vi... .5! 1 5.11. 1!; f. .x...a.... . 293-: .§. )1! 1:- 0. .1 Iv! a! 8.7.. it. . £1 22‘}... awnéfimafina em 451:? t '5‘“ ”Err—u” p. ‘91... 3. 5.... :1: :3. .. x. , . , 1.... 5...? 3mm...) . .. i. , it ... .. was.» 2% R. if? Ema... : 32%. . an“? . 5.... ....... 1.5... 1.21:3? . an... . 3.x}... Vt Hana“... a ; hi ‘ if}: S. Lfifio: ... .........a..n. Jo: . :unfikzflvi . and. $1.» in. .fii . , .53. 1:... 5‘1: I v z. 4: . i ,uK)». a)! ‘Xu Alan. , .zmmwmw luau“. .3. {muniulliv .x t it} «Hui Eat $3. . {‘4 . Pia-ii; .v .1! .77 .4 v. 7‘ 1:1! 111-9311: Al... .. vfiflgu . 1:1! mmfiéu . 1|?! L2. flag {1! I'll ; h.{l..t§£§f .(‘ta: :9: a V1? I: an '53!- .mmum‘eix1vdmmu 655... u . 3153.”. 3. t .. it In. 3 3%}? _. . s .1 . . -3 l . ...vu~lvi: .uotoldnciifllnavné .4 . .. a- Elia . l! mu... :1 l ”EM?! lb VI a.“ . v ... 10.“. a ‘ 543-5.! ......-.:: 2 r...” .uvxu.‘ n. 1.1.01. ’0'. \VKKI. A... u D IQ... This is to certify that the thesis entitled SPATIAL PATTERNS OF GENETIC VARIATION: OLD GROWTH WHITE PINE presented by PAULA E . MARQUARDT has been accepted towards fulfillment of the requirements for M. S . 11 . FORESTRY egree 1n :3 Major professor Date MM 0—7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 c:/ClRC/DateDue.p65-p.15 w..- _ .——_ SPATIAL PATTERNS OF GENETIC VARIATION: OLD GROWTH WHITE PINE BY Paula E. Marquardt A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Forestry 2002 ABSTRACT SPATIAL PATTERNS OF GENETIC VARIATION: OLD GROWTH WHITE PINE BY Paula E. Marquardt Natural populations of tree species are dynamic systems: population demographics such as age, density and dispersal distance interact with evolutionary processes to determine spatial genetic structure. Old growth and second growth populations were evaluated, located within Hartwick Pines State Park, Grayling, MI. From each population, 120—122 contiguous trees were sampled for genetic analysis at seven SSR (simple sequence repeat) DNA loci. Genetic diversity was high and inbreeding levels were low for both populations. There was little divergence in allele frequency between populations. Spatial autocorrelation analysis suggested that individual genotypes were randomly distributed in the second growth population whereas weak positive spatial structure was observed at short distances for individuals in the old growth population. Logging may have decreased the degree of spatial structuring at the second growth site, suggesting that silvicultural practices may alter “natural" spatial patterns. Spatial structure is the main factor increasing levels of biparental inbreeding. DEDI CATION To Emily and Jennie in ACKNOWLEDGMENTS I appreciate my major professor, Dr. Bryan Epperson for his advice, guidance, and support in conducting this research and preparing this thesis. I also thank the rest of my thesis committee, Dr. Daniel Keathley and Dr. Kim Scribner, for their helpful comments in the final version of the manuscript. I am grateful to Dr. Jud Isebrands, Dr. Mark Kubiske, and Dr. Charles Michler, all of the USDA Forest Service, for their support throughout all aspects of my Masters program. The USDA Forest Service RWU-NC—4152 and 4155, and the McIntire - Stennis Cooperative Forestry Research Program (project #1774), funded this project. I thank Ann Stephens, of the Hartwick State Park, for her help and aid in collecting sample materials. I thank my daughter Jennie May, our students Jenny Dionne and Rebecca McFadden, and my coworker Ray Lange for helping to maintain my literature database and reprint files, and for various other thesis related support. I thank my friends for their support of my research and studies, in particular Joeseph Zeleznik and Kenichro Shimatani, for their help in sample collection, and Robert Bloye, Eric Myers and Erin Smith, for their discussions and moral support. I gratefully appreciate my friends and iv colleagues, Dr. John Erpelding and Dr. Paul Anderson, for their encouragement and mentorship. I am extremely grateful to my friend, Jerry Rayala, for his support and encouragement during the final few months of my thesis and defense preparations. Finally, I would especially like to thank my daughters, Emily and Jennie, for their love and confidence, and for standing by me through the difficult times. TABLE OF CONTENTS LIST OF TABLES ................................... viii LIST OF FIGURES .................................... Xi INTRODUCTION ........................................ 1 LITERATURE REVIEW ................................... 7 Genetic Diversity .............................. 7 Methods Of Measurement ......................... 8 Spatial Processes ............................. 12 Population And Spatial Genetic Structure ...... 15 Maintenance Of Genetic Diversity .............. 19 METHODS ............................................ 24 Site Description .............................. 24 Sampling ...................................... 29 DNA Isolation ................................. 34 Marker Analysis ............................... 34 Quality Control ............................... 39 Statistical Analyses .......................... 41 General Diversity ........................ 41 Population Structure ..................... 43 Spatial Structure ........................ 45 RESULTS ............................................ 50 Characterization Of Loci ...................... 50 Polymorphism ............................. 50 Allele Frequency ......................... 56 Perfect Repeat Diversity ................. 59 Characterization Of Populations ............... 61 Allele Frequency And Class ............... 61 Genetic Diversity ........................ 64 Population Structure ..................... 64 Spatial Structure ........................ 67 DISCUSSION ......................................... 72 Characterization Of Loci ...................... 72 Polymorphism ............................. 72 Allele Frequency ......................... 75 Perfect Repeat Diversity ................. 77 Null Alleles ............................. 82 Locus Rps6/34b ........................... 89 vi Characterization Of Populations ............... 91 Allele Frequency And Class ............... 91 Genetic Diversity ........................ 94 Population Structure ..................... 96 Spatial Structure ........................ 99 CONCLUSION ........................................ l O 3 APPENDIX .......................................... 1 O 5 Spatial Autocorrelation Coefficients ......... 106 LITERATURE CITED .................................. 119 vfi Table LIST OF TABLES Page Primer sequences for 8 nuclear (CA)n simple sequence repeat markers. Forward then reverse sequences are listed for each primer pair. Rps6 and Rps34b amplify the same locus ........ 36 Seven nuclear (CA)nsimple sequence repeat (nSSR) loci characterized for mature white pine. A total of 242 trees were assayed. Reported are numbers of chromosomes sampled, observed number of alleles per locus (k), observed allele sizes in base pairs (bp), change in number of base pairs between the smallest and largest alleles, amplified size of most common allele in basepairs (bp), and frequency of most common allele (Xi) ......................................... 51 a = OG, b = SS. Allele frequencies measured in two mature populations of Pinus strobus. The number of individuals sampled was 122 and 120 individuals in CG and SS respectively. Alleles are ordered according to size. The smallest allele is listed first (allele 1) ............................................ 57 Repeat sequence, category, and number of repeat units (m), in seven eastern white pine nSSR markers .................................. 60 Diversity measures for four nSSR perfect CA repeats: number of repeat units (u), Nei's genetic diversity (He), observed heterozygosity Ho, observed number of alleles per locus (k), and effective number of alleles (A9) .......................................... 6O vfii 10 Distribution of allele frequencies in mature white pine populations: total alleles, common and uncommon alleles, low frequency alleles, and rare alleles. The haploid sample size was 244 and 240 for CG and SS respectively. There were seven loci sampled. The total number of alleles present in both populations was 54. The number of alleles present in each class is in parentheses ................................. Private alleles and common alleles present in two mature populations of white pine. The haploid sample size was 244 and 240 for 0G and SS respectively. There were seven loci sampled. The number of alleles is in parentheses .................................... General diversity measures for Old Growth and Second Growth Pinus strobus: number of trees sampled, percentage polymorphic loci, total number of alleles, allele richness (A), effective number of alleles per locus (As), observed heterozygosity (Ho), and expected heterozygosity or Nei’s genetic diversity (He). Standard deviations are in parentheses ......... Population indices F and Fm_for each nSSR locus and as a mean across all loci as measured in two populations of white pine. Seven loci were sampled. Sample size was 122 and 120 individuals in OG and SS, respectively ......... a = old growth (OG), b = second growth (SS). Mean spatial autocorrelation coefficients (Moran’s I) in two mature populations of Pinus strobus for 9 distance classes, 2n = 2441 for 0G and 2n = 240 for SS ......................... ix 62 63 65 66 69 11 12 13 14 a = CG and b = SS. Markers RpsG and 34b amplify the same locus. Characterized are the large allele heterozygotes for Rps6 (alleles 177, 179, 180 and 186) with the corresponding genotypes for Rps34b where the large allele is a null allele. Rps34b homozygotes are 17 bp smaller than the small Rps6 alleles ............ 88 Estimating the rate of miscalls for the locus defined by markers Rps6 and Rps34b. Characterized for the Second Growth (SS) population are the homozygous genotypes scored for marker Rps6 that present heterozygous genotypes when scored for marker Rps34b. Rps34b peak 1 alleles are 17 bp smaller than Rps6 peak 1 alleles. The alleles with discrepancies are in bold typeface. ........... 90 Spatial autocorrelation coefficients (Moran's I) for 8 loci in the Old Growth population (OG) of Pinus strobus for 10 distance classes, n = 1221 ........................................... 107 Spatial autocorrelation coefficients (Moran's I) for 8 loci in the Second Site population (SS) of Pinus Strobus for 10 distance classes, n = 1201 ........................................ 113 Figure LIST OF FIGURES Page Map of the Old Growth (OG) and Second Growth (SS) study sites in the Hartwick Pines State Park in Crawford County, Michigan ............. 26 Diameter at breast height distributions for both adult populations of white pine at Hartwick Pines State Park. (mean i 1 SD) ..... 28 a = Old Growth (OG) and b = Second Growth (SS). The spatial distributions of individual trees located within two mature populations of white pine. Study sites were approximately 1.2 hectares in size. The distance between successive tick marks is 20 meters ............ 33 a = Rpslb, b = Rps2, c = Rps6, d = Rps39, e = RpsSO, f = Rp584, and g = Rp5127. Distribution of allele frequencies detected at seven nSSR loci in 2 mature populations of white pine .................................... 52 Mean Moran's (I) indices measured spatial structure in two populations of white pine. The I values for the first three distance class were plotted ............................ 71 Sequence data for two cloned white pine fragments containing (CA)H repeats. Marker Rps6 was derived from the plasmid clone pPs6 and Rps34b from pPs34 with a 17 bp difference in sequence size. Forward and reverse primers are underlined. The repeat sequences are in bold typeface. Lower case letters indicate differences between the sequences ............. 84 a = Rps6 and b = Rps34b. Displayed are the allele frequency distributions for markers Rps6 and Rps34b. These markers amplify the same (CA)nlocus, with Rps 34b having four null alleles corresponding to the Rps6 alleles 177, 179, 180 and 186 .............................. 85 xi INTRODUCTION Natural populations of tree species are dynamic and evolving systems influenced by numerous genetic and environmental factors. Population genetics examines changes in gene frequencies on the shortest of evolutionary time scales. Furthermore, species can be subdivided into local breeding units, which allows us to evaluate the amount and distribution of genetic diversity. Population demographics such as age, density and dispersal distance interact with evolutionary processes such as genetic drift, natural selection, mating system, gene flow and mutation to determine spatial structure, i.e. the spatial distribution of genotypes and their degree of relationship. Because the magnitude and direction of correlation between distance and genotype is influenced by stand demographics, silvicultural practices such as thinning, harvesting, or no treatment can alter stand level spatial patterns. Spatial structure in turn influences other aspects of population genetic structure such as inbreeding. An understanding of the level of genetic diversity, how it is partitioned, and the spatial patterns of its distribution in both natural and managed stands, will aid foresters in making informed decisions about stand management and reforestation. Thus, population genetic evaluation provides them with additional tools for maintaining genetic diversity and reducing inbreeding. For example, spatial structure can help anticipate bi-parental inbreeding levels and quantify population genetic structure since limited seed and pollen dispersal may result in mating between related individuals. Eastern white pine (Pinus strobus L.) is a species of both historical and economic value in the northeastern United States and Canada. It graces the landscape of our public and private forests with aesthetic beauty and is a valuable tree species for the lumber and paper industry. Despite its importance, little is known about the population and spatial genetic structure of white pine, its mating system, and how its genetic diversity is affected by changes in the forest ecosystem. Although white pine is primarily out-crossing, self- fertilization and bi—parental inbreeding can occur, resulting in a reduction in genetic diversity from values expected under random mating. The loss of genetic variation may reduce the fitness of the present population adapted to the current or past environmental conditions that have contributed to the development of the population. Moreover, loss of genetic diversity may lead to the loss of latent genetic potential, which could degrade adaptation to future environmental change, and sustainability of the forest resource. Future management will need to consider potential losses to the diversity of regenerating progeny when designing and applying silvicultural prescriptions to the parent population. One major concern is the loss of alleles present in low frequency (i.e., rare alleles). Rare alleles could be a crucial class of alleles for natural selection to act upon for future adaptation. Human activities have fragmented the landscape, introduced exotic pests, and created environmental changes such as pollution, which alters environmental chemistry (acid rain), the earth’s temperature and precipitation patterns (global warming), and plant physiology (rate of plant growth, leaf chemistry affecting nutrition and defense to insects, and tree health). All of these changes in the forest ecosystem will require adaptive changes on the part of natural populations in order for life to continue to evolve and exist. The design of effective management strategies can be facilitated by the use of population and spatial genetic structure to predict the effects of silvicultural treatments on genetic variation. Silvicultural systems can minimize inbreeding and maximize genetic diversity by three methods: 1) ensuring the effective breeding population is large enough to retain sufficient numbers of pollen and seed parent trees representing most gene variants including sampling of gametes containing rare alleles; 2) minimizing related neighborhoods by maintaining proper spacing and reducing patchiness by keeping migration routes open; and 3) keeping stand densities high enough to ensure cross- fertilization while reducing self-fertilization and mating between related trees. Because of its economic value, white pine will be more actively managed and manipulated in the future to improve growth, form, and disease resistance. Therefore it is imperative to study evolutionary relationships before they are lost with domestication. Of equal importance to its value as a crop, is its value for habitat, aesthetics, and history when maintained as old growth forests. Although the old growth in the United States is not in danger of becoming extinct, it has been made extremely fragmented, and if not managed using sound population genetic principles, losses in the unique variation found in these untouched gene pools will occur. Therefore, this study was conducted to offer preliminary insight into identifying appropriate resource management practices for maintaining levels of genetic variation characteristic of old growth forest ecosystems. The study involved the evaluation of two P. strobus populations located within Hartwick Pines State Park, near Grayling, Michigan. The old growth population is approximately 240 years old and has been unmanaged, except for fire exclusion. The second growth population regenerated naturally after being harvested in the late 1800’s. Both populations were minimally disturbed. From each population, 120-122 contiguous trees were sampled for genetic evaluation at seven SSR (simple sequence repeat) DNA loci. The objectives in this study were to: 1) assess fine scale genetic diversity, population structure, and underlying spatial genetic patterns in the populations; and 2) explore the possible effects of logging on the spatial genetic structure of a naturally regenerated second growth stand as compared to a virgin old growth stand. Specific research hypotheses were: 1) There will be little genetic difference between the two populations of white pine, i.e. they will have similar allele frequency distributions. 2) Genetic diversity will be high for both populations. 3) The second growth population may have less genetic diversity than the old growth population. 4) Population genetic structure and spatial genetic structure will be weak for both populations. 5) Spatial patterns will be different for the two study populations. L I TERATURE REVI EW Genetic Diversity Plants are genetically very diverse. For allozymes, average heterozygosity within populations (He = 0.11) is greater than that for invertebrates (He = 0.10) or vertebrates (He = 0.05) (Hamrick & Godt 1989). Gymnosperm is one of the most genetically diverse group of plant taxa (He = 0.15) (Hamrick et a1. 1992). Eastern white pine is quite variable among individual trees and across its geographic range. Its diversity is maintained by mechanisms favoring an outcrossing mating system and other components of the pine genetic system i.e. the reproductive, recombination and mutation systems (Ledig 1998). There is abundant variation in white pine for morphological characters (Beaulieu & Simon 1995) in provenance seed trials (Pauley et al. 1955)(Mergen 1963)(Fowler & Heimburger 1969)(Wright 1970)(Li et al. 1997) allozyme loci (Eckert et a1. 1981)(Eckert & Ryu 1982)(Brym & Eckert 1986)(Beaulieu & Simon 1994a)(Chagala 1996), DNA loci (Echt & Nelson 1997), quantitative traits (Cornelius 1994), and secondary plant compounds (Smith et al. 1969)(Zavarin et al. 1969). High heterozygosity is found for allozyme loci in eastern white pine (Beaulieu & Simon 1994b)(Buchert et al. 1997)(Rajora et al. 1998)(Epperson & Chung 2001), and for DNA loci (Echt et a1. 