of variance. Ecology 65:626-63 1.Tyson, H. 1989. Genetic control of seed weight in flax (Linum usitatissimum) and possibleimplications. Theor. Appi. Genet. 77:260-270.107Wheeler, N.C. and W.B. Critchfield. 1985. The distribution and botanical characteristics of lodgepolepine: 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 morphologicalvariation 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 polymorphismsamplified by arbitrary primers are useful as genetic markers. Nuc. Acids Res. 1 8:6531 -6535World Resources Institute (WRI). 1990. World Resources. A Guide to the Global Environment. TheWorld 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 tosystems of mating. Evolution 19:395-420.Wright, S. 1978. Evolution and Genetics of Populations, vol 4. Variability Within and Among NaturalPopulations. 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 GeneticImprov. 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 itsimplications 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 itsmanagement 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.108APPENDICESAppendix 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 eightpopulations(B). for progeny trial, seed was collected in 1987 from seven populations(A). Number of familiesPopulation Seed Seedling IsozymeUndisturbedML1 18 14 17ML2 13 11 11DisturbedKKI 7 6 6KK2 10 8 5MKI 4 4 4MK2 4 4 4SKi 7 6 7SK2 6 3 3TOTAL 69 56 57(B). Number of familiesPopulation Progeny trialUndisturbedDL 20ML 15SKH 5DisturbedKH 9MS 19SM 5SK 12TOTAL 85109Appendix B Seed trait means for individual families in eight populationsPOPULATION GERMINATION R50 SEED WEIGHT% (day) (mg)OVERALL MEAN 80.5870 9.2517 24.1748UNDISTURBED POPULATIONPOPULATION ML1 88.4167 10.6628 24.3676FAMILY 1 87.5000 12.4875 26.4850FAMILY 2 79.0000 8.2425 21.5450FAMILY 3 7 1.5000 7.6775 20.0200FAMILY 4 95.0000 12.7000 25.0275FAMILY 5 88.5000 11.0150 27.5250FAMILY 6 86.5000 8.1575 24.7400FAMILY 7 90.0000 9.0325 25.3325FAMILY 8 92.0000 11.1700 27.2575FAMILY 9 92.5000 13.5050 22.7125FAMILY 10 94.5000 9.7125 22.7950FAMILY 11 93.5000 9.0175 16.9200FAMILY 12 89.0000 10.6075 24.9150FAMILY 13 87.5000 13.7825 29.3325FAMILY 14 90.5000 12.7150 23.4600FAMILY 15 93.0000 8.6350 23.3725FAMILY 16 91.5000 11.3400 28.0850FAMILY 17 78.0000 13.5275 26.8600FAMILY 18 91.5000 8.6050 22.2325POPULATION ML2 78.9615 8.7677 23.6936FAMILY 1 88.0000 12.4075 19.0200FAMILY 2 92.5000 10.3550 25.0850FAMILY 3 93.5000 6.3750 23.6250FAMILY 4 69.0000 6.6925 23.8925FAMILY 5 92.5000 9.9575 24.9025FAMILY 6 77.0000 11.2725 20.4850FAMILY 7 86.5000 10.6700 28.6850FAMILY 8 72.5000 7.5625 22.3550FAMILY 9 97.5000 8.8375 18.2410FAMILY 10 95.0000 8.7525 25.2620FAMILY 11 70.5000 7.3250 25.1520FAMILY 12 52.5000 10.8100 21.5975FAMILY 13 59.0000 9.3500 32.0860DISTURBED POPULATIONPOPULATION KK1 75.0714 11.1564 24.6226FAMILY 1 88.0000 10.2625 30.3710FAMILY 2 26.0000 13.4775 16.4490110Appendix B (continued)FAMILY 3 81.5000 14.6275 