UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The impacts of forest harvest on the persistence and colonisation potential of pacific giant salamanders… Curtis, Janelle Marie 2000

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

Item Metadata

Download

Media
831-ubc_2000-0044.pdf [ 6.55MB ]
Metadata
JSON: 831-1.0074848.json
JSON-LD: 831-1.0074848-ld.json
RDF/XML (Pretty): 831-1.0074848-rdf.xml
RDF/JSON: 831-1.0074848-rdf.json
Turtle: 831-1.0074848-turtle.txt
N-Triples: 831-1.0074848-rdf-ntriples.txt
Original Record: 831-1.0074848-source.json
Full Text
831-1.0074848-fulltext.txt
Citation
831-1.0074848.ris

Full Text

T H E IMPACTS O F F O R E S T H A R V E S T O N T H E P E R S I S T E N C E A N D COLONISATION POTENTIAL O F PACIFIC G I A N T S A L A M A N D E R S (DICAMPTODON  TENEBROSUS)  IN BRITISH COLUMBIA.  BY  Janelle Marie Curtis Hons BSc, University of Toronto, 1997 A T H E S I S S U B M I T T E D IN PARTIAL FULFILMENT O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F MASTER OF SCIENCE  in T H E F A C U L T Y O F G R A D U A T E STUDIES (DEPARTMENT O F ZOOLOGY)  We accept this thesis as conforming to the required standard  University of British Columbia March 2000 ©Janelle Curtis, 2000  In  presenting  degree  at  this  the  thesis  in  partial  fulfilment  University  of  British  Columbia, I agree that the  freely available for copying  of  department publication  this or of  reference and study.  thesis by  this  for  his thesis  scholarly  or  her  for  Department  of  The University of British C o l u m b i a Vancouver, Canada  Date  DE-6  (2/88)  tZ%  H-OACJ^  ZpQO  I further  purposes  the  requirements  agree that  may  representatives.  financial  permission.  of  It  gain shall not  be is  for  an  Library shall make  permission for  granted  advanced  by  understood  the that  be allowed without  it  extensive  head of  my  copying  or  my  written  ABSTRACT  The Pacific Giant Salamander {Dicamptodon  tenebrosus)  is considered  vulnerable to local extirpation from British Columbia by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) and is red-listed by the B C Ministry of the Environment, Lands and Parks. The impacts of forest practices potentially threaten the long-term persistence of Pacific Giant Salamanders in Canada. I used microsatellite and A F L P markers to indirectly assess the impacts of forest harvesting on the population structure and colonisation potential of Pacific Giant Salamanders. Levels of genetic variation and population differentiation were compared among eight sub-populations in three coastal forest types (old-growth, secondgrowth and clear-cut) in the Chilliwack River Valley, British Columbia, and other populations across D. tenebrosus'  biogeographic range.  Patterns of genetic  variation and heterozygosity revealed that populations at the northern extent of D. tenebrosus' range have lower allelic richness and heterozygosity than more central and southern populations. Comparisons of genetic variation among forest types in B C revealed that recently clear-cut sites have less genetic variation than secondgrowth and old-growth sites, suggesting that clear-cutting may cause genetic bottlenecks. The level of genetic variation (allelic richness and percent polymorphic loci) and heterozygosity were significantly correlated with stand age. There was no relationship between geographic distance and genetic differentiation within the Chilliwack Valley.  Analyses of molecular variance (AMOVA) and estimates of  population structure F t and <3> confirmed that there was slight to moderate s  sit  differentiation among sub-populations of D. tenebrosus in B C . The colonisation potential of Pacific Giant Salamanders appears to be sufficient to re-establish locally extirpated sub-populations or to recover lost genetic variation from surrounding streams, particularly among clustered streams within drainages. However, long-term studies are required to assess whether the recovery of sub-populations is occurring faster than they are being disturbed, and whether Pacific Giant Salamanders are numerically stable in B C , or whether they are in decline.  ii  TABLE OF CONTENTS  Abstract  ii  Table of Contents  iii  List of Tables  vi  List of Figures  viii  Acknowledgements  xi  C H A P T E R 1: G E N E R A L INTRODUCTION  1  2  C H A P T E R 2:  ISOLATION A N D CHARACTERISATION O F MICROSATELLITE LOCI IN T H E  PACIFIC G I A N T S A L A M A N D E R (DICAMPTODON  TENEBROSUS).  2.1  Introduction  7  2.2 2.2.1 2.2.2  Background Microsatellite Mutation and Evolution Applications of Microsatellite Markers  9 9 10  2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5  Methods Microsatellite Cloning and Isolation Microsatellite Characterisation Geographic Variation Cross-Species Amplifications Phylogenetic Inference Using Microsatellite Data  11 11 13 14 15 15  2.4 2.4.1 2.4.2 2.4.3 2.4.4 2.4.5  Results Microsatellite Cloning and Isolation Microsatellite Characterisation Geographic Patterns of Variation Cross-species Amplifications Phylogenetic Inference Using Microsatellites  16 16 16 17 18 18  2.5 2.5.1 2.5.2 2.5.3 2.5.4  Discussion Isolation and Characterisation of D. tenebrosus Microsatellites Geographic patterns of variability Cross-Species Amplifications Phylogenetic Inference Using Microsatellites  19 19 20 22 23  3  C H A P T E R 3 : E V I D E N C E O F R E C E N T B O T T L E N E C K S IN S U B - P O P U L A T I O N S O F T H E PACIFIC  G I A N T S A L A M A N D E R (DICAMPTODON  TENEBROSUS)  IN BRITISH C O L U M B I A .  43  3.1  Introduction  43  3.2 3.2.1 3.2.2 3.2.3 3.2.4  Methods Study Sites and Sample Collection Microsatellite and AFLP Allele Frequencies and Heterozygosity Comparisons of Genetic Variation Among Forest Types Detection of Recent Bottlenecks  46 46 47 49 49  3.3 3.3.1 3.3.2 3.3.3  Results Microsatellite and AFLP Allele Frequencies and Heterozygosity Comparisons of Genetic Variation Among Forest Types Shifts in Allele Frequency Distributions  50 50 51 53  3.4 3.4.1 3.4.2 3.4.3  Discussion Comparisons of Genetic Variation Among Forest Types Shifts in the Distribution of Allele Frequencies Implications for Management of Pacific Giant Salamanders in BC  53 55 56 57  4  CHAPTER 4  T H E POPULATION G E N E T I C S T R U C T U R E A N D COLONISATION POTENTIAL O F  PACIFIC G I A N T S A L A M A N D E R S IN T H E CHILLIWACK RIVER V A L L E Y , BRITISH C O L U M B I A .  4.1  Introduction  72  72  4.2 Methods 4.2.1 Study Sites and Sample Collection 4.2.2 Analysis of Isolation-by-distance 4.2.2.1 Within the Chilliwack River Drainage 4.2.2.2 Across the Northern Range of D. tenebrosus 4.2.3 Analyses of Molecular Variance 4.2.4 Comparisons of genetic distance among forest types 4.2.5 Population structure and Gene Flow  75 75 76 76 77 77 78 78  4.3 Results 4.3.1 Analysis of Isolation-by-distance 4.3.1.1 Within the Chilliwack River Drainage 4.3.1.2 Across the Northern Range of D. tenebrosus 4.3.2 Analyses of molecular variance 4.3.3 Comparisons of Genetic Distance among Forest types 4.3.4 Population structure and Gene Flow  80 80 80 80 81 81 82  4.4 Discussion 83 4.4.1 At which spatial scales(s) are sub-populations genetically structured within the Chilliwack River Valley: among streams, among drainages, or not significantly structured? 83 4.4.1.1 Analyses of Isolation-by-distance 83 4.4.1.2 Across the Northern Range of D. tenebrosus 84 4.4.1.3 Analyses of Molecular Variance 84 4.4.1.4 Estimates of Population Structure 84 4.4.2 Is there evidence from pair-wise genetic distances that clear-cutting causes population bottlenecks?86  iv  4.4.3 Is dispersal potential among critical nursery streams sufficient for the re-establishment of disturbed or extirpated sub-populations? 4.4.3.1 Interpreting estimates of gene flow from F 4.4.3.1.1 Concordance Among Marker Classes and Loci 4.4.3.1.2 Inferences of Dispersal from Ecological and Demographic Data 4.4.3.1.3 Direct Estimates of Gene Flow from Mark-Recapture Studies 4.4.3.2 Colonisation Potential of Pacific Giant Salamanders in BC st  5  C H A P T E R 5: G E N E R A L C O N C L U S I O N S  6  LITERATURE CITED  86 86 87 88 89 89  99  105  V  LIST O F T A B L E S  Table 2-1 Sample size (n), latitude, longitude, and location of 14 D. tenebrosus  populations screened with six polymorphic microsatellite loci (see map in Fig. 2.2) 26  Table 2-2 Sequences of 25 cloned D. tenebrosus microsatellite-bearing inserts.  Microsatellite repeat arrays are in italics and complementary sequences of designed PCR primers are underlined. N denotes an ambiguity in sequence.. 27  Table 2-3 Primer sequences, repeat array of cloned allele, number of detected  alleles (A), sizes of alleles in base pairs, and annealing temperature (Ta) at fourteen microsatellite loci isolated for D. tenebrosus. Monomorphic loci (A=1) were screened with 20-25 individuals from across D. tenebrosus' range. Polymorphic loci (A>1) were screened with 92 D. tenebrosus individuals. Genbank accession numbers are given for polymorphic loci 29  Table 2-4 Expected H e and observed H 0 (direct count) heterozygosities at 14  sample sites. Mean H and H are calculated across all 12 sites and within sites across all loci. Sample sizes are given in parentheses. The three single samples from sites BROO, THCR and LACR were pooled for comparisons and are renamed as OR (see Table 2.1) 31 e  0  Table 2-5 ANOVA table comparing a linear model incorporating both latitude and  sample size as predictors of microsatellite richness against the benchmark model which uses sample size as the only predictor. The squared multiple correlation coefficient (R ) and the partial correlation coefficient for the contribution of latitude to the prediction of allelic richness, given sample size (r i.n) are given (see text for description) 32 2  a  Table 2-6 Sizes of alleles successfully amplified in 18 closely and distantly related  species of salamanders from 5 families. N refers to the number of individuals assayed per species 33  Table 3-1 Sample sites of eight Pacific Giant Salamander nursery streams in BC. N  indicates the total number of genetic samples collected. The estimated larval density (LD) (1998 data, Ferguson & Johnston 1999) sampled reach length, drainage creek, UTM (Universal Transverse Mercator) location, elevation, stand age (in 1998), forest class and Ministry of Forests (MOF, Chilliwack Forest District, Rosedale, BC) map and opening numbers are included. Sites that are located within the same drainage are located in different headwater streams. Each opening number corresponds to an area with a unique harvest and treatment history and must be used in conjunction with map numbers for identification. No opening numbers had been assigned to NesF4 and NesF5 in 1999. No opening number was available for VedB. The exact stand age of  vi  VedB is unknown but was estimated to be approximately 30 years (OG = oldgrowth, SG = second-growth, CC = clear-cut) 59 Table 3-2 The product size of 41 AFLP markers scored using three pairs of AFLP primers (Gibco BRL) 60 Table 3-3 Microsatellite allele frequencies at three polymorphic microsatellite loci Dte4, Dte8, and Dte11, in eight sub-populations of D. tenebrosus in BC. Observed heterozygosity, H , was calculated within sub-populations over all three loci 61 0  Table 3-4 Estimated frequencies of the dominant AFLP alleles (coding for the "presence" of a band) within eight sub-populations of D. tenebrosus in BC. Frequencies are also estimated across all sub-populations. The AFLP marker indicates the Mse\ primer used to amplify the fragment (all Mse\ primers were paired with the same Eco RI primer, see text), as well as the fragment size 62 Table 3-5 The percent polymorphic AFLP markers (PPM) within eight subpopulations of D. tenebrosus in BC. Heterozygosity was estimated from the proportion of recessive alleles (see text)  63  Table 3-6 Spearman's Rank Correlation Coefficient and p-values for correlations between levels of genetic variation, microsatellite allelic richness and AFLP PPM, and six spatial and sampling factors (n=8) 64 Table 4-1 Hierarchical analysis of molecular variance (AMOVA) tables for microsatellite (A) and AFLP (B) data. The hierarchical levels (sources of variation), degrees of freedom (DF), sums of squared genetic distances (SS), variance components and the percentage of the total amount of genetic variation explained by each source (%) are given. Estimates of F- and O-statistics and respective p-values are also included. Italicised numbers represent AMOVAs excluding recently clear-cut sites 92 Table 4-2 Global Fst (or O t) values (calculated across all eight BC sites, or among forested sites only) and derived estimates of the effective number of migrants, N m (* indicates significant estimates of population structure) 93 s  e  vii  LIST O F FIGURES  Figure 2-1 Autoradiographed clones. The large circles are hybond filters used to  bind single-stranded DNA that was cloned in approximately 5000 bacterial colonies (smaller dots on the filters). The darker clones (11, 13, 14, and 18) represent colonies bearing microsatellite sequences. In these cases, radiolabelled oligonucleotides hybridised to the microsatellite sequences in the bound DNA and exposed radiation-sensitive film 34  Figure 2-2 Range of D. tenebrosus from southwestern British Columbia to  northwestern California (light shading). Locations are numbered according to sample sites where 1=Nes20, 2=MICR, LDCR, and CUCR, 3=MACR, 4=MIRI, 5=COCR, 6=PRCR, 7=WIRI, 8=ONGO, 9=LACR and THCR, 10=BROO and 11 =PTAR (range is based on Daugherty et al. 1983 and Good 1989) 35  Figure 2-3 Autoradiograph of a polyacrylamide denaturing gel used to size the alleles of three D. tenebrosus microsatellite loci: Dte11, Dte14 and Dte16. The  four lanes labelled GATC represent the standard M13 sequence used to size the alleles. Dte11 and Dte14 are polymorphic with 3 and 8 alleles shown, respectively. Dte16 is a monomorphic locus with a single 149 bp allele. Numbers indicate allele sizes, and short arrows point to three heterozygotes. Stutter bands - several lighter bands beneath the true allele - are characteristic signatures of microsatellites and are particularly pronounced in Dte11 36  Figure 2-4 Latitudinal gradient in the number of microsatellite alleles detected in 14 populations of D. tenebrosus. Sub-samples of 3 individuals were randomly selected from each population ten times. The open circles represent single samples from BROO, LACR and THCR. Error bars represent the standard error of the mean 37 Figure 2-5 Geographic pattern of observed heterozygosity (H0) in D. tenebrosus.  H o was averaged across individuals within populations. Closed circles represent samples with 3-14 observations. Open circles represent single observations from BROO, LACR and THCR. Error bars indicate the standard error of the mean of the 10 sub-samples 38  Figure 2-6 Variation in D. tenebrosus allozyme richness over latitude. Allelic  richness (number of alleles detected) was averaged over 31-34 allozyme loci (data from Daugherty et al. 1983 and Good 1989) 39  Figure 2-7 UPGMA cluster of Dicamptodon populations. Nei (1978) genetic  distance, D, is on the horizontal axis. Percent bootstrap replicates supporting nodes are indicated. No tied trees were generated. Please refer to Table 2.1 for location of D. tenebrosus populations 40  viii  Figure 2-8 UPGMA cluster of 19 salamander species constructed with 3  microsatellite loci. Nei (1978) genetic distance (D) is on the horizontal axis. Percent bootstrap replicates supporting two of the nodes are indicated. Numbers 1-5 correspond to the families shown to the left of the tree. Thirteen tied trees were produced 41  Figure 2-9 Phylogenetic tree of nine extant families of Caudata (salamanders)  based on 27 non-paedomorphic (adult) characters. Species in families with bold type were used for the cross-amplification assays using microsatellites developed in D. tenebrosus. Modified from Duellman & Trueb (1994) 42  Figure 3-1 The geographic distribution of Dicamptodon tenebrosus ranges along  the Pacific Coast from northern California to the extreme southwestern corner of British Columbia (light shading). The Chilliwack River Valley is adjacent to the Canada - United States border (indicated by a white arrow). (Range from Daugherty et al. 1983 and Good 1989) 65  Figure 3-2 The location of eight D. tenebrosus sample sites in the Chilliwack River  Valley. The site names are indicated in the legend. Circles mark the location of old-growth sites, triangles indicate second-growth sites, and squares represent recently clear-cut sites. Site 1 (VedB) is located on the west slope of Cultus Lake at the West end of the Chilliwack River Valley. Chilliwack Lake is located at the extreme East end of the Chilliwack Valley. The eight sites are located in five drainages: Ascaphus Creek (1), Tamihi Creek (2,3), Nesakwatch Creek (4,5), Foley Creek (6,7), and Centre Creek (8). Sites within the same drainage were sampled from different tributary streams 66  Figure 3-3 The size distribution and location of 258 D. tenebrosus individuals  captured in FolQ during the summer of 1998. Closed circles represent individuals that were marked during a mark-recapture study conducted by WE Neill & JS Richardson (UBC, unpublished data) and tail-clipped for this genetic study (n=65). Open circles (n=188) are individuals that were marked only 67  Figure 3-4 Autoradiograph of a polyacrylamide denaturing gel used to size  amplified fragment length polymorphisms (AFLPs). The four lanes labelled GATC represent the standard M13 sequence used to size the bands. Lanes 1-5 represent individuals from TamC (clear-cut), lanes 6,8 and 9 are from NesF4 (old-growth), lanes 7, 12 and 13 are from NesF5 (old-growth), and 10 and 11 are from Cen23 (second-growth). Arrows indicate markers that are monomorphic throughout the eight sub-populations of D. tenebrosus in BC 68  Figure 3-5 Comparisons of genetic variation among forest types: (top) mean  microsatellite allelic richness calculated over three polymorphic microsatellite loci and (bottom) mean percent polymorphic markers (PPM) calculated over 39 AFLP markers. Error bars represent the standard error of the mean (Old Growth N=2; Second Growth N=3; Clear-cut N=3) 69  ix  Figure 3-6 Mean microsatellite allelic richness (A) and percent polymorphic AFLP  markers (AFLP PPM) (B) increase with the logarithm of stand age (in years). Squares denote recently clear-cut sites, triangles represent second-growth sites, and circles represent old-growth sites. The diamond in (A) represents two oldgrowth sites at the same co-ordinates. The solid lines represent linear regression lines. The equations of the lines are (A) y=0.71+0.85x and (B) y=19.23+48.63 70  Figure 3-7 The distribution of microsatellite allele frequencies in two subpopulations of D. tenebrosus: (A) FolQ and (B) NesF4. FolQ has a qualitative shift in its microsatellite allele frequency distribution, while NesF4 conforms more closely to a typical L-shaped distribution. N refers to the number of alleles with a given frequency (frequencies were rounded to the nearest multiple of 0.1) 71 Figure 4-1 Pair-wise genetic distances among sub-populations in the Chilliwack  Valley measured as F t (A) and Nei's D (B), is plotted against geographic distance (the minimum distance within the 1200 m elevation limit between two sites (ELD)). Closed circles represent pair-wise genetic distance estimates between two forested sites (old-growth and / or second-growth). Open circles represent pairs including one or two recently clear-cut sites) 94 s  Figure 4-2 Isolation by distance across the northern extent of D. tenebrosus' range  (457 - 49 °N). F was calculated according to Weir & Cockerham (1984) using three polymorphic microsatellite loci. Closed circles represent forested sitepairs, while open circles represent pairs with at least one clear-cut site from the Chilliwack Valley (all sites in Washington and Oregon were forested: Jacqueline Brinkman, Redpath Museum, Montreal, pers. comm.). The regression line represents the sample of forested sites only (the equation of the line is shown, r =0.117, p<0.0001) 95 st  2  Figure 4-3 The proportion of D. tenebrosus genetic variation accounted for by three  hierarchical levels: (A) within-streams, (B) among-stream-within-drainages and (C) among-drainages 96  Figure 4-4 Pair-wise genetic distances estimated with (A) three microsatellite loci  (F ; Weir & Cockerham 1984) and (B) 39 AFLP markers (Nei's D) according to forest type. Forested pairs represent comparisons among forested sites. Clearcut pairs represent pairs with 1-2 clear-cuts 97 st  Figure 4-5 Pair-wise F s t (A) and Nei's D (B) between the mean of the old-growth  sites and previously harvested sites plotted against the age of the harvested sites  98  ACKNOWLEDGEMENTS  I owe many thanks to Dr. William Neill for inviting me to participate in his lab at UBC, and for encouraging me to pursue my interests both within and outside the huts. Dr. Neill, thank you for all the academic guidance, support, insight, patience, laughs, encouragement, comforting chats, and your interest in my academic and personal growth. Thank you to Dr. Eric Taylor who invited me to carry out my genetic work in his lab. There I found encouragement, support, constructive feedback, intellectual challenges, and an infinite string of pearls of wisdom (as well as much deserved teasing). Thank you to Dr. John Richardson for all of his encouraging words and constructive comments on my work. Thank you also to Dr. Michael Whitlock for his thoughtful insights on the theory of population genetics and its applications to questions about population persistence and movement patterns. Drs. Neill, Taylor, Richardson, and Whitlock were my research committee members. I very much appreciate their help in shaping my thoughts and guiding me through this learning process. Thank you to the "Neills" and the "Tay-Macs" for showing me the ropes: Alon Altman, Lance Barrett-Leonard, Dawn Cooper, Heather Ferguson, Julie Hofer Drew Hoysak, Ernest Keeley, Jason Ladell, Steve Latham, Megan McCusker, Jen McLean, Dr. JD McPhail, Dave O'Brien, Dorothee Schreiber, Mike Stamford, Patrick Tamkee, and Zoe Redenbach. In particular, Dave, Mike, Dorothee, Dawn and Steve shared their insights on the "greater picture" and helped me take life (and myself) a little less seriously. Thanks! Many thanks to Merran Hague who, during the summer of 1998, challenged me each morning to catch as many salamanders as she did! Merran was a pleasure to work with, and without her help, I could not have accomplished my goals of that summer. Thank you also to Dora Repard and Karl Mallory for help with the collection of Pacific Giant Salamander tissue samples in BC. Thanks to Kyle Young for helping me measure, weigh, sample and mark hundreds of salamanders at DFO in Chilliwack. Thanks also to my sister Joelle and her friend Steve for spending three weeks helping me in the field and having so much fun with me! Thank you to Leo Frid and Brent Matsuda for help collecting terrestrial Pacific Giant Salamanders.  Thanks also to Glenys Webster for her help with DNA extractions, restrictions and ligations in the lab. Thank you to Tanya Trepanier, Ross McCullough, and Dr. Bob Murphy at the University of Toronto and Royal Ontario Museum for sharing D. tenebrosus tissue samples from California, as well as samples from 15 other salamander species. Thank you to Dr. Tom Titus from the University of Oregon (in Eugene) for samples of D. tenebrosus from Oregon. Thank you also to Jacqueline Brinkman and Dr. David Green of the Redpath Museum and McGill University for sending D. tenebrosus samples from Washington and Oregon. I thank the University of Toronto and the University of British Columbia for scholarship support during my MSc, The Department of Zoology (UBC) for the opportunity to TA, and WWF, HCTF and FRBC for personal and project-related financial support. Thank you also to Jens Hauser and Alistair Blachford for their help through the ZCU. Thank you to Dr. Bill Milsom for listening. And of course, many thanks to Mom & Dad for all their love and support!  xii  1  C H A P T E R 1: GENERAL INTRODUCTION  The Pacific Giant Salamander (Dicamptodon tenebrosus) is the largest semiaquatic salamander in North America, with adults reported as long as 35 cm (Petranka 1998). Its distribution ranges from Mendocino County, California to part of the Lower Fraser Valley in British Columbia (Fig. 2.2) (Daugherty et al. 1983; Nussbaum et al. 1983; Good 1989; Petranka 1998). Dicamptodon tenebrosus is generally considered an obligate old-growth species (Nussbaum et al 1983; Leonard et al. 1993) and is found most commonly in or along small, fast-flowing headwater streams (Richardson & Neill 1998). Pacific Giant Salamanders depend on these streams for breeding and larval development.  