1996)(Rajora et al. 2000). Methods Of Measurement Genetic evaluation with all marker techniques suffers limitations. Morphological traits are common first descriptors of phenotype, but often show little variation and can be difficult to measure. Quantitative traits are used for the genetic improvement of forest trees. They are often polygenic (Hamrick et al. 1992) so do not provide simple allele frequency data for calculating genetic diversity parameters. Early biochemical methods used secondary plant compounds such as terpenes to measure genetic variation in forest trees, but their genetic basis was and remains poorly understood (Hamrick et al. 1979). In addition, these three genetic systems (morphology, quantitative traits, and secondary chemistry) are influenced by environmental heterogeneity, and this decreases their usefulness in population genetic studies. Because estimates of genetic diversity and genetic structure are based on gene frequencies, the most appropriate marker systems are those that determine allele frequencies directly. More recent molecular based biochemical techniques accomplish this objective by surveying for Mendelian variation in protein expression (allozymes) or at the DNA level. In addition, DNA markers are regarded as selectively neutral, serving as valuable measures of gene flow. High levels of genetic variation are correlated with plant life history traits (Hamrick et al. 1979)(Hamrick et al. 1992). White pine is highly diverse, and combines many of the traits associated with greater diversity: gymnosperm taxonomic status, long lived perennial, widespread geographic range, wind pollination and seed dispersal, high chromosome number, long generation times, sexual reproduction, high fecundity, and late successional status. One of the more discriminating molecular based marker systems is required for genetic analyses of eastern white pine, where high levels of variation are anticipated. Allozyme studies (Hubby & Lewontin 1966) are simple, inexpensive, and robust, but are limited by few enzyme systems, tissue specific expression, low levels of variability for some species, and variation revealed only in protein coding genes. In addition, they often require destructive sampling and therefore are not useful for evaluating endangered species or other populations when study organisms cannot be sacrificed i.e. mark and recapture studies. Allozymes evaluate the products of gene expression, but it is better to evaluate genetic diversity at the DNA level, which offers greater polymorphism. Four popular DNA based marker systems that overcome the limitation of few variable loci, include RFLPs (Restriction Fragment Length Polymorphism), RAPDs (Random Amplified Polymorphic DNA), AFLPs (Amplified Fragment Length Polymorphism), and SSR's (Simple Sequence Repeat). RFLP’s (Botstein et al. 1980) examine size differences among restriction fragments of DNA, where differences in banding patterns reflect genetic differences. RFLP’s are simple, reproducible and co—dominant (heterozygotes can be distinguished from homozygotes), but require large amounts of high quality DNA, and are time consuming and expensive. The RAPD technique (Williams et al. 1990) is fast, simple, and an inexpensive PCR (Polymerase Chain Reaction) based marker system (Mullis & Faloona 1987), using arbitrary primers and little DNA. However, it suffers from problems of reproducibility. AFLP’s (Vos et al. 1995) combine the RFLP and PCR based marker systems, creating highly informative fingerprints. Although AFLP markers are sensitive and reproducible, they are expensive to develop and technically demanding to use. Another shortcoming of both RAPDs and AFLPs is that they are dominant marker systems. Since a dominant gene (A) is expressed whether a 10 tree is homozygous (AA) or heterozygous (Aa) for the gene, dominant markers are unable to distinguish between these two genotypes, making them decidedly less powerful for population studies. For this reason, SSRs i.e. microsatellite markers (Litt & Luty 1989)(Weber & May 1989) are attractive for stand level population studies since they are co-dominant markers. Microsatellite sequences are preferentially located in non-coding regions of the genome, within introns or between genes (Weber & May 1989)(Hancock 1999) and are assumed to be neutral markers. Moreover, microsatellites are single locus markers with high mutation rates, offering high levels of reproducible polymorphism. Length variations in microsatellite sequences can easily be detected from small amounts of DNA by PCR amplification of the repeat region with unique flanking primers. Microsatellite markers do have some limitations. They are expensive and time consuming to develop, technically demanding, and because of their sensitivity, suffer from cross-contamination problems. PCR primers bind a specific target; therefore, cross-contamination between species would not be a concern unless they were very closely related. Generally, sources of cross-contamination would be intra—specific (mixing tissue samples from the same 11 species) or from PCR products (previous amplification reactions with complimentary priming sites). These shortcomings do not negate the strengths and usefulness of microsatellites. Thus, they remain a popular marker system for population genetic studies, which evaluate allele frequency changes on a micro-evolutionary scale. Since successful amplification is based on relatedness (Fields & Scribner 1997) (Echt et al. 1999) time scale needs to be considered when addressing macro-evolutionary questions. For eample, microsatellites did not amplify across more distantly related species in pine (Echt et al. 1999)(Karhu et al. 2000)(Mariette et al. 2001). An exception is when they occur in highly conserved loci, which has resulted in successful cross amplification for more closely related species in various taxa, including waterfowl (Fields & Scribner 1997), salmon (Scribner et al. 1996), mammals (Moore et al. 1991), whales (Schlotterer et al. 1991), and primates (Blanquer-Maumont & Crouau-Roy 1995)(Garza et a1. 1995), as well as in pines (Echt et al. 1999). Spatial Processes A major assumption in population genetics theory is that individuals close to one another in space are more genetically alike (Wright 1943). Therefore, individual genes in populations may not be randomly distributed, but 12 distributed spatially in a patchy structure. Population geneticists examine changes in allele frequencies among populations, the spatial patterns of these genes, and the micro—evolutionary forces and population demographics that determine genetic patterns of spatial variation. Papers that emphasize population genetic inferences include Sokal & Oden (1978a), Sokal & Oden (1978b), and Sokal et al. (1989). Spatial autocorrelation analyses test for the presence, sign and strength of spatial structure in ecological and genetic data. These methods also provide description by predicting the underlying cause of structure based on the shape of the correlogram i.e. plot of correlation coefficients (y) against distance (x) (Sokal & Oden 1978a)(Sokal 1979)(Legendre & Fortin 1989). Two common correlation coefficients used by geneticists for estimating spatial autocorrelation are the Moran’s I index and the join counts (Sokal & Oden 1978a)(Sokal 1979)(C1iff & Ord 1981). The most popular is the Moran's coefficient (Moran 1950) for interval data. Genotypes are converted to allele frequencies (0 if the genotype doesn’t contain the allele, 0.5 if it is a heterozygote, and 1.0 if it contains two copies). The analysis of nominal data (individuals 13 representing different genotypes) is performed using the join counts. Many interacting factors may underlie observed patterns of spatial variation in allele frequencies. Stand demographics such as low stand density, short dispersal distances, and population age; all increase spatial genetic structure. Low stand density serves to enhance spatial structure by increasing homozygosity through the effects of inbreeding and genetic drift. This is accomplished in three ways: 1) decreasing the effective population size; 2) increasing self-fertilization and mating between relatives; and 3) decreasing gene flow (fewer migration routes). Short distance seed dispersal increases spatial structure through the formation of family groups (Schnabel & Hamrick 1990)(Berg & Hamrick 1995). Spatial structure can either increase or decrease with population age. Spatial structure builds through successive generations of consanguineous mating (Schnabel & Hamrick 1990). This generational effect results in the newest (youngest) generation having more positive structure (larger groups of related individuals) than older generations. Over time, structure increases in the population as a whole. In comparison, if the genetic load is high, natural selection will remove spatial structure. Genetic load is the number 14 of deleterious recessive genes carried in a population. When inbreeding brings two copies of a deleterious recessive gene together in one individual, this homozygous individual will have lower fitness, and may be removed by natural selection. Because natural selection has more opportunity to eliminate inbred individuals from the population as it ages (Yazdani et a1. 1985), structure will decrease over time, with the older trees being less structured than younger age classes. As mentioned above, the evolutionary forces of genetic drift, mating system, gene flow, and natural selection all interact with population demographics, influencing the patterns of spatial variation. Mutation is the final evolutionary force to consider. High levels of mutation and recombination will decrease spatial structure through the introduction of novel genotypes. Finally, in other mating system examples, it has been shown, that small amounts of clonal reproduction increases spatial structure (Schnabel & Hamrick 1990) (Berg & Hamrick 1995) (Chung & Epperson 1999). Population And Spatial Genetic Structure The genetic system of conifers is both remarkable and complex. Three separate genomes with three different modes of inheritance provide a model system for investigating 15 population and evolutionary genetic questions. Nuclear DNA markers follow Mendelian inheritance, whereas haploid organelle inheritance occurs primarily through one parent. The chloroplast DNA (chNA) is passed through the male lineage (pollen), whereas the opposite, sole maternal inheritance occurs for mitochondrial DNA (mtDNA) through seed (Stine & Keathley 1990)(Wagner 1992)(Dong & Wagner 1994). Because mutation rates, dispersal distances, migration rates and effective population sizes interact and vary among the genomes, measures of diversity and population and spatial structure also vary accordingly. For example, the haploid nature of the mitochondria and chloroplast organelles would tend to create more structure than the nuclear genome. The smaller effective population size decreases diversity through increases in genetic drift, and inbreeding, although high levels of gene flow can disrupt and reduce this structure. In three studies evaluating organelle and nuclear markers in pine, it was evident that the mitochondrial genome has more structure than either the chloroplast or nuclear genomes. Population differentiation for maternal mtDNA markers is greater than for the more similarly structured paternal chNA markers and nuclear allozymes. Differences in allele frequencies among 16 populations were 68% for mtDNA, 1% for chNA, and 2% for the nuclear allozymes (Latta & Mitton 1997). These results are consistent with two other evaluations that showed mtDNA differentiation varied from 87-93% (Wu et al. 1998) and chNA ranged from 2—4% (Dong & Wagner 1994). The authors conclude the difference between the mitochondrial, and the chloroplast and nuclear population structure measures, indicate evidence of high gene flow through wind-dispersed pollen. Additional allozyme studies have also found little relative population divergence, with just 1.5 — 7.6% of diversity partitioned among pine populations. Ffl_values were 0.024 for P. rigida (Guries & Ledig 1979, 1982), 0.04 for P. pondersosa (Linhart et al. 1981), from 0.015 — 0.061 for P. strobus (Beaulieu & Simon 1994a)(Rajora et al. 1998)(Epperson & Chung 2001), 0.044 for P. thunbergii (Kim et al. 1997), 0.076 for P. pinaster (Salvador et al. 2000), and 0.021 for P. brutia (Panetsos et al. 1998). In addition to reducing population structure, high gene flow also reduces spatial genetic structure and minimizes biparental inbreeding. Pines primarily outcross through wind-dispersed pollen. Inbreeding levels in non- isolated natural populations of adult pine are typically close to zero (Guries & Ledig 1979, 1982)(Linhart et al. 1981)(Yazdani et al. 1985)(Beaulieu & Simon 1994a)(Kim et 17 al. 1997)(Panetsos et al. 1998)(Rajora et al. 1998)(Epperson & Chung 2001). However, structuring of genotypes and inbreeding can still occur over time due to limited seed dispersal distances and mating between relatives. Spatial structure would negatively affect the population through inbreeding induced decreases in genetic diversity, which in turn may reduce the fitness of the population. Conifers in general have a high genetic load i.e. number of recessive lethal genes (Ledig 1998). Therefore eastern white pine, like other conifers, suffers from strong inbreeding depression (Johnson 1945)(Fowler 1965), resulting in slower growth, reduced vigor and greater chlorosis. Other effects of inbreeding in pine are aborted embryos, low germination, stunted seedlings and reduced survival (Williams & Savolainen 1996)(Wu et al. 1998). In the only mating system study of eastern white pine (Beaulieu & Simon 1995c) the mean single locus outcrossing rate was close to 1.0 in two old growth populations, suggesting absence of self-fertilization and consanguineous matings (inbreeding resulting from mating between relatives). These results are consistent with average single locus estimates reported in lodgepole pine (Epperson & Allard 1984). In comparison, outcrossing rates were 18 slightly lower for western white pine with a mean single locus estimate of 0.94 (El—Kassaby et al. 1993) indicating possible spatial family structure, a major factor in biparental inbreeding. Few studies have examined the spatial genetic structure of pines within stands. Although spatial genetic structure is expected to be weak in continuous populations, it can be the most important factor controlling levels of inbreeding. Epperson & Allard (1989) reported lodgepole pine (P. contorta ssp. latifolia) genotypes to be nearly randomly distributed suggesting absence of patch structure in the presence of high gene flow. These findings are consistent with Knowles (1991) spatial evaluation of old growth black spruce (Picea mariana (Mill.) B.S.P.). In comparison, Epperson and Chung (2001) provided evidence for reduced structure in old growth white pine, with weak positive structuring of alleles at 15 m (Moran's I = 0.05). Maintenance Of Genetic Diversity The sustainability of genetic variability will be instrumental in the ability of forests to adaptively respond to the often rapid, human induced changes in the environment. For example, pollution causes changes in plant physiological processes such as rate of tree growth, (Isebrands et al. 2001) photosynthesis (Kubiske & Pregitzer l9 1996)(Kubiske et al. 1997)(Takeuchi et al. 2001) tree health (Rundel & Yoder 1998) the introduction of exotic pests and changes in tree defenses to insects (Lodge 1993)(Niemela & Mattson 1996)(Mattson 1998) and leaf chemistry affecting nutrition and levels of insect herbivory (Herms et al. 1996). However, range expansion through plant migration in addition to adaptation has been crucial to the ability of plants to respond to past climate change (Davis & Shaw 2001). It is likely that habitat fragmentation, again a result of human activity, will interfere with plant migration (Lodge 1993)(Davis & Shaw 2001). Moreover, future predicted rates of climate change are much more rapid than historical climate changes of the past (Etterson & Shaw 2001). Together, these two factors may not provide adequate opportunity for plant adaptation to occur through range shifts and refugia. Examining the changes in gene frequencies of natural populations in response to the current changes in climate and habitat fragmentation is an exciting and relatively untouched area of research. Previous management goals have been focused at the community or species levels. Now we have the tools necessary to manage populations at the genetic level, for the sake of maintaining species health in the face of a continually changing environment. 20 Wright (1943) summarized the levels of neighborhood differentiation that could be anticipated in a continuous population as a result of isolation by distance only, from neighborhood size and predicted levels of seed and pollen dispersal associated with various mating systems. Wright's (1946,1969) definitions of neighborhood, neighborhood area, and neighborhood size are useful in understanding potential applications of spatial autocorrelation measures to stand management. Wright defined a neighborhood as the surrounding area from which parents are likely to have been drawn. Wright's neighborhood may not be panmictic, i.e. all male—female pairs have equal probability of mating. Although a population may be distributed over a large area, parents are restricted to a neighborhood of limited distance, which often results in the chance mating between relatives. Mathematically, the neighborhood area is contained in a circle with radius 20, where o is the parent-offspring (combined seed and pollen) dispersal distance. Finally, the neighborhood size is the number of trees within a neighborhood, and is calculated from the standard deviation 0 of the parent-offspring dispersal variance oz. Both dispersal distance and density influence the neighborhood size. For example, neighborhood size will be small due to short pollen or seed dispersion. Shorter 21 dispersal (slower migration) will allow greater opportunity for genetic differentiation to occur (and spatial structure to build). Low population density will also slow migration rates (fewer paths) with a decrease in neighborhood size. Approximating Wright's neighborhood size from a measure of spatial autocorrelation, the Moran's I statistic (Epperson et al. 1999) affords a method to estimate the combined levels of seed (0;) and pollen (op) dispersal (Epperson & Chung 2001) using Crawford's (1984) dispersal variances. Upon empirical validation, this approach shows promise for selecting pollen and seed sources for maintaining the diversity of regenerating progeny. This is the first published microsatellite study to quantify spatial structure within stands of eastern white pine, using Wright’s neighborhood size and Moran’s I statistic. The complementary allozyme study (Epperson & Chung 2001) will allow comparisons between allozyme and microsatellite markers. As the number of alleles for the microsatellites will be much greater than for the allozyme markers, their power to distinguish genetic structure would be expected to be greater. Because stands are the management units of foresters, by examining the spatial patterns of genetic variation within natural pine populations, we can better anticipate inbreeding levels in 22 eastern white pine. Greater knowledge of inbreeding will aid in the design of effective management strategies for reducing family structure and maintaining in-situ eastern white pine genetic diversity. 23 METHODS Site Description Two mature populations of eastern white pine (Pinus strobus L.) from Hartwick Pines State Park (Hartwick Pines) were selected for genetic analysis. The park study site was chosen for several reasons: (1) the forest is unmanaged; (2) the trees are over 100 years in age with some being over 250 years old; and (3) both old growth and second growth forests are within close proximity to each other. Hartwick Pines is located in Crawford county of Michigan’s lower peninsula, 11.3 km northeast of the town of Grayling (44°39’N / 84°42'N). The park is situated on an outwash plain, formed 10,000 to 15,000 years ago by the Wisconsin Pleistocene glaciers (Sommers et al. 1984), (Whitney 1986). The elevation of the region is 347.5 m above sea level (NCDN, June 2000). Spodosol soils are common to Crawford County, which are sandy, well drained and infertile (Sommers et al. 1984), (Wheeler 1898). The climate of Grayling is described as humid continental (Strahler 1965). Mean annual precipitation is 833 mm and fairly evenly distributed throughout the year. The average annual temperature is 5.6°C with an average growing season of 98 days. (NCDC, June 2000). 