23.1135FAMILY 4 74.5000 6.1100 25.8590FAMILY 5 83.0000 13.4450 30.5540FAMILY 6 82.5000 8.8925 23.7440FAMILY 7 90.0000 11.2800 22.2680POPULATION KK2 79.7500 9.6355 23.9697FAMILY 1 75.5000 10.4275 21.0120FAMILY 2 87.5000 8.5125 20.0990FAMILY 3 93.0000 5.7925 23.6550FAMILY 4 86.0000 6.4775 21.0370FAMILY 5 68.5000 7.3250 24.5820FAMILY 6 78.0000 10.5475 25.67 10FAMILY 7 84.0000 10.0850 33.0500FAMILY 8 79.5000 9.7375 16.5720FAMILY 9 35.5000 6.5575 25.1200FAMILY 10 90.5000 14.5050 26.5275POPULATION MKI 88.5000 8.2013 25.4536FAMILY 1 87.0000 6.8200 28.0330FAMILY 2 86.0000 9.3850 25.4820FAMILY 3 90.0000 8.9750 21.7225FAMILY 4 91.0000 7.6250 26.5770POPULATION MK2 80.7500 7.8906 20.2830FAMILY 1 90.5000 7.9975 20.8780FAMILY 2 81.5000 6.9175 18.6670FAMILY 3 76.0000 7.9450 17.53 10FAMILY 4 75.0000 8.7025 24.0560POPULATION SK1 71.4286 7.1871 23.0084FAMILY 1 85.0000 8.2400 24.6700FAMILY 2 27.5000 8.3325 27.4460FAMILY 3 92.0000 5.6225 19.7610FAMILY4 87.5000 5.1175 27.1670FAMILY 5 44.5000 8.8025 13.0070FAMILY 6 89.5000 6.1900 24.4600FAMILY 7 74.0000 8.0050 24.5480POPULATION SK2 73.7500 7.2212 27.6025FAMILY 1 75.0000 4.9375 25.3370FAMILY 2 78.0000 5.9425 29.0630FAMILY 3 88.0000 5.4425 24.8480FAMILY 4 49.0000 13.5000 27.0120FAMILY 5 65.0000 6.3625 29.4650FAMILY 6 87.0000 7.1425 29.8900111Appendix C Seedling-trait means for height (HT), diameter (DM), total dry weight (TDW), androot-shoot ratio (RS).Trait HT DM TDW RS(cm) (cm) (gm)Overall mean 53.3090 .6099 15.6205 .9577Undisturbed populationML1 AVE. 63.6896 .6386 17.0650 .7809FAM 1 62.4417 .6762 20.4225 .6956FAM 2 62.3334 .6418 18.3300 .6662FAM 3 59.9667 .6269 18.2400 .7107FAM 4 65.4167 .6250 14.2950 .6918FAM 5 65. 1500 .6493 14.4125 .5461FAM 6 65.8750 .6229 17.0200 .6581FAM 7 64.7500 .5975 17.4400 .6706FAM 8 63.0168 .6064 16.6925 .8950FAM 9 58.9167 .6059 17.5425 .9210FAM 10 71.6875 .6463 15.6675 .8692FAM 11 70.7500 .7321 18.1950 .7371FAM 12 60.6417 .6214 16.3725 .9763FAM 13 57.8333 .6696 15.9400 1.0135FAM 14 62.8750 .6187 18.3400 .8821ML2AVE. 59.2837 .6210 16.4659 .7523FAM 1 61.2917 .6129 19.4025 .6421FAM 2 63.9250 .6395 16.3900 .7942FAM 3 43.1667 .5200 14.3475 .8720FAM 4 70.8667 .7089 16.4550 .7887FAM 5 49.6667 .5954 16.4250 .7793FAM 6 66.7917 .6907 16.9625 .7958FAM 7 72.8959 .6694 17.4550 .6345FAM 8 45.3167 .5321 12.8675 .7395FAM 9 59.2084 .6085 14.5600 .8653FAM 10 56.4167 .6528 18.9300 .5854FAM 11 62.5750 .6011 17.3300 .7783Disturbed populationKK1 AVE. 49.3056 .6013 15.6509 .8199FAM 1 42.5500 .5582 14.1200 .9081FAM 2 50.7500 .6331 22.1467 .8033FAM 3 55.7167 .5962 10.6375 .5951FAM 4 54.5667 .5745 13.9200 .7545FAM 5 45.3333 .6394 17.9875 .9921FAM 6 46.9167 .6066 16.7175 .8621112Appendix C (contKK2 AVE. 52.1146 .6237 15.4468 .9484FAM 1 54.7083 .6569 12.4875 1.2542FAM 2 48.8333 .6246 15.4725 .8778FAM 3 45.9083 .6000 14.3900 .9858FAM 4 58.5000 .6400 18.0667 1.0848FAM 5 54.4000 .5934 17. 