Larvae spend 3-4 years within  streams, after which they transform into terrestrial juveniles, or mature as streamdwelling neotenes (Nussbaum et al. 1983; Richardson & Neill 1998). Although little is known about the reproductive behaviour of Pacific Giant Salamanders, breeding is thought to be asynchronous (non-seasonal) in BC (Haycock 1991), with females laying clutches of 100-200 eggs in stream beds or subsurface springs (Nussbaum 1969). Where Pacific Giant Salamanders are present, they are considered the primary vertebrate predator in fish-less streams (Murphy & Hall 1981), and feed on land snails, invertebrates, other amphibians (including conspecifics), reptiles, birds and small mammals (Nussbaum et al. 1983; Blaustein et al. 1995). Natural mortality is thought to be primarily due to cannibalism, predation and desiccation (Nussbaum & Clothier 1973). In 1989, the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) recognised the Pacific Giant Salamander as being "vulnerable" to local extirpation from Canada. The BC Ministry of Environment, Lands and Parks also assigned Pacific Giant Salamanders to its red list, a list of species that are considered endangered or threatened in British Columbia (BC Ministry of the Environment, Lands and Parks 1993). The justification for listing Pacific Giant Salamanders as vulnerable or endangered in BC was based on its limited distribution in Canada, potentially poor dispersal ability, and rapid rates of habitat loss and alteration aggravated by widespread logging activities in the Chilliwack  l  River Valley (Farr 1985, 1989; Haycock 1991). Moreover, BC populations of Pacific Giant Salamanders are on the northern periphery of their range, where habitat may be marginal, and therefore may already be more susceptible to local extirpation than central populations. Recent studies have shown that peripheral populations in other taxa have a greater probability of extinction (e.g. Lomolino & Channel 1995; Nathan et al. 1996) due to lower population densities, lower survival rates (Randall 1982; Rogers & Randolph 1986) and lower fecundity (Caughley et al. 1986). Global amphibian declines have been linked to increased habitat loss and fragmentation (Blaustein et al. 1994; Blaustein & Wake 1995). Amphibians are intolerant of major environmental changes, such as deforestation and particularly to changes in moisture and temperature regimes (Duellman & Trueb 1994). Ash (1997) and Petranka et al. (1993) report decreases in salamander density, or local extirpation of salamanders in recently clear-cut sites. The absence of amphibian species in clear-cut sites is generally attributed to direct mortality during logging operations, increased physiological stress due to micro-climatic changes, disruption of stream flow, or increased sedimentation (Bury 1983; Connor et al. 1988; Corn & Bury 1989; Petranka et al. 1993; Dupuis et al. 1995; Vesely 1996). Blaustein et al. (1994) suggest that physiological constraints and low recolonisation rates make amphibian populations particularly susceptible to habitat disturbances because they are unlikely to recolonise locally extirpated sites. In general, population size (Pimm et al. 1988) and colonisation ability (Burbidge & Mackenzie 1989; Laurance 1991) are among the best predictors of the long-term persistence of species.  Moreover, the degree of connectivity among  populations is a key determinant of the potential for recolonisation after the occurrence of a local extinction (Lamberson et al., 1992). Therefore when species are at risk, conservation efforts are often directed at maintaining or augmenting population sizes, and understanding and preserving movement patterns among subpopulations. Several studies conducted throughout the range of Pacific Giant Salamanders have reported lower larval densities in clear-cuts relative to forested stands (Bury 1983; Bury & Corn 1988; Connor et al. 1988; Corn & Bury 1989; Cole et al. 1997)  2  while other studies report no difference in larval density among forest types (Hawkins et al. 1983; Kelsey 1995; Vesely 1996; Neill & Richardson 1997; Neill 1998; Richardson & Neill 1998) or increases in density within deforested sites (Murphy et al. 1981; Murphy & Hall 1981). At the most northern extent of D. tenebrosus' range, forest harvesting has been ongoing since the late 1800s (McCombs & Chittenden 1990) particularly at low elevations. Agricultural practices have also considerably impacted the landscape over the last century (e.g. Drainage of Sumas Lake, channelisation of Vedder River) at lower elevations. Old growth forest is generally restricted to fragments at high elevation, >1000m, which also corresponds to D. tenebrosus elevation limit 1  (Richardson and Neill 1995). My thesis is part of an ongoing study to assess the impacts of current forest practices on the persistence and movement patterns of Pacific Giant Salamanders in BC.  This ongoing study, conducted by WE Neill (Dept of Zoology) and JS  Richardson (Faculty of Forestry) at the University of British Columbia, suggests that recently clear-cut sites have larval densities that are similar to or greater than second-growth and old-growth sites (Neill & Richardson 1997; Neill 1998). However, comparisons of larval density among forest types may be confounded by elevation, with density increasing as elevation decreases (Neill WE, pers. comm.). Prior to this ongoing study, little was understood about the movement patterns of Pacific Giant Salamanders. Neill and Richardson (1997) and Ferguson (1998) have recently shown that larvae and stream-dwelling neotenes have extremely high site fidelity and can be found from year to year beneath the same rocks or logs within a stream. As such, aquatic Pacific Giant Salamanders are probably poor dispersers.  Indeed, Ferguson (1998) estimated that the full  recolonisation of experimentally barren stream reaches by larvae immediately adjacent to the reach would take several decades. Conversely, Johnston (1998) and Johnston and Frid (unpublished manuscript) used radio-telemetry to show that terrestrial adults are capable of moving several hundred metres within a few months. Although Ferguson (1998) and Johnston (1998) concluded from their studies that Pacific Giant Salamanders were poor dispersers, their assessments probably  3  underestimated dispersal potential because neither study considered the movement patterns of recently-transformed juveniles.  Previous studies have shown that  amphibian juveniles often represent the dispersing stage in amphibians (Berven & Grudzien 1990; Duellman & Trueb, 1994). Radio-telemetry and mark-recapture techniques for examining the movement patterns of juveniles were not feasible, therefore I chose to gain a more historical and landscape level perspective on the population structure of Pacific Giant Salamanders using molecular genetic markers. My thesis focuses on the patterns of genetic variation within and among eight sub-populations of D. tenebrosus that have been intensively studied since 1996 by WE Neill and JS Richardson. My objectives were to indirectly assess the impacts of clear-cuts on population structure, as well as assess the potential of D. tenebrosus to re-establish locally extirpated populations. My thesis is partitioned into three research chapters (Chapters 2, 3, & 4) and a section of general conclusions. The objectives of each research chapter are summarised below. Chapter 2  Isolation and characterisation of microsatellite loci in the Pacific Giant Salamander {Dicamptodon  tenebrosus).  The need to develop and characterise the level of polymorphism of cloned microsatellites in Pacific Giant Salamanders gave me the opportunity to examine questions that are not directly related to the questions outlined in Chapters 3 and 4. The first objective of this chapter relates to hypotheses regarding the mutation rates and evolution of microsatellites. The second aim of this chapter is to compare levels of genetic variation throughout the range of Pacific Giant Salamanders. Extant levels of genetic variation reflect both current and historical processes. In order to address questions relating to the level of genetic variation among forest types in BC (Chapters 3 & 4), it is important to have a better understanding of the historical processes that may have influenced populations at the northern extent of this species' range. Several studies report lower genetic variation and heterozygosity  4  among northern populations relative to more central, or southern populations and interpret this genetic pattern as reflecting recent colonisations subsequent to glacial retreat -10-12000 years ago (e.g. Templeton et al. 1995; Murdoch & Hebert 1997). If genetic variation at the northern extent of D. tenebrosus' range is consistent with this pattern, it would suggest that populations may have recently recolonised the Southwest corner of BC. The third objective is to assess the potential usefulness of D. tenebrosus microsatellite markers in closely and distantly related species. Because microsatellite development is time and cost intensive, such an assessment could prove useful to other researchers. The last objective is to test whether two phylogenies (the first of salamander families, the second of species within Dicamptodon) constructed with microsatellite data are consistent with phylogenies constructed using allozymes or morphological characters. Chapter 3 Evidence of recent bottlenecks in sub-populations of the Pacific Giant Salamander (Dicamptodon tenebrosus) in British Columbia. Population bottlenecks can reduce the adaptive potential of populations and increase the risk of local extinction by reducing population size. Populations that have recently passed though bottlenecks are also expected to lose rare alleles (Luikart 1997). Casual observations of lower larval densities, or absence of larvae, immediately subsequent to clear-cutting suggest that current forest harvest practices may cause populations to pass through bottlenecks, or become locally extinct (WE Neill, pers. comm.). Moreover, studies of the impact of clear-cuts on other salamander species show that salamanders decrease in abundance or disappear from clear-cuts altogether (Petranka et al. 1993, Ash 1997). The objective of this chapter is to compare levels of genetic variation among forest types in order to test whether sub-populations in recently clear-cut sites have lower genetic variation than either second-growth or old-growth sites. I used three polymorphic microsatellite markers described in Chapter 2, as well as 39 AFLP (amplified fragment length polymorphism) markers for this comparison. In order to  assess whether there is a potential for recovery of locally disturbed sub-populations, I also test whether there is a relationship between genetic variation and stand age.  Chapter 4 The  population genetic structure and colonisation potential of Pacific Giant  Salamanders in the Chilliwack River Valley, British Columbia. The objective of this chapter is to indirectly assess the spatial level(s) at which Pacific Giant Salamander populations are genetically structured: withinstreams,  among-streams-within-drainages, or among-drainages.  I used 3  microsatellite and 39 AFLP markers to estimate F t and its analogue O . I also test s  s t  whether genetic distances (based on microsatellite and AFLP data) among forested sites are lower than among recently clear-cut sites. Lastly, I discuss the strengths and weaknesses of inferring colonisation potential from measures of population structure among sub-populations of Pacific Giant Salamanders and relate these to the management implications for this species in BC. Chapter 5 General Conclusions Here I summarise my principle findings and discuss the implications of this study for the management of Pacific Giant Salamanders in BC.  6  2  CHAPTER  2:  ISOLATION A N D C H A R A C T E R I S A T I O N  O F MICROSATELLITE  L O C I IN T H E P A C I F I C G L A N T S A L A M A N D E R (DICAMPTODON  2.1  TENEBROSUS).  INTRODUCTION  The intra- and inter-specific phylogenetic relationships among North American salamanders,  including Dicamptodon spp. have largely been examined by  morphological studies (see Duellman & Trueb 1994). Although modern molecular techniques have been used to resolve complex relationships among other salamander taxa (e.g. Hay et al.1995; Shaffer & McKnight 1996; Good & Wake 1997), few surveys of genetic variation have been carried out within the genus Dicamptodon. Dicamptodon is an ancient lineage known from the lower Pliocene in California (Nussbaum 1976). Until recently, Dicamptodon was considered to have a single extant species: Dicamptodon ensatus.  However, Nussbaum (1976)  recognised morphologically distinct populations in the Olympic Peninsula, southwestern Washington and north-western Oregon, which he assigned to D. copei. Two more recent studies (Daugherty et al. 1983; Good 1989) report the use of 31-34 allozyme loci in assigning species status to four groups in Dicamptodon: D. copei, D. ensatus (south of Mendocino Co. California), D. atterimus (allopatric groups in Idaho and Montana,) and D. tenebrosus (south-western British Columbia to north-western California). No previous studies report the use of genetic markers in assessing population-level processes in Dicamptodon. I considered using allozymes for this study; however, no allozyme variation was detected at 31-34 loci north of 47 °N (data from Daugherty et al. 1983 and Good 1989). Therefore I chose to develop microsatellite markers which are often hypervariable even in populations that have no allozyme variation (Queller et al. 1993). The advantages of using microsatellite markers in population genetic studies are numerous.  Microsatellites provide sensitive markers for resolving population  structure and have minute tissue requirements (Bruford et al. 1996). They are abundant throughout the genome and thus easily cloned and isolated.  Once  developed, microsatellite length polymorphisms generate discrete genotypes that  7  are easily distinguished by electrophoretic separation. Microsatellite alleles are amplified under stringent PCR conditions such that results are repeatable and reliable. The primary disadvantage of using microsatellite markers lies in the costly and time-consuming isolation, development and optimisation of suitable loci, particularly if no previous markers have been developed for the species or genus in question (see Bruford et al. 1996). No salamander microsatellite markers were published in the literature prior to this study, therefore I cloned and isolated microsatellite loci directly from D. tenebrosus DNA. Once isolated, microsatellite loci were characterised in order to  assess their suitability for studying the population-level questions addressed in Chapters 3 and 4. In doing so, I addressed the following questions: 1)  Is allelic variation within loci correlated with repeat size or the level of  interruption within the repeat array?  Previous studies have suggested that  longer, uninterrupted repeat arrays have higher mutation rates (and therefore more alleles) than shorter, imperfect repeats due to a greater probability of slippage (see Background, below). 2)  How does microsatellite variation in D. tenebrosus vary across its  geographic range? Levels of genetic variation in populations of D. tenebrosus  were compared across forest types in British Columbia in order to assess the impact of current forest practices on these populations (see Chapter 3). However, the extant levels of genetic variation may reflect both present and historical processes. Comparisons of genetic variation across the geographic range of D. tenebrosus may help determine whether low levels of neutral genetic variation are due to current forest management practices, or whether they reflect historical events such as a recent expansion of the northern periphery of D. tenebrosus range. 1  3)  Are the microsatellites developed for D. tenebrosus potentially useful in  studies of closely and distantly related salamanders? In order for microsatellites  to be useful markers for population genetic studies, they must be variable and  8  amplify non-ambiguously. However, previous studies suggest that microsatellites developed in a focal species may not be as long or variable in related species or may not amplify at all. 4)  Are phylogenies constructed with D. tenebrosus  microsatellites as  reliable as phylogenies constructed using allozyme loci or morphological characteristics?  Because microsatellite alleles are distinguished by length, it is not possible to know whether alleles are similar because they share a common ancestral allele, or because they have converged on the same length through mutation. Consequently, microsatellites may be inappropriate for examining the relationships among distantly related taxa.  2.2  2.2.1  BACKGROUND  MICROS A TELLITE MUTA TION AND EVOLUTION  Microsatellite loci, also referred to as simple sequence repeats (SSR) and variable number tandem repeats (VNTR), are short DNA sequences (typically 1-6 base pairs; Bruford et al. 1996) that are repeated in tandem.  Microsatellites are  ubiquitous throughout eukaryotic genomes (Tautz & Renz 1984; Stallings et al. 1991). Although these markers are typically distributed among DNA sequences that code for genes (euchromatin) (Stallings et al. 1991, Glenn 1995), they are generally considered to be non-coding, selectively neutral markers.  Different taxonomic  groups appear to differ in the size, number and variability of microsatellites (Crawford et al. 1998). For example, Lagercrantz et al. (1993) found that the most abundant type of dinucleotide repeats in plants and mammals were (AT) and (GT) , n  n  respectively, and microsatellites in plants were 5 times less abundant than in mammals. Some fish genera have microsatellite alleles that are twice as long as alleles described in mammals (Brooker et al. 1994). Hyper-variability of microsatellite loci is attributable to high mutation rates (Bruford et al. 1996). Weber and Wong (1993) estimate that microsatellite mutation  9  rates range from 10" to 10" per locus per generation, while other studies report 2  4  mutation rates of 10" to 10" (Shlotterer & Tautz 1992; Edwards et al. 1992) and 10" 4  5  3  to 10" (Glenn 1995). This variation in estimated mutation rates reported in the 4  literature probably reflects mutational differences among repeat types, as well as locus-specific characteristics. For example, dinucleotide repeat arrays tend to have faster mutation rates than tri- and tetranucleotide repeats due to a greater probability of slippage (see below) (Chakraborty et al. 1997; Kruglyak et al. 1998; Schug et al. 1998). Two mutation mechanisms are thought to account for most of the mutations at microsatellite loci: unequal crossing over between sister chromosomes during meiosis (Valdes et al. 1993) and slippage during DNA replication (Levinson & Gutman 1987; Schlotterer & Tautz 1992). During DNA synthesis, a transient bulge may form due to misalignment between the two DNA strands when the polymerase complex disassociates from the DNA. The repair mechanism eliminates the bulge by either lengthening or shortening one of the strands, resulting in either no mutation, or a change in repeat number (Schlotterer & Tautz 1992). Most mutations within repeat arrays tend to result in alleles that differ from parent alleles by one or two repeat units, either increasing or decreasing the repeat length of the allele (Weber & Wong 1993).  Differences in alleles are thus observed as length  polymorphisms and can be easily detected by electrophoretically resolving PCR amplified alleles on polyacrylamide denaturing gels. As microsatellites evolve, Taylor et al. (Dept. of Biological Sciences, Simon Fraser University, unpublished manuscript) suggest that substitutions, insertions and deletions increasingly interrupt the repeat arrays. These mutations within the repeat array are thought to stabilise the microsatellite locus such that imperfect microsatellite loci tend to be shorter, have lower mutation rates and be less polymorphic than perfect repeat arrays (see Weber 1990, Garza et al. 1995). 2.2.2  APPLICA TIONS OF MICROSA TELLITE MARKERS  Microsatellite markers were initially used for linkage analyses (Goldstein et al. 1995) and reconstruction of human phylogenetic relationships (Bowcock et al. 1994). More recently however, variable microsatellite loci have been used to quantify the level of genetic variation within and among populations (Bowcock et al. 1994; Gottelli 10  et al. 1994, Taylor et al. 1994; Allen et al. 1995;) and to assess the level of gene flow among sub-populations (Allen et al. 1995, Fitzsimmons et al. 1997; Miller & Kapuscinski 1997). These markers have also proven to be powerful in estimating the degree of relatedness among individuals (Queller et al. 1993; Dow & Ashley 1996; for a review of microsatellite use in population genetics see Bruford & Wayne 1993). Microsatellites have been less successful in phylogenetic analyses of distantly related taxa (Goldstein et al. 1995; Goldstein & Pollock 1997; Orti et al. 1997; Paetkau et al. 1997).  The limitations of microsatellites in phylogenetic  analyses relate to the discrete number of possible allele states (Goldstein et al. 1995; Goldstein & Pollock 1997) and the amount of variation masked by distinguishing among alleles by size alone (Angers & Bernatchez 1997; Orti et al. 1997). Microsatellites can increase and decrease in repeat number such that similarly sized alleles may be either identical by descent (descended without mutation from a common allele) or strictly by state (not necessarily sharing a common ancestral allele, but exhibiting the same length). Although microsatellites appear to mutate primarily via slip-strand mis-pairing, deletions, insertions and substitutions are common within microsatellite repeat arrays and in flanking regions (Angers & Bernatchez 1997; Orti et al. 1997). Such hidden variation (variation that is not detectable with size analysis alone), combined with the ambiguity of geneological relationships among alleles may confound the ability to infer evolutionary relationships among taxonomic groups based on allele size alone (Angers & Bernatchez, 1997; Orti et al. 1997).  2.3  2.3.1  METHODS  MICROSATELLITE CLONING AND ISOLATION  Methods for isolating microsatellite markers employed standard cloning techniques (see Bruford et al. 1996; Glenn 1995). Genomic DNA was extracted from ethanol-preserved tail fin tissue from 3 larval D. tenebrosus using standard Pronase and phenol/chloroform extraction (Taggart et al. 1992). Approximately 150  n  jxg of DNA were digested with Alu\, Haelll, HincW and  Rsa\ restriction  endonucleases, and electrophoretically separated on a 1% low melting point agarose gel. Fragments ranging in size from 200-700 bp were recovered from the gel with |3-Agarase I (New England Biolabs).  DNA fragments were cloned and  isolated as in Glenn (1995) using an M13 vector (Bluescript SK+) and a supercompetent cloning bacterial strain (DH5alpha, Stratagene). Approximately 4950 32  recombinant clones from 2 partial genomic libraries were probed with six [Y P]dATP labelled oligonucleotides: (GT)15, (GTGA)s, (GACA)s, (AAG)io, (TAA)io, and (TAG)io- The libraries were then autoradiographed (Fig. 2.1). Seventy-one putative microsatellite-bearing clones were picked from the original genomic libraries and stored in 20 ju.1 of autoclaved distilled water at -20C. Each cloned fragment was amplified using the polymerase chain reaction (PCR) (30 cycles of 1 min denaturation at 94C, 30 s annealing at 50C and 30 s extension at 72C) with M13 forward sequencing and reverse sequencing primers. Products were electrophoretically separated on 2% agarose gels in order to verify the presence of a 200-700 bp cloned insert. Of the seventy-one clones, 61 yielded a single band in the expected size range and were sequenced. All sequences were obtained using an ABI 377 Prism Automated Sequencer. Primers for PCR amplification were designed using the programs OSP (Hillier and Green 1991) and Primer3 (Rozen and Skaletsky 1998). During primer design, an attempt was made to reach consensus between the two programs.  PCR  conditions were initially optimised on 1% agarose gels using MgCI and annealing 2  temperature (T ) gradients. a  Final PCR assays were performed on a PTC-100™  Programmable Thermal Controller (MJ Research), with 100 ng template DNA and 0.5 U Taq DNA polymerase (Gibco/BRL) in 10 ul of 20mM Tris-HCI (pH 8.4), 50 mM KCI, 1.5mM MgCI , 800 U M dNTPs, 0.5 pmol of [y P]-dATP labelled forward primer, 32  2  2.5 pmol unlabelled forward primer, and 6 pmol of reverse primer in dH 0. The PCR 2  profiles for each locus consisted of 1 cycle of 2 min denaturation at 95C, 1 min annealing at T +2C (see Table 2.3 for T ), and 1 min extension at 72C, 5 cycles of a  a  1min at 94C, 1 min at T +1C, and 1 min at 72C, and 25 cycles of 45 s at 92C, 30 s a  12  at T , and 30 s at 72C. The last cycle was followed by a final 10-40 min extension at a  72C.  PCR products were separated on 6% denaturing polyacrylamide gels and  autoradiographed. Alleles were scored using a standard M13 sequence. Primer pairs were only considered for further screening if the amplified alleles were in the expected size range (one allele being of the expected size), if non-target bands did not obscure target alleles, if the genotype frequencies conformed to Hardy-Weinberg expectations, and if the alleles exhibited stutter bands (2-4 fainter bands beneath the allele) typical of microsatellites (see Fig. 2.3). 2.3.2  MICROSATELLITE CHARACTERISATION  Tissue samples of 3 -14 larvae from 13 populations were acquired from the Redpath Museum (Montreal) and the Royal Ontario Museum (Toronto) (Fig. 2.2, Table 2.1).  Thirteen individuals were also sampled from a population near  Nesakwatch Creek in British Columbia (Fig. 2.2, Table 2.1).  Unambiguous  microsatellite loci were initially screened for polymorphism with 20-25 D. tenebrosus samples selected from populations across its range. Polymorphic microsatellite loci were used to assay all 92 D. tenebrosus samples in order to characterise the level of polymorphism and calculate expected (H ) and observed (H ) heterozygosities. Each locus was tested for deviations from e  0  Hardy-Weinberg expectations across all samples using the software  GENEPOP  (V3.1d, Raymond & Rousset 1995). Expected H and H were also calculated within e  0  all populations at each locus (samples from BROO, THCR, and LACR were pooled into a group "OR" for comparisons across the geographic range, see Table 2.1). The total number of alleles detected at each locus was recorded. A t-test was used to test whether monomorphic loci had fewer units in the repeat array than polymorphic loci. A Fisher exact test was used to determine whether there was an association between the level of polymorphism (monomorphic, polymorphic) and the amount  of  non-repeat  sequences  (perfect,  imperfect)  in D.  tenebrosus  microsatellites.  