24 Hartwick Pines includes 16 hectares of virgin old growth forest (NW % of Sec. 15, T.27N, R.3W). The old growth forest will be referred to as OG. The spodosol soil underlying this population can be further described as Rubicon sands (Robertson & Tiedje 1984). The OG stand is located along the old growth forest trail of the state park (Figure 1) and the dominant trees are primarily white pine with some red pine (Pinus resinosa Ait.). Although unmanaged, it is not entirely untouched by human activity. There are unpaved trails in addition to the major paved trail through the forest. It is thought that this virgin stand was established from a few seed trees following a wind disturbance in the 1750’s (Wackerman 1924). Based on this history, the 0G population was approximately 240 years old at the time of sampling. In 1978, increment cores (Rose 1984) placed the mean age of the CG white pine at 177 years with a range from 101 to 229 years, which agrees well with Wackerman’s (1924) estimate. The mean diameter at breast height (DBH) was 58 cm and ranged from 25 to 110 cm. Today the age distribution of trees would be 123—251 years. 25 White pine study areas Scenic drive Hwy 93 N Figure 1. Map of the Old Growth (OG) and Second Growth (SS) study sites in the Hartwick Pines State Park in Crawford County, Michigan. 26 The mean DBH of the trees sampled in the study was 62 cm with a range from 33 to 99 cm (Figure 2). Sixty—three percent of the trees ranged from 50 to 69 cm in diameter. Other woody species found in a dense lower canopy of the forest are sugar maple (Acer saccharum Marsh), American beech (Fagus grandifolia Ehrh.) and eastern hemlock (Tsuga canadensis L.). There was no successful pine regeneration in the sparse undergrowth. Although white pine seedlings were numerous, they grew poorly and were less than 30 cm in height. Individuals greater than 3 years of age (Farrar 1995) and less than 33 cm in DBH, were completely absent. Red pine regeneration was also absent since no seedlings or small trees were observed in the study site. The second growth stand, which will be referred to as SS, is located approximately 2 km east of the CG stand (SW X of Sec. 11, T.27N, R.3W) and covers an area of approximately 6 hectares (Figure 1). This stand regenerated naturally following logging in the early 1890's and has been unmanaged since. White pine is the dominant species. Although all diameter size classes are represented in the population, we chose to sample trees larger than 20 cm in DBH. This approach allowed us to sample contiguous adult trees on the same scale (size of area sampled and density) as the CG, while avoiding the 27 Anm H H Gamay .thm manum mmnfim xoflsuumm um mafia muss? wo mnofludeQom Daupm Suon MOM mcoflusnfluumfio usmflwa ummmnn um kumamfla .N mhsmflm Ammo nu3ouo vacuum HHHU Soc nusouo “So I 50 ca :mo mmnom mmnom mbuos. mwnow mmuom m¢uo¢ anion mmuom ELlI' Ch-i F p p P p p o 1 0H m i 3 i. J l. I. 1 cm 0 13 3. [ a | ONHHG 1 cm a so ma + 56 m¢ mmo cams mm s I «Wan a Eu NH + so am mma name 60 I o¢ om Ammnv panama umumun um Houoauan 28 many, smaller trees that would not be contributing pollen and seed. The mean DBH of the sampled trees was 48 cm, ranging from 22 to 94 cm (Figure 2). The size distribution of the sampled trees was fairly uniform for the various diameter classes. Red pine was also present in the upper canopy of the SS stand. In addition to those woody species identified in the lower canopy at the CG study site, the lower canopy of the SS site included white pine, red pine, balsam fir (Abies balsamea (L.) Mill) and black spruce (Picea mariana (Mill.) BSP). Sampling The total number of sample genotypes is the main factor controlling experiment—wise statistical power for spatial analysis (Epperson & Li 1996). Sample size and proportion of the total number of trees sampled from the population (porosity) were chosen to optimize fine scale spatial statistical analyses for a wind—pollinated species with limited seed dispersal (Epperson et al. 1999). Our sampling scheme included a sample size of 120 individuals, a porosity of 1 (all adult individuals with DBH > 20 cm were sampled), and approximately 50 alleles surveyed at seven loci. The total number of alleles sampled per population was approximately 6000 (120 trees and 50 alleles). Statistical power is high for detecting spatial 29 structure at this level of sampling, based on theoretical simulations. Referring to Table 4 (Epperson et. al. 1999) the sampling situation (porosity = 1, sample size (n) = 10,000, and Wright's neighborhood size (Ne) = 125.7), the standard deviation is small (0.02) and statistical power high for rejecting the null hypothesis of a random distribution (99%). This situation is a good estimate for our study, where the sampling was identical to the above scheme except for a decrease in total sample size from 10,000 to approximately 6,000 (120 trees times 50 alleles). The predicted values of Moran's I statistic for individual genotypes are insensitive to sample size (Epperson et al. 1999). These authors found that for an identical sampling scheme, decreasing the size of the sample increases the standard deviation by the square root of the ratio of the original sample size to the decreased sample size. For example, a 100—fold decrease in sample size from 10,000 to 100, results in an approximate 10-fold increase in the standard deviation (100001/2 / 100“?) (Epperson et a1. 1999) . White pine is a shade intolerant species. Individuals less than 40 to 50 years old would not be expected to survive (Dallimore & Jackson 1984) or contribute to the gene pool of future generations. To determine the genetic 3O structure of the reproducing individuals in both stands, we sampled adult trees capable of producing both seed and pollen. The trees at the CG site were all very large with DBH's 3 33 cm. Adult trees from the SS were arbitrarily defined as individuals with DBH 2 20 cm. As mentioned above, this minimum DBH guideline allowed us to sample adult trees from an area approximately the same size and density as the CG, while avoiding the large number of non— reproducing trees. Sampling areas (Figure 3a and 3b) were approximately 1.2 hectares in size for 0G (100m x 120m) and for SS (110m x 110m). The white pine tree density was approximately 100 adult trees per hectare for both populations. One plot was sampled per site, and within each plot, all contiguous adult individuals were tagged and mapped. Metal identification tags were attached near the base of the tree with a small aluminum nail. Trees were mapped to a neighboring study tree or a park location marker. The angle (9) was measured to the nearest degree, distance (r) to the nearest 0.1 m, and DBH to the nearest cm. The x and y map coordinates were determined from the geometric relationships x = rcosO and y = rsinO. The x and y coordinates were used when quantifying spatial structure. Triangulation was used to crosscheck the map. 31 Needle and bud tissues were collected from 122 CG and 120 SS trees from late August to November 1998, with total of 242 individuals being sampled. Harvesting tools included a pole pruner, flexible saw, and 222 rifle. Small branches were sampled from the tips of the lowest exposed branch, which sometimes was 30 meters or more up. Harvested tissue ’was kept on ice until transported to the laboratory and stored at 4°C for 1-3 weeks until bud and needle tissue could be separated. These tissues were then stored at —20°C prior to DNA extraction. 32 .mnmume on ma mxums xowu 0>Hmmmoozm cmw3umn moGMDmflU use .mufim as mmHMuoon N.H hamumeflxoummm mumz mmuflm >U5um .mcflm mufi£3 mo mcofluwasmom madame o3u cflnuflz UmumooH momma Hmscfi>fiocfl m0 maofiusnfinuwflo Hmuummm was .mec :u3ouw cacomm u a vcm Loov nuzouo 6H0 u m .nm 6am mm musmflm _ F L U L L L L L P L L L I 4 4 I 4 4. I 4t9nV. 4 4 4 44 I 4M4 I 4 4 LI 4 4 1'4 4 14 4h144 «W I 4 4‘ 4 £4A£V1€4. I 4 4444 4444 I 4 4 fl 44 44 44 44‘ I 4 4 a 4444 A I 4 4 4 44 4 ¢44 fl 4 444 44 44. I I 4.6. 4 4. I 4 4 4. LV4. 44 4 4 .4114“ AN 444 4 (4 I I 4 4 44 I 4 4 4. 4 4 4 444444 («44 I 44“ «444«4 I 4 44 4‘ .444. I 4 4. 4 4 44 44 I I 4 44.44 ONH u G 4444 I I NNH n a 4 I mm hm mo am L L L L T 4 4 L L a L L L mqofiusnwuunwv Hufluunm «man 33 DNA Isolation Total genomic DNA was purified from 20-40 mg bud tissue with Qiagen Plant DNeasy isolation kits (Qiagen Inc., Valencia, CA). In situations where bud material was insufficient, 70 mg needle tissue was substituted for DNA extraction. Yields ranged from 20—160 pg DNA. DNA was quantified by DNA fluorometry and diluted to 2.5 ng/ul in 'FmEl (10 mM Tris-Cl pH 8.0, 1 mM EDTA) for use in PCR amplifications. All DNA stocks and dilutions were kept at -20°C for long-term storage. Marker Analysis Eight microsatellite markers (Rpslb, 2, 6, 34b, 39, 50, 84 and 127) were used to survey genomic SSR variation at seven loci. After the initial analysis was completed, it was discovered that primer pairs for markers Rps6 and Rps34b were designed from the same sequence, and characterized the same locus; therefore data for Rps34b was removed from the final statistical analyses. In addition, it was determined that marker Rp834b contained null alleles. Echt et al. (1996) describes the primer sequences (Table 1) and marker characterization for the seven nuclear (CA)n P. strobus loci. Unlabeled reverse primers were purchased from Research Genetics Inc., Huntsville, AL. The fluorescently labeled forward primers were purchased from 34 the Children’s Hospital of Philadelphia, Philadelphia, PA or PE Biosystems, Foster City, CA. The amplification reaction for each primer pair was conducted separately. The PCR reaction mixture contained 2 ng/ul DNA template in 10 pl of reaction buffer. The reaction buffer consisted of 20 mM Tris—Cl pH 8.75, 10 mM (NH4)2SO4, 10 mM KCL, 2 mM M9804, 0.1% Triton X-100, 100 pg/pl BSA, 6% sucrose, 0.1 mM cresol red, 200 pM each dNTP, 200—800 nM each primer, and 0.025 U/ul AmpliTaq Gold DNA Polymerase (PE Biosystems, Foster City, CA). The final Mg concentration was adjusted to 4.5 mM with 25 mM MgC12 to improve yield of the amplified PCR products. The increased Mgm'concentration also promotes non-template nucleotide addition to the 3’ end of target sequences (Clark 1988)(PE Applied Biosystems 1996). The plus-nucleotide form 35 Table 1. Primer sequences for 8 nuclear (CA)n simple sequence repeat markers. Forward then reverse sequences are listed for each primer pair. Rps6 and Rps34b amplify the same locus. Rps locus PCR primers lb GCCCACTATTCAAGATGTCA GATGTTAGCAGAAACATGAGG 2 CATGGTGTTGGTCATTGTTCCA TGGAGGCTATCACGTATGCACC 6 TTTTCTAATCAGTGTGCGCTACA CACCGCTGCCCTATTTTACA 34b CAGTGTTCTCTTATCACAGCG GCACTATAATGAAATAGCGCA 39 GCCAGCTCCAACCAGAATC GGCTCGCTGACCCAATAA 50 CCCAGAAATCTGTTTTAGAGC ACACATGAAATGTCAGAATGC 84 CCTTTGGTCATTGTATTTTTGGAC CTTCCTTTTCCTTCTTGCTCCAC 127 ACTTCCTCCAAGTTACTATTGTCA CCTTGTCTTCTAAAAAACACTTTT 36 simplifies peak patterns and improves allele scoring by Genotyper software (PE Applied Biosystems 1994) The cresol red and sucrose were included in the amplification reaction mix to eliminate the addition of loading dye prior to agarose gel electrophoresis (Routman & Cheverud J. 1994). The sucrose also improves amplification of weak reactions (Erpelding, personal communication). A touchdown amplification protocol described by Echt et al. (1999) with a modified target annealing temperature of 55°C, was performed using PTC—100 or PTC—200 thermocyclers (MJ Research Inc., Watertown, MA). Other adjustments to the protocol included an initial denaturation of 92°C for 9 minutes to activate AmpliTaq Gold and a final extension of 70°C for 20 minutes to promote blunt ended, plus-nucleotide addition (Clark 1988)(PE Applied Biosystems 1996). The touchdown protocol included an initial denaturation step at 92°C for 9 minutes. The first two cycles of the protocol consisted of a denaturing step at 94°C for 1 minute, an annealing step at 65°C for 1 minute, and an extension step at 70°C for 35 seconds. The next 18 cycles comprised a denaturing step at 93°C for 45 seconds, an annealing step at 64°C for 45 seconds (which subsequently was decreased by 0.5°C every cycle until a final temperature of 55.5°C was reached), and an extension 37 step at 70°C for 45 seconds. Conditions for the last 20 cycles were 92°C for 30 seconds, 55°C for 30 seconds, and 70°C for 60 seconds, followed by a final extension at 70%: for 20 minutes. Touchdown PCR (Don et al. 1991)(Hecker & Roux 1996) and AmpliTaq Gold (Kebelmann-Betzing et al. 1998)(Moretti et al. 1998) were used to increase specificity and yield of amplification products. Agarose gel electrophoresis was used to check the quality and quantity of the PCR amplified products. A 4 pl sample of each amplification reaction was electrophoresed through 2% LE agarose gels using 1X TAE running buffer (Tris-acetate-EDTA) containing 0.2 pg/ml ethidium bromide to determine the amplification efficiency for marker analysis. PCR products were diluted in ddeD in order to standardize and normalize fluorescent signal for accurate measurement of sample data peaks. Final PCR dilutions ranged from 1:1.2 to 1:15. Product dilutions for 2 to 3 loci were pooled for multiplexing on polyacrylamide gels, based on the fluorescent color of the marker label and size of amplified PCR product. The pooled, fluorescently labeled amplified fragments were resolved using denaturing polyacrylamide sequencing gels (6% acrylamide, 8.3M urea) with 1X TBE running buffer (Tris—borate-EDTA). Pre- prepared acrylamide gel mixes (Burst-Pac 6% sequencing gel) 38 were purchased from Owl Scientific, Portsmouth, NH. The resolved fragments were sized by GeneScan Analysis software. (PE Applied Biosystems 1996). Alterations to the manufacturer's GeneScan protocol included cleaning of glass plates with cerium oxide (Bunville et al. 1997) to remove fluorescing background haze and the use of CXR fluorescent ladder (Promega Corp., Madison, WI) as the internal size standard for assigning PCR fragment sizes. To aid in data interpretation, fragment lengths, scored as alleles and reported to 0.01 base pairs (bp) by GeneScan Analysis, were binned or grouped to 1 bp categories and assigned ordinal labels using Genotyper software. Quality Control To ensure consistency of PCR reactions, the mixing of PCR reaction buffer and aliquoting of DNA templates were conducted just once per population. For each primer pair, one large PCR master mix was constructed with enough reagents to amplify all DNA’s for one population. These master mixes contained primer, enzyme and all reaction components necessary for PCR amplification except DNA and were stored at —20°C. DNA templates were aliquoted to the wells of the utiter reaction plates, dried in a food dehydrator and stored at -20°C. Sample DNA’s were aliquoted in a 2X-offset format using an eight—channel pipettor. 39 Enough DNA reaction plates were aliquoted at one time to complete the study. To assess well-to—well gel variation in fragment mobility and fragment molecular weight size assignments, the CXR size standards from all lanes of all gels were combined in one Genotyper file. The fragment peaks were then binned to survey for accurate base—pair calling of the size standards by GeneScan analysis. Several PCR amplification controls were run throughout the study. Running a negative PCR control, which was lacking template DNA and included for each primer pair master mix, monitored PCR product contamination. Internal and external PCR controls were developed to assess reproducibility of PCR generated genotypes. To measure PCR to PCR variability, an internal control procedure was designed in which 13% of the original PCR reactions were replicated. I was 95% confident that this level of re- sampling could detect a 1% departure from expected frequencies. Given the critical value (CV) for x2045,1,and x2 = (Y - p)2 / np(1 — p), then 11 = (Y — p)2 / p(1 — p)CV, where (Y — p) is the difference between the observed and expected values for the number of alleles, p the allele frequency, and n the re-sample size needed to detect a type I error, or rejection of HO when H0 is true (Steel & Torrie 4O 1980). A random number generator determined the DNA templates to be re-assayed and assigned the duplicate well positions in the ptiter reaction plate. The repeated PCR products were amplified, pooled, and evaluated under the same conditions as the original products. An external PCR control was evaluated to survey for lane variation between gels. The external control was constructed by amplifying loblolly pine (Pinus taeda, L.) DNA primed with the chloroplast marker Pt30204 (Vendramin et al. 1996). This amplification reaction generated a single PCR product for use in quality control, and was analyzed in the same lane on all GeneScan gels. Statistical Analyses Genetic diversity, population structure, and spatial structure were evaluated using conventional genetic measures. Each population was considered separately for all statistical analyses. POPGENE software, version 1.31 (Yeh et al. 1999), was used to calculate the diversity measures and Fa values. The spatial structure indices were computed by the SAAP - Spatial Autocorrelation Analysis Program written by Daniel Wartenberg (1989). General Diversity: Five indices measured genetic diversity: (1) observed number of alleles per locus (k); allele richness (A); (3) 41 allele frequency (xi); (4) genetic diversity (He) or Hardy— Weinberg expected heterozygosity; and (5) effective number of alleles per locus (Au). Observed number of alleles (k) was simply an arithmetic count of allelic classes for each locus. Allele richness (A) was the mean number of alleles over all loci. Allele frequency (x3) was determined by xi = y1/2n where yi.is the number of occurrences in the ith allelic class and n is the total sample size. To evaluate the distribution of allele frequencies within and between populations, allele morphs were arbitrarily assigned to three frequency classes. A two-tiered system first classified the allele frequencies as common (xi 3 0.05) and uncommon (xi < 0.05). The uncommon class of alleles was further partitioned into alleles having low frequency (0.005 < xi < 0.05) or rare alleles (xi_5 0.005) (Hartl & Clark 1997). The two populations were analyzed separately. A frequency of 0.005 meant there was only one copy of the allele in one population. Genetic diversity (He)‘was measured for each locus by'Hg = 1—fo:(Nei 1973), where xi is the frequency of the ith allele, fo the total homozygosity, and heterozygosity being 1 - fo. Individual PM values were averaged across all loci. The following arbitrary guidelines for He‘were used to assess the level of marker polymorphism: 42 He‘z 0.75 high polymorphism 0.25 < He < 0.75 moderate polymorphism He.E 0.25 low polymorphism The effective number of alleles (As) for each locus was calculated byA.3 = 1/fo (Kimura & Crow 1964) or the reciprocal of homozygosity, where the X; are defined as for He. Individual Au values were averaged across all loci Population Structure: Deviations from Hardy-Weinberg equilibrium were measured for each locus by Wright's (1921, 1922) fixation index (F). F = lrlg/Hg (Nei 1977), where H0 is the observed heterozygosity and He the expected heterozygosity. Ho = 1- ZXfi/n, where Xfi_is the number of occurrences in the ii homozygous class, n is the total sample size, and Xfi/n is the proportion of homozygotes. Non-zero F values indicate either a deficit of heterozygosity (positive F) or an excess of heterozygosity (negative F) from the values expected when mating is random (F = 0). These indices were tested for significant differences from zero, given the CV = XZOJHJH, by x? = FhiUc—l), df = k(k-1)/2, where n is the total sample size and k is the number of observed alleles (Li & Horvitz 1953). Summing the xf-values and degrees of freedom provides a test of the mean F value over all loci. 43 Relative population divergence was measured at each locus by Fa (Wright 1943, 1951, 1965) using Fm,= 1-Hmflh (Nei 1977, 1987)(Hartl & Clark 1997), where Hg and Hi are the expected average heterozygosities of the subpopulations (CG and SS) and total population (OG plus SS), respectively. HS is calculated by averaging Nei's genetic diversity (He) over all subpopulations with HS = 1- ZQQXIK/n, where xm_is the frequency of the ith allele in the kth subpopulation, and n is the number of subpopulations. Ht is calculated from the mean allele frequency across all subpopulations with Ht = 1—Zx12, where xi is the mean frequency of the ith allele in the total population. The null hypothesis FM,= 0 was tested for significant differences in allele frequencies between populations by the x°—test of heterogeneity given the CV = xfmoLdf, with x2: 2nFm(k-1), df = k(k-1)(a-1), where n is the total sample size, k is the number of observed alleles, and a is the number of populations (Workman & Niswander 1970). Summing the x°4values and degrees of freedom tested the mean Fm value over all loci. Ffl_values range from 0 for no divergence (populations share the same allele frequencies) to 1.0 for complete fixation of alternate alleles in different populations. Qualitative guidelines 44 for interpreting genetic differentiation are given by (Hartl & Clark 1997): 0.00 — 0.05 little differentiation 0.05 - 0.15 moderate differentiation 0.15 — 0.25 great differentiation > 0.25 very great differentiation Spatial Structure: The degree of relation between gene frequencies in individuals separated by distance, or spatial structure, (autocorrelation) was measured by the Moran's I-statistic (Sokal & Oden 1978a) for each of 10 distance classes. Moran's I (Cliff & Ord 1981) is defined as I = I12(2) (wijzizj) ' (2(2)wijzzi2)-l Where: (3 n is the number of individuals. (3 w“ is a join in the binary connectivity matrix. I is weighted by the distance between individuals, with w“ set to 1.0 if the ith and jth individuals are in the distance class and zero if they are not both present. 