1275 .8956FAM 6 64.1917 .7429 23.0275 .8925FAM 7 36.0000 .5188 12.0550 .9203FAM 8 54.3750 .6133 11.6025 .7107MK1 AVE. 46.0802 .5890 16.5547 1.3141FAM 1 49.5000 .6394 15.6675 1.3854FAM 2 52.8708 .5968 12.6300 1.0118FAM 3 41.5750 .5636 22.5933 1.7439FAM 4 40.3750 .5563 16.8375 1.1154MK2 AVE. 40.5781 .5579 13.6306 1.1854FAM 1 42.8667 .5552 13.6400 1.3042FAM 2 37.5417 .5346 19.9125 .9055FAM 3 54.4875 .6797 17.7200 1.0700FAM 4 27.4167 .4621 13.2500 1.4620SKi AVE. 38.3437 .5672 11.9413 1.4791FAM 1 27.9375 .5119 10.9700 2.0420FAM 2 47.8750 .6346 13.7000 1.3569FAM 3 41.2083 .5915 14.1725 1.4624FAM 4 44.4083 .6032 9.6625 1.3549FAM 5 34.5000 .4900 13.3050 .8566FAM 6 34.1333 .5721 9.8375 1.8019SK2 AVE. 50.6944 .5987 15.0133 1.0014FAM 1 43.6667 .5475 11.8700 .9979FAM 2 58.4584 .6496 19.1875 1.0786FAM 3 49.9583 .5990 13.9825 .9277113Appendix D Scoring method ( Keiding et a!, 1984) for stemformClass scorePersistence of axisDouble or multiple stems from ground level 1Axis branches out in 1st quarter of the tree 2Axis branches out in 2nd quarter of the tree 3Axis branches out in 3rd quarter of the tree 4Axis branches out in 4th quarter of the tree 5Complete persistence 6StraightnessCrooked - more than three serious bends 1Crooked - one or two serious bends 2Slightly crooked, many bends 3Slightly crooked, few bends 4Straight 5Branch heaviness*Very heavy - branches from 1/2 to 3/4 of stemHeavy - branches about 1/2 of stem 2Medium - branches from 1/4 to 1/2 of stem 3Light - branches about 1/4 of stem 4Very light - branches less than 1/4 of stem 5* Branch heaviness is a relative measure of branch sizein proportion to the stem at the base of the branches.114Appendix E Average field progeny trait measures for height (HT) and diameter (DM) at 4 yearsTRAIT HT DM(m) (cm)Overall mean 3.4823 5.9480Undisturbed populationDL AVE. 3.6920 6.3918FAM 1 3.8050 6.7048FAM 2 3.7225 5.8952FAM 3 3.1438 6.0303FAM 4 3.6110 6.4035FAM 5 3.2688 5.7410FAM 6 3.4413 6.0445FAM 7 3.8650 6.4695FAM 8 3.7760 6.9902FAM 9 3.6912 5.9440FAM 10 3.7187 6.4260FAM 11 3.6662 6.8827FAM 12 3.9588 6.8180FAM 13 3.5835 5.8777FAM 14 4.0175 6.9828FAM 15 3.6625 6.0760FAM 16 3.7162 6.1863FAM 17 3.9037 6.7200FAM 18 3.4775 6.2398FAM 19 3.9713 6.5377FAM 20 3.8400 6.8650ML AVE. 3.5040 5.7200FAM 1 3.6613 6.3618FAM 2 3.6800 6.0422FAM 3 3.4600 5.4875FAM 4 3.0250 5.0188FAM 5 3.5263 5.9737FAM 6 3.2900 5.8387FAM 7 3.1831 4.8785FAM 8 3.4038 5.5681FAM 9 3.1825 4.6802FAM 10 3.9100 6.3468FAM 11 3.5050 5.8160FAM 12 3.9200 6.2628FAM 13 3.8338 6.0752FAM 14 3.5525 5.4412FAM 15 3.4113 5.9765115Appendix E (continued)SKH AVE. 3.7045 5.9882FAM 1 3.5062 5.8483FAM 2 3.9650 6.2088FAM 3 3.8775 6.1260FAM 4 3.8063 6.5073FAM 5 3.3675 5.2508Disturbed populationKH AVE. 3.5093 5.7343FAM 1 3.0775 5.2963FAM 2 3.8800 6.5940FAM 3 2.9625 4.8375FAM 4 3.8500 6.2435FAM 5 3.6100 6.2110FAM 6 3.4938 5.5878FAM 7 3.4962 5.6438FAM 8 3.6687 5.5375FAM 9 3.5450 5.6575MS AVE. 3. 