13  2.3.3  PATTERNS OF GEOGRAPHIC VARIATION  Prior to determining the relationships between latitude and the dependent variables allelic richness and heterozygosity, I calculated Pearson's correlation coefficient between sample size and each of the two dependent variables. Because corrections for sample size could seriously bias estimates of allelic richness in the smallest samples (e.g. N=1, 3), the partial correlation coefficient for latitude (when sample size is held constant) was estimated using a benchmarking approach, where the general (i) and benchmark (ii) models below were compared: i)  Y = Po + P1X1 + M  ii)  Y = p + Pixi+E  2  +E  0  In these models, Y is the predicted number of alleles detected within populations (averaged over 6 microsatellite loci), p is the estimated intercept, P i is the 0  regression coefficient for the independent variable sample size (xi), p is the 2  regression coefficient for the independent variable latitude (X2), and E is the residual. Benchmark analysis determines the unique contribution a predictor (in this case latitude) makes to account for variation in the predicted variable (allelic richness or H ), when other independent variables (in this case sample size) are held constant 0  (see Stevens 1996).  In order to visualise the qualitative relationship between  latitude and allelic richness, I randomly sub-sampled 3 individuals from each sample (except LDCR, BROO, LACR, and THCR) ten times and then plotted the mean number of alleles detected against latitude. The problem of low sample size does not affect estimates of heterozygosity in the same manner as allelic richness because H is first calculated over all loci within individuals, and then averaged over 0  all individuals within populations. D. tenebrosus allozyme allelic richness (derived from Daugherty et al. 1983;  Good 1989) was plotted against latitude for comparison with microsatellite data. Because the mean per locus number of allozymes detected was not correlated with sample size (Daugherty et al. (1983) and Good (1989) data combined: r = 0.039, p 2  14  = 0.55), benchmark fitting was not used to assess the importance of latitude (i.e. sample size was omitted from the model altogether). 2.3.4  CROSS-SPECIES AMPLIFICATIONS  Two to six individuals from 18 species in the five families, Ambystomatidae, (Ambystoma  califomiense,  A. gracile, A. laterale, A. mexicanum,  A. texanum, A.  tigrinum), Cryptobranchidae (Andrias davidianus), Dicamptodontidae (Dicamptodon atterimus,  D.  Desmognathus  copei,  and  D.  ensatus),  Plethodontidae  (Aneides  monticola, Eurycea bislineata, Plethodon vehiculum,  ferreus,  Speleomantes  flavus), and Salamandridae (Paramesotriton sp. and Triturus carnifex) were assayed  with six polymorphic loci (DteA, Dte5, Dte6, Dte8, Dte11 and Dte14). All assays were performed as described above for D. tenebrosus (e.g. same PCR profiles) and no attempt was made to optimise conditions for cases where homologous alleles were not amplified (with the exception of D. ensatus, see below). 2.3.5  PHYLOGENETIC INFERENCE USING MICROSATELLITE DATA  Nei's (1978) genetic distance (D), calculated over 6 polymorphic microsatellite loci using TFPGA 1.3 (Miller 1997), was used to construct UPGMA' (un-weighted pair-group method of arithmetic averages) trees of the populations of Dicamptodon (including populations of D. tenebrosus, D. copei, D. atterimus and D. ensatus) in order to compare the use of microsatellites in phylogenetic analyses with allozymes (Daugherty et al. 1983; Good 1989) and morphological characters (Nussbaum 1976). Note that no alleles were amplified at three loci (Dte6, Dte11 and Dte14) in D. ensatus in spite of several attempts to optimise PCR conditions (MgCb and DNA  concentration, annealing temperature, and PCR profile times) for amplification. Therefore the missing alleles were treated as null alleles (alleles that do not amplify due to mutations in the priming site). TFPGA 1.3 (Miller 1997) cannot differentiate between null alleles and missing data, therefore D. ensatus was assigned null alleles for these three loci. A UPGMA phenogram of all 19 species assayed was also constructed using Nei's D for comparison to a dendrogram of salamander families (Duellman & Trueb 1994) based on 27 morphological characters.  The former dendrogram was  15  constructed with 3 microsatellite loci only (Dte4, Dte5 and DteS). Dte6, Dte11 and Dte14 were not used to generate this tree, because they did not amplify in species from the  families Ambystomatidae, Cryptobranchidae,  Salamandridae.  Plethodontidae,  and  Null alleles could not be assigned to these species (as with D.  ensatus) because no attempt was made to optimise conditions for amplification at  these loci. In order to infer the strength of both UPGMA trees, bootstrap sampling of the data sets was performed with 1000 permutations using TFPGA 1.3 (Miller 1997). The percent number of tree replicates supporting a given node was used as an index of confidence for that node.  2.4  2.4.7  RESULTS  MICROSATELLITE CLONING AND ISOLATION  Forty-four distinct sequences were obtained from the sixty-one sequenced clones (17 microsatellite sequences were cloned in multiple colonies). Of these, 12 had no microsatellite repeats, 5 yielded microsatellite arrays that were too close to the end of the cloned sequence for primer design in one of the flanking regions, and 2 sequences were not interpretable, probably due to contamination during colony incubation and isolation (e.g. more than one clone in a single colony). Eight loci were dinucleotide arrays of AG (=TC=GA=CT) ranging from 9-32 repeat units (32, 27, 16, 17, 17, 10, 9, 19), 14 were dinucleotide repeats of GT (=CA=TG=AC) ranging from 9-45 repeat units (16, 15, 18, 13, 45, 18, 9, 15, 29, 29, 24, 12, 16, 16, 13); 2 loci were compound arrays of AG and GT (24 and 40 dinucleotide repeat units), and one locus was a tetranucleotide repeat of CACT (14 repeat units, Table 2.2). A search of Genbank (a nucleotide sequence database) with BLAST 2.0 (Altschul et al. 1990) revealed that none of the D. tenebrosus microsatellite sequences matched any sequence deposited in Genbank. 2.4.2  MICROSATELLITE CHARACTERISATION  Unambiguous PCR products of the expected size were successfully amplified in D. tenebrosus at 14 loci: Dte1, Dte2, Dte3, DteA, Dte5, Dte6, Dte8, Dte11, Dte12, Dte13, Dte14, Dte16, Dte18, and Dte19. Of these, 6 were polymorphic with 4-17 16  alleles: Dte4, DteS, DteQ, Dte8, Dte11, and Dte14 (Table 2.3, Fig. 2.3). All six polymorphic loci were dinucleotides (three (GT) repeats and three (AG) ). Two of n  n  the polymorphic loci were perfect repeat arrays (DteQ, Dte~\4: no interruptions or substitutions within the microsatellite sequence), two were near-perfect repeat arrays (Dte5, DteQ: 1-3 substitutions within the repeat array) and the most polymorphic loci were imperfect repeat arrays (Dte4, Dfe11: several insertions or substitutions) (Table 2.2). Among the monomorphic loci, one was a perfect repeat (Dte3), five were near perfect repeats (Dte1, Dte12, Dfe13, Dte16, and Dte19) and  two were imperfect repeats {Dte2, Dte18). Monomorphic loci ranged in size from 830 repeat units, while polymorphic loci ranged in size from 16-31 repeat units. There was no significant difference between mean repeat numbers in polymorphic and monomorphic loci (two-sample t-test; p=0.63). I also found no association between the level of polymorphism (monomorphic, polymorphic) and the amount of nonrepeat sequences (perfect, imperfect) detected in D. tenebrosus microsatellites (Fisher exact test, n=14, p=0.53). All six polymorphic loci appeared to conform to Hardy-Weinberg expectations (0.089<p<1.0). Per locus mean H and H calculated e  0  across populations ranged from 0.056 - 0.464 and 0.045 - 0.413, respectively (Table 2.4). 2.4.3  PATTERNS OF GEOGRAPHIC VARIATION  There was a significant correlation between the number of alleles detected and sample size (Pearson correlation coefficient, r=0.442, p=0.05).  There was,  however, no correlation between sample size and mean heterozygosity (Pearson correlation coefficient, r=-0.188, p=0.121). tenebrosus  The level of polymorphism in D.  microsatellite loci declined with increasing latitude.  Three of the  polymorphic loci (Dfe5, DteQ, and Dfe14) were fixed for the most common allele in all populations north of Pierce County (WA, -47.3 °N), and of these, 2 were fixed within and north of Multnomah County (OR, -45.5 °N). The benchmark analysis revealed that latitude contributes significantly to the explanation of microsatellite allelic richness when sample size is held constant (partial correlation coefficient r=0.63, Fi,n=7.16, 0.02<p<0.05) The squared multiple correlation coefficient for the  17  general model including both latitude and sample size as predictors was 0.52 (Table 2.5). Allelic richness is plotted against latitude in Figure 2.4. Mean heterozygosity at microsatellite loci also decreased significantly with latitude (1^=0.64, p<0.001) (Fig. 2.5). Similarly, linear regression revealed a significant relationship between allozyme richness and latitude in the combined allozyme data (excludes populations with sample size <3, r = 0.37, p = 0.013) (Fig. 2.6). Populations above -47 °N were 2  fixed for single allozymes at all 31-34 loci (see Daugherty et al. 1983; Good 1989). 2.4.4  CROSS-SPECIES  AMPLIFICATIONS  Homologous alleles at DteA, Dte5, and Dte8 were successfully amplified in all 18 species of related salamanders.  Successful cross-species amplifications of  homologous products at Dte6, Dte11, and Dfe14 were restricted to species in the genus Dicamptodon (Table 2.6). All species except D. copei, and D. atterimus shared the same most common allele as D. tenebrosus at DteA, Dte5 and Dte8. DteA and Dfe8 were polymorphic within all families, while Dte5 was monomorphic across all families except Dicamptodontidae (Table 2.6). 2.4.5  PHYLOGENETIC INFERENCE USING MICROSATELLITES  A UPGMA tree (constructed with 6 polymorphic microsatellite loci) of populations of Dicamptodon spp. is shown in Figure 2.7. The tree suggests that D. tenebrosus is most closely related to D. ensatus. Dicamptodon atterimus is more closely related to D. ensatus and D. tenebrosus than to D. copei. Dicamptodon  copei is the most distantly related among all four species with Nei (1978) genetic distances (D) of approximately 2.4. No tied trees were generated. The relationships among all salamanders were poorly resolved using microsatellites. When the nineteen salamander species from five families were clustered in a UPGMA tree (Fig. 2.8), D. atterimus and D. copei clustered together while all others fell within a separate cluster. Within the latter cluster, there was no apparent grouping among families: the seventeen species shared Nei genetic distances of < 0.05. Thirteen tied trees were generated, however all tied trees clustered D. copei and D. atterimus apart from the rest of the species including D.  18  tenebrosus and D. ensatus.  None of the remaining tied trees (not shown)  successfully resolved the relationships among the salamander species either.  2.5  2.5.1  DISCUSSION  ISOLATION AND CHARACTERISATION OF D. T E N E B R O S U S MICROSATELLITES  The isolation and development of microsatellite markers is straightforward, albeit costly and time consuming, and has been further simplified by the availability of microsatellite manuals (e.g. Glenn 1995; Estoup & Turgeon 1996). Although amphibians generally have unusual genomes with sizes 1-2 orders of magnitude greater than other taxa (Duellman & Trueb 1994), the density of microsatellites in the D. tenebrosus genome appears to be similar to those of other vertebrates. Approximately 1% (47 / 4950) of the recombinant clones bore microsatellite loci. This is consistent with the findings of Scribner et al. (1994) who reported that 1.6% of clones bore microsatellites isolated from the common toad Bufo bufo. Glenn (1995) reported that typical eukaryotic genomic libraries yield microsatellite sequences in < 1% of the recombinant clones and that 25% of suitable microsatellite sequences yield unambiguous,  amplifiable loci.  I successfully amplified  unambiguous alleles in 14/25 loci (56%). However, only 6 of the 14 suitable loci (43%) were polymorphic in D. tenebrosus.  The remaining 11 loci, for which  microsatellite alleles were not successfully amplified, yielded several tens to hundreds of products indicating that the primer pairs designed for these loci were amplifying non-target sequences throughout the genome. Redesigning primers with greater specificity for the target sequences would be required to amplify these loci. Although the D. tenebrosus genomic libraries were screened with 1 dinucleotide repeat, 3 trinucleotide repeats, and 2 tetranucleotide repeats, 24/25 (96%) of the distinct microsatellite loci were dinucleotide repeats and 1/25 (4%) was a tetranucleotide repeat (Table 2.2). Of the dinucleotide microsatellite loci, 64% of them had (GT) repeat arrays. This is consistent with results reported in the n  literature: the most commonly isolated and used type of microsatellite is a (GT)  n  repeat array (Glenn 1995). Trinucleotide and tetranucleotide repeat arrays are  19  considered less common due to a lower probability of slippage during replication (Kruglyak et al. 1998).  In humans, some dinucleotide repeats appear to have  mutation rates 1.5-2 times higher than tetranucleotides (Chakraborty et al. 1997). Schug et al. (1998) also report that tri- and tetranucleotide repeats mutate at rates 6.4 and 8.4 times slower than dinucleotide repeats in Drosophila melanogaster. Previous studies suggest that microsatellites need to have >10 repeat units in order to be hypervariable (Weber 1990). The repeat number of the 14 amplified microsatellite loci ranged from 8-31. J. Taylor et al. (Dept. of Biological Sciences, Simon Fraser University, unpublished manuscript) suggested that as microsatellites evolve, the repeat arrays increasingly become interrupted by substitutions, insertions and deletions such that imperfect microsatellite loci tend to be shorter, have lower mutation rates and be less polymorphic than perfect repeat arrays (see Weber 1990, Garza et al. 1995). However, I found no significant difference between mean repeat numbers in polymorphic and monomorphic loci, and no association between the level of polymorphism (monomorphic, polymorphic) and the amount of non-repeat sequences (perfect, imperfect) detected in D. tenebrosus microsatellites. The observed heterozygosity (H ) and number of alleles at microsatellite loci 0  varies substantially across populations and taxa (Brooker et al. 1994; Bruford et al. 1996). The mean H at each locus in D. tenebrosus were slightly lower than the 0  observed heterozygosity of 0.625-0.750 at a single (CA) locus in Bufo bufo n  (Scribner et al. 1994), but within the range of reported values in other vertebrates (generally between 0.2 - 0.7) (e.g. Bowcock et al. 1994; Gotelli et al. 1994; Allen et al. 1995; Fitzsimmons et al. 1997; Miller & Kapuscinski 1997; Paetkau et al. 1997). 2.5.2  GEOGRAPHIC PATTERNS OF VARIABILITY  The low microsatellite and allozyme variation at the northern extent of D. tenebrosus' range may reflect recent patterns of D. tenebrosus distribution and range expansion. Post-glacial range expansion has been inferred as the underlying mechanism for lower neutral genetic variation in some northern peripheral populations (Sage & Wolff 1986; Hewitt 1996). As glaciers retreat, the unoccupied habitat is gradually recolonised through a series of dispersal events.  Such  colonisation events are predicted to be associated with founder effects such that 20  recently colonised populations should have less genetic variation than more central populations (Chakraborty & Nei 1977). During the last glacial maximum (Wisconsin period), the Cordilleran Ice Sheet advanced as far south as approximately 48 °N in the Pacific Northwest (McPhail & Lindsay, 1986, Williams et al. 1993). This ice sheet retreated northward approximately 10-12000 years before present.  The  relatively depauperate microsatellite and allozyme variation at the northern extent of D. tenebrosus' range may reflect a recent and gradual expansion northward into the Skagit, Snohomish, and Whatcom Counties as well as the Chilliwack Valley in BC (Figs. 2.4, 2.5, and 2.6). Alternatively, the pattern of decreasing variation at the northern extent of D. tenebrosus range may simple reflect a more general edge 1  effect: peripheral populations may have a lower probability of acquiring and maintaining genetic variation through migration than more central populations. However, populations at the southern extent of D. tenebrosus range do not appear 1  to share this decline in genetic variation as latitude decreases (see Figs. 2.4, 2.5 & 2.6). In an analysis of the geographical distribution of allozyme and quantitative variation in the pitcher plant mosquito, Wyeomyia smithii, Armbruster et al. (1998) found that allozyme heterozygosity decreased at the northern extent of its range. Armbruster et al (1998) suggested that the decrease in heterozygosity reflected a recent post-glacial range expansion northward into Manitoba.  Several other  phylogeographic studies have inferred recent post-glacial range expansions and the location of glacial refugia from the patterns of genetic variation across a species' range (e.g. Templeton et al. 1995; Redenbach & Taylor 1999; see Sage & Wolff 1986; Hewitt 1996). Populations that have acted as recent glacial refugia should have higher levels of neutral genetic variation than recently colonised, and gradually expanding populations. Fossils of Dicamptodon are known from the lower Pliocene ( 1 . 8 - 5 million years B.P.) in California and are thought to represent ancestral populations of the four extant species of Dicamptodon (Nussbaum 1976; Duellman & Trueb, 1994).  Interestingly, the greatest amount of microsatellite and allozyme  variation was found in Northern California and Southern Oregon.  21  2.5.3  CROSS-SPECIES AMPLIFICATIONS  One of the potential advantages of microsatellites is the use of primers in related taxa (Bruford et al. 1996). Most studies of microsatellite development report the successful cross-species amplification of homologous products in closely related taxa (e.g. Gotelli et al. 1994; Fitzsimmons, Moritz & Moore 1995; Rico, Rico & Hewitt 1996; Valsecchi & Amos 1996; Huang et al. 1998). Studies by Moore et al. (1991) and Glenn et al. (unpublished, as in Glenn 1995), of birds and mammals indicate that microsatellite loci amplify successfully among species within a genus. Seventyfive percent of primer pairs were successfully amplified in 8 closely related species of the genus Actinidia (dioecious perrenial plants) (Huang et al. 1998). However, some authors (e.g. Glenn 1995; Peakall et al. 1998) report a decline in the proportion of successful cross-species amplifications as divergence time among taxa increases. Up to 65% of primer pairs developed for soybeans successfully amplified alleles in other species within the genus Glycine. However, amplification outside the genus was restricted to 3-13% of primer pairs (Peakall et al. 1998). Microsatellites isolated in one family are less likely to be amplified in other families, and are rarely amplified in different orders (Glenn 1995, see also Bruford et al. 1996). Three of the six polymorphic microsatellite markers (DteA, Dte5 and Dte8) developed for D. tenebrosus were successfully amplified in 18 related species from 5 families of salamanders (including Dicamptodontidae). The other three loci (Dte6, Dte~\ 1 and Dte14) were successfully amplified within the genus Dicamptodon only. Several cross-species amplification studies also report that alleles tend to be longer and more variable in species in which the microsatellites were initially isolated (Ellegren et al. 1995; Fitzsimmons et al. 1995; Rubinsztein et al. 1995). Ellegren et al. (1995) argue that the screening and development protocol of microsatellites favours longer than average repeats resulting in a length ascertainment bias relative to other species. Hutter et al. (1998) performed a reciprocal test of ascertainment bias in Drosophila melanogaster and D. simulans and found no difference in mean product length, but reported greater number of alleles and heterozygosity in focal species (species from which the microsatellites were isolated). Hutter et al. (1998) suggested that the low variability in non-focal species is an artefact of the cloning,  22  isolation, and optimisation procedure and that the magnitude of ascertainment bias is a function of distribution of microsatellite sizes in different species genomes and preference of researchers for longer repeat arrays. I found no evidence of length ascertainment bias. The alleles amplified in the 15 species of the Ambystomatidae, Cryptobranchidae, Plethodontidae, and Salamandridae were all within the size range of alleles detected in D. tenebrosus and there were no differences in allele sizes among all families (Tables 2.3 & 2.6). Although several alleles in D. ensatus, D. atterimus, and D. copei fell outside the range of alleles detected in D. tenebrosus,  there was no tendency for the alleles to be larger or smaller than those in D. tenebrosus.  The number of alleles and heterozygosity observed in 18 non-focal  species were lower than those observed in D. tenebrosus. However, this is probably in part an artefact of sample size (92 D. tenebrosus versus 2-6 samples of each of the other species). Although one cannot easily compare the level of variability at DteA and DteQ among species due to low sample sizes, one can infer the usefulness of these markers from the observed heterozygosity. H calculated over all samples from the 0  families Ambystomatidae, Cryptobranchidae, Plethodontidae, and Salamandridae was calculated as 0.341 and 0.317 for DteA and Dfe8, respectively.  These  estimates of H are similar to those of microsatellite markers successfully used in 0  intra- and inter-population studies throughout a wide variety of taxa (see section 2.4.1). 2.5.4  PHYLOGENETIC INFERENCE USING MICROSATELLITES  Goldstein et al. (1995) proposed the use of microsatellites for phylogenetic studies of species that had independent evolutionary trajectories for up to several million years.  They modelled the behaviour of distance statistics based on the  stepwise mutation model (SMM, Kimura & Crow 1964) and the infinite alleles model (IAM). Goldstein et al. (1995) found that as microsatellite mutation rates increased and the number of available allele states decreased, SMM-based statistics lost their reliability. An empirical comparison of distance measures found that statistics that were not based on either the SMM or IAM models performed well in phylogenetic reconstructions of bear populations (Paetkau et al. 1997).  In addition, statistics 23  based on the SSM model had high variances and did not resolve known relationships between two sister bear taxa. Paetkau et al. (1997) suggested that the inability to resolve phylogenetic relationships among taxa with SSM-based statistics was due to allele size and distribution constraints of the microsatellites (few repeat arrays have more than 60 units, implying that there is a constraint on allele size; Goldstein & Pollock 1997). These studies assume that the mutations leading to changes in allele length are due to changes in repeat number only. However, sequencing of microsatellite alleles has demonstrated that differences between alleles are more complex than previously assumed (Angers & Bernatchez 1997; Grimaldi & CrouauRoy 1997; Orti et al. 1997). Orti et al. (1997) sequenced alleles (all of the same length) at one locus in 66 horseshoe crabs and found 34 distinct alleles with repeat numbers varying from 5-11. They also found variation at 22 sites in the flanking region of the microsatellite. Angers and Bernatchez (1997) and Grimaldi and CrouauRoy (1997) also found incongruences in the non-repeated flanking regions of similarly sized alleles. These studies suggest that the use of microsatellite length polymorphisms for the reconstruction of phylogenies may be problematic because the phylogenetic relationships among alleles are masked by the effect of hidden variation. Comparisons of the phenograms constructed with microsatellites at the family level (within Dicamptodontidae), and  among-family level (Ambystomatidae,  Dicamptodontidae, Cryptobranchidae, Plethodontidae, and Salamandridae) are consistent with these observations.  Phylogenetic inference using microsatellites  was unsuccessful in resolving the phylogenetic relationships among nineteen salamander species (Fig. 2.8) as compared to the presumed phylogeny of salamander families depicted in Figure 2.9. The estimated genetic distance D among all species except Dicamptodon atterimus and D. copei was extremely low. Without knowledge of the allele sequences in each sample, it is not possible to determine whether the discrepancy is due to undetected variation, or whether the microsatellite alleles shared by all species were conserved throughout the evolutionary history of those five families.  Given the abundant evidence of  incongruence between allele sequence and length, particularly over longer  24  evolutionary times, the latter is probably false. Of course, it is probable that the relationships were poorly resolved in part due to small sample sizes (2-6) and the small number of loci (3) sampled. Conversely, within the family Dicamptodontidae, microsatellites performed well in reconstructing the phylogeny of four species.  Morphological (Nussbaum  1976) and allozyme (Daugherty et al. 1983; Good 1989) characters were used to reconstruct the phylogeny of populations of Dicamptodon spp. and assign those populations to four species. Using six polymorphic microsatellite loci, I successfully constructed  a UPGMA phenogram that corroborates  the  classification of  Dicamptodon into 4 major groups (Fig. 2.7). There are however, slight differences  among the trees constructed by Daugherty et al. (1983), Good (1989) and the one I present. Daugherty et al. (1983) and Good (1989) found that D. copei was most similar to D. tenebrosus, while D. atterimus was the most distantly related of the four species.  I  found that D. ensatus was most closely related to D. tenebrosus, followed  by D. atterimus. D. copei was the most distantly related within the genus. My results are more similar to those of Nussbaum (1976) who recognised D. copei as a separate species and D. atterimus as being a geographic variant of D. ensatus. Nussbaum (1976) did not recognise differences between populations now assigned to D. tenebrosus and D. ensatus.  