0 zhzj is given by 21 = xi — x and Zj = Xj - x where xi and Xj are the genotypes for the ith and jth individuals expressed as frequencies (described below), and x is the mean score for all individuals in the population. 45 <3 Wtflth is the matrix cross product or co-variance and ZZK the total variance. Individual genotypes were converted to allele frequencies (Dewey & Heywood 1988) where 1.0 was assigned to Xfig homozygotes of the ith allele, 0.5 to X99 heterozygotes of the ith allele (j ¢ 1), and 0 to genotypes with zero copies of the ith allele. The Euclidean distance between sampled trees was used to assign all pairs or joins of individuals to one of ten distance classes. The upper bound of the first distance class was set at 15 m, with upper bounds increased by 10 m for each successive class. To ensure that most pairs of nearest neighbors were included in the first distance class, the relationship that spacing is equal to the inverse of density (Epperson 1989) was used to approximate the minimum distance between any pair of trees. The upper bound of the first distance class was determined by 2” * (A/T)” (Epperson & Chung 2001) where A/T is the inverse of the adult tree density. The bounds for mutually exclusive distance classes were chosen to include 7-10 % of the total joins. Including approximately the same number of joins in each distance class is desired for uniformity of statistical power between classes (Epperson 1989). The total number of pairs or joins (Jt) is n(n-1)/2, where n is the total sample size. 46 After removing individuals with missing genotypes, separate I indices were calculated for each allele of each locus for each of the ten distance classes. Moran's I values roughly range from +1 for strong positive autocorrelation for pairs of trees that have identical genotypes, to -1 for strong negative autocorrelation for genotypes that are dissimilar. For each distance class, a 95% confidence interval was constructed from the standard error of the mean and compared to E(I) to determine if the study populations differed significantly from the null hypothesis of no spatial structure. E(I) = -1/n—1 (Cliff & Ord 1981), where n is the sample size, and E(I) approaches zero for a random distribution with no autocorrelation. To control for the accumulated risks associated with multiple type I errors, the Bonferroni's approximation was applied to each allelic class to adjust comparison-wise error and ensure overall experiment-wise (correlogram) protection. The Bonferroni's correction (Kuehl 1994) is given as dc = Ge / k, where GC is the comparison-wise error rate, de is the experiment-wise error rate, and k is the number of independent tests. Many of the microsatellite alleles are low frequency alleles with p < 0.05, or less than 12 allele copies per population. The Moran's I values for the individual 47 alleles were highly variable and may have been influenced by these small sample sizes. Although the Moran I values for individual alleles were highly variable, for the final analyses, the mean population I indices provided meaningful indicators of spatial structure. Excluded from final analysis were alleles present in less than five copies. These low frequency alleles were considered to contain insufficient information for spatial analyses. The second allele of a bi-allelic locus was also not considered for further analysis since these alleles are correlated (p = 1 - q) and contribute identical information. The remaining individual I values were averaged over all alleles for each locus separately. Also calculated was the un-weighted average over all alleles and loci to generate a single Moran's I statistic for each distance class for each population. I compared the mean values of the first distance class to look for significant differences in spatial structure between the two populations. In two closely related populations, allele frequencies are historically correlated through their ancestors. To compare population means in dependent samples, a paired t — test (Gonick & Smith 1993) was used to control for variability between loci by comparing the structure within a single locus (i). I calculated di, or the difference 48 between the mean I indices in the ith loci pair, as a single measure of difference for each locus. For example, the di for Rps2 (d2) = ($06) - (£53), where (106) = the mean Moran's I for the old growth population at locus Rps 2 and (iafl = the mean Moran's I for the second growth population at locus Rps 2. The expected value of the difference is (di) = 0. The paired t-test statistic was constructed from the mean and standard deviation of the differences and is defined as tmnq = n.“2 * d / Sd, where g = the sample mean of the differences, n = number of allele pairs, and Sd = the standard deviation of the differences. Autocorrelation is not the same for all distance classes. For a graphical display of the spatial patterns in each population, the mean I statistics were plotted against distance to produce a correlogram. Inferences can be formed about a population’s structural pattern through the shape of its correlogram. 49 RESULTS Characterization Of loci Polymorphism: The seven nuclear (CA)n microsatellite loci used to evaluate the two white pine populations in the study detected polymorphisms in both populations. Two to 14 alleles were detected for the loci assayed for the 242 individual trees sampled, with an average of 7.7 alleles per locus (Table 2). The locus with the largest number of alleles (14) was RpsSO. In contrast, the Rps127 locus had only two alleles. There were a total of 54 alleles for all loci combined. Allele sizes ranged from 144 to 213 base pairs (bp). Locus Rps50 had the largest range of allele sizes (32 bp) among the seven SSR markers and Rps127 had the smallest (2 bp). The loci for Rps markers 1b, 2, and 6 produced variants with multiples of either 1 bp or 2 bp differences (Figure 4a-c). In contrast, for Rps loci 39, 50, 84 and 127, allele size differed by multiples of two bases, i.e. one repeat unit, as expected (Figure 4d-g). In most instances, the observed number of alleles was less than the total number of alleles possible based on the size range of the marker and predictions based on step-wise changes in the repeat sequence. For example, locus Rps39 exhibited 50 cofiumfl>mp UHMUGMum H mH.o mw.o mb.0 Mb.o mN.o 4m.o bw.o mm.o mm.o Luxv >0G03qwum NmH m¢H me mwa Hma mma PON lane mHmHHm COEEOU umoz N ma Nm NH bN NN 0N momma as 0mcmno ¢mH NwH mmH mbH mmH MPH MHN Lane I NmH I vwd I mmH I Fwd I mmH I HmH I mmH woman OHOHH< 4m [‘03 omw mb¢ ¢m¢ 4H ¢m¢ m om¢ m ¢m¢ OH 4&4 m Nm¢ \DNI‘M va .0: mufim 0H0HH¢ mamfimm Hmuos Ham «HMO—)— sad am om mm m N ha mSUOH mam .Luxv mamaam 208800 umOE Mo Nocmsqmuw can .Lmnv whammmmmn ca mamaam GOEEOU umOE Mo 0Nflm Umflmaamsm .mmawaam ummmumH cam ummflamsm 03p C003umn muflmm 0mmn “0 McQESG CH mocm£o .LQQL mufimm 0mmn CH mmmflm mamaam Um>ummno .va mSUOH H00 mmamaam m0 Honesa cm>u0mno .UmamEMm mmeomoeouno m0 mH0QEda 0mm pmuuommm .U®%mmwm ®H03 mOOHU N¢N MO HMUOU 4 undume MOM Umufluwuumumno HUGH memcv ummmmu mucosqmw wameflmn£¢dv H00H05G cm>mm .mcfim mufl£3 .N OHQMB 51 .mnam muw£3 wo mcofluwasmom musums N ca flood mmmc cm>0m um Umuomump mmfiocmsqmnm mamaam mo coflusnfiuumfla .bmammm u m cam .4mmmm u m .ommmm n 0 .mmmmm u U .meM u 0 .mmmm u n .Qammm u m umIMv musmfim SDSOHO pcoomm D 530.8 30 I mufimm 0mmn as mmufim mamaad mba Aha mmH boa mma ¢mH mmH mmH mma Hma mam mom son m::u How mma wmfi mmH rnk'I‘ lbll ‘l-ll _ r ‘L :1 o c 0 CI?- U_ h Cpl C-l Cf “r C— O o o IH.O IH.O IN 0 In o ,m 0 Im o W m. I4 0 I8 o.L 3 Im 0 Im o J I I m o I w o a . m I b o I b O U A OH H x ,m o m n x f Im o mmmm Qw -m.o gamma 04 Im.o o H 0..“ maofiuflnfiuumac honmnumum 0H0HH4 52 Lvmscflucoov w musmflm nu3ouo vacuum mHHu 538.0 30 I whammmmmn ca 0Nflm mamaafl m: m: H: m3 SH 63 03 29 SH m3 «3 SH 9: m3 4%. tf . I. 0.0 igl— cr _ - 0.0 I H.o I H.O IN 0 IN o V -m 0 Im o n 9 .¢ 0 Iv o m. Im 0 Im o n. a ,6 0 Im o.m C u a I v.0 I b o w A m H v— I m o m H x I m 0 SEE 3 . m6 mmmm Us I ma o.H o.H unawusnwuumfiu hononvmuu maoaad 53 mescflucoov 4 wusmflm nusouo ccoomm HHHU cuzouo pao I whammmmmn GA muwm mamfiad mmH mmH HmH mbH hm...“ mbH MbH .2...“ mmH wa HmH hmH mmH mmH Cf Ci _ LIL_I _ CL _ p I 1 LI _ L‘ O . O _ I H . o Aouanbex; GIBTIV ¢Hu_x Lm.o ommmm m4 ommnm maooa How uofiuanwnumfiv hononumuu mamaad 54 mmflmmmmmn as mmuflm mamaad ¢mH NmH NmH va NmH m¢a mescflucoov ¢ mndmflm zuzouo ccoomm flHHu 330.6 30 I mwa ¢vH p I o o 0 III: IEF——_ _ L — smflmmm m6 .m.o wwwdm unawusnwuunfiu honmfluonu mamaad Aouanbex; aIeIIV 55 variants that differed in size by multiples of two base pairs (Figure 4d). The change in size from the smallest allele of 167 bases to the largest allele of 179 bases was 12 bp (Table 2). The maximum number of alleles expected from a non—interrupted, two-step distribution is seven, but the observed number was only five. Only the locus Rp3127 exhibited a non-interrupted allele frequency distribution, having two alleles separated by one repeat unit (Figure 4g). The other six distributions contained interruptions of two to seven repeat units. For example, locus Rps6 had a gap of 14 bp or seven repeat units between morphs 163 and 177, for both populations (Figure 4c). This was the largest distribution break found among the seven loci. Allele Frequency: Allele frequencies for each population were estimated from 232 to 244 sampled chromosomes (Table 3a-b). The frequency of the most common allele varied from 0.28 (CG and SS) for locus RpsSO to 0.87 (SS) and 0.92 (OG) for Rpslb, with an average frequency of 0.65 for all loci. For most loci the most common allele was in majority (xi = 0.5656-0.9180). The only exception was Rps50. This locus had both the largest number of alleles and the most even distribution of alleles frequencies, which ranged 56 pmamEMm m0§0m0§0u£0 m0 HunEdz mamaam wum>flnm « H mmao.o 4H mHmHH4 mono.o ma mamaam vmao.o «H mHmHH¢ mmoo.o HH mHmHH4 nmmH.o «mHo.o oH mHmHH< mmao.o Hwoo.c fivoo.o m mamfiam m¢HH.o avoo.o mm¢o.o m mHmHH4 mmma.o ammo.o mmam.o s mHmHH4 mmso.o mvsm.o avoo.o mmmm.o omHm.o m mamaam mmmo.o Hmsa.o mmmo.o mmmo.o mmoo.o m mamaad m¢HH.o mmao.o H¢oo.o oa¢o.o mmao.o a mamaam mmoo.o .H¢oo.o mumm.o 66mm.o H¢oo.o mmHo.o m mHmHH4 mmmm.o «mms.o .mmflo.o smam.o pmma.o .Hvoo.o m mHmHH4 msas.o Hwoo.o mono.o H¢oo.o Hm¢o.o H mamaad ¢¢m ¢¢m 64m 04m ¢¢m ¢vm H¢¢m umammm «mmmm ommmm mmmmm mmmm «mam nammm msooq / mHmHH4 00 .mm .LH mamaamv umuflm omumfia 57 ma mawaam um0HHMEm one .MNHm 0u mcflpuooom Umumouo mum mmamaad .>H0>flu00mmmn mm cam 30 ca mamspfl>fipcfi ONH cam «NH mus UmamEMm mausofi>flpcfl m0 Homes: 058 .03Q0Mum msmwm m0 mGOHumHnmom muzums 03a a“ UmnsmMGE mmfiocmsvmum 0H0HH4 .mm Q .00 n 0 “Am cam Mm wands omNo.o «H mHmHH< omNo.o mH mHmHH4 NONo.o NH mHmHH4 NONo.o HH mHmHH4 mm¢H.o mmoo.o OH mHmHH< smHo.o N¢oo.o N«oo.o m mHmHHm Nvmo.o .mmoo.o N¢oo.o msmo.o m mHmHH4 mNmH.o mmmo.o mmmH.o .Hmoo.o s mHmHHN mmmo.o omsN.o N¢oo.o NHmm.o smmm.o m mHmHH< NH¢o.o NmNN.o NmNo.o NH¢o.o .NHoo.o smoo.o m mHmHHm meH.o .mNHo.o NmNo.o smHo.o oomo.o m¢mo.o 4 mHmHH4 mmoo.o mNmN.o N¢mm.o N¢oo.o Hmoo.o m mHmHHm mva.o mmmm.o omsm.o MNON.o .N¢oo.o N mHmHH4 smvs.o N¢oo.o mNHo.o .Nwoo.o Nvoo.o oomo.o .vmoo.o H mHmHH4 NHN 04N 64N o¢N OHN ovN mmN sNHmdm wmmmm ommmm mmmmm mmmm dem nHmmm msuoq / mHmHH4 mm .Qm LUODSHDGOUV m mHQMB 58 from 0.0041 to 0.2746. The frequency of RpsSO's four most common alleles ranged from 0.1458 to 0.2750. Perfect Repeat Diversity: For the seven (CA)n loci assayed, four were categorized as perfect repeats, one an imperfect repeat, and two were compound perfect repeat sequences (Table 4). Upon review of the text sequence for Rps39, I reclassified this locus as an imperfect repeat (AC)n(TC)(AC)3, which differs from (AC)m, the original motif reported for this locus (Echt, May—Marquardt, et al. 1996). The chromatogram was not available for review. For all categories combined, the number of dinucleotide repeat units (u) ranged from 11 to 21 with an average of 15.7 units. Nei's genetic diversity (He), in addition to other diversity measures, was used to evaluate if polymorphism increased as perfect repeat length increased among loci (Table 5). As the number of perfect repeat units (u) increased from 11 to 17 (g: 14.2), the level of polymorphism also increased. Genetic diversity or EH increased from 0.20 (low polymorphism) to 0.82 (high polymorphism). The average H; of 0.52 indicates moderate polymorphism for all loci. In addition, the observed heterozygosity (Ho) values increased from 0.18 to 0.80 (Ho = 0.49), the total number of alleles (k) from 8 to 14 59 Table 4. Repeat sequence, category, and number of repeat units (0), in seven eastern white pine nSSR markers. Rps Repeat Category Repeat locus sequence units (p) 1b (AC)11 perfect 11 2 (AC)u5 perfect 15 6 (AC)M perfect 14 39 (AC)B(TC)(AC)3 imperfect 17 50 (AC)y7 perfect 17 84 (CT)m(AC)u_ compound (perfect) 21 127 (AC)m(AT)5 compound (perfect) 15 mean 15.7 SD 3.1 Table 5. Diversity measures for four nSSR perfect CA repeats: number of repeat units (u), Nei's genetic diversity (He), observed heterozygosity Ho, observed number of alleles per locus (k), and effective number of alleles (AS). Rps )4 H8 H0 k Ae locus 1b 11 0.20 0.18 8 1.2 6 14 0.50 0.50 9 2.0 2 15 0.54 0.49 10 2.2 50 17 0.82 0.80 14 5.7 mean 14.2 0.52 0.49 10.2 2.8 SD 2.5 0.25 0.25 2.6 1.9 60 (k = 10.2), and the effective number of alleles (As) from 1.2 to 5.7 (g... = 2.8). Characterization Of Populations Allele frequency and class: The distribution of allele frequencies at each of the seven loci was similar between the two study populations (Figure 4a-g). Allele frequencies ranged from 0.0041 to 0.9180 for OG, and from 0.0042 to 0.8697 for SS (Table 3a-b). The old growth population (OG) contained 87% of the total allelic diversity in both populations, and the second growth population (SS) contained 94% (Table 6). Thirty-three and 37% of alleles were in low frequency (p < 0.05) for CG and SS, respectively, with an additional 18% distributed as rare alleles (p < 0.005) for both populations. Eighty-one percent of all alleles were shared between populations (Table 7). More private alleles were found in SS than OG, 13% of the 54 alleles were unique to SS, and only 5% unique to the CG population. The frequency of private alleles ranged from 0.0041 (one copy) to 0.0123 (three copies) in OG, and 0.0420 (one copy) to 0.0125 (three copies) in SS (Tables 3a—b), classifying them as either rare (xg_5 0.005) or low frequency alleles (0.005 < xi< 0.05). Rpslb and RpsSO had the largest number of 61 LOHV me LOHV me Lmoo.o w.Hxv whom SNL “Km 3: Nmm 30.0 v .x v 30.3 655va sou Lomv wmm mev wmm Lmo.o v Hxv COEEOUCD LHNV Nam a: Nmm 36.0 M :3 202500 LHmL wvm LNNL wbm mmHmHHm Mo .02 Hmuoe mmmHo mm 00 COHDMHDQOQ .mmmmnucwumm CH mH mmMHU 3000 SH ucmmmum mmamaam m0 HmQEdG one .¢m mm3 mcoHDMHSQOQ anon :H ucwwwnm mmamaam m0 HmnEdc HMDOD 059 .pwamEMm .>H0>Huommmmu mm cam 00 no“ owm paw v¢m mm3 muHm 0HQEMm UHOHan HUGH G0>0m 0H03 whose 0:8 .mmamaam mumu cam .m0H0HHm >ocmnqmum 30H .mmHmaam GOEEOUG: cam GOEEOO .mmHmaHm HMDO» "mGOHumHsmom mcflm 0DH£3 mnsumE GH mmflocmsqmum mamaam m0 nofiusnflhumHQ .m 0HQMB 62 Table 7. Private alleles and common alleles present in two mature populations of white pine. The haploid sample size was 244 and 240 for CG and SS respectively. There were seven loci sampled. The number of alleles is in parentheses. Population OG SS Private1 5% (3) 13% (7) Shared2 81% (44) 81% (44) Total 87% (47) 94% (51) L2 Percentage of 54 alleles detected in both populations 63 private alleles, with each locus containing three. For Rpslb, all three private alleles were found in SS. At locus RpsSO, two private alleles were found in OG with one in SS (Figures 4a and 4c). There were no private alleles for Rps84 and Rp8127; the only two loci categorized as compound (perfect) repeat sequences (Table 4). Genetic Diversity: Each of two populations maintains a high level of polymorphism at the seven loci examined (Table 8). Diversity measures were similar for the two populations. All loci were polymorphic in both populations, with 47 and 51 alleles detected in CG and SS respectively. Respective values for the per-locus mean number of alleles or allele richness (A), and the effective number of alleles (AC), were 6.7 and 2.4 for OG, and 7.3 and 2.3 for SS. On average, levels of heterozygosity were moderately high. Mean observed heterozygosities (H§)'were 2% to 6% lower than the probabilities expected under Hardy-Weinberg equilibrium (He) . Respective values for Ho and He were 0.47 and 0.48 for OG and 0.46 and 0.49 for SS. Population Structure: The mean fixation indices F, were 0.01 (OG) and 0.05 (SS), indicating little inbreeding in either population (Table 9). For the CG genotype frequencies, 29% differed 64 LmH.ov m¢.o Lom.ov m¢.o mev huHm0m>N0H0u0£ pmuommxm LmH.oL mv.o LHN.oV bv.o Lomv >uHm0m>N0H0umn U0>Hmmno L¢.HL m.N Am.HL «.N Lymv mmamaam mo .02 0>Hu00mwm Lm.mv m.b Lm.mv b.w Lmv mmmccofiu mamaad Hm b4 HUOH n H0>0 mmamaam Lo .02 Hmuoa NOOH NOOH 0H£QHOE>HOQ HUOH unmoumm omH NNH mmmuu mo .02 oHumHDMDm huHmH0>HQ nuzouw ocoomm £u30uw UHO coHumHsmom .mmmmsucmnmm CH 0mm mGOHumH>0p pumpcMDm mamaam .wwamaam m0 Hogans anuou .mev >uHmH0>HU 0Humcmm m.Hmz H0 >uHm0m>N0H0u0£ Umuommxm cam .Lomv Nuwmomhnoumumn p0>kmmno .LvHuommm0 .Ldv mmmcnofln .HUOH UHSQHOE>HOQ 0mmucmoumm .U0HQEMw mmmuu mo Hones: "WDQOHQQ QSQWQ £U3OHU UGOUOW Uflm £930.50 UHO HON mOHDmMOE kaflwHONVHU HmHOGOG .m OHQMH. 65 Table 9. Population indices F and Fm for each nSSR locus and as a mean across all loci as measured in two populations of white pine. Seven loci were sampled. Sample size was 122 and 120 individuals in CG and SS, respectively. Population OG SS Statistic F F Fm Rps locus 1b -0.06 0.16 0.0056" 2 0.15 -o.02 0.0156*‘ 6 0.01 0.01 0.0011 39 0.26" 0.17 0.0048" 50 -0.04 0.09 0.0033" 84 -0.04 0.05 0.0064” 127 -o.23*' -0.11 0.0010 Mean 0.01 0.05 0.0054“ SD 0.16 0.10 0.0050 ' Significant at a = 0.01 66 from Hardy-Weinberg expectations. Locus Rpsl27 had a significant excess of heterozygotes while locus Rps39 showed a significant heterozygote deficient. SS genotype frequencies did not vary significantly from values expected under Hardy—Weinberg equilibrium. There was little divergence between the two populations (mean §m_of 0.0054). Differences in allele frequencies between populatiosns were statistically significant for 71% of loci. Spatial Structure: Spatial structure within both populations was weak. For each of two populations, Moran's I spatial autocorrelation coefficients (I) were calculated for each of 10 distance classes. The indices for the first nine distance classes for 0G and SS are reported in Table 10a and 10b, respectively. Shown are mean indices for individual loci, and the overall mean values. Although no precise tests of significance are available for multiallelic - multilocus data, by comparing the overall mean I values for distance class one, the structure at CG (1 = 0.015) is 15 fold greater than that for SS (1 = 0.001). The theoretical expected value of I, when there is no spatial structure, is -0.008. The shape or pattern of spatial structure in the CG stand differs from that in the SS stand (Figure 5). OG had weak positive genetic correlation or similarity between 67 individuals separated by 15 meters or less, followed by negative correlations or dissimilarity at longer distances. This spatial pattern is in contrast to SS, which appears to have a completely random distribution. 