1288 5.7650FAM 1 3.4638 5.7117FAM 2 3.5512 6.0432FAM 3 2.8738 5.3323FAM 4 3.7188 7.1325FAM 5 3.4013 6.2620FAM 6 3.1388 5.6053FAM 7 3.2512 6.2217FAM 8 2.8225 4.9893FAM 9 2.9943 5.7915FAM 10 3.3563 6. 1063FAM 11 3.4662 6.7655FAM 12 2.9850 5.1958FAM 13 2.3913 4.9705FAM 14 2.9538 5.3356FAM 15 2.7238 5.2590FAM 16 3.2075 5.9940FAM 17 2.8450 5.0707FAM 18 2.7449 5.4008FAM 19 3.5300 6.2710SM AVE. 3.4653 6.1108FAM 1 3.6063 6.3267FAM 2 3.2712 6.0305FAM 3 3.6800 6.2970FAM 4 3.4500 5.8898FAM 5 3.3188 6.0097Appendix E (continued)SK AVE. 3.5556 5.8543FAM 1 3.5487 6.1018FAM 2 3.1700 5.7785FAM 3 3.4925 5.5467FAM 4 3.4562 6.0670FAM 5 3.9005 6.9512FAM 6 3.7612 6.0827FAM 7 3.7975 6. 1480FAM 8 3.7603 5.8307FAM 9 3.5783 5.5698FAM 10 3.2400 5.3647FAM 11 3.7875 5.7330FAM 12 3. 1750 5.0772116117Appendix F Average measures of persistence of axis, stern straightness, and branch heavinessPOPULATION PER STN BROverall mean 2.3531 2.5024 2.8319Undisturbed populationDL AVE. 2.3 150 2.5363 2.7925FAM 1 2.4750 2.7000 2.8500FAM 2 2.6250 2.7500 2.8750FAM 3 2.2750 2.4750 2.8500FAM 4 2.3500 2.6500 2.6750FAM 5 2.4500 2.4750 2.8500FAM 6 2.3000 2.4500 2.8000FAM 7 2. 1750 2.4500 2.6750FAM 8 2. 1750 2.3750 2.9000FAM 9 2.4000 2.5750 2.8500FAM 10 2.3750 2.7750 2.9000FAM 11 2.1750 2.3500 2.8000FAM 12 2.2750 2.5250 2.7500FAM 13 2.2250 2.4250 2.8000FAM 14 2.2000 2.5250 2.8250FAM 15 2.3250 2.5500 2.8500FAM 16 2.3000 2.4250 2.6000FAM 17 2.2250 2.7000 2.8750FAM 18 2.2750 2.3750 2.7750FAM 19 2.4000 2.7500 2.7000FAM 20 2.3000 2.4250 2.6500ML AVE. 2.3859 2.5654 2.8557FAM 1 2.1500 2.3750 2.5500FAM 2 2.5500 2.6000 3.0000FAM 3 2.3750 2.4750 2.8750FAM 4 2.4500 2.4750 2.7750FAM 5 2.2000 2.4500 2.7750FAM 6 2.4000 2.5500 2.8500FAM 7 2.2821 2.6667 2.8974FAM 8 2.4054 2.7297 2.9189FAM 9 2.4750 2.7000 2.9250FAM 10 2.3500 2.6000 2.9250FAM 11 2.5000 2.6500 2.5750FAM 12 2.4250 2.6250 2.9750FAM 13 2.3500 2.4750 2.9500FAM 14 2.5500 2.6750 2.9250FAM 15 2.3250 2.4500 2.9250SKH AVE. 2.4850 2.5950 2.9500FAM 1 2.3250 2.6500 2.8250FAM 2 2.5750 2.7000 2.9750FAM 3 2.5000 2.5250 3.0000118Appendix F (continued)FAM 4 2.3750 2.5000 2.9250FAM 5 2.6500 2.6000 3.0250Disturbed populationKH AVE. 2.4444 2.5194 2.8667FAM 1 2.2250 2.5000 2.8500FAM 2 2.3750 2.4750 2.8500FAM 3 2.3250 2.3500 2.7750FAM 4 2.3000 2.4500 2.8250FAM 5 2.3750 2.5250 2.8500FAM 6 2.4250 2.4750 2.9000FAM 7 2.6250 2.5750 2.8250FAM 8 2.6250 2.4750 2.8750FAM 9 2.7250 2.8500 3.0500MS AVE. 2.2493 2.3886 2.7573FAM 1 2.3250 2.5000 2.6500FAM 2 2.3750 2.6250 2.8250FAM 3 2.3000 2.2000 2.8750FAM 4 2. 1250 2.2500 2.6500FAM 5 2.3250 2.5000 2.9750FAM 6 2.2750 2.3500 2.7750FAM 7 2. 1250 2.4750 2.9000FAM 8 2.1500 2.2750 2.5750FAM 9 2.1000 2.2500 2.8250FAM 10 2.2750 2.4500 2.6750FAM 11 2.3250 2.3250 2.7250FAM 12 2.2778 2.3889 2.8889FAM 13 2.1750 2.3250 2.7750FAM 14 2.2821 2.3590 2.7949FAM 15 2.3000 2.4500 2.6000FAM 16 2.3000 2.4000 2.7500FAM 17 2.2250 2.3750 2.6500FAM 18 2.1538 2.3590 2.7949FAM 19 2.3250 2.5250 2.7000SM AVE. 2.3400 2.4550 2.8800FAM 1 2.2750 2.4750 2.8250FAM 2 2.3750 2.5750 2.9000FAM 3 2.5000 2.5500 