25  Table 2-1 Sample size (n), latitude, longitude, and location of 14 D. tenebrosus  populations screened with six polymorphic microsatellite loci (see map in Fig. 2.2). Sample Site  Code  n  Nesakwatch 20 Cumberland Creek Little Deer Creek Miller Creek Mallardy Creek Miller River Prairie Creek Cold Creek Wind River Oneata Gorge Thompson Creek Lake Creek Brookings Point Arena  Nes20 CUCR LDCR MICR MACR MIRI PRCR COCR WIRI ONGO THCR LACR BROO PTAR  13 8 3 14 13 4 10 6 5 6 1 1 1 7  Latitude <°N)  -49.00 48.31 48.23 48.29 48.00 47.43 47.26 47.22 -46.00 -45.50 44.04 44.07 42.06 -39.50  Longitude (°W)  -121.5 -121.6 -121.5 -121.4 -121.4 -121.2 -123.6 -121.2 -121.0 -122.1 -123.5 -123.3 -124.3 -123.7  District / County  Province / State  Chilliwack Skagit Skagit Skagit Snohomish King Pierce Kittitas Skamania Multnomah Lane Lane Curry Mendocino  BC WA WA WA WA WA WA WA WA OR OR OR OR CA  26  Table 2-2 Sequences of 25 cloned D. tenebrosus microsatellite-bearing inserts. Microsatellite repeat arrays are in italics and complementary sequences of designed PCR primers are underlined. N denotes an ambiguity in sequence. Locus  Sequence (5'-3')  Dtel  Cr.Ar.r.TCTCTGCTAGGTGGGGACATGTCCCCATTATGCTTCCAAAATCAGAACGCATGGCTCTAAGTTCGTT GATGGCTCCTGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGAGGG AGACA CTCTCTCCGCATTGTGGGGACCCCTCCCTGCCTAACAGAGAGGGGG ACAGATTACTGCCTCTTTCACCGCCATTTTTTCTGATCCGTGGGATTGGCAAACTACTCCCTGTGAAATGAA ATCTTTCGCAATCTCTCTCTCrCTCTCTCGCTCTCTCTCTCTCTCTCTCTCTCTTTCTCTCTCTATCTCCCACACAC AAACACACATTCCTCCCTCTGCTCCCTCTGGTATCGGAAATGAAAATCGTATCAACATCATATAATCAAATGTA TTT AT ATGTATTAG A ATGP ArTnGGrrrGTGTGACGAATGACATGGTCCTGGTATGCCTAACCTGTTATCCTT TGTCTTCCCCCTTTCTGTATAATGGGTAAGAATCAGTCTAAGCAACCACCACTCCCCGAAAAAGGTCGCCTG CTAACTACATGT CTTAGTCATTCAAGGCTACTTCTTTCCCATCCCTGCCACCTAACATGGGAGAGCAAAAAACTTGACGCCATG GGTCGTAATGTTTTTAATTTAGACAGCCTGGCTCCCACCAACCCGCTCCCCATC4CACAC/4C4C4C4CACAC4C ACACACACACACGGGAGCCAACTTTTAAATGT CCCATCTAATGAGTGGGTTGCTTCTGCCACCATAGCCTATTTAGGTGAAAGCAGGCTGGCGTTGCTTTTCTG AAGCAACGGCTGGTCTGGTTCCAGTGCGCGCGrGrGrGCGrGCGrGrGGGCGTGCGCGTGCG7'GrGrG7G7'G7G7*G7'G7'GrGCG7'GCG7'GrGCCCCCCTGTTCTGTTCTCCATGACCGCAACAAAGGCTGGCTCTGAGGCAGGGC AGGTTACACCCTGTGGAGTGGCAATCACCCTCTAGAATGCTAATCGACCCCATGGGGTAACTCCCTGTCCTCA CCCCATCTAGGGATCAAAGAATTCTGCTAGACTTTTCGGCACCCCCATCCCAG CTTGGAGGTTGGGAGGAGTTTTTGAAGTTGGAGGTTTGGAGGACAGTGATTACATCTTGGAGGGGAGTTTG TTATAGTGGAGGTTGGAGGGAGAAAGTAGAGAGCGAGAGAGAGAAAGAGAGAGAGAGAGAGAGAGAGAGAG AGAGAGAGA ATCACCACTTTGTGAGAGAGGAGGACTGGGAGAATGTTTGGAGAATTGT CTTGTTTGTTTGTCCACTCCCCTATTCTCCCTACTAACATACTCACA CTCACATCTAAAATCTCATACAACTTA A CA CTA CA CA CA CA CA CA CA CA CA CA CA CA CA CACACACACAG A AATCCGTTCCAC ATTCAGC ATC ACCATG A CTACC CCAAGGGAGCAGGGAAGGGACACATCTCTCCAACTGGGTCTCCATGTTGTGATGTGGGGAACTTGATTGTG TCACAGTCGCAGTCAGTGTTCGTTTAGTTAATTAGTTGAGrGrG7"G7TGCTCTGTAGGTGTGCTTTGGATGGGG TC,GrTAGGrGTG4 rG7"G7"G A AGG A AGGrGrGrGrG rGrGrGrGrGrGrGrNTCATATGGCCATAGCCCTAGG CAAAATGTGTCA rr.TGTTT ATG.A A A AGCTAGTTTTTTGAGGATGACTTTCTTTAGATCTGCATACATTGCATCTCCGGTTCGATT GGCTGCCTGATAAGGCAATTGCATCTTACTGACCAGTAAGCGAGTGGTATATATTGACTCCAGCGTTCTTCAG GTCAAGTAGCATTTCTAGC.C.ACTCTCTCTCTCTCTCTCTCTCTCTCTCTCTCTNNGTTAGTGAAGATGACCTTG CGG CTCCCACGTCTTCAACAGTCTGCTTCCAAATTATATGrGrGrGrorGrGrGrGAGACTGCCTGTGCACGAGA CAATGTGTGC AT ATCTG AG AC AGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGT GTGTGTGTGTGTGTGTGTGTGrGTGTGTGTGTGrGTGTGTGTGTGCGCAAGCGCATGC AC ACAAGGTCGC A CTTAGACTGGAACCCAATCATCTTCGGTGCTGATCCTTCCTTAAATCTCGTCTGTTTCAGGACAGTTGGTTG GTTTCCTGTTTGTTCATCTCTCCTCTGCAGTGATGATCAATGCCAAGAGGCGATGTAGTTTGCAGTGTGCACTT TATAAAGGGTCACA CTCACTCCTCACTCACTCACTCACTCACTCACCCATCCACTCACTCACTCACTCACATGAT TTAAAACAGATTATGGCCTAAGANGGGAAGTGTTCCCAAAATGGAGACAAAGAAAGAGAATGGTAGTGGG CCCACTGTCTCTTACACATGGTTGCTCACTCT AACCCAGTGTGTCACACACACACACGCACACACACGCACAC ACACACrcrcrCCCTTAATGCCACACACTATCCCTCATATTGTCTCAGATTGCCACACC7"A7"CACACACGCAC ACACACA CACACACACACATGAATTTTTCACACTCACACACTATCATA CTGGAAACTGCTGAGCCAACATGGACATCTGTTTGATATTATAAGACGAGGACCAGCAACCTTTGGCACTATG TTTTGAACGACCACTCCCATGATCCCACACCATTGACTATGCTGCCTTGGGCTGATGGGAGTTGTATTCCAAA ACATCTGAGAAGCAAAGGTCGCAGACCTCGGTCGTAAGGGGTGATATCTTGTCACGCATAAGCGTTACACA CACACA CACACATCATTTTGCCTTTGTTTGTGGATCAGGGGCAGACAGATCTGTGCCGCTCACGANCACACCG CCGCCCCAATACACATGAGAATTAAGTT CTGTGGAACAGCGGCAGAGAGCCCAAGGGGAAGGGTGGAGAGAGGAGACAGTTGCAGAACAGCGGCGGAGA GCCTAAGGAAGGGCGGAGGAAGGAGATAGCCGCGGAGCAGCAGTGGCGAGGGAAGACGCCCAAGGGCAG 4r.4G4G4G4G4G4r,4G4G4G4G44G4G4G4TATAGCC.GCGGAGGAGCAGTGAGAAGGGAACTCCCTAGAG AGGGGTGAGAACCCANGCCCACNGGT CTACAGGTCTTTCAAAAGAAGAGGAGTGAGACAGGGTGAGCGCCTGAACGAAGATCGCCTGGAGAGAGAG 4GAGAGAGAGAGA GAGAGAGAGAGATGGCGGGNGTGTGAGCATTTGATGGTGAAGATGGAGATATNCTGAC CTGGAAGAGGAGGAGAGGTGCCNACATGANACCCGCCATGCGGGGGACACCAAGACCCTGGNTGCCTACTA ACAGAGTGGAGAGAGGGT CTATTGTTGCATCTAGCATTTTAAGCATTAGTGCTGGCTTACTCCTCGCCCCACTCGGCTTGAGTCAGTATAT TGGAGTCT ACACATCAAAGCAATGTATCT ACTGAGAGAGAGAGA GAGA GAGAGATTTGACCATTGGAAAAGT GAGAAAGTGAAATGGGAGTGGTCTGTGGAGGGTGGTGGGTTTGCAAAGAAGANATCAGGTTATTTTGCCAT GTTAGGCCCAACATTGGCAGTG  Dtel  Dte3 DteA  Dte5 Dte6 Dtel  Dte8  Dte9 DtelO (Tetrarepeat) Dt ell DteYl  Dtell  DtelA  DtelS  ,  27  Table2.2 Continued Dte 16  £>tel7  Dtel8  Dte\9  CCCCAGGATCTTTAAAGGTATGAAAAATGTTTGTGTATGTGTGTGTTTGTGTGTGTGTAACTAAAATACGTG rr,TATr,Tr,r,ArTAATr,rAr,rrAGTrrATr,r,ATr.TTTrTTTTr,ACCTCATGGCTCTGGCATTTGTCCTTCCCTTG TATTTGTGTGTGACAGTGTGGCAGTGTGAATTAAT TACCCACTACTGTATTCTATGTGAAAGTTCAGTCCTTATGATCAAAGAAGGGAACCTTTAAGAAACTGTATAC ACCCACTTGCGTTTAAATGATCAACCAAATTATTGATTGCCTATTTCAAGTGCTTTCCCATTCAGCCATGATGA rrTrTATTTArTTAGTAATGTTAATrrCTGTTTTATCTATCAGACATTTGGTGCGTTGCGCCTTTTTCTGTATC CTTCGCACAATCCTGATGGGGTATCGACTGGGGGGGGGGCTGGATTTGACTTATACAGTGAGGAAAAATAT GTCATCCCTACACGTGANAGGATCTTTATCTCGACTGATTTACAGATTCAGGTTCAAAAAGGCACCTATGGTA GGATAGTCCCTAATTCAGGGTTAGCCGCTAACCCATATGTCGACAATTGATCCCCAATTATAAGGAATGTTAA TGGTTCTCCCTGAACCAACAATAGTGCTCACCCTTACCTATTCTNGGAATTTAACCCCAATCCCCCAGG7'GrGr GTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTGTTCT ACAGTG ACT ATTTACGA GTTGGAAGTGTTTGACTGTTCCCGTGCAGTATCAAAG CTGTTGGAGTAGATGGGGAGATGGAGAGATCCATGAGGAGAGGAAGGGAGAGAGGAGAGrAGCAGACAGG AATGTAGAGGTGAGAGGGAGGCACCTGCACAGTGCATGTCGAACTCACCAGGCGACCTAGTAGGCTGGAGG GTAAGrAGATGTGCGrGAGGGCAGTAGCAGGAGCCGGGCAGGAGAGTTCAGGTGTTGGGCTGAGGGTCCAT CATCATAACACTTGTCAGCCGTCAGAGGGCGAGCCAGGCGAAGAGGCTA CCACACATCACCAGGTGACTTCTTACTTGCATGATATATAGTGCATGACTTGCCTAGCATCCACGTCCCCCTGT TGTGTATTGGArGr ATGATGGAGCTCAGGGCGATGGGGTGTAACTGATTACCATGTNACACAAAGGAGAAGA AACATAGGGGATTAAGACAAATTGGACAGAGCAAGACAAGTCCACAAA CACACACACACACACACACACACA C4C4CACACACACACACACACACACACACAAAACAAGTGATTTTTAAAGCAACAAACTCCACTTTAGAATAA  DtelO Dte2\  Dte22  Dte23  Dte2A  Dte25  ATAATGT CTGGAGCGTTCCTGGTAGCGGTGTGGCTTTGGTCCTCGGGAGGCTGCAGTGGTGAACAAGTAAGAAACAAT GGTGTGTGTGTGTGTATGCGTGTCTGCGAACAGTTGTGTATGTGTGTGTGTTTGTCTG AAAGGGTGCGCATGT GAAATGAAAGCTCATGATGCTCTCATACTTAAGTTTGGGANGA CTGTTCCTGCATCAGATCCATCTGGAGCATGTCCAGGTCCAGTGCTCCGCCGAACTGCATGGTOGrGrGrGA GGrGrGGrGrGrGrGGTTTGGCCTCTCGAATATTGTTGCGTGACTTTTTGACACCTCTACAAATCATTTTCTC GGGTATCCTTTGGCTTTGAG rrrGAGATCGTAATTCTATTGCCTTCTTACTATTGCATATTTTTTATGCCCTTCCCCAAGGAAAACACATACT GAATTGTCCGGTATATCCTAGGTCTCACATTCGTGTGGTTTACGTTCGrG7TC7'GrG7'GrGrGrGrGrArA7'G7'A  TACA rGArGrGrArACArAGCATTGTGAATAGTTAATGAGACCGGTTTATGTAGAAT ACACACACACACACACCTGCACAGACACACACACACCACCCAGGAGCATTCGGAAACACCJGCATGTGTGCG TGCACArACArACTTGCATGTATAGAGGCACACACACACGCACACArACACACACCTCTGCACAAGCAAGCAT GCTCACAAACCTCTATTCGAATGAGTGCTCTCACACCTCTGTGGGTACGTGATCACACArGCACACrCATACA rAGCGGTGCGCACACCCACAAACATACCTGAGCATTCATGCTCACACACCTGAATGTGTGCACArACACACAC ATCCACACACACAG CAGGCAACTAGCGATGGAAGAAGGCTTAAGTGGGAGAGTTTTAGGGCAGATGTCTTGGTGGGAGGTATAA TTTCTGTAGCCCTCCCACTCTATGGTGCTGTCCAGCACCTCTTTCTCTACACCAAGGTTGTGCATTGTTGGGGC TTCTGTTAATACAAATAAACTTAAGCCTTAAGAAAGAAAGANGGAGAGAGGGAGAGAGAGAGAGAGAGAGAG AAGAAGACCNTTCCTTCAGANAAGGTGAGrGrGAGAGAGAGGGGGAGGGAGAGACGTCTCTGTAACTGGTTAT TGGTTAGAGCACTCTGTAACTGGTTATTGGTGACAGCGCCCTGTAACTAGTTATTGG CCACATGCCCACTTAACAGACACCCATCCATACACACACAATCACACGCGCATACACACAGGCTCACTCGCG rArArArATAr.CGCATGrATA<?ArAAACA01CACACAATCACAAGCGCATGCACACAGGCTCGGGGGTCATCT TACCTCCATTCCAGCGCCGACAGCCCAGCACCTCAGCAACAG  28  c o £ « a> 2 W o £ ° c3 co o  TJ  <D  CO CO JD  CO  0 0  "cO O  o _o o CL  O  0 c 0 0 o  (3  N, 'CO  to  CO to  0 0  to  0 0 "co  c  CD  CD CD  TD  CD JO  JD  E  CD  CD c  O  E  0  TD  o to « o o  o  CL  D_  CD  •+-*  o  o J=  o  TJ  CD C  CO to  s •Q CD C  TD  v.  0 c o o  CO = CO CO CD Q. CD  o co 0 0  t  o CO CO  H 0  13  to to O  b CO  E o  _co CO TD >  -*—>  CD  E  CO I CM CD — (0 I-  c "CO  0  c c CO TD c CO CO  CO CL  CL  o E  6  CM  cu  0 0  5  T<  CM  i- ~  CM CM  co" of  Is  CO 00  00 CM  t-  i-  Is  LO" CM CM CM T-"  i-  CO  - <o  iOT t -  co" O " i„1  OJ  o oo  05 00 O  oo oo  CO CD CO LO  £ C M  CM"  co oo  oo co ^t-  CD CD  o"° . CM co o oo  Is  oo  ^  CM  c  0 >  < o o o <  0 CO  0  to  I-  <  CO  CD O CD  CO  0 o o  CO c CO  CD O CD  O  c c  o  o So a £ CD  CM  (0  co CO a.  O CJ)  ^ 5 CD I-  o O O  jz. <  <  < CD  CO OC  O P n CD O O CD CD " O O O  O  < o O O  O <  < o o o  < <  O <  o o CO o <  JD  < < — I o < o  CO  — I  o  o < < o < o  c  0  O < CD  5 < 5 CD OCD  TD '>  C  CO  •^t CO CO CO CO LO  i— ^:  'K  </)  CM 00 CM CM  CM  co £2 t—  (A CU N  18  CL  TD  0 c 0 0 o  co  CM W CO LO" _" CM iCM CM CO ~ C CD  o  OJ  E 0  CD  JZ  _co CO  CD to  cr  y  CD  c lO  cfl Q.  o °  oo CM  co £2 CM  TD  to" CD O c CD  0  LO  CM CD  s  E  0~ — t- » CO to  s  LO  I - ro CO  JD  CD  a) "cc  o  I-  5  o  •a CD  CM  a>  oo  I-  < c  CO  A <  s  s  O LO  1  E o c o  CO  CO CM  CO  CO  0  in o  CO  CO  u c  CO  3  s -Q CD C CD  C\J CD  O CD CD CD  CO  cr co CO del co £T Oo ho E «? a.  IT)  9*7  Li. DC co  -5  a  o « 9O - CD - CD  < o < CD < < CD CD CD I  DC  < CD O O ° t  o  • < <  F £  9° LL  o 9, o^ o  n  c3 o o < o o < CDO <  DC ll DC  t=o o Io < o < < < o o  ^8 <p  ob cot o < £o o o r- O  o o o o o < o o<  Ocj < h<— I O t—I-  DC  DC  m CD  CD  co  O I< O I-  o oo  P8 LL  h-  o  o o o H  O  o < o o P o O Io o <o O— I <o o< DC  DC  53  Q  c  CD  ^ o 5 w C  oo m  o  o> u O< O o  CD  co in  CO  CO  CM  m  o oo CM  co ? C\J  c\f  2  W  0) N  o co CO  i - co  <7j  00 CM  in  t -  CO  CD  CO  CD  CM  <  CD  <  CD  CM  CD <  CD <  <  CD < < CD CD < CM CD < CD  CM  IS  (0 0)  a  < O H  00  <  CD O < CD < CD — I O OCD 2 CD ^DCD C CD 0)  3  c c o o CO CM l-g  0) 3  cr <D </>  E " ~ w  a. —"  CD  CO  CC  a> o c  CO  I < -  o p  O CD <— I CD CD < h-  O < CD O  &  CD <  H CD < — I CD  <  CM  CD <  < CD  < S3  O  <  CD U CD CD  <o  << O CD <D O C  CD CD b OO CD O CD O CD < <CJ  <^ u. CC  LL. DC  £8 St  o  c3 < h-  CDO < O CD — I I- O CD — I < O CD O CD <  <9 LL DC  O < CD °  CD O CD O  <%  |!  o o CD o  is  < CD CD— I < < CD CD < < CD CD  < CD P CD h tD  CDP CD < < <> O <  99  L L DC  CM  CO  CO  Q  Q  Q  CD O CD < CD  ^ < I I I DC  < O  < < <  CD CD  O P < H  S3  CD CD  CD < CD < < H < O Li.Lt  |oo Q  CD  CD  CC o X T3 C CC CD  X c CO CD CO  CD CO 0 CL  E  cc to  c 'to CD CD L_  .C CD -C H to CD to CD C CD CO  CL  c 0  > 'co 0  0  CO to CD N 03  ' CO  0  •ti CC tO  O  CL  N O  CC  CD  0  CD  C  ZJ  o o o  CD  E  CO o o CO to to  X  _Q O •a c cc _£>  T3 0  O CD  a. x UJ  CM a>  (0 IXI  CC  d  .Q  I— 0 0  DC cc •a CO c 0  cc to 0  -t—'  to CM  CO to to  O i  CJ  cc •o 0  -4—»  DC  .92  O  to CZ O  CC CL  E  o o  CO co CO  1  o  m m  d  d  d  co  d  o  o  o o  in  in d  CM CD  d  d  d  co CO  co CD  d  d  d  o  o  o  o  CM  d  cn Cfl CM  CO  d  d  o  co  y—  CM  CM CD CO  co co co  i— co  d  o  d  o  o  o  o  d  d  o CM  d  d  co  d  in d  8 s!  •a 0  o o CL  0  DC  o o  d o m CM  d  in in  in d  m  r-d  in d  ao  OC  o o m d  o  to  2. o  to  E o i _  to  0  CO to  E  OC  o  d  ,_  o o DC +—>  o  o  o  o  o q  o CM  d  o  1—  oo  d  00 i—  d  o m CM  d  CN <> / £ Hin  d  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  o  CT>  CM  d  o  o  o  o  o o in d  d  o o co  to CM  m  oo co  d o  m •i—  d  d  o  o  o  o  d m  co  d o m  CM  m d  00 oo CO  d  d  CD CO CM  hco  CO  d  d  d  in  00  d  o d  00  oo  d  CO CM CM  d  oo  d  o  d  o in CM  r-  o m CM  oo CM CO  o m CM  a>  d  d  0 0 CM d  CM  CM CM  d  o CM  d  o CM  d o  o in CM  d  o in d  1  CO co  co CO  o  o  d  d  o CO  o o CM  d  d  d  o  CO  d  o d  CO o m d d  CM co  to  d  d  o  CO eo  oo  d CO  to  d  d  d  T— T—  m o  CM  i—  d  d  o  o  d  d  d  o  d o  d  o  CM  d  o  o  d  m d  O) CO  CO CO  d  d  m CM  cn CM  d  d o  00 00  i—  CM CM  d  T—  o  d  d  T -  o  CM  oo co  T—  o  o CO  CO  o> to  o  d  CM m cn d d  00  o  co CM  CM  T -  o  co  o d o d m  un  CD  o  CO CO co  o  CM  d  CO  o  CO  m  d  o  o  •tf  CD  oo  o o  o  OC  T3 CZ  0  o  CD  d  _l  m  d  SSI d  O <  O X H  CO CO  LO CD  CO  cc DC  o  DC  OC <=>CD  o o  d  o  •a tz cc  Q.  CO  m CM  CC  JC ZJ  o d q d  to CM  C5  0  DC  TJ c  to LO  O  t_  to to c  d  d  eo d  0  E  to •<*  cn to  O to  0  -o CD CD to  o o  o o cc 0  o  CM a>  in CM  d d o a> o CD o a> CD CD o a > o o a > o X X X X X I X X X X X X X X *t m in to to 00 00 co co 0) 0) 0) CO £ a > a > a > a > a > tt) a ; a ) to Q a S Q Q • a 6 S a a 5 T—  c o  CM LO  o  o o  •4—»  TJ  fc—  o  eto  CO CD  k_  mpl  ize as  CL  o H JZ  TJ 0 CL  c o  _C0 H—  O c  o  cCO  CO CO  CD  0  JZ  0  4—»  o JD CD  oC L  JZ  cor  hmar modi wh  CZ "•+-» CO  CO  near lodel  c  E  —  CO CD c b  g  0  CJ c 0  JD 0  co CL  Eo o 0 XJ  CO -4—»  < >  o 0c o it  o CL  c  o  •«—'  CO  "0  L_  o  CO  cz CO  DC  CO 0 TJ  0  -»—• 0 0  0  c JZ  in evi 0  -4—• CO CO  0  o g E  CL *•«—»  ZJ  E  "<*  T—-  CM  LU  l_  +  0  X  CM  N CO  E c  co"  <N GZL  LJJ  X  X  +  0  CO  0  CT)  LO  CO  _cg  oo  LO CO LO  '"c B  oo c o 0  O CO LO  CM  0 > 'CD  0  (0  T —  0  CO CO  H—  CO  CT)  CO CM CO  CO LO  0 >  0  CO CO CO  LL  c  'g  CJ)  Z3  JO_  'CO  c  CO  o  CL  7? ®  I—  oo 0  c  o  (0  CD V CD-  CO  O  XI  o  coe  -*—'  CL  LO  » cz  the p rtial c  -«—* <  _C0  N CO  ises am  ZJ  CD  -•—' "•4—•  JZ -*—'  73  co  0  co CO "O  0  d  pn  CD  Q.  the  I—  CO  buti  CO  CD  sq  o  TJ 0  GZL 0  V  o cz  u  0  3  0  +  o cd  +  ca + o  ca ll  O CO  0  cz JZ  g g 0  "co  o  i co  0  TJ O  CO I0  c 0  (3  E  JZ  o cz 0  CQ  32  LO  E ,5  o CO CM  Q  O CM  O CD CM  i-  Q  00 CM  oo  00 O  O OO  oo oo o oo oo I - r- oo I - I s  s  CD CD  oo  CO 00 00  oo I-  co I-  I-  oo o oo I - oo I -  co oo  oo co I- I- I-  0 0 oo  s  s  mm  s  — I r-~ r-~ r-~ r— r — - r — s  o  s  s  o oo  r--  r-.  s  s  s  oo col I- I-  s  s  s  oo  Q  00 CM CM CM  o CM CM CO CO CO CO CO CO CO oo oo oo oo oo oo oo  Q  iCD- CCOM 2  CD  CM  CM I-  in  CD CD CD O) CD CD CD oo oo oo oo oo oo oo  CD oo  o  i—CD  CO CO CO CO CO oo oo oo oo oo  i-  CD CO CO T - CM CM CM  Ol  co cd" co oo oo CM T— T—  CO oo  T—  CD  T—  CD i 00 -  CO  OJ  CO CO CO  .« ILL  CD CO TO  CD RO  a E CO o >. ,1 "i il  CO CO CO CO CO  CO CM  .CJ  5  co  OD CD  oo oo I  •S o  c  ,  CM  CD CD CD" CD CD oo oo oo oo oo  T"  CD CO  CO "5 CD  CM •>CM CM CM CM  s  I "  to  oo oo I  iCM- CiCMM- CCOMM CO iCM-  CM  CO CO CM CM CM CO CO  CD CO I  fc co  m  ^  ca  3 fc  .ca  S  to .cc  c  I CD CO  32 c  •  o  cc  c  £  "OD I  I  co to  c CO  •fe  a  o  E  1  ca to CD =3 .C c  . , c "E5.2Q Q 9K co Q o  < < < < < <  s c  CD CO  32  I o sz £ 0.  § 53 •c -= c c -° CO 6 o Dl CO O CD  £ &  CD CD  to cr a. co a. CD 3  Q Uj  CD CO TO  "i  O "3 CO CD  ^  to  ~u£ 5 c co 5 3 , E ^ a. ;c haj CO  CO  33  *  *  Figure 2-1 Autoradiographed clones. The large circles are hybond filters used to bind single-stranded DNA that was cloned in approximately 5000 bacterial colonies (smaller dots on the filters). The darker clones (11, 13, 14, and 18) represent colonies bearing microsatellite sequences. In these cases, radio-labelled oligonucleotides hybridised to the microsatellite sequences in the bound DNA and exposed radiation-sensitive film.  34  Figure 2-2 Range of D. tenebrosus from south-western British Columbia to northwestern California (light shading). Locations are numbered according to sample sites where 1=Nes20, 2=MICR, LDCR, and CUCR, 3=MACR, 4=MIRI, 5=COCR, 6=PRCR, 7=WIRI, 8=ONGO, 9=LACR and THCR, 10=BROO and 11=PTAR (range is based on Daugherty et al. 1983 and Good 1989).  35  Figure 2-3 Autoradiograph of a polyacrylamide denaturing gel used to size the alleles of three D. tenebrosus microsatellite loci: Dte11, Dte14 and Dte16. The four lanes labelled GATC represent the standard M13 sequence used to size the alleles. Dte11 and Dte14 are polymorphic with 3 and 8 alleles shown, respectively. Dfe16 is a monomorphic locus with a single 149 bp allele. Numbers indicate allele sizes, and short arrows point to three heterozygotes. Stutter bands - several lighter bands beneath the true allele - are characteristic signatures of microsatellites and are particularly pronounced in Dte11.  36  2.8 co cu O  < o  k_  CD .Q  E Z  ^  {  2.6 2.4 2.2 2  1.8 1.6  i  1.2  h  1 38  43  48  Latitude (degrees N)  Figure 2-4 Latitudinal gradient in the number of microsatellite alleles detected in 14 populations of D. tenebrosus. Sub-samples of 3 individuals were randomly selected from each population ten times. The open circles represent single samples from BROO, LACR and THCR. Error bars represent the standard error of the mean.  37  0.9 0.8 0.7  o X  0.6  CO CD  0.4  c  0.5  co  0.3 0.2 0.1 -  0 38  43  48  Latitude (degrees N)  Figure 2-5 Geographic pattern of observed heterozygosity (H ) in D. tenebrosus. 0  Ho was averaged across individuals within populations. Closed circles represent samples with 3-14 observations. Open circles represent single observations from BROO, LACR and THCR. Error bars indicate the standard error of the mean of the 10 sub-samples.  38  1.40 -, 1.35 CO CO CD  c o  Lr  1.30 1.25 1.20  o  1.15 -\  <  1.10 1.05 1.00 39  41  43  45  47  49  Latitude (N) Figure 2-6 Variation in D. tenebrosus allozyme richness over latitude. Allelic richness (number of alleles detected) was averaged over 31-34 allozyme loci (data from Daugherty et al. 1983 and Good 1989).  39  — I  2.426  1  1 820  I  1  —I  1.213  0.000  0.607  D- tenebrosus (MACR) D. tenebrosus (MtCR)  97  D. tenebrosus (NES20)  91  D- tenebrosus (MtRI)  63  D. tenebrosus (PTAR)  66  D, ensatus D- atterimus 99  [  D. copei D. copei  Figure 2-7 UPGMA cluster of Dicamptodon populations. Nei (1978) genetic distance, D, is on the horizontal axis. Percent bootstrap replicates supporting nodes are indicated. No tied trees were generated. Please refer to Table 2.1 for location of D. tenebrosus populations.  40  1.600 1 2 3 4 5  1  1.200  0.800  0.400  0.000  Dicamptodontidae Ambystomatidae Cryptobranchidae Plethodontidae Salamandridae  100  83  4 2 5 3 2 1 4 2 4 4 5 2 2 2 2 4 1 1 1  Speleomantesflavus Ambystoma gracile Triturus carnifex Andrias davidianus Ambystoma californiense Dicamptodon tenebrosus Plethodon vehiculum Ambystoma tigrinum Desmognathus monticola Eurycea bislineata Paramesotriton deloustali Ambystoma cingulatum Ambystoma mexicanum Ambystoma texanum Ambystoma laterale Aneidesferreus Dicamptodon ensatus Dicamptodon copei Dicamptodon atterimus  4  Figure 2-8 UPGMA cluster of 19 salamander species constructed with 3 microsatellite loci. Nei (1978) genetic distance (D) is on the horizontal axis. Percent bootstrap replicates supporting two of the nodes are indicated. Numbers 1-5 correspond to the families shown to the left of the tree. Thirteen tied trees were produced.  41  Sirenidae Hypnobiidae Crvptobranchidae  Proteidae  Dicamptodontidae  Amphiumidae  Salamandridae Ambystomatidae  Plethodontidae  Figure 2-9 Phylogenetic tree of nine extant families of Caudata (salamanders) based on 27 non-paedomorphic (adult) characters. Species in families with bold type were used for the cross-amplification assays using microsatellites developed in D. tenebrosus. Modified from Duellman & Trueb (1994).  42  3  C H A P T E R 3 : E V I D E N C E O F R E C E N T B O T T L E N E C K S IN S U B - P O P U L A T I O N S O F T H E PACIFIC G I A N T S A L A M A N D E R (DICAMPTODON  TENEBROSUS)  IN  BRITISH C O L U M B I A .  3.1  INTRODUCTION  Population size (Pimm et al. 1988) and colonisation ability (Burbidge & Mackenzie 1989; Laurance 1991) are among the best predictors of the long-term persistence of populations. Large, widely distributed populations are less likely to go extinct over the short term than small, isolated populations (Frankel & Soule 1981; Pimm et al. 1988; Berger 1990; Primack 1993; Frankham 1995a). Fluctuations in census size are common among natural populations due to demographic, genetic, and environmental stochasticity. Population bottlenecks are rapid declines in effective population size, N (the size of an ideal population that e  would experience the same amount of genetic drift as the natural population is experiencing) (Hartl & Clark, 1997). The demographic outcome of a population bottleneck is usually a reduction in census size (although Luikart (1997) notes that rapid decreases in N are not always reflected in census size), which may increase a e  population's risk of extinction (Pimm et al. 1988; Berger 1990; Primack 1993). The genetic outcomes of a population bottleneck include a reduction in genetic variation, an increase in the probability of fixation of slightly deleterious alleles (Lande 1994) and consequently, a loss of evolutionary potential (Frankel & Soule 1981). Bottlenecks can cause a rapid loss of alleles, particularly rare alleles, accompanied by a less rapid loss of heterozygosity (Wright 1931; Allendorf 1986). Theoretical and empirical evidence suggests that changes in the genetic structure of populations induced by bottlenecks, coupled with a decrease in population size can reduce the probability of long-term persistence (Lande 1988; Mills & Smouse 1994; Lande 1994; Frankham 1995a, 1995b).  The impact of bottlenecks on genetic  variation and the long-term persistence of populations is related to their severity such that populations passing through small bottlenecks have a greater probability of extinction than populations passing through larger bottlenecks. Natural populations  43  commonly pass through bottlenecks due to environmental (e.g. glaciations) and demographic factors (e.g. founder effects), as well as human disturbances (e.g. over-exploitation, habitat loss, exotic species introductions). In general, the effects of human disturbance are to reduce and fragment habitat, thereby destabilising populations and causing them to decline. Habitat fragmentation and loss are believed to be among the leading causes of global amphibian declines (Phillips 1990; Pechmann et al. 1991; Blaustein et al. 1994; Blaustein & Wake 1995). It is difficult to infer population-level characteristics that are of conservation concern (e.g. N , rate of inbreeding, loss of genetic variation, levels e  of gene flow) from census and demographic data alone (Harris & Allendorf 1989; Frankham 1995c), particularly because historical population sizes and levels of genetic variation are unknown.  However, such characteristics have been  successfully assessed indirectly using molecular genetic markers. Genetic markers have been used to help identify cryptic species, assess levels of pair-wise relatedness, population structure and differentiation, understand behaviour and ecology, describe mating systems, detect bottlenecks, infer historical processes, and design captive breeding programs (Milligan et al. 1994; Hedrick et al. 1996; Mace et al. 1996; Cornuet & Luikart 1996; Luikart 1997; Redenbach & Taylor 1999; Taylor et al. 1999). The Pacific Giant Salamander, Dicamptodon tenebrosus, is a semi-aquatic species that breeds predominantly in cool, fast flowing mountain streams from southern British Columbia (BC) to northern California (Nussbaum et al. 1983; Blaustein et al. 1995). In BC, D. tenebrosus is presently recognised as vulnerable by COSEWIC and is red-listed by the BC Ministry of the Environment, Lands, and Parks because habitat destruction and fragmentation threaten its long-term persistence in Canada (Farr 1985, 1989; Haycock 1991). The northern distribution of D. tenebrosus extends into the lower Fraser River Valley, but is mostly restricted to the Chilliwack River Valley, an area of approximately 250 km immediately north of the Canada - United States border (Fig. 2  3.1). The consequences of current forest practices on the long-term persistence of D. tenebrosus populations are poorly understood. They may be negligible, or the  44  may include changes in demographic parameters (e.g. recruitment, survival, and growth rates), a loss of genetic variation, population bottlenecks, lower dispersal rates, and local extirpation. Several studies conducted throughout their range have reported lower larval densities of Pacific Giant Salamanders in clear-cuts relative to forested stands (Bury 1983; Bury & Corn 1988; Connor et al. 1988; Corn & Bury 1989; Cole et al. 1997) while other studies reported no difference in larval density among forest types (Hawkins et al. 1983; Kelsey 1995; Vesely 1996; Neill & Richardson 1997; Neill 1998) or increases in density within deforested sites (Murphy et al. 1981; Murphy & Hall 1981). My research is part of an on-going study to assess the impacts of current forest practices on the persistence and colonisation potential of Pacific Giant Salamanders in British Columbia. The detection of recent bottlenecks in clear-cut sites of Pacific Giant Salamanders in BC may signal the need to implement forest management practices that reduce the impacts of forest harvesting on the population structure of the salamanders.  In this chapter, I report the use of 3 polymorphic microsatellite loci  and 39 amplified fragment length polymorphism (AFLP) markers in addressing two related questions: 1  Do populations in recently harvested sites exhibit signs of bottleneck  effects (lower allelic variation and heterozygosity, greater proportion of fixed loci, shifts in allele frequencies)? If clear-cut logging causes D. tenebrosus  sub-  populations to pass through bottlenecks, then we should expect to observe lower levels of genetic variation in recently clear-cut sites than in recently undisturbed sites. 2  Is the level of genetic variation correlated with stand age?  populations of  D. tenebrosus  Sub-  are scattered throughout the Chilliwack River Valley  and are probably linked to various degrees by migration (see Chapter 4). If clearcuts reduce the amount of genetic variation of local sub-populations, most of this lost genetic variation may be recovered over time through migration from surrounding sub-populations.  