68 HUGH H0>0 c008 Umuanmz .mwamaam H0>o c008 0Hums£uHu¢ m mmHmHHm u0>0 memnm>m mmsHm> H capo: « mmmHo 00c0umHQ m Lmumumev mpcson monmumHo N mNHm 0HQEMm UHOHQmm H Hm0.0 mm0.0 0N0.0 mN0.0 Nm0.0 0N0.0 Nm0.0 HN0.0 0N0.0 Om 0H0.0 0H0.0I 000.0 000.0I 0H0.0I mN0.0I mH0.0I 000.0I mH0.0 mammz 0m0.0I 000.0I 0N0.0 0N0.0I 0H0.0I 000.0 0H0.0I 0m0.0I 0m0.0 bNHmmm 0H0.0 mN0.0I m00.0I 000.0: 000.0I mm0.0I m00.0 0N0.0I mH0.0I wmwmm 0H0.0 000.0: 000.0 0H0.0I 000.0 0N0.0I N40.0I 0H0.0 0H0.0I 0mmmm 0b0.0 bN0.0 0H0.0I bm0.0I 0m0.0I mm0.0I 000.0 0H0.0I 000.0 mmmmm m00.0 wm0.0I mN0.0 000.0 0H0.0I mH0.0I m00.0I m00.0I m00.0 mem 000.0 MN0.0 m00.0I 0H0.0 mm0.0I m¢0.0I 0H0.0 000.0I mm0.0 Nmmm m00.0n 040.0I mvo.0 0H0.0 0m0.0 mm0.0I 000.0I 0H0.0I 0m0.0 «Qmem m m b 0 m w m N H mmmmHUumHQ mm mm mb mm mm m0 mm mm ma Npcsomumflm UO 00H .mm MOM o¢m H am 000 00 H00 H¢¢N n Gm .mmmmmHo conunmflp m ~00 0:00Hum mscwm mo mnoHumHsmom musume 03p 0H LH m.nmuozv mucmonmuwoo COHumHmunooousm Hafiummm saw: .mev £u30Hm pcoomm u D .LOOV £u30nm 0H0 u m .QOH paw 00H 0HQMB 69 000.0 HN0.0 0N0.0 mHo.o NNo.0 mHo.0 0N0.0 0N0.0 mH0.0 Om 0N0.0: MH0.0: mH0.0: MH0.0: N00.0 bH0.0I m00.0: N00.0 H00.0 Emwz ovH.0: 0m0.0: 000.0: 0m0.0: 0m0.0 000.0: 0m0.0 0m0.0 0m0.0 bNHQO 000.0: 000.0 0m0.0: m00.0: 0m0.0: mN0.0: 000.0: mm0.0 MH0.0: vmmmm 000.0: 400.0: 0H0.0: bm0.0: 000.0 000.0 m00.0: NH0.0: H00.0: 0mmmm 0¢0.0: 000.0 0H0.0 000.0: mHo.0 m¢0.0: m00.0: MN0.0: m00.0: mmwmm mv0.0: 0N0.0: m¢0.0: m00.0: 0N0.0: mH0.0: mN0.0 0N0.0 0H0.0 mmmm 000.0 m¢0.0: 000.0: MN0.0 0H0.0 0m0.0: mm0.0: m00.0 MH0.0 Nmmm MN0.0: 0H0.0: m00.0 MH0.0: 0H0.0 0N0.0: MN0.0: 000.0: 000.0: QHQO m m b m m w m N H mmMHUumHQ mm mm m0 mm mm m¢ mm mN mH Undomumfln mm 90H mescHucoov 0H 0HQMB 7O Spatial Correlograms 0.020 0.015- 0.010: 0.005— 0 0 E(I)=-0.008 .000~ .0053 -0.010* -0.0154 “0.020 I I I 15 25 35 Moran's I Distance in meters ———— Old Growth ———- Second Growth Figure 5. Mean Moran's (I) indices measured spatial structure in two populations of white pine. The I values for the first three distance classes were plotted. 71 DISCUSSION Characterization Of Loci Polymorphism: The number of alleles was low to moderately high for the seven SSR loci studied. In total, fifty-four alleles were identified for both populations with an average of 7 alleles per locus and a range of 2 to 14 alleles. These measures were lower than values reported by Rajora et al. (2000), where two old growth white pine stands in Ontario, Canada, were evaluated at 13 loci, including 5 of the 7 loci assayed in this study. The latter study had an average of 10 alleles per locus and a range of 2 to 21 alleles. A similar number of trees were sampled for both studies: an average of 121 trees were sampled from the Hartwick Pines populations and 119 trees from the Ontario populations. The mean number of alleles per locus is a measure of allele richness, independent of allele frequency. It is an important measure for surveying populations for purposes of conservation of unique genes (Marshall & Brown 1975)(Le Corre et al. 1997) or for quantifying effects of tree harvesting on gene pools (Rajora et al. 2000). For example, because allelic richness was moderate for the Ontario study, harvesting had a substantial impact on 72 allelic composition. A 75% reduction in the size of the breeding populations resulted in a loss of alleles of approximately 26%, and 44% of the lost alleles had frequencies less than 0.05 (Rajora et al. 2000). It is yet unknown how these genetic changes will affect future regeneration. Allele frequency distributions at microsatellite loci are often predicted using the stepwise mutation model (SMM) (Valdes et al. 1993)(Goldstein et al. 1995)(Slatkin 1995)(Goodman 1997). SMM generally oversimplifies the mutational processes occurring at SSR loci. The strict SMM assumes that mutation only causes the repeat number to change by one unit, and thus alleles with similar sizes are much more related than alleles with very different sizes. However, mutations of SSRs can also result in large changes in allele size (Weber & Wong 1993). Transversions, transitions, indels and duplications occur both within and around the repeat sequence (Callen et al. 1993)(Estoup et al. 1995)(Paetkau & Strobeck 1995)(Grimaldi & Crouau—Roy 1997)(Karhu et al. 2000). The genotypes observed in the Hartwick Pines populations do not include all alleles predicted by SMM theory for neutral alleles. For instance, based on the model and the distribution of allele sizes, the total expected number of alleles would be approximately 73 73, which is considerably larger than the 54 alleles observed. Five premises can help explain the missing alleles or “distribution breaks" observed at all but one locus (Rps 127) of the study: (1) missing alleles occur in very low frequency and thus were not sampled; (2) missing alleles were lost by genetic drift; (3) more than one repeat was lost or gained during the mutation process; (4) other non—SMM mutations occurred; or (5) missing alleles are under selection or linked to genes under selection. In addition, 1 bp allele size differences were observed and these cannot be produced under a dinucleotide SMM model. Moreover, it is unlikely that the 1 bp polymorphisms present in Rpslb, Rps2 and Rps6 initiated from the SSR repeat region of these loci. Thus, the number of alleles that are the result of mutation at the repeat region would be further reduced. For example, Rpslb would be characterized as having only six alleles rather than eight, if only size variants with multiples of 2 bp morphs were considered. Another mutation process is called plus- A. During plus-A activity, DNA polymerase catalyzes the addition of a single, non—template nucleotide to the 3’ hydroxyl terminus of blunt ended template regions (Clark 1988). The extra nucleotide (usually adenosine) is added to both the stutter peaks and allele peaks of PCR products. 74 PCR amplification produces a characteristic pattern of 1 bp peaks, which is in contrast to the patterns of variation observed in Rpslb, Rps2 and Rps6. The primer pairs for these three loci produced 2 bp stutters with both 1 bp and 2 bp allele morphs, therefore, plus—A activity does not explain the observed 1 bp polymorphisms. A plausible explanation is the 1 bp variation originated from insertions or deletions of adenosine in the short poly A sequences that flank the repeat CA.n region (Primmer et al. 1998), but only perfect repeat sequences were affected. Since all of the sequences flanking the Rps markers contain poly A regions (data not shown), it is possible that the perfect SSR repeat, and / or the flanking sequence, increases the mutability of the poly A region. Allele Frequency: Allele frequency must be considered in addition to total number of alleles when evaluating the effectiveness of molecular markers for quantifying various aspects of population genetic structure. Diversity measures based on evenness of allele frequency are useful for measuring effective heterozygosity levels (He), effective population size (Hartl & Clark 1997), inbreeding (Nei 1973, 1977), and for identifying individuals and assigning parentage based 75 on DNA fingerprinting (Amos & Hoelzel 1992)(Adams 1992)(Blouin et al. 1996). He is a popular diversity measure since it equates diversity with heterozygosity, assigning a direct genetic interpretation to diversity. Ideally for fine scale population genetic studies, a marker should posses a large number of alleles present in equal frequency. For example, Rps50 presented many alleles (k = 14), with the four most common alleles having similar frequencies. This locus had the largest Ac and He'values of the study (5.7 and 0.82 respectively). RpsSO is the best suited of the 7 loci for fine scale population genetic analyses. But equally frequent alleles are usually not the case with microsatellite loci. Although SSR markers can be very polymorphic, a SSR locus is often characterized by one or a few high frequency alleles and many low frequency alleles (Streiff et al. 1998), which makes the allele richness index sensitive to sample size. In the Ontario study (Rajora et al. 2000) for example, allele richness was affected by harvesting, but He showed less than a 5% reduction. The mean He for both populations was approximately 0.5 both before and after harvest. For these reasons, expected heterozygosity is the diversity measure 76 of choice when study objectives are to maintain allele frequency evenness of common alleles. Perfect Repeat Diversity: Nei's genetic diversity measure, He, increased across loci as the perfect repeat length increased. These results agree with human studies in which a Similar trend (Weber 1990) is observed for a related measure, the PIC (polymorphism information content) value (Botstein et al. 1980). The imperfect and compound perfect repeat sequences in white pine contain less information than the perfect repeat sequences. These loci have fewer alleles and lower heterozygosity levels than do perfect repeat sequences with the same number of repeat units. Weber (1990) suggested that the higher polymorphism of long, uninterrupted CAn repeats was the result of higher mutation rates. Strand slippage is hypothesized as the mutational mechanism responsible for many of the new mutations formed in SSR repeat regions (Levinson & Gutman 1987)(Weber 1990), occurring more frequently when the number of repeats is larger. 'Increased slippage explains the higher He values for alleles with larger repeat lengths. In 1992, Schlétterer and Tautz supported this hypothesis by in vitro experiments. They synthesized repetitive di- and 77 trinucleotide repeats sequences from short primers, dNTP’s, and polymerase, estimating synthesis rates for the various motifs. The authors found that the rate of growth for DNA fragments was independent of the length of the template DNA fragment. Length independence suggests formation of transient single stranded regions within the simple repetitive sequence. Schlétterer and Tautz hypothesized that the unpaired regions could be substrate for repair mechanisms in vivo. The creation and elongation of simple sequence regions could contribute to the length variation observed in these repeat regions. In non-perfect and compound sequences, the interruptions to the main repeat sequences may stabilize the repeat area by decreasing the number of tandem CA units (Weber 1990)(Jin et al. 1996). Interruptions should reduce slippage and improve proofreading (Taylor et al. 1999), lowering the mutation rate for interrupted sequences with fewer new alleles being formed. Messler et al. (1996) proposes that the “birth” of a microsatellite is a two-step process. The first mutations are point mutations, creating enough repeat units for expansion to occur by replication slippage. The second mutations are strand slippage, producing repeat length variation. For substantial strand slippage to occur, a 78 minimum of 10 tandem CAnrepeats is necessary for the replicating DNA polymerase to falter and introduce more or less repeats into the newly synthesized strand than is found in the template strand (Weber 1990). All of the loci in this study have a sufficient number of tandem repeats, ranging from 10 to 17 cmm units, to mutate further. In addition to higher mutation rates, perfect repeat sequences may also have higher rates of homoplasy (Estoup et al. 1995). Homoplasy occurs when alleles are identical in state (IIS) but not identical by descent (IBD). Two alleles are IBD if they descend from the same ancestral allele without mutation. For microsatellite analyses, two alleles are homoplastic if they are the same size without being IBD. PCR products of the same size but different sequences (convergence) or even the same sequence (parallelism) can arise from independent mutational pathways. Two IIS alleles may also descend from the same ancestral allele if one of the alleles mutates and then back mutates to its original state (reversion), while the second allele remains unchanged. When allele size morphs contain a mix of IIS and IBD alleles, this has a homogenizing effect on genetic differences, decreasing the resolving power of SSR markers for detecting divergence of SSR loci (Nauta & Weissing 1996). 79 Although homoplasy commonly occurs at microsatellite loci (Estoup et al. 1995)(Grimaldi & Crouau—Roy 1997), it is unlikely to be a concern for the genetic analysis of closely related populations. Homoplasy could be problematic, for instance, if the study populations had been isolated for a long time and were not exchanging migrants. In one study of honey-bee (Apis mellifera), homoplasy was not seen at the population level, or in populations belonging to the same subspecies, or in closely related subspecies, but length convergence was detected among distantly related subspecies (Estoup et al. 1995). Sequencing revealed that homoplastic alleles varied both in the number of TC repeats, and in the position and number of imperfections. In contrast, Grimaldi and Crouau-Ray (1997) demonstrated length convergence at the population level, in humans. Two allele morphs were identical in size but differed greatly in sequence. The unique size variation at this locus results not only from differences in the number of CA repeats, but also from differences in the flanking region. For example, alleles 7 and 8 differ in size by 2 bp. The expected variation is the addition of 1 repeat unit. The actual variation originated from a 2 repeat unit deletion and a 6 bp duplication occurring in the flanking 80 region, causing the observed 2 bp difference. Together, these studies suggest that in addition to operating in distantly related subspecies, homoplasy can also occur at the population level. However, it is important to note, that in both of these studies homoplasy only involves a fraction of the same size alleles common to the populations under study. It is unknown if homoplasy operates at the population level for white pine. If it occurs at all, it is most likely rare, and should not affect data interpretation of this study. It may be that SSRs with moderate polymorphism, such as RpsZ, Rps6, and Rp539 (and therefore often moderate rates of mutation and homoplasy), may be best suited for studying differences between populations. Markers with the highest polymorphism are the most useful for examining parentage or individual differences on the shortest of time scales. Estoup et al. (1995) suggests that interrupted SSRs, such as Rps84 and Rps127, may have lower homoplasy rates than perfect repeats making them better suited for investigating changes in allele frequency and evolutionary relationships between more distantly related populations. 81 $75} 9.. 0f Nu11.Alleles: At the population level, a greater area of concern for accurate data interpretation is the presence of null alleles (Callen et al. 1993)(Paetkau & Strobeck 1995)(Pemberton et al. 1995)(Jarne & Lagoda 1996)(Gullberg et al. 1997)(Becher & Griffiths 1998)(Van Treuren 1998). Null alleles result from mutations in the sequence at the priming site. Mispriming prevents PCR amplification of the target sequence; therefore genotypes are no longer visible through gel electrophoresis. The primer pairs chosen for marker analysis can greatly affect the results, since the sequences flanking the SSR repeat region may be highly variable in addition to the repeat region. After the initial analysis was completed, it was discovered that the primer pairs for markers Rps6 and Rps34b were designed from the same sequence and flanked the same nSSR locus (Figure 6). Thus they amplify the same locus. I also suspected that marker Rps34b contained null alleles, which would explain the observed differences in the allele frequency distributions between the two markers. For these reasons, marker Rps34b was dropped from the final analyses. Nonetheless, the data on Rps6 and Rps34b was quite informative about the nature of alleles in P. strobus. Very little of the variation at 82 the Rps6/Rps34b locus appears to be in the dinucleotide repeat region. For example, the allele distribution pattern for Rp834b is straightforward and is characterized by five 1 bp morphs (142-146) (Figures 7a-b). In contrast, Rps6 has a greater allele number (k = 9) with a 14 bp break between the small alleles (159—163) and the large alleles (177-186). The allele frequency distribution for Rps6 is more complex than that of Rps 34b. Both 1 bp morphs (alleles ranging from 159 to 163) in addition to alleles differing by three repeat units (allele pair 180/186) were observed. It is also possible for one allele to contain both repeat (2 bp) and non-repeat (1 bp) polymorphisms. For example, allele 180 of Rps6 differed in size from the 179 bp allele by 1 bp, while allele 179 differed by 2 bp or 1 repeat unit from the 177 bp allele. One premise for the observed genetic differences between these two primer pairs is the large alleles (177/179/180/186) of Rps6 are artifacts. A second explanation is mispriming at the Rps34b locus producing non-amplifying null alleles. If the large alleles of Rps6 were artifacts, we should occasionally see samples with three peaks, but this did not occur. If the suspected large alleles (160/162/163/169) of Rps34b are absent 83 pPsG1 (marker sequence size 162 bp) CCTATGACAACTAACCCATGGGACGACTTACATAGTCGGATAATCCATGCCGGCCCTTG CATGAATTTTTAAAACACTGATTTTTTCTAATCAGTGTGCGCTACATAACCTAGCGCAC CAGTGTTCTCTTATCACAGCGCACCAAGCACATTTTGTTATAAACACACACACACACAC ACACACACACACATATTATTTTATTTATTTTTAATATGTGCACCATATGTAAAATAGGG CAGCGGTGCGCTATTTCATTATAGTGCATTACGCAaGT pP334 (marker sequence size 145 bp) CCTATGACAACTAACCCATGGGACGACTTACATAGTCGGATAATCCATGCCGGCCCTTG nATGAATTTTTAAAACACTGATTTTTTCTAATCAGTGTGCGCTACATAACCTAGCGCAC CAGTGTTCTCTTATCACAGCGCACCAAGCACATTTTGTTATAAACACACACACACACAC ACACACACACACATATTATTTTATTTATTTTTAATATGTGCACCATATGTAAAATAGGG CAGCGGTGCGCTATTTCATTATAGTGCATTACGCAtGT Figure 6. Sequence data for two cloned white pine fragments containing (CA)M repeats. Marker Rps6 was derived from the plasmid clone pPs6 and Rps34b from pPsB4 with a 17 bp difference in sequence size. Forward and reverse primers are underlined. The repeat sequences are in bold typeface. Lower case letters indicate differences between the sequences. 1 plasmid Pinus strobus clone 6 84 .mmH can 00H .mbH .be mmeHHm mmmm man 00 mcHUcommmuuoo mmHmHHm HHdn 830m mGH>mn nvm mam £DH3 .msooancdv 0800 van NMHHQEM mnmxHME 0m0£B .Q¢mmmm 080 000% 080x808 800 mGOHuanHumHU >onmsvmnm 0H0HHm mnu 0H0 Um>mHmmHQ .Q4mmmm n a new mmmm u m "Db 080 mu musmHm 803080 ccoumeHHHL 5380 30 I mHHmmmmmn 8H 0NHm 0HOHH4 04H mvH 46H maH N¢H mmH omH mbH 68H mmH NmH HmH omH mmH p Cpl P I Io.o ILI JHWL q-i‘flq-Ii—LLI H . II 0.0 IH.o IH.o IN 0 IN 0 V In 6 IN 0 H m. .¢.o I4 0 a 1: Im 0 Im o 1 a Im.o I6 0 w r a W Hb.o In 0 w .A m u I x Im o m : x IN 0 mm 96m m Db Im.o meN 00 IN 0 o.H o H mnowuanHuch hoamsuouu 0H0HH4 85 b: 01' si of su pr 0:: (TI 001 mi( Rps leI usz' anc nu] mut Sit ShO‘ Call because they are null alleles, then treating the large allele heterozygotes of Rps6 as small allele homozygotes and subtracting 17 bp from the Rps6 alleles should produce genotype scores identical to RpsB4b. The expected 17 bp differences in allele sizes were based on the sequence sizes of the two markers (Figure 6), with a sequence size of 162 bp for Rps6, and 145 bp for Rps34b. The data supports the Rps34b null allele hypothesis with the second primer pair (Rps6) providing complete genotypic data for the locus. For the 28 trees that have large Rps6 alleles (Table 11a-b), all corrected genotypic assignments correspond to the genotypes for Rps34b. As with many microsatellite markers, within population variation for the Rps markers is not well characterized at the sequence level. The markers were isolated from genomic libraries using oligonucleotide probes (Echt et al. 1996). Cloning and sequencing the Rps34b null alleles would verify our null allele interpretation and determine the exact mutational changes that occurred at the Rps34b priming site. In addition, sequencing would enable us to look for homoplasy at the population level. Several studies have shown that variation at the priming site has occurred resulting in null alleles due to mispriming. For example, Callen et al. (1993) determined that an 8 bp deletion 86 within the priming sequence prevented the binding of one primer. Paetkau and Strobeck (1995) identified a G to C transversion at the 3' end, which prevented binding of the primer sequence and resulted in null alleles. For this study, a possible DNA insertion is suggested by the difference of 14 bp between the small and large alleles of Rps6. Since the Rps34b forward priming site lies interior to the Rps6 forward priming site (Figure 6), an insertion occurring downstream of the Rps6 site and interrupting the Rps34b forward priming site could produce the null allele pattern that we observed in our data set. Moreover, the mutation underlying the Rps34b null alleles seems to be ancient, since the large alleles of Rps6 vary in size according to predicted microsatellite patterns, with morphs changing by multiples of 1 or 2 bp repeat units. The observed frequency of the Rps34b null alleles was 0.059 (29 of 484 alleles, see Table 11a-b)) for both populations combined. In addition, one null homozygote in 242 trees was observed (p = 0.004), which agrees with the Hardy-Weinberg expected frequency of 0.003 (0.0592) or 0.7 null homozygotes (0.003 * 242). It is estimated that 25- 30% of all microsatellite markers contain null alleles 87 Table 11a and 11b: a = CG and b = SS. Markers Rps6 and 34b amplify the same locus. Characterized are the large allele heterozygotes for Rps6 (alleles 177, 179, 180 and 186) with the corresponding genotypes for Rp834b where the large allele is a null allele. Rps34b homozygotes are 17 bp smaller than the small Rps6 alleles. 