2.9250FAM 4 2.4000 2.4250 2.9250FAM 5 2.1500 2.2500 2.8250SK AVE. 2.4208 2.5146 2.8896FAM 1 2.3500 2.5500 2.9250FAM 2 2.3250 2.5750 3.0500FAM 3 2.3250 2.4750 2.7500Appendix F (continued)FAM 4 2.4000 2.6250 2.8750FAM 5 2.3250 2.5000 2.9750FAM 6 2.7000 2.4500 2.9500FAM 7 2.2750 2.4500 2.7750FAM 8 2.4500 2.6500 2.8500FAM 9 2.4750 2.5000 2.9750FAM 10 2.4000 2.4500 2.9000FAM 11 2.5250 2.5500 2.8250FAM 12 2.5000 2.4000 2.8250119120Appendix G Extraction buffer (Liengsiri et al, 1990)Borate pH 8.0 0.05 M MgC12 0.2 %Ascorbic acid 0.01 M CaC12 0.2 %Cysteine-HCL 5.4 mM PEG 20 M 1.0 %Sucrose 0.5 M Tergitol 1.0 %Tween 80 1.0 % B-mercaptoethanol 0.3 %pH adjusted to 8.0 with iN NaOHAppendix H Running buffer systemsSystem Electrod buffer Gel bufferH Tris 0. 125 M Histidine-HCL 0.05 MPitel and Cheliak (1984) EDTA 1.4 mMpH adjusted to 7.2 pH adjusted to 7.2with 1M citric acid with 1 M TrisB Lithiumhydroxide 0.06 M Tris 0.03 MRidgeway ci at (1970) Boric acid 0.3 M Citric acid 0.005 MElectrode buffer B 1 %pH adjusted to 8.1 pH adjusted to 8.5with iN NaOH with iN NaOH1212. Aspartate aminotransferase (AAT) E.C.2.6. 1.1Pyridoxal-5-phosphaseFast blue BB saltAAT substrate solutionNote:Incubated in the dark about 30 mm. at 37oc untilbackground.3. Diaphorase (DIA) E.C. M Tris-HCL, pH 8.02,6-dichiorophenol-indophenolB-NADHMTTNote:Incubated in the dark about 45 mm. at 37 oc25.0 ml125.0 mg38.0 mg150 units0.5 ml0.5 ml0.5 ml4. Esterase/colorimetric (EST) E.C.3. M phosphase buffer, ph 6.4 25.0 mlL-naphthyl acetate 50.0 mgB-naphthyl acetate 50.0 mgFast blue RR salt 100.0 mgNote:L-naphthyl acetate and B-naphthyl acetate were dissolved in 2.5 ml acetone before addingother substrates.:Incubated in the dark about 60 mm. at room temperature until bands appear.5. Isocitrate dehydrogenase (IDH) E. C. 1. 1. 1.420.2 M Tris-HCL, pH 8.0DL-isocitric acid (trisodium salt)1 % MgCI2 (wlv)ADPNBT25.0 ml200.0 mg0.5 ml0.5 ml0.5 ml0.5 mlAppendix I Enzyme staining recipes1. Aldolase (ALD) E.C.4.i.2.30.2 M Tris—HCL, pH 8.0Fructose-1-6-diphosphase (tetrasodium salt)Asenic acid (disodium salt)Glyceradehyde-3-phosphase dehydrogenaseNADMTTPMSNote:Incubated in the dark about 60 mm. at 37 oc until dark blue bands appear.appear on a pink4.0 mg100.0 mg25.0 mlblue bands25.0 ml0.5 mg13.0 mg0.5 mlPMSNote:Incubated in the dark about 60 mm. at 37 oc until dark bands appear.122Appendix I (continued)6. Leucine-amino peptidase (LAP) E.C. 0.5 M boric acid 50.0 ml(b) 0.2 M Tris-HCL, pH 8.0 25.0 ml0.2 M maleic anhydride 25.0 mlMixed and adjusted pH to 5.3 with 1 N NaOH just prior to staining(c) 10 % MgC12 0.5 mlL-leucine B-naphthylamide-HCL 50.0 mgFast black K salt 50.0 mgNote: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.370.2 M Tris-HCL, pH 8.0 12.5 ml0.5 M DL-malic acid, pH 7.0 12.5 mlNAD 0.5 mlNBT 0.5 mlPMS 0.5 mlNote: Incubated in the dark about 45 mm. at 37 oc until dark blue bands appear.8. Malic enzyme (ME) E.C.1.1.l.40Electrode buffer H 12.5 