45  3.2 3.2.1  METHODS STUDY SITES AND SAMPLE COLLECTION  A genetic survey was conducted during the summer of 1998 in the Chilliwack River Valley, British Columbia, at the northern periphery of  D. tenebrosus'  range  (Fig. 3.1 & Fig. 3.2). The study area falls within the Coastal Western Hemlock biogeographic zone. Eight sample sites were selected for this study: 2 sites were within old-growth stands (>250 years, NesF4, NesF5); 3 sites were within secondgrowth stands (30-60 years, Cen23, TamSGI, and VedB); and 3 sites were within recently clear-cut sites (3-9 years, FolQ, Fol12, and TamC) (Figure 3.2, Table 3.1). It was not possible to randomly select sites because few streams are occupied by sufficiently high densities of salamanders for reasonable comparisons of larval density and genetic variation among forest types (WE Neill, Dept of Zoology, University of British Columbia, pers. comm.; pers. obs.). In searching for suitable sampling sites in May and June 1998, I surveyed 60-150m reaches in 2 old-growth, 7 second-growth, and 8 clear-cuts streams for larval salamanders.  With the  exception of 3 adjacent second-growth sites (near TamSGI, see Fig. 3.2), I found fewer than 9 salamanders in each of those streams (surveys entailed 4-8 personhours of intensive stream searching). The sub-populations of D.  tenebrosus  chosen  for this study have been intensively studied since 1996 by WE Neill and JS Richardson (Faculty of Forestry, University of British Columbia) and afford the opportunity of future comparisons between genetic and demographic data. The two old-growth sites are within the same drainage (Nesakwatch Creek), as are two of the recently clear-cut sites (Foley Creek) (see Fig. 3.2). All samples sites were in different headwater streams. Each sample site is referred to as a putative subpopulation throughout the text. Tissue samples were collected from 31-65 individuals at each of the eight sites (Table 3.1). In order to reduce the potential for sampling families, samples were collected from 3-5 size classes (<60mm, 61-80mm, 81-100mm, 101-14 mm, >141mm total length), throughout the length of the stream reach (85-200m) (Fig. 3.3). Minute tissue samples were collected from tail fins of larvae and tails of adults  46  and preserved in 95% ethanol. DNA was extracted from the tissues using standard phenol/chloroform extraction protocols and stored in TE (Tris-EDTA) buffer at -20C (Taggart et al. 1992).  3.2.2  MICROSA TELLITE AND AFLP ALLELE FREQUENCIES AND HETEROZYGOSITY  Fourteen microsatellite loci were developed for this population-level study: Dte1,  Dte2, DteQ, DteA, Dte5, DteQ, DteQ,  Dte11 Dte12, Dte13, Dte14, Dte16, Dte18  and Dte19 (Table 2.3). My survey of genetic variation over the geographic range of D. tenebrosus  (Chapter 2) revealed genetic polymorphisms at  at the extreme northern extent of  D. tenebrosus'  DteA, DteQ  and Dte11  range, while no polymorphisms  were detected at any of the remaining loci in BC. Therefore microsatellite alleles were amplified at the three polymorphic loci  (DteA, DteQ,  and Dte11) using standard  microsatellite protocols (see Chapter 2) in all 31-65 individuals from the eight subpopulations in BC. Randomised sub-samples (5-8 individuals) from each subpopulation were also used to screen for rare alleles in Dte~\,  Dte2, DteQ, Dte5, DteQ,  Dte12, Dte13, Dte14, Dte16, Dte18 and Dte19. Statistical tests designed for detecting bottlenecks have low power (<20%) with three polymorphic loci (Luikart 1997). Therefore, I also optimised three pairs of AFLP primers for DNA fingerprinting using the AFLP™ Analysis System I and AFLP Starter Primer Kits (Gibco BRL, Life Technologies). AFLPs provide a rapid method for screening genetic variation within and among populations (Vos et al. 1995). They combine the advantages of screening many loci with the high stringency and reliability of selective PCR (polymerase chain reaction) primers. Briefly, AFLPs are generated by (1) restricting the template DNA (both nuclear and mitochondrial) with EcoRI and Msel restriction enzymes, (2) ligating short double-stranded DNA sequences (20 bp) to the restriction sites, and (3) amplifying small fragments (typically 100-300 base pairs) using the PCR primers designed to match the ligated sequences (Vos et al. 1995; Blears et al. 1998). AFLPs generate band profiles, typically with 50-150 bands (Vos et al. 1995) that are similar to randomly amplified polymorphic DNA markers (RAPDs) in that they are non-discrete, dominant markers (see Blears et al. (1998) for a review of the AFLP procedure and its applications) 47  (Fig. 3.4). I screened 20-28 randomly selected individuals from each sub-population with three A F L P primer pairs (Table 3.2) following the recommended protocol (Gibco BRL, Life Technologies). The A F L P markers were pre-amplified using non-selective primers as suggested by Vos et al. (1995) and Life Technologies. The EcoRI primer (EcoRI-ACC) was end-labelled with [y- P]-dATP. 33  A F L P band profile reproducibility trials (AFLP protocol repeated three times for each primer pair using two individuals) showed that A F L P banding profiles were reproducible  in  bands  greater  than  approximately  100  base  pairs  (bp).  Approximately 4 0 % of all the A F L P markers appeared to be monomorphic. The aim of my study was not to determine absolute levels of genetic variation, but rather compare levels of variation among sites. Therefore, I selected 41 polymorphic and easily scored A F L P markers from a total of approximately 900 bands in the three A F L P band profiles. The 41 markers ranged in size from 139 - 298 bp (Table 3.2). A F L P fragments were separated on 6% polyacrylamide denaturing gels and scored using a standard M13 sequence. Microsatellite allele frequencies and observed heterozygosity were calculated within each sub-population.  Each locus was tested for deviations from Hardy-  Weinberg expectations within sub-populations using GENEPOP Rousset 1995).  3.1D (Raymond &  Discrete genotypes are not generated in A F L P band profiles,  therefore each A F L P marker was treated as a single locus with two alleles: one coding for the "presence" of a band, the other being an "absence" allele. A F L P s are considered dominant markers, such that one cannot distinguish between dominant homozygotes and heterozygotes (both are expressed as the "presence" phenotype). A F L P allele frequencies and heterozygosity were estimated with TFPGA  1.0 (Miller  1997) from the proportion of recessive homozygotes ("absence" phenotype) as in Weir (1990). Pair-wise comparisons of both microsatellite and A F L P markers within sub-populations were also used to test for linkage disequilibrium with  GENEPOP  3.1 D (Raymond & Rousset 1995). Microsatellite allelic richness (number of alleles) and the percentage of polymorphic A F L P markers (PPM) were calculated within each sub-population. Generally, the percentage of polymorphic loci is calculated under the 9 5 % criterion,  48  (i.e. excluding alleles with frequency < 0.05) in order to focus on loci where variation is common (Hartl & Clark, 1997). However, I calculated microsatellite richness and PPM without a frequency criterion in order to detect the presence or absence of rare alleles. 3.2.3  COMPARISONS OF GENETIC VARIA TION AMONG FOREST TYPES  I used Spearman's rank correlation coefficient (r ) to test whether genetic s  variation (microsatellite allelic richness and AFLP PPM) was correlated with sample size, the estimated larval density within sites, or four spatial attributes of the sites: elevation, northing and easting (Universal Transverse Mercator co-ordinates), and sample reach length. One-way Analyses of Variance (ANOVAs) were performed to test for differences in mean microsatellite allelic richness, mean AFLP PPM and mean heterozygosity (microsatellites and AFLPs) among three forest types: oldgrowth, second-growth and recently clear-cut. Microsatellite allelic richness, AFLP PPM, and heterozygosity estimated from both marker classes were also regressed against stand age. The two old-growth sites (NesF4 and NesF5) are clustered within the same drainage (Nesakwatch Creek), while two recently clear-cut sites are clustered within the Foley Creek drainage (although all sub-populations are in different headwater streams) (Fig. 3.2). It could be argued that these sample sites represent pseudoreplication within the Nesakwatch and Foley drainages, which could bias my results towards finding differences in genetic variation among forest types. Therefore, I repeated the linear regressions and comparisons of mean microsatellite allelic richness and AFLP PPM between forested (old-growth and second-growth sites combined) and recently clear-cut sites. Here, the Nesakwatch sites were pooled as one site, as were the two Foley sites. 3.2.4  DETECTION OF RECENT BOTTLENECKS  When populations are in mutation-drift balance, they tend to have many rare alleles with fewer intermediate and common alleles that generally conform to an Lshaped frequency distribution (Luikart 1997). Populations that undergo bottlenecks are expected to lose rare alleles such that the frequency distribution of the remaining  49  alleles is shifted towards greater frequencies than expected from a typical L-shaped frequency distribution (Luikart 1997; Luikart et al. 1998). Microsatellite and AFLP data were used to test for significant shifts in allele frequencies, which may signal the recent loss of rare alleles. This test, the mode-shift indicator, is part of a genetic analysis program called BOTTLENECK  (Cornuet & Luikart 1996; Luikart 1997). The  program assigns alleles to one often frequency classes (i.e. 0.001 - 0.100; 0.1001 0.200; ...; 0.901 - 1.000) and identifies samples with shifted allele frequencies as those in which fewer alleles are found in the lowest frequency class than in one or more of the intermediate classes (0.1001 - 0.9). The mode-shift indicator is not a proper statistical test in that the type I error rate varies with sample size and number of loci, however, this qualitative graphical method has been successfully used to identify recently bottlenecked populations (Luikart 1997; Luikart et al. 1998).  3.3 3.3.1  RESULTS MICROSATELLITE AND AFLP ALLELE FREQUENCIES AND HETEROZYGOSITY  Six microsatellite alleles were detected among the eight sub-populations in BC at DteA (189, 221, 223, 225, 229, 231), while only 3 alleles were detected at Dte8 (176, 178, 184) and Dte11 (90, 96, 98). No polymorphisms were detected in  the random sub-samples of any of the eight sub-populations at the remaining 11 loci (Dtel, Dte2, Dte3, Dte5, Dte6, Dte12, Dte13, Dte14, Dte16, Dte18 and Dte19). Microsatellite allele frequencies at DteA, Dte8 and Dtel 1 varied from 0-1 within the eight sub-populations and are listed in Table 3.3 along with observed heterozygosity (H ) estimated over all three loci. The common allele of each microsatellite locus 0  within all sub-populations had a frequency > 0.7. Of the 41 AFLP markers that I selected for this study, I found that one (CTA276) was in fact fixed for the dominant allele across all sub-populations. Estimated frequencies of the dominant alleles of the 41 AFLP markers are summarised in Table 3.4. The AFLP PPM and estimated heterozygosity within subpopulations are summarised in Table 3.5.  50  The three microsatellite loci conformed to Hardy-Weinberg expectations within sub-populations (0.0252 < p < 1.000). At a=0.05, the critical p-value was set a priori  to 0.0018 using a Bonferroni correction for 27 simultaneous multiple  comparisons (Miller 1991). AFLP marker allele frequencies cannot be directly counted, therefore one cannot test for deviations from Hardy-Weinberg proportions. However, because the three microsatellite loci did not violate this assumption, I assumed that the 41 AFLP loci also conformed to Hardy-Weinberg expectations. Pair-wise comparisons among Dfe4, Dfe8 and Dte11 within each subpopulation were used to test for linkage disequilibrium  V3.1 D, Raymond  (GENEPOP  & Rousset 1995). Using the same critical p-value for 27 multiple comparisons, I found no evidence of linkage disequilibrium among the three microsatellite loci (0.0255 < p < 1.000). Therefore, I did not exclude any of the microsatellite loci from subsequent analyses due to linkage disequilibrium. However, pair-wise tests for linkage disequilibrium among AFLP markers across all pooled sub-populations revealed three significantly linked pairs (820 simultaneous comparisons, p-critical = 0.00006): 1  CTA 199 / CTA 200 (p < 0.00001)  2  CTA 200 / CTA 274 (p < 0.00001)  3  CTA 246 / CTA 267 (p < 0.00001)  Therefore I excluded AFLP markers CTA 200 and CTA 267 from all subsequent analyses, reducing the total number of AFLP markers to 39. 3.3.2  COMPARISONS OF GENETIC VARIATION AMONG FOREST TYPES  The amount of genetic variation at microsatellite and AFLP loci was not correlated with the estimated density of larvae in streams, the sample size, elevation, easting and northing (Universal Transverse Mercator co-ordinates), or stream length sampled (Spearman's rank correlation: n=8; I r | always < 0.590; p s  always > 0.123) (Table 3.6). Mean microsatellite allelic richness was greater in oldgrowth (NesF4 and NesF5) and second-growth (Cen23, TamSGI and VedB) sites than in recently clear-cut (FolQ, Foil2 and TamC) sites.  The total number of  51  microsatellite alleles detected was 8 in both old growth sites, 5-8 in second growth sites, and 3-5 in clear-cut sites with means (± standard error) of 2.66 (± 0.000), 1.99 (± 0.333) and 1.33 (± = 0.192), respectively (Fig. 3.5A).  A one-way ANOVA  revealed a significant difference in the mean number of microsatellite alleles among the three forest types (p=0.046). Old-growth sites also had more polymorphic AFLP markers (mean = 92.7% ± 4.88) than second-growth sites (mean = 80.5% ± 8.57) and recently clear-cut sites (mean = 64.2% ± 5.33) (Fig. 3.5B), however the differences in mean PPM among the three forest types were not statistically significant (p=0.097). Heterozygosity, calculated over three microsatellite loci was greater in old-growth sites (mean = 0.197 ± 0.047) than in second-growth sites (mean = 0.119 ± 0.027) and recently clear-cut sites (mean = 0.080 ± 0.042). Similarly, heterozygosity estimated from 39 AFLP markers was greater in old-growth sites (mean = 0.383 ± 0.003) than in second-growth (mean = 0.317 ± 0.033) and recently clear-cut sites (mean = 0.252 ± 0.030).  However, one-way ANOVAs  revealed no significant differences in mean microsatellite heterozygosity (p=0.212) or estimated AFLP heterozygosity (p=0.082) among the three forest types. Mean microsatellite allelic richness was significantly correlated with logtransformed stand age (simple linear regression, one-tailed test, 1^=0.59, p<0.025) (Fig. 3.6A). I also found a significant correlation between the logarithm of stand age and the PPM of AFLPs (simple linear regression, one-tailed test, 1^=0.72, p<0.005) (Fig. 3.6B). The regressions between the logarithm of stand age against AFLP and microsatellite heterozygosity were statistically and non-statistically significant, respectively (simple linear regressions, one-tailed tests; AFLPs: r = 0.75, p<0.0025; 2  Microsatellites: r = 0.373, 0.05<p<0.10). 2  When the Nesakwatch and Foley sites were pooled into their respective groups, microsatellite and AFLP variation remained higher in the forested sites (n=4) than in clear-cut sites (n=2). Mean microsatellite variation was 2.17 (± 0.289) in the forested sites and 1.41 (± 0.252) in the recently clear-cut sites.  The mean  percentages of polymorphic AFLP markers were 83.53 (±6.781) in the forested sites and 61.58 (±7.930) in the recently clear-cut sites. One-way ANOVAs, however,  52  revealed that the differences were no longer significant (microsatellite; p=0.177: AFLP; p=0.123).  Similarly, mean differences in heterozygosity were greater in  forested sites that in recently clear-cut sites, but the differences were no longer significant (microsatellite; p=0.249: AFLP; p=0.136). Microsatellite and AFLP variation also increased with the logarithm of stand age when the Nesakwatch and Foley sites were pooled into their respective groups. However, the slope of the regression lines remained significantly different >0 only with the AFLP markers (simple linear regression, one-tailed tests; microsatellite: r =0.44, 0.05<p<0.1; AFLP: r =0.70, p<0.025). 2  2  3.3.3 SHIFTS IN ALLELE FREQUENCY DISTRIBUTIONS  Using microsatellite data, the mode-shift test (Cornuet & Luikart 1996) indicated a significant shift in allele frequencies (biased towards intermediate frequencies) in FolQ only (Fig. 3.7).  The allele frequencies in all other sub-  populations did not have more alleles in intermediate frequency classes than expected from similarly sized populations in drift-mutation balance. Similar modeshift tests using AFLP markers did not signal any qualitative allele frequency shifts.  3.4  DISCUSSION  Since Farr's (1989) Pacific Giant Salamander status report was submitted to COSEWIC, 133 new cut-blocks, extending over 1500 ha, have fragmented the known range of Pacific Giant Salamanders in BC (Ferguson & Johnston 1999). The Ministry of Forests (MOF, Chilliwack Forest District, Rosedale, BC) has prescribed an additional 549 ha of clear-cuts and 444 ha of selective harvest and commercial thinning in the present five-year plan (1998-2002). A year subsequent to this genetic survey (October 1999), TamSGI was clear-cut without a buffer strip. NesF4, one of the three old-growth sites surveyed, was also scheduled for clear-cut in 1998-99. Most of the second-growth stands within the known range of Pacific Giant Salamanders in BC will be available for harvest in 2013 (Guy Fried, Ministry of Forestry, Chilliwack Forest District, Rosedale, BC, pers. comm.).  53  The effects of clear-cut harvesting on the persistence and colonisation ability of larval and terrestrial Pacific Giant Salamanders are poorly understood. Logging activities are thought to degrade salamander habitat quality by altering the microclimate, particularly moisture and temperature regimes (Spotila 1972; Maiorana 1978; Jaeger 1980; Welsh 1990).  Higher summer temperatures and reduced  moisture in clear-cuts, perhaps not lethal to terrestrial salamanders, appear to act as a partial barrier to movements by terrestrial Pacific Giant Salamanders and induce changes in behaviour that are consistent with heat and water-stress (Johnston 1998; Johnston & Frid, unpublished manuscript).  Canopy removal increases stream  temperatures (Beschta et al. 1987), which may also be detrimental to larvae if they become heat-stressed. Pacific Giant Salamanders may be prone to heat-stress at temperatures as low as 20-22 C (Ferguson 1998; pers. obs.). Forest harvesting may also increase aquatic larval mortality through increased sedimentation of nursery streams. Sedimentation may increase habitat loss by filling in cracks and interstitial spaces occupied by larval salamanders (WE Neill, pers. comm.). Yearly density estimates from 1995-1999 (WE Neill & JS Richardson, unpublished data) of larval  D. tenebrosus  in TamC, clear-cut in 1994-95, suggest  that this sub-population may have passed through a bottleneck during or subsequent to the harvest event. After the site was logged in 1994, no larvae were found during an intensive search of the stream (WE Neill, pers. comm.). Since 1995, the number of larvae detected and marked in TamC has steadily increased from 27 to 93 in 1999, such that estimated larval density in TamC is now similar to the densities of larvae in second-growth streams (Neill & Richardson 1997). Unfortunately, it is difficult to infer the impacts of clear-cut logging on the persistence and colonisation potential of  D. tenebrosus  in BC using larval density,  growth rates, and survival estimates as indices alone.  Comparisons of these  parameters among different forest types (old-growth, second-growth, and clear-cut) are confounded by at least one spatial attribute of the sample sites: elevation. Larval  D.  tenebrosus  density is negatively correlated with elevation (Neill &  Richardson 1997). Furthermore, forest type is also correlated with elevation: most lower elevation stands were cleared in the 1930s and 1940s (McCombs &  54  Chittenden 1990) and are now second-growth stands; middle elevations are currently being cleared, while old-growth sites are generally restricted to higher elevations. In order to tease apart these confounding effects, I compared levels of genetic variation among forest types using microsatellite and AFLP markers. Before carrying out these comparisons, I first verified that the levels of genetic variation (both microsatellite and AFLP) were not correlated with sample size, larval density, easting, northing, elevation, or reach length. 3.4.1  COMPARISONS OF GENETIC VARIATION AMONG FOREST TYPES  Overall, microsatellite variation was considerably lower in the Chilliwack River Valley than populations of  D. tenebrosus  in Oregon and California (Chapter 2).  Within the Chilliwack Valley, both microsatellite and AFLP variation were significantly correlated with stand age. Approximately 45-60% of the differences in microsatellite richness was explained by stand age. Similarly, 70% of the variation in AFLP PPM within sub-populations was explained by stand age.  Younger stands have less  variation than older stands. Under the assumption that sub-populations in recently clear-cut sites had pre-harvest levels of variation that were similar to surrounding forested sites, these results suggest that clear-cutting negatively impacts aquatic and / or terrestrial Pacific Giant Salamanders, causing them to pass through genetic, and possibly demographic bottlenecks.  Other reasons for considering this  assumption are two-fold. First, estimates of pair-wise F (a measure of population st  differentiation, see Hartl & Clark 1997) among forested sites suggest that D. tenebrosus  has higher dispersal abilities than previously anticipated. Within the  Chilliwack River Valley, no significant isolation by distance was detected (Chapter 4). Moreover, F t estimated over all eight sub-populations was slight to moderate s  (Chapter 4). Low F t values are often interpreted as resulting from high gene flow s  among sub-populations (Waples 1998; Bohonak 1999). With evidence of slight to moderate gene flow among sub-populations in the Chilliwack Valley and no correlation between extant levels of genetic variation and spatial characteristics of the sites, there is no  a priori  reason to anticipate local differences in the amount of  genetic variation among similarly sized sub-populations in BC.  55  The extant level of genetic variation reflects both the phylogenetic heritage of populations and the population-level processes that regulate the structure of genetic variation within and among populations (Avise 1994). The two old-growth sites, NesF4 and NesF5, are clustered within the same drainage (Nesakwatch Creek), while two recently clear-cut sites are nested within the Foley Creek drainage (Fig. 3.2). When I repeated the regressions and comparisons of mean microsatellite allelic richness and AFLP PPM with pooled Nesakwatch and pooled Foley sites, the trends were similar to those found when all sites were treated as independent samples, and remained statistically significant with the AFLP data. It seems unlikely, therefore, that the ANOVA and regression p-values were artificially lowered as a result of non-independence among the sample sites. Levels of genetic variation between old-growth sites were more similar to second-growth sites than to recently clear-cut sites. Without knowledge of the levels of genetic variation prior and subsequent to logging in the second-growth sites, I cannot determine whether variation was depressed (as it appears to be in recent clear-cuts) and subsequently replenished through population-level processes, or whether the forest harvesting had no impact on genetic variation (which could have remained constant throughout the logging and re-growth of the sites). However, the pattern of genetic variation among forest types is consistent with my hypothesis that disturbed sub-populations recover lost genetic variation over relatively short periods (perhaps as little as 60 years) through migration (see also Chapter 4). 3.4.2  SHIFTS IN THE DISTRIBUTION OF ALLELE FREQUENCIES  Theoretical and empirical evidence suggest that populations that are in mutation-drift equilibrium typically have L-shaped distributions of allele frequencies, with few common alleles and many more rare alleles (Hartl & Clark 1997; Luikart 1997). However, recently bottlenecked populations may have lost rare alleles, while retaining alleles at intermediate frequencies with relatively high heterozygosity, causing a characteristic temporary shift in the distribution of allele frequencies at neutral loci (Luikart 1997). Therefore, a qualitative test of the frequency distribution of alleles can be used to detect recently bottlenecked populations.  Tests performed with 56  microsatellite data indicated a significant shift in the allele frequencies of FolQ (clearcut site), while mode shift tests performed with AFLP data did not indicate similar shifts in any of the sub-populations. However, the mode-shift indicator has a type-l error of >0.2, particularly with fewer than 5 polymorphic loci (Luikart 1997). Moreover, Luikart (1997) cautions against its use with samples of fewer than 30 individuals. The mode-shift test may be an inappropriate test for use with my AFLP data because samples sizes ranged from 20-28. Moreover, with a high Type I error rate, it is unclear whether the mode-shift indicator reliably flagged FolQ as a recently bottlenecked population, or whether this shift in allele frequencies occurred by chance. Cornuet & Luikart (1996) suggest that the mode-shift indicator be used in conjunction with quantitative tests that have less variable Type I error rates. They proposed that in addition to comparing traditional measures of genetic variation (allelic richness, heterozygosity, and percent polymorphic loci), one should also use three other tests for detecting bottlenecks: a sign test, a standardised differences test, and a Wilcoxon sign-rank test. These tests detect whether a population has a greater observed heterozygosity than expected from a theoretical population with the same number of alleles that is in drift-mutation balance (Luikart 1997). Unfortunately, the standardised differences test requires a minimum of 20 polymorphic loci, and the sign test and Wilcoxon sign-rank test may require 10-20 polymorphic loci with 4-5 alleles in order to achieve a power of 0.8 (Cornuet & Luikart 1996; Luikart 1997). Although 41 AFLP markers were screened in each of the sub-populations, they are inappropriate for these three latter tests because allele frequencies are inferred from the banding patterns, rather than directly observed. Given the constraint of insufficient numbers of polymorphic microsatellite loci, the most appropriate tests for detecting recent genetic bottlenecks in my study of BC Pacific Giant Salamanders were the comparisons of levels of allelic variation and heterozygosity among forest types. 