11a OG Rps6 Rps34b Sample peak 1 peak 2 peak 1 peak 2 8 161 186 144 144 11 161 179 144 144 12 161 179 144 144 16 161 179 144 144 41 161 179 144 144 46 161 179 144 144 72 160 179 143 143 76 161 179 144 144 100 161 179 144 144 107 161 177 144 144 140 161 179 144 144 11b SS Rps6 Rps34b Sample peak 1 peak 2 peak 1 peak 2 136 161 177 144 144 163 161 179 144 144 176 161 179 144 144 177 161 179 144 144 200 160 179 143 143 201 161 179 144 144 202 161 179 144 144 208 179 180 0 0 212 161 179 144 144 214 161 179 144 144 220 161 179 144 144 223 160 179 143 143 228 160 179 143 143 242 161 179 144 144 244 161 180 144 144 264 161 186 144 144 303 161 179 144 144 88 (Callen et al. 1993) in frequencies up to 0.15 (Jarne & Lagoda 1996). If undetected and unaccounted, a high incidence of null alleles could bias conclusions drawn from a study. For example, allele frequencies will be inflated for the alleles opposite the null allele in a genotype and absent for the null allele itself. In addition, the presence of a null allele reduces the informativeness of a co-dominant marker because the null allele behaves like a recessive allele, making it impossible to distinguish a null allele heterozygote from a non-null allele homozygote. Increases in reported homozygosity levels may result in artificially significant inbreeding coefficients since the population will appear to have more homozygotes than expected at Hardy—Weinberg equilibrium. Locus Rps6/34b: The rate of miscalls for the Rps6/Rps34b locus was estimated by comparing the two complete sets of genotypes for these two markers. There were errors in four genotype assignments for the SS population (Table 12), resulting in a 2% (4 of 242 trees) discrepancy between the two markers. There were no similarly affected trees in the CG population. In all cases, the second larger peak of the Rps6 genotypes were either missing or weakly amplifying and not scored, resulting in the assignment of homozygous 89 genc in R stud expe mark genotypes for Rps6 as opposed to heterozygote assignments in RpsB4b. Rps6 was the weakest amplifying loci of the study. Since this marker is more difficult to assay, it is expected to have a higher error rate than other Rps markers. Table 12. Estimating the rate of miscalls for the locus defined by markers Rps6 and RpsB4b. Characterized for the Second Growth (SS) population are the homozygous genotypes scored for marker Rps6 that present heterozygous genotypes when scored for marker Rps34b. Rps34b peak 1 alleles are 17 bp smaller than Rps6 peak 1 alleles. The alleles with discrepancies are in bold typeface. SS Rps6 Rps34b Sample peak 1 peak 2 peak 1 peak 2 155 160 160 143 144 185 160 160 143 146 254 161 161 144 146 258 160 160 143 144 90 Characterization of Populations Allele Frequency and Class: Populations that are close together exchange more migrants and are more similar in gene frequencies than populations that are far apart (Wright 1943). As expected from their close proximity, the two populations compared in this study had very similar allele frequency distributions. Allele frequencies in a population typically range from near zero to near one. Rare alleles are more likely to be unique to a single population. For this reason, rare alleles (p 5 0.005) (Hartl & Clark 1997) are useful for identifying individuals (Hartl & Clark 1997) and for distinguishing recent immigrants. The two loci with the largest number of rare alleles, and therefore most useful for identification in paternity, forensics and immigration analyses, were Rpsz and Rps6. Both loci are perfect repeat sequences. Five rare alleles were detected for RpsZ and four for Rps6. Rps6 morph 162 was rare in CG (1 copy) but had low frequency in SS (4 copies). Three rare alleles were detected for Rpslb, one each for Rps39, 50 and 84, and none for Rp8127. When rare and low frequency alleles are under selection or are very tightly linked to genes affecting fitness, they may represent the genetic potential necessary 91 for adaptation to future environmental change (Rajora et al. 2000). This is because more common alleles would have been selected for in the current environment or in the past. Thirty-five percent of alleles had frequencies less than 0.01 (2 copies), for both populations combined. These results were 19% higher than those of Rajora et al. (2000). An average of 16% of the alleles in the two Ontario white pine stands prior to harvest had frequencies less than 0.01. The difference in number of rare alleles between the two studies most likely reflects differences in the markers used to evaluate these populations. For example, there could be dissimilarities in allele frequency distributions among loci, some markers will have more rare alleles than others. The Ontario populations were characterized at 6 additional loci above the number used to characterize the Hartwick Pines study. On average, 81% of alleles were shared between populations, similar to the 86% for the Ontario stands (Rajora et al. 2000). Private alleles are rare or low frequency alleles that are present in one population but absent in another. This allele class can be useful for estimating gene flow, or the rate of successful migrants exchanged among local populations (Slatkin 1985). Five of the seven loci contained private alleles. On average, 9% 92 LQ C‘f II f< D1 of alleles were private in the Hartwick Pines study as compared to 7% for the Ontario stands (Rajora et al. 2000). Since both studies conducted a complete census with approximately the same number of trees sampled in each old growth population, the differences in the reported number of private alleles most likely reflects the larger number of loci sampled in the Ontario study. In both Hartwick Pines populations, the number of private alleles is too low to reliably estimate migration rates. Slatkin (1985) recommends 20 or more private alleles per sample for migration estimates to be close enough to the true value to be useful. It is interesting to note that the only loci without private alleles (Rp884 and Rps127) are the only two classified as compound perfect repeats. One explanation for the absence of private alleles is the lack of new mutations. It is generally accepted that interruptions decrease SSR mutation rates (Weber & May 1989)(Taylor et al. 1999). Interruptions are also the hypothesized first step in the “death” of a microsatellite, stabilizing the repeat for subsequent deletion (Taylor et al. 1999). If the compound repeats in these two loci are recent formations, interruptions could help explain the absence of private alleles. 93 Ge pr 10 ch 00' et was 0.4 val gen Opp( The for Popu: Genetic Diversity: Conifers combine several life history traits that promote high genetic diversity: large geographic range, long lived perennial, late successional species, high chromosome number, sexual reproduction, high fecundity, outcrossing breeding system, and wind pollination (Hamrick et al. 1979)(Ledig 1998). Average expected heterozygosity was moderately high for the two populations combined (He:= 0.49) and just 4% higher than the observed random mating values (Ho = 0.47), suggesting little inbreeding. Mean genetic diversity values were 20% lower than values reported by Rajora et al. (2000). For the Ontario study, kg, averaged over both pre—harvest populations, was 0.61, as compared to 0.49 for the Hartwick Pines populations. Two premises could explain the decrease in the measured diversity as compared to the Ontario populations: (1) differences in marker polymorphism and allele frequencies; or (2) the Hartwick Pines populations are younger with less opportunity for selection against homozygous genotypes. The average stand age is 175 years as compared to 250 years for the Ontario populations. Heavy logging occurred in the Hartwick region near the turn of the century. The SS was logged while the CG population was left undisturbed. Since logging reduces the 94 be me wc re be 81' 96 1c be 00 PO: Sir population size, it could be possible, under certain circumstances, for logging to decrease the diversity of the regenerating seedlings. A decrease in diversity would depend on many factors and would not occur every time a forest is logged. For example, a severe reduction in effective population size would leave just a few breeding individuals. If new seed parents were not introduced from outside of the stand, then just a few seed parents would be available to re—establish the stand, reducing the diversity of the regenerating progeny. This is because alleles would be lost since fewer parents would contribute genes to the new generation of white pine. In addition, genetic drift would cause further erosion of genetic diversity through random loss of alleles during gametic sampling. In this study, there is little genetic difference between stands, with diversity measures (A, Ae,IH,and Ho) similar between populations. An explanation for similar genetic diversity between the 0G and SS stands is the logging induced change in Ng'was not severe enough to cause a decline in diversity levels. A second explanation would be the existence of a viable seedbed. Intense fires often occurred following logging. If the SS stand was spared a Post logging burn, genetic diversity would be maintained since bmmh.seedbeds and younger uncut trees, along with the 95 seed trees, could have all contributed seed to the regeneration of the SS. Population Structure: In addition to high genetic diversity, we expect population structure to be weak for both populations. This is because white pine is primarily a wind pollinated, out— crossing species with little bi-parental inbreeding. The losses of heterozygosity from population structure are often measured by fixation indices or Sewall Wright's F- statistics. In support of this hypothesis of weak structure, mean F values indicated little inbreeding in the populations. The mean microsatellite fixation indices were 45% lower than the allozyme analysis of the same tree populations (Table 1)(Epperson & Chung 2001). Mean fixation indices, were 0.05 and 0.11 for the SSR and allozyme marker systems, respectively. These results are consistent with similar findings in the Ontario white pine genetic study. Using their reported H5 and He values (F = 1 — (Ho / He)), the mean microsatellite fixation index (0.141) was 20% lower than the corresponding allozyme value (0.179) (Rajora et al. 2000)(Buchert et al. 1997). Twenty—nine percent of the 0G loci differed significantly from Hardy—Weinberg expectations. The positive F value for Rps39 indicates an excess of 96 homozygotes. Inbreeding from mating relatives or as a result of population subdivision from geographic or temporal isolation (Wahlund effect), seems an unlikely explanation for the observed heterozygote deficiency. In theory, inbreeding affects all loci, but the observed significant deviation above H-W equilibrium for the CG population occurred at just 1 of 14 F values. One possibility for this decrease in heterozygosity would be the presence of a null allele, and there is some evidence to support this premise. Null homozygotes could explain the failed Rps39 PCR reactions for two trees in the CG population. The frequency of the Rps39 null allele was estimated from the observed deficiency of heterozygotes (Brookfield 1996). Using equation 4, individuals with no visible bands were treated as observations with the null allele frequency (r) = (He-IQJ(1 + He)°. The frequency of null alleles is estimated to be 0.07 in both populations combined. If assuming the two failed PCR reactions were the result of double null alleles, then the observed null homozyogte frequency would be 0.008 (2 of 242 trees) with a null allele frequency of 0.09 (0.008”), which is close to the frequency estimated from Brookfield’s (1996) equation. As mentioned above, there is compelling evidence that null alleles also occur in marker Rps34b. Since marker Rps6 97 amplifies the same locus as Rps34b, and did not contain null alleles, Rps34b was removed from the final analysis. Population divergence was negligible. Because statistical power was very high, the small differences in allele frequencies were statistically significant for 71% of loci. Spatial distance between populations or perhaps timber harvest at the SS, are reasonable explanations for the observed differences in allele frequencies between populations. The effects of mutation and migration are similar in that they both introduce new genes into a population, but biologically the effects of these two evolutionary forces are very different. Mutation normally causes differentiation, in contrast to migration, which has a homogenizing affect. Typically, allozyme mutation rates are much lower than migration rates and would not affect Fm, But if mutation rates are high relative to migration, new mutations may increase variance among populations, by negating some of the homogenization of migration. This situation may apply to the relationship between Fm and the higher mutation rates found at some highly mutable microsatellite loci. The mutation rates of microsatellite markers are estimated from 10'4 — 104, and are 100 to 10,000 times higher than for the 10'6 rate estimated for allozyme 98 markers (Voelker et al. 1980)(Jin et al. 1996). In addition, mutation rates vary among alleles at a microsatellite locus (Jin et al. 1996), increasing the variance within loci and may contribute to the heterogeneity in Fa values observed at individual loci. In addition to differentiation, mutation can also have a homogenizing effect, analogous to migration. For example, if mutation rates are high enough, the chance of the same mutation occurring independently in two different populations increases (homoplasy). An increase in homoplasy would decrease the resolving power of microsatellites to detect differences among populations, decreasing the variance among populations. Spatial Structure: Dispersal is distance dependent with individuals proximal to one another being more closely related through local inbreeding (Wright 1943), if they are not removed by inbreeding depression. Wright (1943; 1946) describes these effects of isolation by distance in causing genetic differences between subpopulations of a continuous population. We expected spatial structure for the adults to be weak in both populations because white pine pollen and seed are highly dispersed by wind. Although seed dispersal can be much shorter than pollen dispersal, 99 extensive pollen flow or gene migration would disrupt clustering of related trees from low seed dispersal, preventing neighborhood structure from developing at short distances. In support of this premise of weak spatial structure, Moran’s I analysis indicated a randomly distributed second growth population and weak positive spatial structure at short distances for the CG population. The results suggest that logging decreased spatial structure at the second growth stand. The spatial genetic structure for OG at short distances is consistent with Moran's I values reported for other species with similar wind dispersed pollen and seed, i.e. 0.07 for Maclura pomifera, 0.04 for Gleditsia trianthos (Schnabel et al. 1991), and 0.05 for Quercus laevis (Berg & Hamrick 1995). The overall mean I value for distance class one was 0.02, compared to an isozyme value of 0.06 for this same population (Epperson & Chung 2001). The lower microsatellite Moran's I values could be a function of low frequency alleles, which may decrease measures of autocorrelation (Epperson, personal communication). It may also be that the high mutation rates at SSR loci directly decrease spatial structure (Epperson 1990). 100 Wright's (1946) neighborhood size, or the number of mating individuals (N) drawn at random from within a circle of area 41:0'2 and radius 20, can be used to estimate total dispersal distance. Based on Epperson et al (1999) simulations the Moran's I value of 0.02 corresponds to a Wright's neighborhood size of approximately 300 trees for the CG population (Line 1, Table 4). Standardized for density, the combined seed and pollen dispersal distance (0) was estimated using the neighborhood formula for a monecious population, N = 4n083 (Wright 1952). With a stand density (d) of 100 trees haqy the standard deviation 0 would be approximately 47 m and the average variance of the parent-offspring distance (02) approximately 2210 m2. Seed dispersal can be shorter than pollen dispersal. To allow different offspring dispersal distance estimates for the male and female parents, Crawford (1984) proposed the combined female parent—offspring'(cfis) and the male parent- offspring (02p) dispersal variance to be 0T2 = 1/202p + 02s. Typical empirical measures of dispersal distances for pine pollen range from 17 meters (Wright 1952) to 69 meters (Wang et al. 1960), and for seed from 15 to 30 meters (Epperson & Allard 1989). Given the total parent variance of 2210 n? and the average pollen standard deviation of 43 m, the seed dispersal distance would be 27 m, which agrees 101 well with the average empirical seed dispersal distance of 23 m. Anticipated levels of neighborhood differentiation can also be estimated from neighborhood size. At the CG site there is little differentiation with N approximately 300 trees. For the SS site, the mean M(I) = 0.001, which corresponds to randomness or no structure. With a mean Fm; of 0.0054 indicating little genetic difference between populations, these microsatellite estimates of N and F“; agree reasonably well with Wright's (1943) predictions. Wright concludes that greater differentiation occurs when N=10, moderate differentiation at N=100, and approaching panmixia at N=1000. The allozyme study of these same two populations approximates a smaller neighborhood size of 100 trees for 0G adults (Epperson & Chung 2001), with little genetic difference between populations (§fl_of 0.008). 102 CONCLUS I ON Mutational mechanisms responsible for generating microsatellite polymorphisms are important factors to consider in conducting genetic evaluations and population studies. The complex mutational processes do not negate the usefulness of these markers, but it is important to understand SSR mutation and how it affects the statistics used to describe population genetic and spatial structure. Autocorrelation fits predicted levels for white pine, based on pollen and seed dispersal and density. The results for the microsatellites are slightly lower than for the allozymes, but are not inconsistent. In addition to statistical error, SSRs may directly affect spatial statistics through mutation and allele frequency. Gene flow was high and population divergence negligible with the levels of genetic diversity similar between stands. Spatial distance between populations or timber harvest at the SS, are reasonable explanations for the observed differences in allele frequencies between populations. The weak positive structure at short distances observed for the OS site suggests spatial structure fits the isolation by distance model for predicted genetic differences in a continuous population. 103 Whereas, the results suggest that logging of the second growth stand removed spatial genetic structure. 104 APPENDIX 105 APPENDIX Spatial Autocorrelation Coefficients 106 Table 13: Spatial autocorrelation coefficients (Moran's I) for 8 loci in the Old Growth population (OG) of Pinus strobus for 10 distance classes, n = 1221. DistBoundz 15 25 35 45 55 65 DistClass3 1 2 3 4 5 6 Rpslb Zn}: 244 No. Pairs5 421 649 843 886 883 876 Locusl-03 -0.03 -0.04 -0.02 0.00 -0.04 0.03 Locusl-04 —0.02 -0.03 -0.02 0.01 -0.04 0.03 Locusl—OS -0.01 -0.02 -0.01 -0.00 -0.01 —0.01 Locusl—06 0.06 -0.04 -o.06 -0.03 0.04 0.01 Locusl-08 0.04 0.02 -0.10** -0.04 0.06** 0.01 Average 0.01 —0.02 -0.04 —0.01 0.00 0.01 Rpsz 2n.