ml0.5 M DL-malic acid, ph 7.0 12.5 ml1 % MgC12 0.5 mlNADP 0.5 mlMTT 0.5 mlPMS 0.5 mlNote:Incubated in the dark at 37 oc until dark blue bands appear.9. Phosphoglucose isomerase (PGI) E.C.5.3.l.90.2 M Tris-HCL, pH 8.0 25.0 mlFructose-6-phosphate (disodiurn salt) 12.5 mgGlucose-6-phosphate dehydrogenase 5 units1 % MgCI2 0.5 mlNADP 0.5 mlMTT 0.5 mlPMS 0.5 mlNote:Incubated in the dark at 37 oc until dark blue bands appear.10. Phosphoglucomutase (PGM) E.C.2.7.5. I0.2 M Tris-HCL, pH 8.0 25.0 mlGlucose-i-phosphate (disodium salt) 250.0 mgGlucose-i ,6-diphosphate solution 0.5 mlGlucose-6-phosphate dehydrogenase 25 units1 % MgC12 0.5 mlNADP 0.5 mlMTT 0.5 mlPMS 0.5 mlNote:Incubated in the dark at 37 oc until bands appear.123Appendix I (continued)11. Shikimic acid dehydrogenase (SDH) E.C.l. M Tris-HCL, pH 8.0 25.0 mlShikimic acid 50.0 mg1 % MgCI2 1.0 mlNADP 0.5 mlNBT 1.0 mlPMS 0.5 mlNote:Incubated in the dark at 37 oc.12. 6-phosphogluconate dehydrogenase (6PGD) E. C. M Tris-HCL, pH 8.0 5.0 ml6-phosphogluconic acid (trisodium salt) 15.0 mg1 % MgCI2 1.0 mlNADP 1.0 mlMTT 1.0 mlPMS 0.5 mlNote:Incubated in the dark about 45 mm. at 37 oc until dark blue bands appear.124Appendix J Allele frequencies and heterozygosity for each locus in eight populationsPopulationLocus KKI KK2 MLI ML2 MK1 MK2 SKI 5K26PGD1(N) 118 100 340 220 80 69 140 601 .915 .900 .997 .864 .994 .717 .846 .8922 .085 .100 .003 .136 .006 .196 .071 .1003 .000 .000 .000 .000 .000 .080 .054 .0084 .000 .000 .000 .000 .000 .000 .029 .0005 .000 .000 .000 .000 .000 .007 .000 .000H(b) .155 .180 .006 .236 .012 .441 .275 .195H(unb) .156 .181 .006 .236 .013 .441 .276 .196H(DC) .169 .200 .006 .264 .013 .203 .179 .2176PGD2(N) 119 100 340 220 80 78 140 601 .592 .630 .688 .727 .944 .692 .604 .7002 .408 .370 .312 .245 .013 .244 .239 .2753 .000 .000 .000 .027 .044 .032 .150 .0254 .000 .000 .000 .000 .000 .032 .004 .0005 .000 .000 .000 .000 .000 .000 .004 .000H(b) .483 .466 .429 .410 .107 .459 .556 .434H(unb) .485 .469 .430 .411 .108 .462 .558 .437H(DC) .664 .740 .618 .527 .112 .615 .643 .600AAT2(N) 80 100 260 140 20 80 100 601 1.000 1.000 .660 .489 .150 .250 .160 .5002 .000 .000 .340 .421 .375 .512 .485 .2253 .000 .000 .000 .089 .475 .237 .355 .275H(b) .000 .000 .449 .575 .611 .618 .613 .624H(unb) .000 .000 .450 .577 .627 .622 .616 .629H(DC) .000 .000 .665 .836 .650 .950 .970 1.000ALD3(N) 100 20 300 220 80 80 140 601 .690 .925 .718 .595 .788 .894 .854 .8172 .305 .075 .282 .357 .131 .038 .050 .0753 .005 .000 .000 .048 .081 .069 .096 .108125Appendix J (continued)H(b) .431 .139 .405 .516 .356 .195 .260 .316H(unb) .433 .142 .405 .517 .358 .196 .261 .318H(DC) .620 .150 .563 .809 .425 .213 .236 .367ALD4(N) 100 40 300 220 80 80 140 601 1.000 1.000 .957 .966 .994 .994 1.000 .9252 .000 .000 .043 .032 .006 .000 .000 .0003 .000 .000 .000 .002 .000 .006 .000 .075H(b) .000 .000 .083 .066 .012 .012 .000 .139H(unb) .000 .000 .083 .066 .013 .013 .000 .140H(DC) .000 .000 .060 .059 .013 .013 .000 .117DIA(N) 120 100 340 220 80 80 140 601 1.000 1.000 1.000 .995 1.000 1.000 1.000 .8422 .000 .000 .000 .000 .000 .000 .000 .1583 .000 .000 .000 .005 .000 .000 .000 .000H(b) .000 .000 .000 .009 .000 .000 .000 .267H(unb) .000 .000 .000 .009 .000 .000 .000 .267H(DC) .000 .000 .000 .009 .000 .000 .000 .217EST1(N) 80 100 340 220 80 80 139 601 .656 .605 .501 .586 .650 .688 .698 .6252 .344 .395 .476 .409 .131 .306 .302 .3503 .000 .000 .022 .005 .156 .006 .000 .0004 .000 .000 .000 .000 .063 .000 .000 .025H(b) .451 .478 .521 .489 .532 .434 .422 .486H(unb) .454 .480 .522 .490 .535 .436 .423 .490H(DC) .688 .770 .956 .773 .600 .613 .561 .750EST2(N) 80 100 320 220 80 80 140 601 .731 .635 .547 .841 .944 .806 .825 .7252 .269 .315 .444 .159 .056 .194 .079 .1333 .000 .050 .009 .000 .000 .000 .096 .142H(b) .393 .495 .504 .268 .106 .312 .304 .437H(unb) .396 .498 .505 .268 .107 .314 .305 .440H)DC) .538 .730 .906 .318 .112 .387 .350 .550126Appendix J (continued)EST3(N) 80 100 320 220 80 80 140 601 .819 .895 .841 .530 .894 1.000 .796 .6832 .181 .000 .058 .470 .106 .000 .204 .3173 .000 .105 .102 .000 .000 .000 .000 .000H(b) .297 .188 .280 .498 .190 .000 .324 .433H(unb) .299 .189 .280 .499 .191 .000 .325 .436H(DC) .363 .210 .300 .532 .188 .000 .293 .567IDH(N) 120 100 340 220 80 80 140 601 .954 .840 .815 .791 .762 .606 .729 .4082 .046 .160 .185 .209 .013 .306 .164 .2423 .000 .000 .000 .000 .131 .000 .000 .0004 .000 .000 .000 .000 .025 .006 .000 .0005 .000 .000 .000 .000 .069 .081 .107 .350H(b) .087 .269 .302 .331 .396 .532 .431 .652H(unb) .088 .270 .302 .331 .398 .535 .432 .658H(DC) .042 .200 .306 .282 .325 .438 .300 .483LAP(N) 119 100 340 220 80 80 140 601 .882 1.000 .909 .832 .475 .500 .332 .8172 .118 .000 .091 .168 .500 .500 .554 .1583 .000 .000 .000 .000 .025 .000 .114 .0174 .000 .000 .000 .000 .000 .000 .000 .008H(b) .208 .000 .166 .280 .524 .500 .570 .308H(unb) .208 .000 .166 .280 .527 .503 .572 .310H(DC) .235 .000 .112 .227 .950 1.000 .607 .300MDH(N) 80 100 300 220 80 80 140 601 .700 .990 .870 .773 .950 .931 .986 .9832 .300 .010 .130 .223 .000 .013 .014 .0173 .000 .000 .000 .005 .050 .056 .000 .000H(b) .420 .020 .226 .353 .095 .129 .028 .033H(unb) .423 .020 .227 .354 .096 .130 .028 .033H(DC) .600 .020 .260 .373 .100 .138 .029 .033127Appendix J (continued)MEl(N) 120 100 340 220 80 80 140 601 1.000 1.000 .910 .925 1.000 1.000 1.000 1.0002 .000 .000 .090 .075 .000 .000 .000 .000H(b) .000 .000 .163 .139 .000 .000 .000 .000H(unb) .000 .000 .164 .139 .000 .000 .000 .000H(DC) .000 .000 .138 .105 .000 .000 .000 .000ME2(N) 120 100 340 220 80 80 140 601 .842 .915 .912 .893 .931 1.000 .986 .8752 .158 .085 .075 .084 .063 .000 .014 .1253 .000 .000 .013 .023 .006 .000 .000 .000H(b) .267 .156 .163 .195 .129 .000 .028 .219H(unb) .268 .156 .163 .195 .130 .000 .028 .221H(DC) .317 .170 .165 .195 .112 .000 .029 .250PGI 1(N) 100 100 320 200 60 80 140 601 .965 1.000 1.000 .983 1.000 1.000 .993 1.0002 .035 .000 .000 .018 .000 .000 .007 .000H(b) .068 .000 .000 .034 .000 .000 .014 .000H(unb) .068 .000 .000 .034 .000 .000 .014 .000H(DC) .070 .000 .000 .035 .000 .000 .014 .000PGI2(N) 120 100 320 200 60 80 140 601 .833 .905 .939 .930 1.000 1.000 .961 .9422 .167 .095 .059 .070 .000 .000 .039 .0583 .000 .000 .002 .000 .000 .000 .000 .000H(b) .278 .172 .115 .130 .000 .000 .075 .110H(unb) .279 .173 .115 .131 .000 .000 .076 .111H(DC) .333 .190 .119 .140 .000 .000 .079 .117PGM2(N) 120 100 340 220 80 80 140 601 .975 1.000 1.000 .875 1.000 .969 .879 .5922 .025 .000 .000 .125 .000 .031 .118 .1503 .000 .000 .000 .000 .000 .000 .004 .0174 .000 .000 .000 .000 .000 .000 .000 .242128Appendix I (continued)H(b) .049 .000 .000 .219 .000 .061 .214 .569H(unb) .049 .000 .000 .219 .000 .061 .215 .574H(DC) .050 .000 .000 .205 .000 .063 .014 .300SDH2(N) 120 100 335 220 80 80 140 601 .746 .865 .575 .468 1.000 1.000 .993 .9422 .254 .105 .248 .259 .000 .000 .000 .0583 .000 .030 .178 .273 .000 .000 .007 .000H(b) .379 .240 .577 .639 .000 .000 .014 .110H(unb) .381 .241 .578 .641 .000 .000 .014 .111H(DC) .508 .270 .594 .532 .000 .000 .014 .083H(b) = Heterozygosity (biased)H(unb) = Heterozygosity (unbiased)H(DC) = Heterozygosity (direct counted)129Appendix K Outcrossing rate estimates (t) for individual families from eight populationsPopulation t PopulationUndisturbed populationML1Fam 1 1.97 Fam 15 1.97Fam 2 1.97 Farn 16 1.97Fam 3 1.97 Fam 17 0.90*Fam4 1.97 ML2Fam 5 1.97 Fam 1 0.95*Fam6 1.97 Fam2 1.95Fam7 1.97 Farn3 0.91*Fam 8 1.97 Farn 4 1.95Fam 9 1.97 Earn 5 1.00*Fam 10 1.97 Farn 6 1.95Fam 11 1.97 Fam 7 1.95Fam 12 1.97 Farn 8 1.95Fam 13 1.37 Fam 9 1.95Fam 14 1.97 Fain 10 1.95Disturbed populationKK1 MK2Fam 1 1.01 Fain 1 1.93Fam2 1.94 Fam2 1.93Fam3 1.94 Fam3 0.93*Fam 4 1.94 Farn 4 0.95*FamS 1.94 SKiFam 6 0.20* Fam 1 1.91KK2 Farn2 1.91Fam 1 1.07 Fam 3 1.91Fam2 2.00 Fain 4 1.91Fam 3 2.00 Fam 5 1.91Fam4 2.00 Fam6 Q79*FamS 2.00 Farn7 0.91*MK1 SK2Fam 1 1.11 Fain 1 1.97Fam 2 2.00 Fam 2 1.97Farn 3 2.00 Fain 3 1.97Fam4 0.69** Family with outcrossing rate 1.0 or smallerCD > CD C CD C CD 0• - CD - CD CD j) C CD (10 C C.L.)—Q’.c.i.)———‘aQo-C’L.L’J——‘C00C’tJ—7’——1C’c,c-—c—CI-.)——.——to———————k)—(0 a_——()t-.)———D—)L-)1-)tJI-.)I-.)D-———t—)-.)C-——--)1-_)—--)t.,)0---D-)—--)t-JL-.)—0————(‘Jt-_)I_J—t-I-JI-Itkt-lI)(--I)———L--)———-—-L)t’-)—(—)I-J—I->)(—)L.tJ——(-I—)——(-—————(-a———(-J.)—(-J——t-Jt->————————tJ———--—I_)I--)——L-(.(()—k)—-”--———I_)I)l-)——(_)t)—————(_)————L.a———().—.—I-.)—4i(_)l_)——I-——-Jk)k)I——t)L.——)————t-)———I——I-)-3—I)—I-———I—)L-JI(J—(-)————I-_)——————J——i-)———“--I_J-(->k)-—U——0( 0 00 a. a C..)a a C U a I-.)C, C, U G I—)G 00 a I’-)C) C., 0 00 U 0 00 0 00 C.-)Cç_(/———————————————————CEEEEEEEEEEEE————————13—————————————il)k))3)——$.)3k)———L’3l)L’3(3(9(9(9(9(9(9(9(9(9——t)1331313($—(9I3—--($-—-l3I3——t3--1313—13——131313—1313——13—————1313I3I3—l3($($(3(3—1313(9—————-—1313(3—13———13—


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items