3.4.3  IMPLICATIONS FOR MANAGEMENT OF PACIFIC GIANT SALAMANDERS IN BC  It is important to detect recent bottlenecks and monitor populations of conservation concern for changes in genetic variation and N , in order to assess the e  57  impacts of deterministic factors (e.g. anthropogenic habitat loss and fragmentation) on them, and mitigate the factors most likely to negatively impact their long-term persistence (Luikart 1997). Evidence of forest harvest induced genetic bottlenecks presented in this study of  D. tenebrosus  in BC suggests that local sub-populations  may be also passing through demographic bottlenecks and as such, may be susceptible to local disturbance and possibly extirpation. This may be of particular concern because the sub-populations that were sampled in this study are probably among the most densely populated streams and, as such, are most well buffered against fluctuations in population size. Numerous surveys of larval  D.  tenebrosus  throughout the Chilliwack Valley have shown that most streams appear to be inhabited by few or no Pacific Giant Salamander larvae (Farr 1985, 1989; Haycock 1991; Richardson & Neill 1995; pers. obs.). Two sites that were forested during my genetic survey are currently, or will soon be, clear-cut. This presents a future opportunity for monitoring the impacts of clear-cutting on census size and genetic variation, as well as testing the hypothesis that current forest practices in BC cause local sub-populations of Pacific Giant Salamanders to pass through population bottlenecks.  58  CO ,_ C 0 0  E  a. 3  CD  JD  E  ZJ  c  "CO -I—•  o  0  to  CL  E  CO CO  oT  1  CO CJ) CJ)  0  E  ZJ  CD CD  o  CO  to  E  TJ C CO "fr LL to  0  CO  c  TJ CD  b  CO  CO  TJ CO 0 CO 0 •*—' ZJ  o o  M—  0_ CO  E  to CD CD  CO  to 0 > F 'cz H  CD  0  CO CD CO 0) CL  E  CO  CO  TJ  0 O 0  CS  I-  b  O  o  0  to  CD  0  CO  CL  E  (0  to  o CO 0 0 cz XI  0 0  CD M—  o  tz (0  •a CD  c 0  o CO 0  0  LU CO  cz CD c 0 CO o C D fZ v_ c CO -t—' Cfl 0 0 CD CO ~a C L •+-• c C O CO cfl o O "D •a to c cfl TJ CO 0 cf c cz o C L o "•^ CO c CO CO ZJ > E 0 O CD 0 O 0 CO 0 0 co c ai" TJ c CO CO o CO g CO CO 0 •a E TJ o to 0 o -<—» CO CO o CD CD DC CO CO o o -2 CD '.£= 0 CO a ^ CO Cfl 0 b ZJ _ l 0 CD c ^ lo C O CD c to TJ 0 2 c o 'cfl c CD £2 a) LL TJ T3 CO £ "CO >CO o 0 CO E as c cfl CO 5 CO -I—»  LO LL  0  JD -*—* CO ZJ  E  TJ c CO  o to Jz c 0 •4—"  0  1 -*—'  TJ CZ CO  CJ  'c  £  ^ to -+—' to  5 TJ  CO  LL  o -+—*  to  'c T3 C CO CO to _C0  o  •4—'  to  0  CO  c  to TJ cz CO o CL CO to 0 co o o 0 CO 0 -*—•  o  CZ  0 1_  CO  CO to CO  c 0 0  E  ZJ CZ CD  c 0  CL  o  0  CO  0 XI  E  £  to  0  CO 3  CO CO  C  to (0 0 to  ZJ JD C  oiS LL O  o cz c  ZJ CO  ZJ  CQ  o  0 >  0 o  TJ  CO  o  O O  CD  CO  -£  E  Sto  CO  2  cz  o  TJ CZ  0  JD ZJ  CO X  0 0  o o 0  to  c o Cfl  O k_  c 0  TJ  CD  JD CO  TJ  CO  II  O  co CD to 0  to CO  \  JD E ZJ  c  CL  CO  E  c o  0  JD  E  ZJ  c  CD  CZ  'cz 0  CL  o o  O  ZJ^ O) O) CD  'c o o  i -  o  ^  to CO  0 >^ o CO .>»  CO CJ) c CD CM 3  o  CO CO  CO CO CO CO CO CM  CD (0 <  T—  T—  o O O O o o o O O O O O o o X X X X X X X X  CM CM CM CM CM CM CM CM CD CD CD CJ) CD O) CD CD  CDOOCDCDOCDCD  C0OOOOOC0C0  o o  5  §  LO LO ^ O LO CM CM i CD CO CO CD CD A A CO LO )  c o CS  > _  o o o o o o o o o  CO CO  -1—  o CD o o o  00 00 CO CD  z -z. z: o o o o o LO o o o LO 00 LO LO CO CO CO o CO CO CO  CD  D CQ C TJ CO 0 > 0  c  CO  TJ  TJ  I  CO  o o 0  TJ 0 TJ _ZJ  CD  ZJ  O 0 o  CZ  -t—' CCZD to 'cz 0 0 CO C L o JZ O 0  cfl 0  »  0  JD TJ CO JZ •4—* CO to  ZJ CO  0  TJ  E  0  O  0  O Z  TJ  I-  c o 'Z3 CS u o  V CD CS  "fr  1—  LO LO CO CO "fr "fr LO LO  LU LU LULU LU LU LU UJ o o o o o o CO o LO o o o o o 00 o CO CM CO CD CM o LO 1^ CO h- LO CO CO o o o o o CD 00 hCO CO CO CO CO LO LO LO Tfr  T—  o o  c  0 0  o  ro  -z. o LO o o o o CM  CO CM "fr "fr LO LO LO LO LO LO 1—  Dr  TJ  O  to  CO CO  0  j |  J ;  F  J= >? CO CO ~ C 0) 03 00 to E  ZJ ~ CO  T=  E o 0 o o a> 0 co co CO O LL l i Z Z — I — I  <  JO 4-1  o CD CS C a> 0  cc _l  o o O O o o o o o O o LO o CD CM CM CM CM 00 CM i —  00 CD CO CO CD o LO CM LO LO CM CO  0  -4—•  CO  T  E  ~ O  "t  x o  1  -~ r  O o  ^_ CM T—  co co 00 CD LO  LO CO LO CO  1  "fr  CO CD CO  CL CL  CO  0  JD O  V 0 TJ .*= O (7) O  LO o C D "fr LL UJ CQ LL to CO E E TJ o 0 o o 0 0 CO CO 0 O LL LL H 1- > 00 CM CM CZ  Table 3-2 The product size of 41 AFLP markers scored using three pairs of AFLP primers (Gibco BRL). Primer Pair  Size of Markers Scored  EcoRI-ACC : Msel-CTA  179, 180, 186, 199, 200, 204, 205, 246, 248, 251, 267, 274, 276, 297, 298  EcoRI-ACC : Msel-CAG  224, 226, 228, 230, 240, 242, 244, 267, 269, 271  EcoRI-ACC : Msel-CAT  139, 140, 166, 167, 168, 194, 195, 196, 197, 199, 200, 202, 203, 240, 244, 247  60  Table 3-3 Microsatellite allele frequencies at three polymorphic microsatellite loci DteA, DteQ,  and Dte11, in eight sub-populations of  D. tenebrosus  in BC. Observed  heterozygosity, H , was calculated within sub-populations over all three loci. 0  Population  DteA Allele  Frequency  DteQ Allele  Frequency Dte11 Allele  Frequency H  Cen23  189 223 225 231  0.794 0.177 0.020 0.010  178  1.000  90 96 98  0.904 0.087 0.010  0.172  FolQ  189 225  0.709 0.290  178  1.000  90  1.000  0.139  Foil 2  189  1.000  178  1.000  90  1.000  0.000  NesF4  189 223  0.903 0.097  176 178 184  0.015 0.970 0.015  90 96 98  0.883 0.100 0.017  0.150  NesF5  189 223  0.737 0.263  176 178 184  0.035 0.930 0.035  90 96 98  0.927 0.049 0.024  0.244  TamC  189 221 229  0.820 0.154 0.025  178  1.000  90  1.000  0.102  Tarn SG1  189 223  0.890 0.109  178  1.000  90 96  0.944 0.056  0.085  VedB  189  1.000  178 184  0.912 0.088  90 96  0.952 0.048  0.100  0  61  Table 3-4 Estimated frequencies of the dominant AFLP alleles (coding for the "presence" of a band) within eight sub-populations of D.  tenebrosus  in BC.  Frequencies are also estimated across all sub-populations. The AFLP marker indicates the Mse\ primer used to amplify the fragment (all Mse\ primers were paired with the same Eco RI primer, see text), as well as the fragment size. AFLP Marker CAT CAT CAT CAT CAT CAT CAT CAT CAT CAT CAT CAT CAT CAT CAT CAT CAG CAG CAG CAG CAG CAG CAG CAG CAG CAG CTA CTA CTA CTA CTA CTA CTA CTA CTA CTA CTA CTA CTA CTA CTA  139 140 166 167 168 194 195 196 197 199 200 202 203 240 244 247 224 226 228 230 240 242 244 267 269 271 179 180 186 199 200 204 205 246 248 251 267 274 276 297 298  AM 0.405 0.695 0.483 0.695 0.598 0.405 0.405 0.054 0.685 0.400 0.269 0.224 0.489 0.639 0.842 0.889 0.600 0.700 0.365 0.245 0.700 0.258 0.207 0.788 0.584 0.440 0.131 0.522 0.756 0.230 0.546 0.278 0.624 0.506 0.189 0.734 0.639 0.366 1.000 0.792 0.115  Cen23 Fol12 FolQ NesF4 NesF5 TamC TamSGI VedB 0.333 0.529 0.570 0.728 0.615 0.306 0.491 0.019 0.615 0.362 0.333 0.423 0.280 0.570 0.728 1.000 0.660 0.723 0.380 0.266 0.660 0.123 0.145 1.000 0.660 0.240 0.139 0.667 0.808 0.423 0.423 0.230 0.728 0.491 0.362 0.808 0.808 0.456 1.000 1.000 0.139  0.184 1.000 0.592 1.000 1.000 0.592 0.293 0.000 1.000 0.293 0.184 0.000 0.592 1.000 1.000 1.000 0.293 0.592 0.293 0.355 0.423 0.423 0.293 0.592 0.423 0.500 0.065 0.441 1.000 0.339 0.209 0.171 0.750 0.339 0.171 1.000 0.567 0.209 1.000 1.000 0.250  0.236 1.000 0.236 0.711 1.000 0.592 0.293 0.043 1.000 0.500 0.134 0.043 0.711 1.000 1.000 1.000 0.456 0.657 0.358 0.160 0.580 0.314 0.160 0.657 0.515 0.515 0.051 0.553 1.000 0.258 0.408 0.106 0.553 0.408 0.106 1.000 0.452 0.368 1.000 1.000 0.134  0.544 1.000 0.423 0.711 0.711 0.460 0.592 0.000 0.711 0.423 0.388 0.323 0.592 0.796 0.796 1.000 0.631 0.631 0.631 0.293 0.787 0.436 0.262 0.787 0.574 0.574 0.293 0.608 0.804 0.191 1.000 0.481 0.520 0.481 0.321 0.660 0.723 0.350 1.000 0.804 0.101  0.293 0.592 0.460 0.646 0.796 0.158 0.500 0.043 0.460 0.388 0.323 0.388 0.236 0.460 0.796 0.796 0.489 0.583 0.248 0.192 0.705 0.166 0.091 0.792 0.534 0.375 0.175 0.368 0.654 0.225 0.434 0.200 0.553 0.600 0.225 0.800 0.553 0.307 1.000 0.800 0.084  0.537 1.000 0.466 1.000 1.000 0.537 0.402 0.155 0.733 0.622 0.036 0.155 1.000 1.000 1.000 1.000 1.000 1.000 0.484 0.184 1.000 0.423 0.184 1.000 0.742 0.635 0.106 0.635 1.000 0.144 1.000 1.000 1.000 1.000 0.034 1.000 1.000 0.484 1.000 0.742 0.034  0.684 0.776 0.613 0.776 0.452 0.684 0.368 0.025 1.000 0.408 0.329 0.106 0.776 0.613 1.000 1.000 0.608 0.723 0.168 0.216 0.723 0.123 0.216 0.723 0.380 0.445 0.071 0.436 0.787 0.047 0.787 0.147 0.699 0.478 0.121 0.631 0.478 0.360 1.000 0.787 0.000  0.423 1.000 0.544 1.000 0.592 0.423 0.264 0.043 1.000 0.544 0.209 0.158 1.000 1.000 1.000 1.000 1.000 1.000 0.423 0.293 1.000 0.264 0.388 1.000 1.000 0.544 0.134 0.646 1.000 0.236 0.646 0.388 0.592 0.711 0.134 1.000 0.796 0.460 1.000 0.796 0.209  62  Table 3-5 The percent polymorphic AFLP markers (PPM) within eight subpopulations of D.  tenebrosus  in BC. Heterozygosity was estimated from the  proportion of recessive alleles (see text). Population Cen23 FolQ Foil 2 NesF4 NesF5 TamC TamSGI VedB  Sample Size 28 24 20 27 25 24 24 24  (%)  Estimated Heterozygosity  90.24 70.73 68.29 87.8 97.56 53.65 87.8 63.41  0.367 0.285 0.279 0.380 0.386 0.192 0.332 0.254  PPM  Table 3-6 Spearman's Rank Correlation Coefficient and p-values for correlations between levels of genetic variation, microsatellite allelic richness and AFLP PPM, and six spatial and sampling factors (n=8). Factor  Larval Density Sample Size Elevation Easting Northing Reach Length  Spearman's Rs p-value (Microsatellite allelic richness) 0.077 -0.111 0.417 0.335 -0.590 -0.166  0.856 0.794 0.304 0.417 0.123 0.695  Spearman's Rs (AFLP PPM)  p-value  0.055 0.240 0.491 0.551 0.108 0.576  0.898 0.568 0.217 0.157 0.799 0.135  64  Figure 3-1 The geographic distribution of Dicamptodon  tenebrosus  ranges along  the Pacific Coast from northern California to the extreme south-western corner of British Columbia (light shading). The Chilliwack River Valley is adjacent to the Canada - United States border (indicated by a white arrow). (Range from Daugherty et al. 1983 and Good 1989).  65  "a  0 >  CD  2 CL  CO  o  CD ^_  CT  CO  CD  CC ZJ  tx  E  CT  c  CD  CD  co  c  O 00  to" 0 to  CD  JZ  CM  "5 LL IS-  0 CT) "a  "CC >  c  o o 0 to 0 *cO o  CD >  2 o LL CO  o  CC  5 LO LL co CO Z  CT £Z  JZ  CJ  CO  CD  O E  12  CL  JZ  to  E cc  o  CO CO  CT) CT o  s  CO  g  c  ro o o  CD  E  CN  JZ  CO  >  c  0  o LZ  o CO  CJ  o 0  ICM i CO  0 l_ ZJ CO  CO  to _0 o  O "D C 0 CT) 0  0  c O  o  0  ro CT C  to o  CO  CO  CO  0  CT)  0  0 0  0 0  >, 0  0 o  ro >  LL  CO  0 CO  to ZJ  ±ZJ  O o to to 0 £  c o 0  o  ro 5  o 0  o o  CO  CO CT  0 >  ro  0 -•—•  x  0 ^ o ro  o  O 0  to ro ZJ  J2 C 0  1—  0 H—  CT  0 0  CT 0  E ro  ro  E  co c\f  ro  CO  -4—'  0  o  0  ro 0 o  CO  0 0  CO  ZJ  o JZ  ro  "D 0 ro o o  to 0 to -*—•  © 0  O CT  to co LU  to ro 0  o  E •  LO  JZ  CD  CO  E  CL  E  ro to CD  0 0  CD CO  0 0  c  to  'ro — i "D 0  =3 JZ CL CO  ro to 0  6 o to  E  < to 0  CT) CO  ro c 0 'ro > 2CJ0 •a c  .E  0  CO  CT) CD  DQ  "a  o c —  CD  0 ro o 0  CL  to 0  -O  CD  0  C  ro  CO co  O  o  CO CD co  1—  0  CT)  LG  0 0  CT O  c  CC  CO  0 >  CO  1—  CO  OJ  o  0  CC CO CD  co  DC  to 0  CD  66  o No Genetic Sample  200  • Genetic Sample  180 160 E E D) C  ,0) CO  o  140 120 100 80 60 40  o o o o •o o c?  GO  o o  °o  O  o  o o  am •  o  •  o O  • °J«O o  0Q•  o»  o „• o •  oS'cV  • o  .1 o.o 1 ° ••*oo  o  oo^ o o  o  oo  o  oo §  8  •  o  o 'o  %.  'cf  L  o*  o°° «o  <'  > o*b  Q  o  <f o •o  »8.°  20 0 50  100  150  200  Location Within Stream Reach (m)  Figure 3-3 The size distribution and location of 258 D. tenebrosus individuals captured in FolQ during the summer of 1998. Closed circles represent individuals that were marked during a mark-recapture study conducted by WE Neill & JS Richardson (UBC, unpublished data) and tail-clipped for this genetic study (n=65). Open circles (n=188) are individuals that were marked only.  67  Figure 3-4 Autoradiograph of a polyacrylamide denaturing gel used to size amplified fragment length polymorphisms (AFLPs). The four lanes labelled GATC represent the standard M13 sequence used to size the bands. Lanes 1-5 represent individuals from TamC (clear-cut), lanes 6,8 and 9 are from NesF4 (old-growth), lanes 7, 12 and 13 are from NesF5 (old-growth), and 10 and 11 are from Cen23 (second-growth). Arrows indicate markers that are monomorphic throughout the eight sub-populations of D.  tenebrosus  in BC.  68  CO CO  CD  cz  JZ  o  ix  CD +^  3 2.5 2  1.5 ~CD -t—• cd CO  2  1  o  0.5  c  0  CC CD  LL < Q_ Q.  c  cd  CD  Old Growth  Second Growth  Old Growth  Second Growth  Clear-cut  100 90 80 70 60 50 40 H 30 20 10 0  Clear-cut  Figure 3-5 Comparisons of genetic variation among forest types: (top) mean microsatellite allelic richness calculated over three polymorphic microsatellite loci and (bottom) mean percent polymorphic markers (PPM) calculated over 39 AFLP markers. Error bars represent the standard error of the mean (Old Growth N=2; Second Growth N=3; Clear-cut N=3).  69  2.95 -j CD  2.45 -  "CD  CO to 1.95 to to o CD  o c o ir cz CO CD  1.45 -  0.95 0.45 -0.05 1  2  Log Stand Age  Figure 3-6 Mean microsatellite allelic richness (A) and percent polymorphic AFLP markers ( AFLP PPM) (B) increase with the logarithm of stand age (in years). Squares denote recently clear-cut sites, triangles represent second-growth sites, and circles represent old-growth sites. The diamond in (A) represents two oldgrowth sites at the same co-ordinates. The solid lines represent linear regression lines. The equations of the lines are (A) y=0.71+0.85x and (B) y=19.23+48.63.  70  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  Allele Frequency I Dte4  2.5  • Dte8  H Dte11  2 1.5 1 0.5 0  1 0.1  0.2  1 0.3  B  1 0.4  1 0.5  1 0.6  1 0.7  0.8  1— 0.9  1  Allele Frequency  Figure 3-7 The distribution of microsatellite allele frequencies in two subpopulations of D.  tenebrosus:  (A) FolQ and (B) NesF4. FolQ has a qualitative shift  in its microsatellite allele frequency distribution, while NesF4 conforms more closely to a typical L-shaped distribution. N refers to the number of alleles with a given frequency (frequencies were rounded to the nearest multiple of 0.1).  71  4  CHAPTER 4  T H E POPULATION GENETIC STRUCTURE AND  COLONISATION POTENTIAL O F P A C I F I C GLANT S A L A M A N D E R S IN T H E C H I L L I W A C K R I V E R V A L L E Y , BRITISH C O L U M B I A .  4.1  INTRODUCTION  Movement patterns are fundamental elements of a species' ecology. Dispersal distance, the distance an individual moves from its natal site, determines the spatial scale over which a population's self-recruitment occurs and is related to the spatial scale over which populations become genetically differentiated (Slatkin 1985). Patterns of genetic differentiation among neutral loci are determined in part by the relative interactions of genetic drift and migration (Allendorf 1983).  In  fragmented landscapes, the connectivity among sub-populations (through dispersal and gene flow) plays an important role in stabilising small sub-populations that are more susceptible to demographic and genetic stochasticity than larger, continuous populations (Primack 1993). Global declines in amphibian populations have been linked to habitat loss and fragmentation (Blaustein & Wake 1995). Amphibians are particularly susceptible to habitat alterations because they are intolerant of micro-climatic changes including increased air, soil, and water temperatures, and reduced soil and air humidity (Spotila 1972; Jaeger 1980; Welsh 1990). Moreover, locally extirpated amphibian populations may be less likely to become re-established through migration from surrounding sub-populations because amphibians are generally poor dispersers (Corn & Bury 1989; Petranka et al. 1993; Blaustein et al. 1994, but see Berven & Grudzien 1990). The Pacific Giant Salamander is currently recognised as vulnerable by COSEWIC and is red-listed by the BC Ministry of Environment, Lands and Parks because current forest practices are perceived to be threatening its persistence in Canada through habitat loss and fragmentation.  The northern periphery of this  species' range extends into the Chilliwack River Valley and adjacent parts of the lower Fraser River Valley at the extreme south-western corner of BC. Pacific Giant  72  Salamanders breed in small, fast flowing, cool mountain streams and are generally considered an obligate old-growth species (Nussbaum et al. 1983). Larvae spend 24 years in streams before transforming into terrestrial juveniles, or maturing as aquatic neotenes (adults that retain larval characteristics) (Nussbaum et al. 1983). Casual observations in BC of lower larval densities immediately subsequent to clearcutting (WE Neill, pers. comm.), and lower genetic variation within recent clear-cuts relative to forested sites (Chapter 3) suggest that current forest harvest practices negatively impact local sub-populations and may induce population bottlenecks. Current conservation efforts in BC are directed at understanding the impacts of clear-cut logging on the population structure, movement patterns and colonisation potential of larval and terrestrial Pacific Giant Salamanders. Although recent studies have shown that larvae and neotenes exhibit extremely high site fidelity within streams (Neill & Richardson 1997; Ferguson 1998; Richardson & Neill 1998) there is little information regarding the dispersal patterns or population structure of this species. Inferences about the dispersal distance and colonisation potential of Pacific Giant Salamanders, previously considered to be extremely low, have largely been drawn from mark-recapture studies of larvae (Neill & Richardson 1997; Ferguson 1998; Neill 1998), successful stocking attempts of barren streams (WE Neill, pers. comm.) and radio-telemetry studies of terrestrial adults (Johnston 1998; Johnston & Frid, unpublished manuscript). Successful stocking experiments of previously barren streams suggested that the establishment of sub-populations in streams was dispersal-limited (WE Neill, pers. comm.). In her experimental removal study, Ferguson (1998) estimated that re-colonisation of barren stream reaches 25-40 m in length by larvae in adjacent reaches alone could take several decades. Johnston (1998) also concluded that terrestrial Pacific Giant Salamanders were poor dispersers, with a probability of less than 1/1000 of moving > 500 m within a two-month period.  However, the latter two studies may  underestimate the dispersal distance and colonisation potential of Pacific Giant Salamanders because they have not considered the movement patterns of recently transformed-juveniles, which among amphibians, typically represent the dispersing stage (Berven & Grudzien 1990; Duellman & Trueb, 1994). Terrestrial Pacific Giant  73  Salamanders are extremely secretive in nature, spending most of their time burrowed underground (Johnston 1998). Therefore estimates of juvenile dispersal through mark-recapture techniques are not feasible. Moreover, because juveniles begin to transform when they are as small as -90 mm in total length (Nussbaum et al. 1983; Haycock 1991; pers. obs.), implantation of radio-transmitters is not currently possible. Therefore, I indirectly inferred the level of population structure and colonisation potential of Pacific Giant Salamanders in BC using three polymorphic microsatellite loci and 39 amplified fragment length polymorphism (AFLP) markers. My objectives were to address three related questions: 1  To what extent are sub-populations genetically structured within the  Chilliwack River Valley: among streams, among drainages, or are they not significantly structured?  Because dispersal distance plays a key role in the  spatial structuring of genetic variation among sub-populations, evidence of genetic structure at one or more spatial scales may help us understand the colonisation ability of Pacific Giant Salamanders. 2  Is there evidence from pair-wise genetic distances that clear-cutting  causes population bottlenecks? tenebrosus  A comparison of genetic variation in  D.  among forest types in BC suggests that sub-populations may pass  through bottlenecks during clear-cut events (Chapter 3). If recently clear-cut subpopulations have passed through bottlenecks, estimates of genetic distance among clear-cut sites may be greater than estimates among forested sites over similar spatial scales. 3  Is dispersal potential among critical nursery streams sufficient for the  re-establishment of disturbed or extirpated sub-populations?  The successful  re-establishment of locally extirpated populations depends on the colonisation potential, dispersal distance, and fecundity of Pacific Giant Salamanders.  74  4.2  4.2.1  METHODS  STUDY SITES AND SAMPLE COLLECTION  This genetic study was conducted during the summer of 1998 in the Chilliwack River Valley, British Columbia, at the northern periphery of D.  tenebrosus'  range (see Chapter 3: Fig. 3.1). The study area falls within the Coastal Western Hemlock biogeographic zone. Eight sample sites were chosen for this study: two sites were within old-growth stands (>250 years, NesF4 and NesF5); three sites were within second-growth stands (30-60 years, Cen23, TamSGI and VedB); and 3 sites were within recently clear-cut sites (3-9 years, FolQ, Foil2, and TamC) (see Chapter 3: Fig. 3.2 & Table 3.1). The sites that I chose for this study have been intensively studied since 1996 by WE Neill (Dept of Zoology) and JS Richardson (Faculty of Forestry) at the University of British Columbia. The two old-growth sites are clustered within the same drainage (Nesakwatch Creek), as are two of the clearcut sites (Foley Creek). Although the spatial arrangement of sites raised questions of independence among the clustered sites in Chapter 3, they provide an excellent opportunity to examine the spatial genetic structuring of  D. tenebrosus  in the  Chilliwack River Valley. Small tail-fin samples were collected from 31-65 individuals within 85-200 m stream reaches at the eight sites (see Chapter 3: Table 3.1) and preserved in 95% ethanol.  DNA was extracted from the tissue samples  using standard  phenol/chloroform extraction protocols and stored in TE (Tris-EDTA) buffer at -20 C (Taggart et al. 1992). Three polymorphic microsatellite loci developed for  D. tenebrosus  (DteA, Dte8  and Dte11)  (see Chapter 2) were used to assay all individuals  sampled from each of the eight streams. Randomised sub-sets of 20-28 individuals per stream were selected for AFLP assay using 39 AFLP markers (see Chapter 3: Tables 3.2, 3.4, & 3.5), optimised for  D. tenebrosus.  DteA,  Dte8  and  Dte11  conformed to Hardy-Weinberg expectations and were not in linkage disequilibrium (Chapter 3). AFLP markers CTA 200 and CTA 267 were excluded from the analyses because they appeared to be in linkage disequilibrium. The remaining AFLP markers  75  were assumed to be in hardy-Weinberg proportions. Please refer to Chapter 3 for AFLP and microsatellite protocols. 4.2.2  ANALYSIS OF ISOLATION-BY-DISTANCE  Slatkin (1993) predicted that populations close to or in migration-drift equilibrium should conform to an "isolation-by-distance" model: that is, gene flow among populations should decrease as the geographic distances among these populations increase.  Hence populations that are close in geographic distance  should also be more genetically similar than populations that are geographically farther apart. In order to test whether pair-wise genetic distances among sub-populations were significantly correlated with corresponding pair-wise geographic distances, I performed Mantel tests (Mantel 1967). Mantel tests determine whether there is a significant correlation between two matrices (Sokal & Rohlf 1995). I performed these tests at two levels: within the Chilliwack River drainage and across the northern extent of D. tenebrosus' range. 4.2.2.1  WITHIN T H E CHILLIWACK RIVER DRAINAGE  Geographic distances between pairs of streams in BC were estimated using 1:20 000 TRIM (Terrain Resource Information Management) maps (produced in 1992 for the BC Ministry of Crown Lands) and UTM co-ordinates acquired using a hand-held GPS (Garmin 12). Although radio-telemetry studies suggest that clearcuts act as partial barriers to movements by terrestrial Pacific Giant Salamanders (Johnston 1998; Johnston & Frid, unpublished manuscript), it is unclear which natural features of the landscape (e.g. mountain ridges, alpine meadows) act as barriers to dispersing individuals. Therefore, I measured geographic distance three ways. First, distance was measured (using the GPS) as the minimum straight-line distance between site-pairs under the assumption that there were no geomorphological barriers to movements.  Second, distance was measured as the  minimum distance a salamander would disperse between sites if it were restricted solely to stream corridors (TRIM maps). Third, distances were calculated as the minimum distance within the 1200-m contour between sites, because D. tenebrosus  76  is believed to be restricted to elevations < 1200 m (TRIM maps) (Richardson & Neill 1995). The estimated pair-wise geographic distances between sites in BC were arranged into lower-diagonal matrices. A complementary matrix of pair-wise F t s  values was constructed with the three microsatellite loci among site-pairs with  GENEPOP  V3.1D  (DteA,  and Dte11)  DteQ  (Raymond & Rousset 1995), according to  Weir & Cockerham's (1984) multi-locus estimator.  