= 244 NO. Pairs 421 649 843 886 883 876 LOCUSZ-Ol 0.04 -0.02 0.02 -0.05 -0.02 -0.01 LOCUSZ-OZ 0.00 -0.01 -0.0l -0.01 -0.00 *0.01 LOCUS2-03 -0.00 -0.02 -0.00 -0.00 -0.01 -0.01 LOCUS2-04 -0.0l -0.00 0.02 -0.01 -0.01 0.00 LOCUSZ-OG 0.06 0.01 -0.02 -0.06* -0.10** 0.02 LOCUS2-07 0.04 -0.02 0.02 -0.06* -0.09** 0.06* LOCUSZ-OB -0.01 -0.00 -0.0l -0.01 -0.00 -0.0l LOCUSZ-OQ -0.01 -0.00 -0.0l -0.01 -0.01 -0.01 LOCUSZ—lo -0.03 -0.04 0.01 -0.01 -0.01 0.00 Average 0.01 -0.01 0.00 -0.02 -0.03 0.00 1 Sample size of study population 2 Distance bound (meters) 3 Distance class 4 Haploid sample size for each locus 5 Number of joins * p < 0.05 H p < 0.01 107 Table 13 continued DistBound DistClass Rpslb No. Pairs Locusl-03 Locusl-O4 Locusl—OS Locusl-O6 Locusl-08 Average Rpsz No. Pairs LocusZ-Ol LocusZ-OZ Locus2-03 LocusZ—O4 Locus2-06 LocusZ-O? LocusZ-OB LocusZ-O9 LocusZ-lo Average I 000000 75 684 .04* .02 .01 .05* .04 .03 684 .02 .01 .02 .00 .03 .03 .01 .01 .00 .01 -0. -0. -0. -0. -0. -0. 85 8 627 01 -0. 02 -0. 01 -0. 02 -0. O6 -0. 02 -0. 627 .00 0. .00 -0. .01 -0. .06 0. .07* 0. .08* 0. .02 -0. .02 —0. .02 -0. .01 0. 428 03 03 01 06 04 428 06 01 00 03 04 02 01 04 02 11** 1 1 105 10 084 .01 .00 .00 .02 .01 .01 084 .02 .01 .00 .03 .06* .04 .01 .01 .01 .02 OOHOO HHHOOOOOO .498 .812 .000 .498 .023 .638 .584 .650 .689 .023 .047 .000 .000 .000 00000 000000000 .0123 .0123 .0082 .9180 .0492 .0451 .0041 .0041 .0410 .5656 .3156 .0041 .0041 .0164 Overall correlogram significance (Bonferroni approx.) '7 Allele frequency 108 Table 13 continued DistBound 15 DistClass 1 R286 2n.= 244 No. Pairs 421 Locus6—01 -0.01 —O LocusG—OZ —0.05 O LocusG—O3 0.02 -0 Locu36-04 —0.01 -O Locus6-05 0.06 0 LocusG-OG -0.01 -O LocusG-O? -0.02 -O Locus6-09 0.00* -0 Average —0 OO —O Rps34b 2n.= 244 No. Pairs 421 LocusB4-Ol -0.0l -O Locus34—02 -0.06 -O Locus34—03 -0.05 —0 Locus34-04 -0.01 -O Locus34—05 0.06 0. Average -0.01 O. R2539 2n = 240 No. Pairs 412 Locus39-02 0.08* —O. Locu339-03 0.10** —O. Locus39-O4 0.02 -O. Locus39-05 -0.00 -0. Average 0.04 -O. 649 .00 .00 .03 .01 .07* .01 .05 .00 .00 649 .00 .01 .02 .01 07* 01 630 00 02 01 01 01 843 .01 0.01 0.03 .01 .06 .01 0.00 .00 .01 843 .01 0.01 0.07** .01 .06 0.00 821 0.01 .01 .00 .00 .00 -0. -0. -0. -0. -0. -0. 886 .01 .05 .01 .01 .03 .01 .01 .00 .01 886 01 05 O4 01 03 03 865 .04 .02 .01 .01 .02 -0. -0. -0. -0. -O. -0.06* -0 -0 883 01 .01 01 01 01 01 .01 .01 883 .01 .01 .00 .01 .01 .00 868 .05 .01 .03 .03 .03 0. -0. -0. 876 .01 .01 .01 .01 .00 .01 01 00 00 876 .01 .01 .02 .01 .00 .00 848 .05 .05 .01 .07* .05 109 Table 13 continued DistBound 75 85 95 105 DistClass 7 8 9 10 Rps6 No. Pairs 684 627 428 1084 p q LocusG—Ol -0.01 -0.01 —0.00 -0.01 1.000 0.0041 Locus6-02 0.00 —0.05 0.04 -0.01 0.715 0.1967 Locus6-03 0.00 -0.08* -0.00 -0.03 0.267 0.6844 Locus6-04 -0.01 —0.00 0.00 -0.01 1.000 0.0041 LocusS-OS 0.02 -0.06 —0.04 -0.01 0.119 0.0656 LocuS6-06 -0.01 -0.00 0.00 -0.01 1.000 0.0041 Locus6-07 0.08** -0.04 0.01 -0.00 0.074 0.0369 Locu86-09 -0.02 -0.02* —0.03** -0.01 0.083 0.0041 Average 0.01 —0.03 —0.00 -0.01 Rps34b No. Pairs 684 627 428 1084 p q Locus34-01 -0.01 -0.01 -0.00 -0.01 1.000 0.0041 Locus34—02 -0.01 —0.04 0.04 -0.00 0.689 0.2008 LocusB4-03 -0.01 —0.06 0.02 -0.02 0.076 0.7254 Locus34-04 -0.01 —0.00 0.00 —0.01 1.000 0.0041 Locu834-05 0.02 -0.06 -0.04 -0.01 0.119 0.0656 Average -0.01 -0.04 0.01 —0.01 Rps39 No. Pairs 660 598 409 1029 p q Locus39—02 0.01 0.04 0.05 -0.03 0.253 0.6167 Locus39—O3 -0.04 0.05 0.03 —0.02 0.094 0.3375 Locus39-04 —0.03 —0.04 0.04 0.02 0.698 0.0125 Locus39-05 -0.00 -0.01 0.13** —0.01 0.006 0.0333 Average -o.02 0.01 0.06 -o.01 110 Table 13 continued DistBound 15 25 DistClass 1 2 R2350 2n = 244 No. Pairs 421 649 LocusSO-Ol 0.01 —0.03 LocusSO-02 -0.02 -0.01 Locus50-03 0.01** 0.01** LocusSO-OS -0.03 -0.02 LocusSO—06 0.02 0.05 LocusSO-07 —0.05 -0.00 LocusSO-08 0.04 0.03 LocusSO-09 -0.03 —0.05 LocusSO-lo -0.03 0.03 LocusSO-ll —0.02 —0.02 LocusSO-12 -0.02 -0.03 LocusSO-13 -0.05 -0.01 LocusSO-14 —0.02 -0.02 Average -0.02 —0.01 R9584 2n.= 244 No. Pairs 421 649 LocusB4-01 0.00 —0.01 Locus84-02 0.01 -0.03 Locus84—03 —0.03 -0.02 Locus84-04 -0.03 -0.02 Locus84—05 -0.03 —0.00 Locus84—06 -0.01 —0.03 Average -0.01 —0.02 Rpslz7 2n.= 244 No. Pairs 421 649 Locu3127-01 0.03 —0.05 Locu3127-02 0.03 —0.05 Average 0.03 -0.05 843 -0.01 0.00 0.01* -0.00 -0.03 0.01 -0.09** 0.01 -0.10** -0.02 0.02 -0.03 -0.03 -0.02 843 -0.00 0.03 -0.02 0.03 -0.02 -0.03 -0.00 843 -0.01 -0.01 -0.01 886 -0.02 -0.00 0.00* 0.00 -0.06* 0.01 -0.03 0.04* -0.02 -0.03 -0.03 -0.03 0.02 -0.01 886 -0.00 -0.04 —0.02 -o.07* 0.02 -0.04 -0.03 886 0.04 0.04 0.04 883 -0.03 -0.00 -0.00 -0.02 -0.03 -0.01 0.04 -0.04 0.02 -0.02 0.00 0.02 0.02 -0.00 883 -0.01 -0.02 0.05** -0.01 -0.04 0.04* 0.00 883 -0.01 -0.01 -0.01 876 0.01 -0.01 0.00 -0.05 0.02 -0.02 -0.08* -0.03 0.05* 0.06** 0.02 -0.03 -0.04 -0.01 876 -0.04 -0.02 0.01 0.00 -0.00 -0.01 876 —0.02 -0.02 -0.02 111 Table 13 continued DistBound DistClass R2950 No. Pairs LocusSO-Ol LocusSO-OZ LocusSO-03 LocusSO-OS LocusSO-06 LocusSO-07 Locus50—08 LocusSO-09 Locus50-10 LocusSO-ll Locus50-12 LocusSO-13 LocusSO—14 Average R2384 No. Pairs Locus84-01 LocusB4-02 LocusB4—03 Locus84-04 Locus84-05 Locu584-06 Average R2812? No. Pairs Locu8127-01 Locu3127-02 Average -0. 0. -0. -0. -O. -O. -0. 0. 0. 0. 75 7 684 .04 -0. .01 0. .01 -0. .01 -0. .00 -0. .02 -0. .06* 0. .02 -0. .02 -0. .02 -0. .04 0. .06* -0. .04 -0. .01 -0. 684 01 -0. 01 -0. 00 -0. 00 0. 02 -0. 01 —0. 01 —0. 684 02 -0 02 -0 02 -0 627 00 05** 01 00 06 00 05 02 03 01 01 04 04 01 627 02 04 01 00 01 05 02 627 .06 .06 .06 95 428 0.09* -0.03 -0.02* 0.05 -0.02 0.02 0.06 0.01 —0.02 -0.00 -0.03 -0.03 0.07* 0.01 428 -0.01 0.08* -0.01 0.05 -0.05 -0.01 0.01 428 -0.05 -0.05 -0.05 105 10 1084 -0.02 -0.05* -0.05** -0.01 0.02 -0.01 -0.04 0.01 0.01 0.00 0.02 0.00 -0.01 1084 -0.01 0.00 -0.02 0.02 0.03 1084 0.00 0.00 0.00 OOI—‘OOOOI—‘OOOOO 000000 .107 .040 .028 .936 .360 .000 .029 .215 .028 .024 .000 .149 .185 .584 .235 .045 .239 .800 .481 .639 .639 OOOOOOOOOOOOO 000000 .0205 .0123 .0041 .1721 .2746 .1352 .1148 .0123 .1967 .0082 .0164 .0205 .0123 .0041 .7664 .0082 .1148 .0328 .0738 .7172 .2828 112 Table 14: Spatial autocorrelation coefficients (Moran's I) for 8 loci in the Second Site population (SS) of Pinus strobus for 10 distance classes, n 1201. DistBound2 15 25 35 45 55 65 DistClass3 1 2 3 4 5 6 R2slb 2n? = 238 No. PairsS 519 743 924 910 796 810 Locusl-Ol -0.02 -0.01 -0.01 -0.02 -0.01 -0.01 Locusl-OZ -0.01 -0.02 -0.02 -0.01 -0.01 -0.00 Locus1-03 -0.01 -0.02 -0.02 -0.02 -0.01 -0.01 Locus1-04 0.01 —0.02 -0.02 -0.03 0.01 -0.01 Locusl-OS —0.04 -0.02 -0.03 0.05** -0.02 —0.02 Locusl-06 -0.04 -0.01 -0.03 -0.02 -0.03 -0.03 Locusl-07 -0.03 -0.03 0.04* -0.02 —0.01 -0.02 Locusl-08 0.01 0.01 -0.02 —0.03 0.01 -0.00 Average -0.02 -0.01 -0.01 -0.01 -0.00 -0.01 R282 2n = 240 NO. Pairs 524 753 933 921 804 822 LocusZ-Ol 0.13** -0.04 -0.07* —0.09** 0.01 0.08** Locus2-03 0.00 0.00 0.01* 0.00 0.00* -0.00 LocusZ-04 —0.06 0.01 —0.02 -0.03 0.03 -0.03 LocusZ-OS -0.01 —0.02 0.00 -0.00 -0.01 —0.01 LocusZ-06 0.01 -0.02 -0.00 —0.02 0.01 0.01 LocusZ-07 -0.03 0.06* -0.05 0.02 -0.01 0.03 LocusZ-OB -0.02 -0.01 -0.02 -0.02 -0.01 -0.01 LocusZ-09 0.00 -0.00 -0.00 -0.01 -0.02* -0.02* Locu82710 -0.00 —0.00 -0.02 —0.03 0.03* -0.02 Average 0.00 -0.00 -0.02 -0.02 0.00 0.00 1 Sample size of study population 2 Distance bound (meters) 3 Distance class 4 Haploid sample size for each locus 5 Number of joins * p < 0.05 “ p < 0.01 113 Table 14 continued DistBound DistClass R2slb No. Pairs Locusl-Ol Locusl-02 Locus1—03 Locusl-04 Locusl-OS Locusl-06 Locusl-07 Locusl-08 Average R232 No. Pairs Locusz-Ol LocusZ-03 LocusZ-04 Locus2-05 LocusZ-06 Locus2-07 Locus2-08 Locus2-09 LocusZ—lo Average 75 85 7 8 768 663 -0.02 -0.03 -0.01 0.00 0.06** -0.02 -0.01 -0.01 -0.02 -0.01 -0.01 0.02 -0.01 -0.02 0.03 -0.04 0.00 —0.01 780 678 -0.01 -0.03 -0.00 -0. 0.01 -0.00 -0.01 -0.01 0.00 -0.04 -0.03 -0. -0.00 0.00 -0.01 -0.01 -0.01 -0.00 -0.01 -0.03 03** 11** 95 458 0.09** 0.00 -0.02 -0.02 0.01 -0.02 0.00 0.03 0.00 472 0.01 -0.03* 0.02 -0.01 0.02 0.01 -0.00 -0.02 -0.00 1 05 10 430 -0. -0. —0. 0. 0. —0 —0. 0 02 00 03 .05 01 .02 .01 .04 .00 453 02 .07** 01 .00 .02 .00 .01 .00 .00 .01 HOHOOOOO OOOOHOOOO .013 .743 .016 .709 .067 .000 .114 .000 .003 .009 .890 .523 .000 .013 .641 .106 .362 00000000 000000000 .0084 .0042 .0084 .0546 .0084 .8697 .0084 .0378 .0500 .0042 .0500 .0042 .6917 .1833 .0042 .0042 .0083 6 7 114 Overall correlogram significance (Bonferroni approx. Allele frequency Table 14 continued DistBound2 15 25 35 45 55 65 DistClass3 1 2 3 4 5 6 R236 2n = 240 No. Pairs 524 753 933 921 804 822 Locu36—01 -0.01 -0.00 -0.02 -0.01 —0.01 -0.00 Locu36-02 -0.10* 0.03 -0.01 -0.01 -0.02 0.01 Locu36-03 -o.05 -0.01 0.00 0.02 -0.02 -0.02 Locus6-04 0.03 0.00 0.00 -0.03 0.02 —0.06* Locus6-05 0.11** 0.11** 0.04* -0.01 -0.03 -0.00 Locus6-06 0.00 0.00 0.01* 0.00 0.00* -0.00 Locus6-07 0.08* -0.05 0.07** —0.06* -0.01 -0.00 Locus6-08 -0.02 -0.03 -0.03 -0.02 0.04* -0.01 Locus6-09 -0.00 -0.00 -0.00 -0.00 —0.00 —0.01 Average 0.00 0.01 0.01 -0.01 -0.00 -0.01 R2334b 2n = 238 No. Pairs 702 702 702 702 702 702 Locus34—01 -0.01 -0 00 —0.02 -0.02 -0 01 -0.00 Locus34—02 -0.04 0.01 0.02 0.04 -0.04 -0.04 Locus34-03 -0.06 -0.01 0.02 0.00 -0.01 -0.06 Locus34-04 0.01 0.05* -0 03 -0.03 0.01 -0.02 Locus34-05 0.06* 0.13** 0.04 -0.01 0.00 —0.08* Average -0.01 0.03 0.01 -0.00 -0.01 -0.04 R2339 2n = 240 No. Pairs 524 753 933 921 804 822 Locu339—01 -0.02 —0.01 -0.02 -0.01 —0.01 -0.00 Locus39-02 0.06 -0.04 -0.01 -0.09** 0.03 -0.02 Locu339-03 0.02 -0.02 -0.04 -0.05 0.09** -0.04 Locus39-04 -0.01 -0.03 0.01 -0.05 -0 00 0.02 Locus39-05 -0.09* 0.00 0.02 0.01 -0 05 0.01 Average -0.01 -0.02 -0.01 -0.04 0.01 -0.01 115 Table 14 continued DistBound DistClass R236 No. Pairs Locus6—01 Locu36-02 Locu36-03 Locus6-04 Locu36-05 Locu36-06 Locu36—07 Locus6-08 Locu36-09 Average R2334b No. Pairs Locus34-01 Locus34-02 Locus34-03 Locu334-04 Locus34-05 Average R2339 No. Pairs Locu339-01 Locus39-02 Locus39-03 Locus39-O4 Locu339-05 Average 75 7 780 .01 -0 .01 -0 .00 0. .01 -0 .14** —0 .00 -0 .04 -0 .01 -0 .01 -0 .02 -0 702 .01 -0 .01 -0 .05 0. .05 —0 .05 -0 .01 -0 780 .01 -0. .05* 0. .02 -0. .01 0. .03 -0. .01 -0. 678 .01 .01 02 .03 .05 .03** .04 .01 .01 .02 702 .00 .02 02 .00 .10** .02 678 01 01 01 01 01 00 95 472 .00 .02 .04 .00 .09* .03* .02 .00 .02 .03 702 .01 .00 .03 .01 .08* .03 472 .00 .07 .02 .04 .03 .03 1 4 -0. -0 —0. -0. 05 10 53 .00 .00 .02 .01 .03 .07** .00 .01 .05* .02 703 01 .02 .00 .01 01 01 453 .00 .03 .05 .02 .02 .01 OOOOOOI—‘OH CDCDCDCDI-I OCDCDCDE-l P .000 .127 .000 .441 .000 .009 .045 .151 .250 .000 .887 .534 .412 .000 .000 .050 .017 .726 .184 00000 00000 q .0042 .2083 .6542 .0167 .0417 0.0042 0.0583 0.0083 0.0042 OOOOO q .0042 .2059 .7227 .0168 .0504 .0042 .6750 .2625 .0292 .0292 116 Table 14 continued DistBound 15 DistClass 1 R2350 2n = 240 No. Pairs 524 Locu350-01 0.00 Locu350—04 —0.03 LocusSO-OS -0.01 Locu350-06 0.02 Locu350-07 -0.02 Locu350-08 —0.02 Locu350—09 -0.02 Locu350-10 —0.00 Locu350-11 0.00 Locu350-12 0.01 Locu350-13 0.04 Locu350-14 -0.03 Average —0.00 R2384 2n = 240 No. Pairs 524 Locu384—01 0.01* Locu384-02 0.00 Locu384-03 -0.01 Locu384-04 0.01 Locu384-05 -0.03 Locu384—06 -0.03 Average —0.01 R23127 2n = 232 No. Pairs 496 Locu3127-01 0.03 Locu3127-02 0.03 Average 0.03 753 .01 .02 .09* .00 0.09** .01 .02 .05 0.02 .06 .00 .01 .01 753 0.00 0.06* .01 0.09** 0.02 .04 0.02 720 0.05* 0.05* 0.05* 35 3 933 -0.02 -0 -0.03 0 0.01 0 -0.03 0 -0.02 0 0.02 0 0.01 -0 0.01 0 0.02 -0 -0.06* 0 -0.01 -0 0.03 -O -0.01 0. 933 0.00* 0. 0.00 -0 -0.01 -0 -0.05 -0 -0.00 -0 0.02 0. -0.01 -0 897 0.05* -0 0.05* -0 0.05* -0 921 .03 .03 .04 .02 .02 .05* .01 .02 .01 .03 .03 .06* 01 921 01* .05 .01 .01 .04 00 .02 887 .04 .04 .04 804 804 0.00* .03 .02 .03 .03 .03 .02 757 0.03 0.03 0.03 822 0.03 .03 .06 .10* .04 0.01 .02 .05 .05 .02 0.00 .02 .03 822 0.00* .03 0.04* 0.01 .00 0.00 0.00 753 -0.03 -0.03 -0.03 117 Table 14 continued DistBound 75 DistClass 7 R2350 No. Pairs 780 Locu350-01 -0.03 Locu350—04 -0.03 Locu350-05 —0.03 Locu350—06 0.02 Locu350-07 -0.04 Locu350-08 -0.03 Locu350-09 -0.03 Locu350-10 0.02 Locu350-11 -0.02 Locu350-12 —0.02 Locu350-13 -0.03 Locu350-14 -0.01 Average -0.02 R2384 No. Pairs 780 Locu384-01 -0.01 Locus84—02 —0.02 Locus84-03 -0.02 Locus84-04 —0.06 Locu384-05 -0.04 Locu384-06 -0.00 Average —0.02 R23127 No. Pairs 719 Locu3127—01 —0.06 Locu3127-02 -0.06 Average -0.06 678 —0.02 -0.03 0.05 -0.02 0.02 0.04 -0.02 -0.03 0.02 -0.02 0.01 -0.01 678 -0.02 -0.02 -0.03 -0.00 0.02 0.00 -0.01 608 -0.05 -0.05 -0.05 95 472 0.05 0.06* 0.05 0.03 ~0.05 -0.06 -0.01 0.03 -0.02 -0.01 -0.03 0.01 0.00 472 -0.05** 0.03 -0.02 -0.08* 0.03 -0.01 -0.02 426 -0.14** -0.14** -0.14** -0. 105 10 453 0.03 -0.02 -0.04 -0.01 -0.03 -0.04 -0.01 -0.04 -0.07 -0.00 -0.04 -0.00 -0.02 453 0.00 -0.01 0.05 0.02 -0.01 -0.01 407 0.03 0.03 0.03 09** P .627 .256 .107 .031 .021 .276 .607 .852 .538 .257 .628 .378 OOOOOOOOOOOO 0.001 0.299 0.195 0.022 1.000 1.000 0.020 0.020 0.020 OOOOOOOOOOOO 000000 00 q .0125 .0125 .2292 .2750 .1625 .0542 .0167 .1458 .0208 .0208 .0250 .0250 .0042 .6958 .0083 .1917 .0417 .0583 .7457 .2543 118 LITERATURE CITED 119 LITERATURE CITED National data climate center (NCDC). Grayling weather station. June 2000. . Adams, W. T. Gene disperal within forest tree populations. New Forests. 1992; 6:217-240. Amos, Bill and Hoelzel, A. Rus. Applications of molecular genetic techniques to the conservation of small populations. Biological Conservation. 1992; 61:133- 144. Beaulieu, J. and Simon, J.-P. Genetic structure and variability in Pinus strobus in Quebec. Can. J. For Res. 1994a; 24:1726-1733. ---. Inheritance and linkage relationships of allozymes in Pinus strobus L. Silvae Genetica. 1994b; 43(4):253— 261. ---. Mating system in natural populations of eastern white pine in Quebec. Can. J. For Res. 1995; 25:1697-1703. Beaulieu, Jean and Simon, Jean-Pierre. Variation in cone morphology and seed characters in Pinus strobus in Quebec. Can. J. Bot. 1995; 73:262-271. Becher, S. A. and Griffiths, R. Genetic differentiation among local populations of the European hedgehog (Erinaceus) in mosaic habitats. Molecular Ecology. 1998; 7:1599-1604. Berg, Edward E. and Hamrick, James L. Fine-scale genetic structure of a turkey oak forest. Evolution. 1995; 49(1):110-120. 120 Blanquer-Maumont, Anne and Crouau-Roy, Brigitte. Polymorphism, monomorphism, and sequences in conserved microsatellites in primate species. J Mol Evol. 1995; 41:492—497. Blouin, M S; Parsons, M; Lacaille, V, and Lotz, S. Use of microsatellite loci to classify individuals by relatedness. Molecular Ecology 1996; 5, 393-401. Botstein. David; White, Raymond L.; Skolnick, Mark, and Davis, Ronald W. Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Human. Genet. 1980; 32:314-331. Brookfield, J. F. Y. A simple new method for estimating null allele frequency from heterozygote deficiency. Molecular Ecology. 1996; 5:453—455. Brym, Petr and Eckert, Robert T. Within-stand clustering of eastern white pine genotypes. in. Proceedings of the 30th NE forest tree improvement conference. Orono, ME: School of Forestry Resources, U of ME; 1986 Jul 22- 1986 Jul 24; pp. 126-138. Buchert, George P.; Rajora, Om P.; Hood, James V., and Dancik, Bruce P. Effects of harvesting on genetic diversity in old-growth eastern white pine in Ontario, Canada. Conservation Biology. 1997; 11(3):747—758. Bunville, Julie; Krakowski, Karen, and Seto, Donald. Cerium oxide removes background fluorescent contaminants, "green smears", from glass plates used for automated fluorescence-based DNA sequencing. BioTechniques. 1997; 22(2):236—238. Callen, David F.; Thompson, Andrew D.; Shen, Yang; Phillips, Hilary A.; Richards, Robert I.; Mulley, John C., and Sutherland, Grant R. Incidence and origin of "null" alleles in the (AC)n:microsatellite markers. Am. J. Human. Genet. 1993; 52:922-927. 121 Chagala, E. M. Inheritance and linkage of allozymes in Pinus strobus L. Silvae Genetica. 1996; 45(4):181-187. Chung, M. G., and Epperson, B. Spatial genetic structure of clonal and sexual reproduction in populations of Andenophora grandiflora (Campanulaceae). Evolution. 1999; 53(4):1068-1078. Clark, James M. Novel non-templated nucleotide addition reactions catalyzed by procaryotic and eucaryotic DNA polymerases. Nucleic Acids Research. 1988; 16(20):9676-9687. Cliff, A. D. and Ord, J. F. Spatial process-methods and applications. Pion, London; 1981. Cornelius, Jonathan. Heritabilities and additive genetic coefficients of variation in forest trees. Can. J. For. Res. 1994; 24:372-379. Crawford, T. J. The estimation of neighbourhood parameters for plant populations. Heredity. 1984; 52(2):273-283. Dallimore, W. and Jackson, A. Bruce. A handbook of coniferae including ginkgoaceae. London: Edward Arnold & Co.; 1948. Davis, Margaret B. and Shaw, Ruth G. Range shifts and adaptive responses to quaternary climate change. Science. 2001; 292:673-679. Dewey, Sarah E. and Heywood, John S. Spatial genetic structure in a population of Psychotria Nervosa. I. Distribution of genotypes. Evolution. 1988; 42(4):834- 838. Don, R. H.; Cox, P. T.; Wainwright, B. J.; Baker, K., and Mattick, J. S. "Touchdown" PCR to circumvent spurious priming during gene amplification. Nucleic Acids Research. 1991; 19(14):4008. 122 Dong, Jinsheng and Wagner, David B. Paternally inherited chloroplast polymorphism in Pinus: estimation of diversity and population subdivision, and test of disequilibrium with a maternally inherited mitrochondrial polymorphism. Genetics. 1994; 136:1187- 1194. Echt, C. S.; May-Marquardt, P.; Hseih, M., and Zahorchak, R. Characterization of microsatellite markers in eastern white pine. Genome. 1996; 39(6):1102-1108. Echt, C. S. and Nelson, C. D. Linkage mapping and genome length in eastern white pine (Pinus strobus L.). Theor Appl Genet. 1997; 94:1031-1037. Echt, C. S.; Vendramin, G. G.; Nelson, C. D., and Marquardt, P. Microsatellite DNA as shared genetic markers among conifer species. Can. J. For Res. 1999; 29:365-371. Eckert, R. T.; Joly, R. J., and Neale, D. B. Genetics of isozyme variants and linkage relationships among allozyme loci in 35 eastern white pine clones. Can. J. For Res. 1981; 11:573-579. Eckert, Robert T. and Ryu, Jang Bal. Foliar isozyme variation in twenty seven provenances of Pinus strobus L.; relationship of three foliar isozyme alleles with height growth and multivariate patterns of range-wide variation. in. Proceedings of the seventh North America biotechnology workshop on physiology and genetics of intensive culture. Lexington, Kentucky: University of Kentucky; 1982 Jul 26—1982 Jul 28: 179- 187. El-Kassaby, Y. A.; Meagher, M. D., and Davidson, R. Temporal variation in the outcrossing rate in a natural stand of western white pine. Silvae Genetica. 1993; 42(2-3):131-135. 123 Epperson, B. K. Spatial patterns of genetic variation within plant populations. in: Brown, Anthony H. D.; Clegg, Michael T.; Kahler, Alex L., and Weir, Bruce 8., editors. Plant population genetics, breeding, and genetic resources. Sunderland, Massachusetts: Sinauer Associates; 1989; pp. 229-253. —--. Spatial autocorrelation of genotypes under directional selection. Genetics. 1990; 124:757-771. Epperson, B. K. and Allard, R. W. Allozyme analysis of the mating system in lodgepole pine populations. The Journal of Heredity. 1984; 75:212-214. —--. Spatial autocorrelation analysis of the distribution of genotypes within populations of lodgepole pine. Genetics . 1989; 121:369-377. Epperson, B. K.; Huang, Z., and Li, T.—Q. Measures of spatial structure in samples of genotypes for multiallelic loci. Genet. Res. Camb. 1999; 72:251-261. Epperson, Bryan (epperson@msu.edu). Low frequency alleles may reduce autocorrelation. 2001. Epperson, Bryan K. and Chung, Myong Gi. Spatial genetic structure of allozyme polymorphisms within populations of Pinus strobus (Pinaceae). American Journal of Botany. 2001. Epperson, Bryan K. and Li, TianQuan. Measurement of genetic structure within populations using Moran's spatial autocorrelation statistics. Proc. Natl. Acad. Sci. USA. 1996; 93:10528-10532. Erpelding, John (mayje@ars—grin.gov). Sucrose improves amplification of weak PCR reactions. 2001. 124 Estoup, Arnaud; Tailiez, Christian; Cornuet, Jean-Marie, and Solignac, Michel. Size homoplasy and mutational processes of interrupted microsatellites in two bee species, Apis mellifera and Bombus terrestris (Apidae). MOl. Biol. Evol. 1995; 12(6):1074-1084. Etterson, Julie R. and Shaw, Ruth G. Constraint to adaptive evolution in response to global warming. Science. 