F t is a fixation index that s  measures the level of population sub-structure (Hartl & Clark 1997) (see below). ISOLDE,  a program that computes Mantel tests, available as part of  GENEPOP  V3.1D (Raymond & Rousset 1995), was used to test for significant correlations between the matrices of geographic distance and pair-wise F t values. s  also used to test for isolation-by-distance among sub-populations of D.  ISOLDE tenebrosus  was in  this fashion, using a matrix of pair-wise Nei's (1978) genetic distances D, which were estimated from 39 AFLP markers using 4.2.2.2  TFPGA  A C R O S S THE NORTHERN RANGE O F  D.  1.0 (Miller 1997). TENEBROSUS  Pair-wise F among all BC sub-populations and eight populations from across st  the northern range of  D. tenebrosus  & Cockerham (1984) using  (47 - 49 °N) were estimated according to Weir  DteA, DteQ  and Dte11 with GENEPOP V3.1D (Raymond  & Rousset 1995). A geographic distance matrix was constructed by converting UTM co-ordinates (Latitude and Longitude) into great circle distances. As above,  ISOLDE  was used to perform the Mantel tests (Raymond & Rousset 1995). In all  ISOLDE  Mantel tests, 1000 Monte Carlo permutations of the genetic  distance matrix were used to define the distribution of the test statistic, in this case, Spearman's rank correlation coefficient. Populations that have recently passed through genetic bottlenecks may upwardly bias estimates of genetic distance (Hedrick 1999). Therefore the Mantel tests were repeated excluding the recently clear-cut sites (FolQ, Foil2, and TamC) for comparison. 4.2.3  ANALYSES OF MOLECULAR VARIANCE  An analysis of molecular variance (AMOVA) is analogous to conventional analyses of variance (ANOVAs).  AMOVAs compute the molecular variance  77  components at different hierarchical levels (typically spatial hierarchies) by using a matrix of squared molecular distances between pairs of individuals (Excoffier et al. 1992). An AMOVA was performed with 3 polymorphic microsatellite loci using ARLEQUIN  1.1 (Excoffier et al. 1992; Schneider et al. 1999) in order to assess the  microsatellite genetic variance at three hierarchical levels: within-streams, amongstreams-within-drainages, and among-drainages.  Similarly, a matrix of squared  genetic distance among all AFLP haplotypes was created with  AmovPrep  1.01  (Miller 1998) using Excoffier et al.'s (1992) metric for distance. This matrix was used to perform an AMOVA using  WinAMOVA  1.55  (Excoffier 1992) in order to assess  AFLP genetic differentiation at the same three hierarchical levels. In order to assess the level of genetic variation within and among drainages, the sites were grouped into the 5 drainages in which they are situated: Ascaphus Creek (VedB), Centre Creek (Cen23), Foley Creek (FolQ, Foil2), Nesakwatch Creek (NesF4, NesF5), and Tamihi Creek (TamC, TamSGI) (see Chapter 3: Fig. 3.2). Recently clear-cut sites may have passed through population bottlenecks (Chapter 3), therefore these sites may downwardly bias estimates of gene flow (Hedrick 1999). Therefore, AMOVAs were repeated excluding the recently clear-cut sites (FolQ, Foil2 and TamC). The significance of the variance components defined by the AMOVAs was tested with 1000 permutations using a Monte Carlo procedure. 4.2.4  COMPARISONS OF GENETIC DISTANCE AMONG FOREST TYPES  I compared mean pair-wise F t (Weir & Cockerham 1984) and D estimated s  between pairs of forested sites and pairs that included at least one clear-cut site. Pair-wise genetic distances (F t and D), estimated between previously harvested s  sites and the mean of the old-growth sites, were plotted and regressed against the age of the second-growth and clear-cut sites. 4.2.5  POPULATION STRUCTURE AND GENE FLOW  Wright (1921) defined a fixation index that quantifies the effects of population substructure on inbreeding within sub-populations. The fixation index is equal to the drop in heterozygosity observed in a set of sub-populations relative to the heterozygosity expected under the assumption of random mating within and among  78  the sub-populations. The fixation index can be applied to several hierarchical levels of populations and combinations thereof. When the least inclusive and the most inclusive levels of the population hierarchy (i.e. levels above the individual level) are compared, the fixation index is called F (Hartl & Clark 1997). In my study of the st  population structure of Pacific Giant Salamanders, the least inclusive level is the level of individual putative sub-populations (Cen23, FolQ, Foil2, NesF4, NesF5, TamC, Tarn SG1 and VedB), while the most inclusive level is the collection of all sampled sub-populations. In this way, F t measures all the effects of population subs  structure between the stream and valley-wide levels. In assessing the general level of genetic structure among all sites (as opposed to performing pair-wise F t s  comparisons for isolation-by-distance analyses), I estimated "global" F t using three s  methods. First,  1.1 (Excoffier et al. 1992; Schneider et al. 1999) was used  ARLEQUIN  to estimate global F from the microsatellite AMOVA variance components (see st  Excoffier et al. 1992; Schneider et al. 1999). Similarly, a global analogue of F t, s  called  was estimated from the hierarchical variance components using the 39  <J>st,  AFLP markers and  WinAMOVA  1.55  (Excoffier 1993).  In order to test for  concordance in estimates of O t among 39 AFLP markers, AMOVAs were also s  performed using each individual AFLP marker (means and 95% confidence limits are reported).  For the purposes of comparison, global F t (estimated from each s  locus and over all three microsatellite loci) and its 95% confidence intervals were also estimated according to Weir & Cockerham (1984) using all sites. In order to assess the colonisation potential of Pacific Giant Salamanders, N m (the effective number of migrants per generation) was derived from the three e  estimates of global F (described above) by the controversial approximation N m = st  (1-Fst)  e  / 4F t. Analyses of population genetic structure and gene flow as described in s  this chapter are performed according the island model of migration which makes several unrealistic assumptions about sub-population sizes and the interactions among them (Hartl & Clark 1997; Waples 1998; Bohonak 1999, Whitlock & McCauley 1999).  While estimates of population structure are generally robust  against deviations from the assumptions underlying the island model of migration,  79  estimates of gene flow derived under these assumptions should be interpreted with scepticism  (Waples 1998; Bohonak 1999, Whitlock & McCauley  1999).  Interpretation of the N m estimates is further discussed in section 4.4 below. e  One of the island model's unrealistic assumptions is that all sub-populations are in migration-drift equilibrium (when the homogenising effects of migration from surrounding sub-populations directly offset the effects of drift on allele frequencies within sub-populations). Although most natural populations are probably rarely at migration-drift equilibrium, evidence of bottlenecks in recently clear-cut sites suggests that these sites may be farther from migration-drift equilibrium than either second-growth sites or old-growth sites. Moreover, these sites may downwardly bias estimates of gene flow. Therefore global estimates of F t and gene flow were s  also obtained from the subset of forested sites (excluding clear-cuts).  4.3  RESULTS  4.3.1  4.3.1.1  ANAL YSIS OF ISOLA TION-BY-DISTANCE WITHIN T H E CHILLIWACK RIVER D R A I N A G E  There was no evidence of isolation-by-distance among the eight subpopulations of  D. tenebrosus  in the Chilliwack River Valley. Mantel tests did not  reveal any correlations between pair-wise microsatellite F t estimates and the three s  measures of geographic distance either with, or excluding recently clear-cut sites (p always > 0.082) (Fig. 4.1 A).  Similarly, no significant correlations were found  between estimates of Nei's D (AFLP markers) and geographic distance within the Chilliwack Valley, either with or excluding recently clear-cut sites (p always > 0.112) (Fig. 4.1 B). 4.3.1.2 A C R O S S T H E N O R T H E R N R A N G E O F D.  TENEBROSUS  There was no significant correlation between pair-wise F t and geographic s  distance when all sites were included in the Mantel test (p=0.167). There was, however, significant isolation by distance when clear-cut sites in the Chilliwack Valley were excluded from the analysis (p=0.009) (Fig. 4.2).  80  4.3.2  ANALYSES OF MOLECULAR VARIANCE  AMOVAs revealed that most of the genetic variation among the eight subpopulations in the Chilliwack Valley was accounted for by within-stream differences both including and excluding recently clear-cut sites (Table 4.2, Fig. 4.3). AMOVAs performed with all eight sites showed that approximately 95% (microsatellites) and 91% (AFLPs) of the total genetic variation was accounted for by differences within sub-populations. Among-population differences within drainages accounted for 1% and 8%, with microsatellite and AFLP markers, respectively. Approximately 4% (microsatellites) and 1% (AFLPs) of the total genetic variation was accounted for by differences among drainages. When recently clear-cut sites were excluded from the AMOVAs, within-population differences accounted for 98.7% (microsatellites) and 92.7%  (AFLPs) of the variation; among-population-within-drainage differences  accounted for 1.3% (microsatellites) and 8.4% (AFLPs) of the variation; and amongdrainage differences accounted for 0.1% (microsatellites) and 0% (-1.1%) (AFLPs) of the total variation. A comparison of the AMOVA variance components with and without the inclusion of recent clear-cuts suggests that recently clear-cut sites augment estimates of the amount of genetic differentiation among sub-populations in the Chilliwack Valley.  The microsatellite within-stream and among-stream variance  components (V and V ) were significant both when recently clear-cut sites were c  D  included and excluded from the AMOVA (p always < 0.003; Table 4.2).  The  microsatellite among-drainage variance component (V ) accounted for a significant a  amount of the genetic variation only when all sites were included in the analysis (p=0.001). The AFLP variance components V and V were significant both when c  D  clear-cuts were included and excluded from the analyses (p=0.001). AFLP amongdrainage differences (V ) did not account for significant genetic differences whether c  clear-cuts were included or excluded from the AMOVAs (p always > 0.331). 4.3.3  COMPARISONS OF GENETIC DISTANCE AMONG FOREST TYPES  Comparisons of pair-wise genetic distances (F t and Nei's D) among site-pairs s  revealed that genetic distances were generally lower among forested sites than among pairs with at least one clear-cut (Fig. 4.4). However, the differences were 81  only statistically significant with microsatellite (F t) data (one-tailed t-tests; s  microsatellite p=0.003; AFLP p= 0.083). The mean pair-wise genetic distances (F t s  and Nei's D) between the mean of the old-growth sites and previously harvested sites decreased with the age of the latter (Fig 4.5) such that second-growth stands were more closely related to old-growth sites than recently clear-cut sites. However, neither correlation was significant (simple linear regression, one-tailed tests: microsatellite r =0.39, 0.05<p<0.1; AFLP: 1^=0.37, 0.05<p<0.1). 2  4.3.4  POPULATION STRUCTURE AND GENE FLOW  Measures of population structure, F , estimated across all loci and among all st  eight sub-populations were 0.053 (from the variance components of the microsatellite AMOVA), 0.084 (global F , Weir & Cockerham 1984), and 0.093 (3> t, st  s  from the variance components of the AFLP AMOVA). The 95% confidence interval for Fst (Weir & Cockerham 1984) among all sites was estimated as 0.014 - 0.119. I found general concordance in estimates of F t among microsatellite loci (F = 0.117 s  st  (Dte4), 0.054 (Dte8) and 0.014 (0fe11)) and 39 AFLP markers (mean = 0.098; 95% confidence interval = 0.051 -0.132). When recently clear-cut sites (FolQ, Foil2, and TamC) were excluded from the estimates, the level of population structure decreased to 0.013 (from the variance components of the microsatellite AMOVA), 0.052 (global F t, Weir & Cockerham s  1984), and 0.073 ( O , AFLP AMOVA). The 95% confidence interval for F (Weir & st  st  Cockerham 1984) among forested sites was 0 - 0.052 (Table 4.3). Global estimates of N m among all sites varied from 2.44 (AFLP Amova) to e  4.56 (microsatellite AMOVA).  Similarly, estimates of N m among forested sites e  (excluding clear-cut sites) varied from 3.17 (AFLP AMOVA) to 18.98 (microsatellite AMOVA). N m corresponding to the upper and lower confidence interval limits for e  F t (Weir & Cockerham 1984) and O t are included in Table 4.3. s  s  82  4.4  DISCUSSION  4.4.1  AT WHICH SPATIAL SCALES(S) ARE SUB-POPULATIONS GENETICALLY STRUCTURED WITHIN THE CHILLIWACK RIVER VALLEY: AMONG STREAMS, AMONG DRAINAGES, OR NOT SIGNIFICANT Y STRUCTURED ?  4.4.1.1  A N A L Y S E S O F ISOLATION-BY-DISTANCE  I found no evidence of isolation-by-distance among eight sub-populations of Pacific Giant Salamanders in headwater streams in the Chilliwack River Valley. The analyses of isolation-by-distance using both microsatellite and AFLP markers suggested that there is very little or no spatial structuring of genetic variation among sub-populations of  D. tenebrosus  in BC. Evidence of isolation-by-distance in other  species has been reported in several studies over similar spatial scales (e.g. littorine gastropod, Johnson & Black 1995; wood frog, Berven & Grudzien 1990; streamside salamander, Storfer 1999). There are several reasons why we might not expect to find a relationship between geographic distance and genetic differentiation among Pacific Giant Salamander sub-populations.  First, Pacific Giant Salamanders probably recently  recolonised the Chilliwack Valley subsequent to glacial retreat which began approximately 10 000 years ago (Chapter 2).  Populations that have recently  expanded into a region are less likely to be close to migration-drift equilibrium and may not have yet had the time to differentiate through drift. Alternatively, I may have sampled sites that belong to a single panmictic population, in which case, all sitespairs should share similar levels of genetic variation. Lastly, it may be that sampling error confounded the correlations between genetic and geographic distance. In order to tease these alternative hypotheses apart, I tested whether there was significant isolation by distance over a wider geographical scale. I also used analyses of molecular variance (AMOVAs) and estimates of population structure (F t s  and O t) to examine the partitioning of genetic variation among three hierarchical s  levels: within streams; among-streams-within-drainages; and among drainages.  83  AAA.2  A C R O S S THE NORTHERN RANGE O F  D.  TENEBROSUS  Significant isolation by distance over the wider scale suggests that although Pacific Giant Salamanders may have recently recolonised the northern extent of their present range (-10 000 years ago), sufficient time has probably passed for these populations to approximate migration-drift equilibrium. The sub-populations of Pacific Giant Salamanders are close enough to each other (within 35 km) that they are probably linked (directly or indirectly) by slight to moderate levels of gene flow, thereby maintaining them qualitatively similar. This is further discussed below. Isolation by distance was not detected when clear-cut sites were included in the analyses. This suggests that recently clear-cut sites are farther from migrationdrift equilibrium than forested sites (also discussed below). 4.4.1.3 A N A L Y S E S O F M O L E C U L A R V A R I A N C E Analyses of molecular variance using microsatellite markers supported the finding of low genetic differentiation among sites at various spatial scales, with approximately 95% of the total microsatellite genetic variation in the Chilliwack River Valley being explained by differences within streams (all sites included). Microsatellite variance components V (among drainages) and Vb (among-streamsa  within-drainages) accounted for less than 5 % of the total microsatellite genetic variation throughout the valley.  In contrast, the AMOVAs using AFLP markers  revealed significant differentiation among streams within drainages.  Both V  c  (approx. 90%) and V (approx. 8%) accounted for a significant proportion of the total b  AFLP genetic variation. The AFLPs were likely more sensitive than microsatellites to genetic structure among the eight sub-populations of Pacific Giant Salamanders because squared genetic distances were calculated over 39 markers, as opposed to only 3 microsatellite loci. 4.4.1.4 E S T I M A T E S O F POPULATION S T R U C T U R E Wright (1978) proposed general guidelines for the interpretation of F t s  estimates with regard to the level of population differentiation. Wright suggested that Fst estimates of 0 - 0.05 should indicate little or no differentiation (although this may not be negligible), estimates of 0.05-0.15 should indicate moderate differentiation,  84  and estimates of F t > 0.25 should represent great genetic differentiation. According s  to Wright's scale, estimates of the level of population structure among all subpopulations of  D. tenebrosus  that were derived from AMOVA variance components  correspond to slight or moderate levels of genetic structure among sites. Estimates of global F t calculated according to Weir & Cockerham (1984) were also consistent s  with these results: F t = 0.084 (all sites) and 0.033 (forested sites only). s  Estimates of moderate genetic structure among these sub-populations of D. are similar to measures of  tenebrosus  F t s  reported in several other amphibian  populations over similar spatial scales (e.g. Berven & Grudzien 1990; Scribner et al. 1994; Storfer 1999). barbouri)  Populations of the Streamside Salamander  (Ambystoma  that were within 5 km of each other were reported to have an  0.162 (Storfer 1999). Three populations of common toads  (Bufo bufo)  F  value of  s t  within 14 km  of each other (and largely separated by agricultural fields) exhibited F  s t  values  ranging from 0.014 (single-locus minisatellites) to 0.090 (allozymes) (Scribner et al. 1994). In a similar study of 20  Bufo bufo  populations,  F t s  was estimated as 0.045  (SE=0.016) (unpublished data, as in Scribner et al. 1994). A study of wood frogs (Rana  sylvatica)  conducted on a smaller spatial scale (0.264 - 3.78 km between  breeding ponds) in a relatively undisturbed area, reported F  s t  as 0.04 using direct  mark-recapture and mating success estimates (Berven & Grudzien 1990). However, Fst estimated from mtDNA (mitochondrial DNA) and allozyme markers was much greater (approximately 0.4-0.43) Salamander  (Ambystoma  Sub-populations of  tigrinum,  in the similarly sized semi-aquatic Tiger Routman 1993).  D. tenebrosus  appear to be moderately differentiated at  the among-stream-within-drainage level. The level of genetic structure suggests that these sub-populations are sufficiently differentiated that any pair of sites probably does not come from a single panmictic population. However, F t estimates are s  moderate, such that sub-populations are probably linked by different degrees of migration (see below).  85  4.4.2  IS THERE EVIDENCE FROM PAIR-WISE GENETIC DISTANCES THAT CLEAR-CUTTING CAUSES POPULATION BOTTLENECKS?  Populations that have recently passed through bottlenecks may be more genetically different from undisturbed populations due to the effects of random genetic drift on allelic richness and allele frequencies (Hedrick 1999). In a study of the population structure of Redback Salamanders  (Plethodon  cinereus),  Gibbs  (1998) found that the genetic distance among six sub-populations in a disturbed and fragmented forest was significantly greater than among six sub-populations within a similarly sized, previously undisturbed, continuous forest. Trends of greater pair-wise F t and Nei's D among recently clear-cut sites and s  between clear-cut and forested sites are consistent with the hypothesis that clearcuts disturb populations. Although the range in F t and D estimates among clear-cut s  sites overlapped with those of forested sites (Figs. 4.4 & 4.5), the genetic distances between harvested (recently clear-cut and second-growth) sites and the mean of the old-growth sites decreased with stand age. This trend suggests that subsequent to clear-cut events, lost alleles may be recovered and disrupted allele frequencies may gradually converge with those of surrounding sub-populations through immigration in as little as 60 years. Comparisons of allelic richness in D.  tenebrosus  among forest  types are also consistent with this hypothesis: populations in recently clear-cut sites have lower genetic variation, but variation is significantly correlated with stand age (Chapter 3). 4.4.3  IS DISPERSAL POTENTIAL AMONG CRITICAL NURSERY STREAMS SUFFICIENT FOR THE RE-ESTABLISHMENT OF DISTURBED OR EXTIRPATED SUB-POPULATIONS?  4.4.3.1  INTERPRETING E S T I M A T E S O F G E N E FLOW F R O M F  S T  Although some studies have reported a general negative correlation between gene flow and the degree of population structure (e.g. Waples 1987; Palumbi 1994), indirect estimates of gene flow derived from F t and its analogues may be s  confounded by other factors (Burton 1983; Hedgecock 1986; Waples 1998). Extant levels of genetic variation and the spatial genetic structuring of sub-populations may reflect selective pressures on the markers of interest, the vicariant history of  86  populations, the amount and patterns of gene flow among sub-populations or temporal fluctuations in N . Moreover, estimates of gene flow are typically derived e  from measures of population differentiation (F t or analogues) under several s  assumptions that are unlikely to be met by natural populations (Waples 1998; Bohonak 1999; Whitlock & McCauley 1999). For these reasons, estimates of gene flow derived from levels of population differentiation are controversial in the literature as to their accuracy, degree of bias, or potential for misinterpretation (Bossart & Prowell 1998; Waples 1998; Bohonak 1999; Whitlock & McCauley 1999). Still, because direct estimates of gene flow among natural populations are generally costly, extremely time consuming, and unfeasible in some cases, most studies that measure dispersal rates among populations infer N m indirectly from estimates of population structure. e  In order to increase the strength of conclusions based on F t-derived s  inferences about gene flow, several recent reviews argue that researchers should address questions of gene flow and population differentiation with testable hypotheses, rather than making post Prowell 1998; Waples 1998).  hoc  inferences (Bohonak et al. 1998; Bossart &  In a recent review of gene flow and population  structure, Bohonak (1999) also recommended an additional three strategies for strengthening conclusions regarding the magnitude of gene flow. First, one should establish whether there is concordance among loci and a variety of markers with regard to estimates of population structure.  Second, one should attempt to  corroborate indirect estimates of gene flow with ecological and demographic data. Third, Bohonak (1999) suggested that one should attempt to use molecular markers as a direct analogue of conventional mark-recapture techniques. 4.4.3.1.1  Concordance  Among  Marker Classes  and  Loci  When all loci and marker classes reveal the same relationships among subpopulations, estimates of gene flow are less likely to be biased by non-selectively neutral loci (although concordance among loci does not rule out the vicariant history of populations or fluctuations in N as potential biases of gene flow estimates). I e  found concordance among two different marker classes, AFLPs and microsatellites, as well as a general concordance among 3-39 loci within marker classes. Although 87  slightly more genetic structure was detected using the 39 AFLPs, F t and <3>t s  s  overlapped considerably in the range of estimates. 4.4.3.1.2  Inferences  of Dispersal  from Ecological  and Demographic  Data  The dispersal ability of amphibians is generally thought to be extremely poor (Petranka et al. 1993; Blaustein et al. 1994). Ferguson (1998) and Johnston (1998) are among the few researchers who have assessed the dispersal ability of Pacific Giant Salamanders. Johnston found that terrestrial salamanders could move as far as 240 m (straight-line distance from release point) over a two-month period. From these data, however, Johnston (1998) concluded that Pacific Giant Salamanders had poor dispersal ability with the probability of salamanders dispersing from one stream to the closest neighbouring stream estimated as 1/1000. However, because Johnston (1998) had to restrict her radio-tracking studies to terrestrial adults > -20 g, her estimates of  D. tenebrosus  dispersal ability may be greatly underestimated.  Although adult amphibians generally have extremely high site fidelity (e.g. Salamandra  salamandra,  Degani & Warburg 1978;  Plethodon  cinereus,  Kleeberger  & Werner 1982), juveniles are typically the dispersers in amphibian species and are capable of moving greater distances than might be expected from adult movement patterns (Duellman & Trueb 1994).  