2001; 294:151-154. Farrar, John Laird. Trees of the Northern United States and Canada. Ames, Iowa: Iowa State University Press; 1995. Fields, R. L. and Scribner, K. T. Isolation and characterization of novel waterfowl microsatellite loci: cross-species comparisions and research applications. Molecular Ecology. 1997; 6:199-202. Fowler, D. P. Effects of inbreeding in red pine, Pinus resinosia Ait. IV. Comparison with other Northeastern Pinus species. Silvae Genetica. 1965; 14:76-81. Fowler, D. P. and Heimburger, C. Geographic variation in eastern white pine, 7-year results in Ontario. Silvae Genetica. 1969; 18(4):97-144. Garza, John Carlos; Slatkin, Montgomery, and Freimer, Nelson B. Microsatellite allele frequencies in humans and chimpanzees, with implications for constraints on allele size. Mol. Biol. Evol. 1995; 12(4):594-603. Goldstein, David B.; Linares, Andres Ruiz; Cavalli-Sforza, Luigi Luca., and Feldman, Marcus W. An evaluation of genetic distance for use with microsatellite loci. Genetics. 1995; 139:463-471. Gonick, Larry and Smith, Woollcott. The cartoon guide to statistics. New York: HarperCollins Publishers, Inc.; 1993. 125 Goodman, S. J. Rst Calc: a collection of computer programs for calculating estimates of genetic differentiation from microsatellite data and determining their significance. Molecular Ecology. 1997; 6:881-885. Grimaldi, Marie-Claude and Crouau-Roy, Brigitte. Microsatellite allelic homoplasy due to variable flanking sequences. J Mol Evol. 1997; 44:336-340. Gullberg, Annica; Tegelstrém, Hakan, and Olsson, Mats. Microsatellites in the sand lizard (Lacerta agilis): description, variation, inheritance, and applicability. Biochemical Genetics. 1997; 35(7/8):281—295. Guries, Raymond P. and Ledig, F. Thomas. Genetic structure of populations and differentiation in forest trees. in. Proceedings of the symposium on isozymes of North American forest trees and forest insects. General technical report PSW-48. Berkeley, CA: US Department of Agriculture, Forest Service, Pacific Southwest Forest and Range Experiment Station; 1979 Jul 27; pp. 42-47. ---. Genetic diversity and population structure in pitch pine (Pinus rigida Mill.). Evolution. 1982; 36(2):387- 402. Hamrick, J. L. and Godt, M. J. W. Allozyme diversity in plant species. in: Brown, A. H. D.; Clegg, M. T.; Kahler, A. L., and Weir, B. S., editors. Plant population genetics, breeding and genetic resources. Sunderland, Massachusetts: Sinauer Associates; 1989; pp. 43-63. Hamrick, J. L.; Godt, Mary Jo W., and Sherman-Broyles, Susan L. Factors influencing levels of genetic diversity in woody plant species. New Forests. 1992; 6:95-124. 126 Hamrick, J. L.; Mitton, J. B., and Linhart, Y. B. Levels of genetic variation trees: influence of life history characteristics. in. Proceedings of the symposium on isozymes of North American forest trees and forest insects. General technical report PSW-48. Berkeley, CA: US Department of Agriculture, Forest Service, Pacific Southwest Forest and Range Experiment Station; 1979 Jul 27; pp. 35-41. Hancock, John M. Microsatellites and other simple sequences: genomic context and mutational mechanisms. in: Goldstein, David B. and Schlétterer, Christian, editors. Microsatellites: evolution and applications. Oxford: Oxford University Press; 1999; pp. 1-9. Hartl, Daniel L. and Clark, Andrew G. Principles of population genetics. Third ed. Sunderland, Massachusetts: Sinauer Associates; 1997. Hecker, Karl H. and Roux, Kenneth H. High and low annealing temperatures increase both specificity and yield in touchdown and stepdown pcr. Biotechniques. 1996; 20(3):478—485. Herms, Daniel A.; Mattson, Wiliam J.; Karowe, David N.; Coleman, Mark D.; Trier, Terry M.; Birr, Bruce A., and Isebrands, J. G. Variable performance of outbreak defoliators on aspen clones exposed to elevated COzanui O3. in: Hom, John; Birdsey, Richard, and O'Brian, Kelly, editor. Proceedings, 1995 meeting of the northern global change program ; Pittsburgh, PA. PA. Gen. Tech. Rep. NE-214: U.S. Deptartment of Agriculture, Forest Service, Northeastern Forest Experiment Station; 1996. Hubby, J. L. and Lewontin, R. C. A molecular approach to the study of genic heterzygosity in natural populations. 1. The number of alleles at different loci in Drosophila pseudoobscura . Genetics. 1966; 54:577-594. 127 Isebrands, J. G.; McDonald, E. P.; Kruger, E.; Hendrey, G.; Percy, K.; Prietzer K.; Sober, J., and Karnosky D.F. Growth responses of Populus tremuloides clones to interacting elevated carbon dioxide and tropospheric ozone. Environmental Pollution. 2001; 115:359-371. Jarne, Philippe and Lagoda, Pierre J. L. Microsatellites, from molecules to populations and back. TREE. 1996; 11(10):424-429. Jin, Li; Macaubas, Claudia; Hallmayer, Joachim; Kimura, Akinori, and Mignot, Emmanuel. Mutation rate varies among alleles at a microsatellite focus: Phylogenetic evidence. Proc Natl Acad Sci. 1996; 93:15285-15228. Johnson, L. P. V. Reduced vigour, chlorophyll deficiency, and other effects of self-fertilization in Pinus. Canadian Journal of Research. 1945; 23, Sec. C.(5):l45-149. Karhu, A.; Dieterich, J.-H., and Savolainen, 0. Rapid expansion of microsatellite sequences in pines. Mol. Biol. Evol. 2000; 17(2):259-265. Kebelmann-Betzing, C.; Seeger, K.; Dragon, S.; Schmitt, G.; Moricke A.; Schild T.A.; Henze G., and Beyermann B. Advantages of a new Taq DNA polymerase in multiplex PCR and time-release PCR. BioTechniques. 1998; 24(1):154-158. Kim, Z. S.; Lee, S. W., and Hwang, J. W. Genetic diversity and structure of natural populations of Pinus thunbergii in Korea. Silvae Genetica. 1997; 46(2- 3):120-124. Kimura, Motoo and Crow, James F. The number of alleles that can be maintained in a finite population. Genetics. 1964; 49(4):725-738. 128 Knowles, P. Spatial genetic structure within two natural stands of black spruce (Picea mariana (Mill.) B.S.P.). Silvae Genetica. 1991; 40(1):13-19. Kubiske, Mark E. and Pregitzer, Kurt 8. Effects of elevated (Kb and light availability on the photosynthetic light response of trees of contrasting shade tolerance. Tree Physiology. 1996; 16:351-358. Kubiske, Mark E.; Pregitzer, Kurt S.; Mikan, Carl J.; Zak, Donald R.; Maziasz, Jennifer L., and Teeri, James A. Populus tremuloides photosynthesis and crown architecture in response to elevated Coz.and soil N availability. Oecologia. 1997; 110:328-336. Kuehl, Robert. Statistical principles of research design and analysis. California: International Thompson Publishing; 1994. Latta, Robert G. and Mitton, Jeffry B. A comparison of population differentiation across four classes of gene markers in limber pine (Pinus flexilis James). Genetics. 1997; 146:1153-1163. Le Corre, Valerie; Machon, Nathalie; Petit, Rémy J., and Kremer, Antoine. Colonization with long-distance seed dispersal and genetic structure of maternally inherited genes in forest trees: a simulation study. Genet. Res. Camb. 1997; 69:117-125. Ledig, Thomas K. Genetic variation in Pinus. in: Richardson, David M., in. Ecology and biogeography of Pinus. New York: Cambridge University Press; 1998; pp. 251-280. Legendre, Pierre. Spatial autocorrelation: trouble or new paradigm? Ecology. 1993; 74(6):1659-1673. Legendre, Pierre and Fortin, Marie-Josée. Spatial pattern and ecological analysis. Vegetatio. 1989; 80:107-138. 129 Levinson, Gene and Gutman, George. Slipped-strand mispairing: a major mechanism for DNA sequence evolution. Mol. Biol. Evol. 1987; 4(3):203—221. Li, C. C. and Horvitz, D. G. Some methods of estimating the inbreeding coefficient. Am. J. Human. Genet. 1953; 5(2):107-117. Li, Peng; Beaulieu, Jean; Daoust, Gaétan, and Plourde, Ariane. Patterns of adaptive genetic variation in eastern white pine (Pinus strobus) from Quebec. Can. J. For Res. 1997; 27:199-206. Linhart, Y. B.; Mitton, J. B.; Sturgeon, K. B., and Davis, M. L. Genetic variation in space and time in a population of ponderosa pine. Heredity. 1981; 46(3):407-426. Litt, Michael and Luty, Jeffrey. A hypervariable microsatellite revealed by in vitro amplification of a dinucleotide repeat within the cardiac muscle actin gene. Am. J. Human. Genet. 1989; 44:397-401. Lodge, David M. Biological invasions: lessons for ecology. Tree. 1993; 8:133-137. Mariette, Stephanie; David Chagne; Stephane Decroocq; Giuseppe Giovanni Vendramin; Celine Lalanne; Delphine Madur, and Christophe Plomion. Microsatellite markers for Pinus pinaster Ait. Amn. For. Sci. 2001; 58(2001):203-206. Marshall, D. R. and Brown, A. H. D. Optimum sampling strategies in genetic conservation. in: Frankel, O. H. and Hawkes, J. G., editors. Crop genetic resources for today and tomorrow. London: Cambridge University Press; 1975; pp. 53-80. 130 Mattson, William J. Exotic insects in North American forests: ecological systems forever altered. in: Britton, K. O., editor. Exotic pests of eastern forest; Nashville, TN. S. For. Res. Sta., P.O. Box 2680, Asheville, NC 28802: USD For. Serv & Tn Exotic Pest Plant Counc i1; 1998: 187-194. Mergen, Francois. Ecotypic variation in Pinus strobus L. Ecology. 1963; 44(4):?16-726. Messler, Walter; Li, Shou-Hsien, and Stewart, Caro-Beth. The birth of microsatellites. Nature. 1996; 381:483. Moore, S. S; Sargeant, L. L.; King, T. J.; Mattick, J. S.; Georges, M., and Hetzel, D. J. S. The conservation of dinucleotide microsatellites among mammalian genomes allows the use of heterologous PCR primer pairs in closely related species. Genomics. 1991; 10:654-660. Moran, P. A. P. Notes on continuous stochastic phenomena. Biometrika. 1950; 37:17-23. Moretti, Tamyra; Koons, Barbara, and Budowle, Bruce. Enhancement of PCR amplification yield and specificity using AmpliTaq Gold“ DNA Polymerase. BioTechniques. 1998; 25(4):7l6-722. Mullis, Kary B. and Faloona, Fred A. Specific synthesis of DNA in vitro via a polymerase-catalyzed chain reaction. Methods in Enzymology. 1987; 155:335-350. Nauta, Maarten J. and Weissing, Franz J. Constraints on allele size at microsatellite loci: implications for genetic differentiation. Genetics. 1996; 143:1021- 1032. Nei, Masatoshi. Analysis of gene diversity in subdivided populations. Proc. Natl. Acad. Sci. USA. 1973; 70(12):3321-3323. 131 ---. F-statistics and analysis of gene diversity in subdivided populations. Ann. Hum Genet., Lond. 1977; 41:225-233. —--. Genes in populations. in. Molecular evolutionary genetics. New York: Columbia University Press; 1987; pp. 149-175. Niemela, Pekka and Mattson, William. Invasion of North American forests by European phytophagous insects. Bioscience. 1996; 46(10):?41-753. Paetkau, D and Strobeck, C. The molecular basis and evolutionary history of a microsatellite null allele in bears. Molecular Ecology. 1995; 4:519-520. Panetsos, K. P.; Aravanopoulos, F. A., and Scaltsoyiannes, A. Genetic variation of Pinus brutia from islands of the northeastern Aegean sea. Silvae Genetica. 1998; 47(2-3):115-120. Pauley, Scott; Spurr, Stephen H., and Whitmore, Frank W. Seed source trials of eastern white pine. Forest Science. 1955; 1(3):244-256. PE Applied Biosystems. Genotyper analysis ver. 2.0 software. User's manual. Foster City, CA; 1994. ——-. Genescan analysis ver. 2.1 software. User's manual. Foster City, CA; 1996. Pemberton, J. M.; J. Slate; D.R. Bancroft, and J.A. Barrett. Nonamplifying alleles at microsatellite loci: a caution for parentage and population studies. Molecular Ecology. 1995; 4:249-252. 132 Rajora, O. P.; Rahman, M. H.; Buchert, G. P., and Dancik, B. P. Microsatellite DNA analysis of genetic effects of harvesting in old—growth estern white pine (Pinus strobus) in Ontario, Canada. Molecular Ecology. 2000; 9:339-348. Rajora, Om P.; DeVerno, Linda; Mosseler, Alex, and Innes, David J. Genetic diversity and population structure of disjunct Newfoundland and central Ontario populations of eastern white pine (Pinus strobus). Can. J. Bot. 1998; 76:500-508. Robertson, G. Philip and Tiedje, James M. Denitrification and nitrous oxide production in successional old- growth Michigan forests. Soil Sci. Soc. A. J. 1984; 48:383-389. Rose, Willard M. Biomass, net primary production and successional dynamics of a virgin white pine (Pinus strobus) stand in northern Michigan. East Lansing, Michigan: Michigan State University; 1984; p. 95. PhD dissertation. Routman, E. and Cheverud J. A rapid method of scoring simple sequence repeat polymorphisms with agarose gel electrophoresis. Mammalian Genome. 1994; 5:187-188. Rundel, Philip W. and Yoder, Barbara J. Ecophysiology of Pinus. in : Richardson, David M., in. Ecology and biogeography of Pinus. New York: Cambridge University Press; 1998; pp. 296—323. Salvador, L.; Alia, R. Agfindez, and Gil L. Genetic variation and migration pathways of maritime pine (Pinus pinaster Ait) in the Iberian peninsula. Theor Appl Genet. 2000; 100:89-95. Schlotterer, Christian; Amos, Bill, and Tautz, Diethard. Conservation of polymorphic simple sequence loci in cetacean species. Nature. 1991; 354:63-65. 133 Schlotterer, Christian and Tautz, Diehard. Slippage synthesis of simple sequence DNA. 1992; 20, (2): 211- 215. Schnabel, Andrew and Hamrick, J. L. Organization of genetic diversity within and among populations of Gleditsia triacanthos (Leguminosae). Amer. J. Bot. 1990; 77(8):1060-1069. Schnabel, Andrew; Laushman, Roger H., and Hamrick, J. L. Comparative genetic structure of two co-occuring tree species, Maclura pomifera (Moraceae) and Gleditsia triacanthos (Leguminosae). Heredity. 1991; 67:357-364. Scribner, Kim T.; Gust, Judy R., and Fields, Raymond L. Isolation and characterization of novel salmon microsatellite loci: cross—species amplification and population genetic applications. Can. J. Fish. Aquat. Sci. 1996; 53:833-841. Slatkin, Montgomery. Gene flow in natural poulations. Ann. Rev. Ecol. Syst. 1985; 16:393-430. ---. A measure of population subdivision based on microsatellite allele frequencies. Genetics. 1995; 139:457-462. Smith, R. H.; Peloquin, R. L., and Passof, P. C. Local and regional variation in the monoterpenes of ponderosa pine wood oleoresin. USDA Forest Service Research Paper PSW-56. 1969. Sokal, Robert R. Ecological parameters inferred from spatial correlograms. in: Patil, Ganapati P. and Rosenzweig, Michael L., editors. Contemporary quantitative ecology and related ecometrics. Fairland Maryland: International Co-operative Publishing House; 1979; pp. 167-196. 134 Sokal, Robert R.; Jacquez, Geoffrey M., and Wooten, Michael C. Spatial autocorrelation analysis of migration and selection. Genetics. 1989; 121:845-855. Sokal, Robert R. and Oden, Neal L. Spatial autocorrelation in biology. 1. Methodology. Biological Journal of the Linnean Society. 1978a; 10:199-228. ---. Spatial autocorrelation in biology. 2. Some biological implications and four applications of evolutionary and ecological interest. Biological Journal of the Linnean Society. 1978b; 10:229-249. Sommers, Lawrence M.; Darden, Joe T.; Harman, Jay R., and Sommers, Laurie K. Michigan: a Geography. Boulder, CO: Westview Press, Inc.; 1984. Steel, Robert G. D. and Torrie, James H. Principles and procedures of statistics. New York: McGraw-Hill Book Company; 1980. Stine, M. and Keathley, D. E. Paternal inheritance of plastids in Englemann spruce x blue spruce hybrids. The Journal of Heredity. 1990; 81(6):443-446. Strahler, Arthur. Introduction to Physical Geography. New York: John Wiley and Sons, Inc.; 1965. Streiff, Réjane; Labbe, Thierry; Bacilieri, Roberto; Steinkellner, Herta; Glossl, Joseph, and Kremer, Antoine. Within-population genetic structure in Quercus robur L. and Quercus petraea (Matt.) Liebl. assessed with isozymes and microsatellites. Molecular Ecology. 1998; 7:317-328. Takeuchi, Y.; Kubiske, M. E.; Isebrands, J. G.; Pregtizer, K. S.; Hendrey, G., and Karnosky, D. F. Photosythesis, light and nitrogen relationships in a young deciduous canopy under open-air'CKb enrichment. Plant, Cell and Environment. 2001; 24:1257-1268. 135 Taylor, John S.; Durkin, Jonathan M. H., and Breden, Felix. The death of a microsatellite: a phylogenetic perspective on microsatellite interruptions. Mol. Biol. Evol. 1999; 16(4):567-572. Valdes, Ana Maria; Slatkin, Montgomery, and Freimer, Nelson B. Allele frequencies at microsatellite loci: the stepwise mutation model revisited. Genetics. 1993; 133:737-749. Van Treuren, R. Van. Estimating null allele frequencies at a microsatellite locus in the oystercatcher (Haematopus ostralegus). Molecular Ecology. 1998; 7:1413-1417. L Vendramin, G. G.; Lelli, L.; Rossi, P., and Morgante, M. A set of primers for the amplification of 20 chloroplast microsatellites in Pinaceae. Molecular Ecology. 1996; 5:595-598. Voelker, Robert A.; Schaffer, Henry E., and Mukai, Terumi. Spontaneous allozyme mutatations in Drosophila Melanogaster: rate of occurrence and nature of the mutants. Genetics. 1980; 94:961-968. Vos, P.; Hogers, R.; Bleeker, M.; Reijans, M.; Theo Van de Lee, M.; Hornes, M.; Frijters, A.; Pot, J.; Peleman, J.; Kuiper, M., and Zabeau, M. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research. 1995; 23 :4407-4414. Wackerman, A. E. Growth of the "Grayling Pine". Journal of Forestry. 1924; 22:796-797. Wagner, David B. Nuclear, chloroplast, and mitrochondrial DNA polymorphisms as biochemical markers in population genetic analyses of forest trees. New Forests. 1992; 6:373-390. 136 Wang, Chi-Wu; Perry, Thomas O., and Johnson, Albert G. Pollen dispersion of slash pine (Pinus elliottii Engelm.) with special reference to seed orchard management. Silvae Genetica. 1960; 9:78-86. Wartenberg, Daniel. SAAP spatial autocorrelation software. Version 4.3. 1989. Weber, James L. Informativeness of human (dC-dA)n 0 (dG—dT)n polymorphisms. Genomics. 1990; 7:524-530. Weber, James L. and May, Paula E. Abundant class of human DNA polymorphisms which can be typed using the polymerase chain reaction. Am. J. Human. Genet. 1989; 44:388-396. Weber, James L. and Wong, Carmen. Mutation of human short tandem repeats. Human Molecular Genetics. 1993; 2(8):1123-1128. Wheeler, C. F. A sketch of the original distribution of white pine in the lower peninsula of michigan. Mich. Agr. Exp. Sta. Bull. 162. 1898:5-6. Whitney, Gordon G. Relation of Michigan's presettlement pine forests to substrate and disturbance history. Ecology. 1986; 67(6):1548-1559. Williams, Clarie G. and Savolainen, Outi. Inbreeding depression in conifers: implications for breeding strategy. Forest Science. 1996; 42(1):102-116. Williams, J. G. K.; Kubelik, A. R.; Livak, K. J. ; Rafalski, J. A., and Tingey, S. V. DNA polymorphisms amplified by arbitraty primers are useful as genetic markers. Nucleic Acids Research. 1990; 18(22):6531- 6535. 137 Workman, P. L. and Niswander, J. D. Population studies on southwestern indian tribes. II. Local genetic differentiation in the Papago. Am. J. Human. Genet. 1970; 22:24-49. Wright, Jonathan W. Pollen dispersion of some forest trees. USDA, Northeastern Forest Experiment Station Paper 46. 1952. ---. Genetics of eastern white pine. USDA For. Ser. Res. Pap. WO-9. 1970. . y____T Wright, Sewall. Systems of mating. I. The biometric relations between parent and offspring. Genetics. 1921; 6:111-178. 7 I -——. Coefficients of inbreeding and relationship. The American Naturalist. 1922; 56:330-338. ---. Isolation by distance. Genetics. 1943; 28:114-138. --—. Isolation by distance under diverse systems of mating. Genetics. 1946; 31:39-59. ---. The genetical structure of populations. Annals of Eugenics. 1951; 15:323-354. —--. The interpretation of population structure by F- statistics with special regard to systems of mating. Evolution. 1965; 19:395-420. ---. Evolution and the genetics of populations. II. The theory of gene frequencies. Chicago: The University of Chicago Press; 1969. Wu, H. X.; Matheson, A. C., and Spencer, D. Inbreeding in Pinus radiata. 1. The effect of inbreeding on growth, survival and variance. Theor Appl Genet. 1998; 97:1256-1268. 138 Wu, Junyuan; Krutovskii, Konstantin V., and Strauss, Steven H. Abundant mitrochondrial genome diversity, population differentiation and convergent evolution in pines. Genetics. 1998; 150(1605-1614):1605-1614. Yazdani, R.; Muona, O.; Rudin D., and Szmidt A.E. Genetic structure of a Pinus sylvestris L. seed-tree stand and naturally regenerated understory. Forest Science. 1985; 31(2) :430-436. Yeh, Francis C.; Yang, Rong-cai, and Boyle, Timothy. Popgene. Version 1.31. 1999. Zavarin, Eugene; Critchfield, William B., and Snajberk, Karel. Turpentine composition of Pinus contorta x Pinus banksiana hybrids and hybrid derivatives. Canadian Journal of Botany. 1969:1443-1453. 139