For example, Berven & Grudzien (1990)  reported that although none of the 11195 marked adult wood frogs  {Rana  sylvatica)  dispersed from the ponds in which they first bred (from 1976 - 1982), the average dispersal distance of dispersing juveniles was approximately 1.2 km. Johnston (1998) and Johnston & Frid (unpublished manuscript) reported that the mean home range size of terrestrial adults was approximately 3000-5000 m , 2  depending on which metric was used (see Johnston 1998), and ranged from approximately 400-35000 m over a 2-4 month period. Johnston (1998) also noted 2  that the home range estimates did not conform to a classic home range model: estimated home range size continued to increase as more location points were included in the analysis. Johnston's (1998) and Johnston & Frid's (unpublished manuscript) estimates of home range size of Pacific Giant Salamanders are generally 1-2 orders of magnitude greater than those of other salamanders: 3.3-29.4 m in Ambystoma 2  sp. (Kleeberger and Werner 1983); 9.8-12.8 m in 2  Salamandra  88  salamandra  (Degani and Warburg 1978); 314-1194 m in 2  (Stebbins 1951) and 10.8-19.9 m in 2  Plethodon  cinereus,  Ensatina  eschscholtzi  (Kleeberger and Werner  1982). For these reasons, my estimates of slight to moderate levels of population structure among sub-populations of Pacific Giant Salamanders across the Chilliwack River Valley were not unexpected. 4.4.3.1.3  Direct Estimates  of Gene Flow from Mark-Recapture  Studies  Unfortunately, the scope and duration of this study are such that direct estimates of gene flow from demographic data or molecular markers were not plausible. The only direct observation of D.  tenebrosus  dispersal in BC involved the  recapture of a marked juvenile on 17 July 1998 in a stream (TamBAW) that was approximately 1.1 km from the stream (TamSGI) in which it had been originally marked as a larva in 1996 (pers. obs.). It should be noted, however, that this observation is not evidence of gene flow as it was impossible to determine whether this individual had, or was going to, successfully breed in a stream other than the one in which it was recruited. 4.4.3.2 COLONISATION POTENTIAL O F PACIFIC G I A N T S A L A M A N D E R S IN BC Because I could not infer levels of gene flow from direct mark-recapture studies, my best assessment of the colonisation potential of Pacific Giant Salamanders must be drawn under the assumption that gene flow among Pacific Giant Salamander sub-populations is related to the level of population structure among sites. Waples (1998) suggested that if estimates of gene flow were to be derived from F , one should use the upper 95% confidence limit as a more st  conservative estimate of N m. When gene flow among all e  D. tenebrosus  sub-  populations was estimated from the upper 95% confidence limits of F t s  (microsatellites: 0.115; AFLPs: 0.137), N m = 1.5-2 individuals per generation. This e  estimate of N m probably greatly underestimates the level of gene flow among e  streams within drainages. The spatial scale at which colonisation potential should be assessed is at the between-stream-within-drainage scale because this is the scale at which cut-blocks occur. Because the probability of moving a distance falls with increasing distance (Johnston 1998), it follows that gene flow between  89  neighbouring streams should be higher than estimates integrated over all sites within the Chilliwack Valley. Indeed, F t is smaller among the three old-growth sites that s  are nested within 5 km of each other in Nesakwatch Creek Drainage. N m among e  these sites was conservatively estimated as approximately 10 individuals per generation. Although estimates of D. tenebrosus  population structure, isolation-by-  distance, and AMOVA are consistent with a set of sub-populations that are moderately linked by migration, interpretations of F t in terms of gene flow are s  problematic for conservation management decisions (Waples 1998). For example, if genetically structured populations of D. tenebrosus are treated as a single panmictic population, then the level of gene flow and colonisation potential among subpopulations may be overestimated and considered great enough to overcome any negative impacts (e.g. reduction in N , loss of genetic variation, local extirpation) of e  clear-cut harvesting on the sub-populations. Conversely, costly and unnecessary management strategies (e.g. retention of buffer strips or dispersal corridors) may be implemented in an effort to reduce the impact of forest harvesting and facilitate movement among nursery streams if F values are misinterpreted as reflecting poor st  colonisation potential and gene flow among sites. Moreover, N m estimates reflect e  the historical movements among streams and may not reflect current population level processes. Apart from the potentially misleading estimates of N m, several lines of e  evidence are consistent with relatively high dispersal ability among sub-populations of D. tenebrosus in BC. First, terrestrial adults, that are expected to have greater site fidelity than dispersing juveniles, have the ability to move over distances that are similar to the distances among nursery streams.  Second, genetic variation  increases while genetic differentiation decreases with stand age, suggesting that sub-populations may recover from clear-cut disturbances through migration from surrounding sub-populations within 60 years. When populations approximate the island model, a migration rate of approximately 1 individual per generation can prevent the fixation of alternate alleles among sub-populations through drift (Kimura & Ohta 1971) thereby maintaining  90  qualitative similarity among sub-populations (Allendorf 1983). As N m increases e  among sub-populations, they become more quantitatively similar, with allele frequencies converging. The sub-populations of Pacific Giant Salamanders do not satisfy several assumptions of the island model: population size estimates vary over two orders of magnitude (Neill & Richardson 1997), and the probability of dispersing to a stream is not constant over the entire valley. However, assuming that N m > 1e  2 individuals per generation among closely situated sub-populations, then dispersal among neighbouring streams is probably sufficient to recolonise streams or restore lost genetic variation to disturbed sub-populations over the short term. Waples (1998) argued that although a N m of 10 is sufficient to have a e  profound homogenising effect on population structure and long term evolutionary processes, the number of effective migrants needs to be on the order of hundreds or thousands to have a significant effect on recolonisation. However, Ferguson (1998) showed that in D.  tenebrosus  nursery streams, an individual female had the potential  to completely repopulate experimentally cleared stream reaches in a singfe breeding season. Similarly, observations of rapid increases in larval density in TamC (clearcut in 1994-95) from 1994 - 1999 suggest that Pacific Giant Salamanders can quickly repopulate a stream (Neill & Richardson 1997, unpublished data). The generation time of Pacific Giant Salamanders is unknown but is probably between 5 & 7 years.  Under the assumption that the true level of gene flow,  estimated over all eight sub-populations, falls within the range of 1-2 individuals per generation, then genetic recovery of extirpated or disturbed sub-populations may span several generations, which in turn, probably translates into several decades. The relationship between the genetic variation within Pacific Giant Salamander subpopulations and the stand age (Chapter 3) suggests that recovery of future disturbances could take as little as 60 years provided that cutting rates and cut-block sizes do not increase and that sources of immigrants remain in similar proximity to what they have been over the last century.  91  Table 4-1  Hierarchical analysis of molecular variance (AMOVA) tables for  microsatellite (A) and AFLP (B) data. The hierarchical levels (sources of variation), degrees of freedom (DF), sums of squared genetic distances (SS), variance components and the percentage of the total amount of genetic variation explained by each source (%) are given. Estimates of F- and O-statistics and respective p-values are also included. Italicised numbers represent AMOVAs excluding recently clear-cut sites. A Source of Variation Among Drainages  DF  SS  Variance Component Va = 0.00256  4  1.926  3  0.641  4 7  0.369  Within Populations  702  41.897  453  41.897  0.09249  Total  709  44.192  0.06300  457  42.718  0.09375  Among Populations within Drainages  0.181  0.00008  Vb = 0.00076 0.00118  Vc = 0.05968  % 4.07  Fixation Index F =0.0407 ct  0.0008  0.08  1.20 1.26  94.73 98.66  p-Value 0.001 0.087  F =0.0407 0.0726  < 0.000  Fst-0.0527 0.0734  < 0.000 < 0.000  Fixation Index  p-Value  sc  0.003  100  B Source of Variation Among Drainages Among Populations within Drainages Within Populations Total  DF  SS  Variance Component  4  63.945  Va = 0 . 0 8 4 7 2  3  44.021  3  35.326  7  16.842  126  97  669.907  563.478  -0.06709 Vb = 0 . 4 6 2 7 1  0.52660  Vc = 5 . 3 1 6 7 2  % 1.44  -0.077  7.89  <t>sc=0.080  8.40  90.66  5.80905  92.67 100  133  769.179  5.8641  707  624.343  6.2685  Oc,=0.014  -7.07  0.083  O t=0.093 s  0.073  0.332  0.513 0.001  0.001  0.001  0.001  92  Table 4-2 Global Fst (or O t) values (calculated across all eight BC sites, or among s  forested sites only) and derived estimates of the effective number of migrants, N m e  (* indicates significant estimates of population structure). / 3>st  Nm  Weir & Cockerham (1984) Global Fst Lower 95% C.I. Limit Upper 95% C.I. Limit  *0.084 0.014 0.117  2.73 17.61 1.89  Fst (Microsatellite AMOVA)  *0.052  4.56  (AFLP AMOVA) Lower 95% C.I. Limit Upper 95% C.I. Limit  *0.093 0.051 0.132  2.44 4.65 1.64  Weir & Cockerham (1984) Global Fst Lower 95% C.I. Limit Upper 95% C.I. Limit  0.033 0.000 0.052  7.33  Fst (Microsatellite AMOVA)  0.013  18.98  *0.073  3.17  Method All sites  F  <Dt s  Excluding Clear-cuts  <E>st  (AFLP AMOVA)  st  e  oo  4.56  93  0.2  n  0.15 CO  LL  CO CL  o  0.1  CD CO  o  °o  8  0.05 -j  o  0  -0.05 10  20  30  40  Distance (Km)  0.16 -| 0.14 0.12  0.1 0.08  Q  oo  0.06 0.04 0.02 0  B  0  10  20  30  40  D i s t a n c e (km)  Figure 4-1 Pair-wise genetic distances among sub-populations in the Chilliwack Valley measured as F (A) and Nei's D (B), is plotted against geographic distance st  (the minimum distance within the 1200 m elevation limit between two sites (ELD)). Closed circles represent pair-wise genetic distance estimates between two forested sites (old-growth and / or second-growth). Open circles represent pairs including one or two recently clear-cut sites).  94  Figure 4-2 Isolation by distance across the northern extent of D.  tenebrosus'  range  (457 - 49 °N). Fst was calculated according to Weir & Cockerham (1984) using three polymorphic microsatellite loci. Closed circles represent forested site-pairs, while open circles represent pairs with at least one clear-cut site from the Chilliwack Valley (all sites in Washington and Oregon were forested: Jacqueline Brinkman, Redpath Museum, Montreal, pers. comm.). The regression line represents the sample of forested sites only (the equation of the line is shown, 1^=0.117, p<0.0001).  95  100  -I  Within Streams  Figure 4-3 The proportion of D.  Among Streams  tenebrosus  Among Drainages  genetic variation accounted for by three  hierarchical levels: (A) within-streams, (B) among-stream-within-drainages and (C) among-drainages.  96  0.225 0.175 co  LL  CD CO  0.125 0.075  s  CO Q_  0.025 -0.025 Forested  0.16  Clear-cut  -i  0.14 0.12 Q  0.1 -  rwi  CD CO  0.08 -  'cd  0.06 -  CL  0.04 0.02 -  B  0 Forested  Clear-cut  Figure 4-4 Pair-wise genetic distances estimated with (A) three microsatellite loci (F ; Weir & Cockerham 1984) and (B) 39 AFLP markers (Nei's D) according to st  forest type. Forested pairs represent comparisons among forested sites. Clear-cut pairs represent pairs with 1-2 clear-cuts.  97  1  0.13  C/3  0.09  CO  CL  CO CD  ^  0.11  0.07  -I  0.05 0.03 0.01 -0.01  20  40  60  Stand A g e  0.12 0.11  0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03  B  0.02 20  40  60  Stand A g e  Figure 4-5 Pair-wise F t (A) and Nei's D (B) between the mean of the old-growth s  sites and previously harvested sites plotted against the age of the harvested sites.  98  5  CHAPTER 5: GENERAL CONCLUSIONS  The ultimate fate of all populations is extinction due to the fluctuating and stochastic processes in the environment.  Natural populations are subject to  demographic stochasticity (e.g. variation in recruitment and survival rates) environmental stochasticity (e.g. extreme temperatures, floods, fires, hurricanes, diseases),  genetic stochasticity (e.g. bottlenecks, drift), and  disturbances (e.g. urban development, harvesting).  anthropogenic  However, some species are  less vulnerable to these kinds of stochastic events than are others. Although there are no hard and fast rules that predict the vulnerability of populations to extinction or local extirpation (extinction on a local scale), population size (Pimm et al. 1988) and colonisation ability (Burbidge & Mackenzie 1989; Laurance 1991) are among the best predictors of the long-term persistence of species. Species that are most likely to be vulnerable to stochastic processes have one or a few small populations of low density, exhibit poor dispersal ability, and sustain low rates of population increase. Habitat fragmentation further reduces population sizes by imposing barriers to migration among habitat fragments. Poor dispersers that are unable to cross natural or anthropogenic barriers are less likely to re-establish populations in unoccupied habitat patches subsequent to environmental change.  For example, some bird,  mammal and insect species will not cross even short distances of open habitat to forested areas due to increased predation risk (Lovejoy et al. 1986; Primack 1993). Populations that are associated with stable environments may also be more vulnerable to environmental disturbances. Slight changes in micro-climate brought about by habitat alteration may affect the ability of populations to persist locally. COSEWIC and the BC Ministry of the Environment, Lands and Parks have identified the Pacific Giant Salamander as vulnerable to extirpation from British Columbia. The justification for listing Pacific Giant Salamanders as vulnerable or endangered in BC was based on its limited distribution in Canada, potentially poor dispersal ability, and rapid rates of habitat loss and alteration aggravated by widespread logging activities in the Chilliwack River Valley (Farr 1985, 1989; Haycock 1991; Ferguson & Johnston 1999). Moreover, BC populations of Pacific  99  Giant Salamanders are on the northern periphery of their range, where habitat may be marginal, and therefore may already be more susceptible to local extirpation than central populations. My thesis is part of an ongoing study to assess the impacts of current forest practices on the persistence and colonisation potential of Pacific Giant Salamanders in the Chilliwack River drainage, BC. My principal objectives were to address four related questions. First, is the level of genetic variation in BC similar to that found throughout the range of Pacific Giant Salamanders? I developed 6 polymorphic microsatellite loci and screened 14 populations from across  D. tenebrosus'  range in order to address  this question. I also consolidated allozyme data from two studies (Daugherty et al. 1983 and Good 1989) of the inter-specific relationships in the genus  Dicamptodon.  found that populations at the northern extent of  range had less  D. tenebrosus"  I  genetic variation (microsatellite and allozyme) than more southern populations. Half of the microsatellite loci were monomorphic north of 47.5 °N, while no allozyme variation was detected at or north of 47 °N. These latitudes coincide approximately with the southern limit of glacial expansion which began retreating northward 1012000 years ago (McPhail & Lindsay 1986; Williams et al. 1993). Pacific Giant Salamanders probably recolonised the northern extent of their range, including the Chilliwack River drainage after the glacial retreat (Chapter 2). Second, does clear-cut harvesting negatively impact populations of D. tenebrosus  by causing reductions in population size? Although the time scale of this  study was not sufficiently long to perform "before and after" comparisons of genetic diversity and population differentiation, my thesis shows that the clear-cuts I sampled (TamC, FolQ and Foil2) have less genetic variation (microsatellite and AFLP) than five forested sites in the Chilliwack River drainage (Chapter 3). Moreover, comparisons of genetic differentiation among forest types revealed that recently clear-cut sites were more genetically differentiated from neighbouring forested sites, than forested sites were to other forested sites even when the latter were in different drainages (Chapter 4). These results are consistent with the hypothesis that clear-cutting causes populations of  D. tenebrosus  to pass through  bottlenecks. Two sites that were forested during my genetic survey are currently, or  too  will soon be clear-cut (NesF4 and TamSGI). This presents a future opportunity for testing the hypothesis (using demographic and genetic data) that current forest practices in BC cause direct and/or indirect mortality in local sub-populations of Pacific Giant Salamanders. The third question I addressed relates to the spatial level at which populations of  are structured. Highly structured populations over small spatial  D. tenebrosus  scales would suggest that Pacific Giant Salamanders have low dispersal distances. However, I found that populations were slightly or moderately structured. Mantel tests suggested that over the northern extent of  D. tenebrosus  1  range, populations  conform to Slatkin's (1993) isolation by distance model (when clear-cut sites are excluded), but that this relationship does not hold over the smaller spatial scale of the Chilliwack River Drainage (Chapter 4). Analyses of molecular variance are consistent with the Mantel tests. AMOVAs show that most of the genetic variance is accounted for by differences within streams, and to a lesser (but significant) extent, by differences among streams within drainages. When clear-cuts were excluded from the AMOVAs, the among-drainage components did not account for any significant differences in genetic variation (Chapter 4). structure,  Ft s  and  4> t, s  Measures of population  among all sites varied from approximately 0.05 to 0.1. When  clear-cut sites were excluded from the analyses, F  st  and <& ranged from st  approximately 0.02-0.08 (Chapter 4). Under the assumption that recently clear-cut sites are farther from migration-drift equilibrium than forested sites (as suggested by results in Chapters 3 & 4), then the estimate of population structure integrated over forested sites only is probably closer to the true level of genetic structure among streams in the Chilliwack Valley.  F t estimates ranging from 0.02 - 0.08 are s  generally interpreted as reflecting slight to moderate differentiation The last question relates to the colonisation potential of Pacific Giant Salamanders.  Most of the Pacific Giant Salamander's range in BC has been  fragmented and disturbed, largely by forest harvesting. Because clear-cuts act as partial barriers to movements by terrestrial Pacific Giant Salamanders (Johnston 1998), landscape level fragmentation by logging roads and clear-cuts may reduce the ability of Pacific Giant Salamanders to recolonise locally extirpated sub-  101  populations or replenish lost genetic variation.  Because there is a non-linear  relationship between F t and N m, even under ideal circumstances, there is s  e  uncertainty in estimates of gene flow derived from F t, particularly when F is small s  st  (0-0.1). However, few natural populations conform to the island model of migration (Waples 1998; Bohonak 1999; Whitlock & McCauley 1999), and indirect estimates of N m may be off by several orders of magnitude (Whitlock & McCauley 1999). e  In this assessment of colonisation potential, it is not so important to know how many  individuals are dispersing to and breeding in new sub-populations. In order to  make suitable management decisions regarding Pacific Giant Salamanders in BC, we need to know whether gene flow and dispersal distance are potentially sufficient to re-establish viable sub-populations or restore lost genetic variation faster than these are being lost. The most conservative estimate of gene flow (derived from the upper confidence limit of F t as suggested by Waples 1998), where 1 < N m < 2, s  e  suggests that dispersal of Pacific Giant Salamanders among streams may be sufficient to recolonise locally extirpated sub-populations as long as these estimates reflect current movement patterns. Moreover, this estimate is integrated over subpopulations distributed over -35 km and probably underestimates gene flow among streams that are clustered within the same drainage. Indeed, the estimate of N m e  derived from F t from two old growth-sites (NesF4 and NesF5) within 2 km of each s  other is approximately 10 individuals per generation.  Because the longest  dimension of a cut-blocks is typically < 1/2 km (most clear-cuts are <40ha, Guy Fried, Ministry of Forests, Chilliwack Forest District, Rosedale BC), there appears to be high potential for neighbouring individuals to disperse into clear-cut sites from neighbouring streams or other reaches of the same stream. Apart from the potentially misleading estimates of N m, several lines of e  evidence are consistent with relatively high dispersal ability among sub-populations of  D. tenebrosus  in BC. First, terrestrial adults have the ability to move over  distances that are of the same order of magnitude as the distances among nursery streams (Johnston 1998).  Second, genetic variation increases while genetic  differentiation decreases with stand age, suggesting that sub-populations may recover from clear-cut disturbances through migration from surrounding sub-  102  populations on the order of 30-60 years.  Lastly, the recapture of a terrestrial  salamander revealed that this individual, recruited and marked in TamSGI, dispersed through a second-growth (and possibly a clear-cut) stand to another stream approximately 1.1 km from where it was originally captured (pers. obs). There are, of course, several caveats to consider with regard to these conclusions. First, molecular markers reflect both current and historical processes, and although my results suggest that the microsatellite and AFLP markers are reflecting processes that have occurred over the last 60 years, it is possible that I am simply describing the legacy of population-level processes that are no longer relevant to  D. tenebrosus  in BC.  Second, my results suggest that Pacific Giant Salamanders have high colonisation potential, particularly among closely nested streams. However, I cannot conclude whether populations are recovering faster (or as fast) as they are being disturbed. If the rate of disturbance is occurring faster than recovery rates, then Pacific Giant Salamanders may be at risk of local extirpation over the long term. Moreover, conclusions drawn from my results are based on the premise that the pool of potential migrants remains relatively stable and accessible to disturbed sites. Only long-term demographic studies can assess whether Pacific Giant Salamanders in BC are numerically stable or in decline. In the absence of such studies, it would be prudent to reduce the impact of clear-cutting on mortality and enhance the potential for recovery. This might be achieved by retaining buffer strips along known nursery streams and increasing the time between harvests. Moreover, the harvest of stands adjacent to each other should be staggered in order to maximise the probability of maintaining a viable pool of potential colonisers. Lastly, the sites I surveyed are probably among the most densely populated sites in the Chilliwack River Drainage. They were chosen partly on the basis of their accessibility, but also on the basis of their larval density, because streams with few larvae are not amenable to reasonable comparisons among forest types.  It is  unclear what the consequences of clear-cutting are on smaller sub-populations, but such sub-populations are likely more vulnerable to demographic stochasticity and local extinction during clear-cutting.  103  Forest harvesting appears to negatively impact numerous salamander species (Petranka et al. 1993, Dupuis et al. 1995; Ash 1997; Johnston 1998). Clearly, if we are to minimise the impact of forest harvesting on long-term persistence of salamander populations, we need to implement mitigative measures that will reduce salamander mortality and ensure movement among sub-populations. The most widely recommended management strategy for reducing the impact of forest clear-cutting on species that breed or rely on stream and riparian habitats is the retention of buffer strips along streams (e.g. Cross 1985; Bren 1995; Vesely 1996; Johnston 1998; Johnston & Frid, unpublished).  In BC, buffer strips are  currently not retained in the steep, 1 and 2 order (MOF S5 and S6) Pacific Giant st  nd  Salamander nursery streams, as these riparian zones are not currently protected under the Forest Practices Code (Anonymous 1995). Wide buffer strips can reduce sediment load and maintain water quality, temperature, humidity, coarse woody debris and invertebrate community composition within and adjacent to streams (Brown & Krygier 1970; Newbold et al. 1980; Beschta et al. 1987; Budd et al. 1987; Davies & Nelson 1994). Johnston (1998) recommended the retention of coarse woody debris in Pacific Giant Salamander habitat subsequent to cutting because 38% of recorded refuges of terrestrial Pacific Giant Salamanders were beneath or within old decaying logs and coarse woody debris.  Johnston (1998) also  recommended minimising soil compaction during forest harvesting by extracting trees with helicopters because terrestrial Pacific Giant Salamanders spend 31% of their time in small mammal burrows and root channels and over-winter in underground burrows and seeps. Maximising the cutting rotation (i.e. number of years between harvests) within sites and minimising cut-block size may also allow the complete demographic and genetic recovery of disturbed sub-populations through recruitment and migration from surrounding sub-populations.  104  6  LITERATURE CITED  Allen PJ, Amos W, Pomeroy PP, and SD Twiss. 1995. Microsatellite Variation in Grey Seals (Halichoerus grypus) Shows Evidence of Genetic Differentiation Between Two British Breeding Colonies. Molecular Ecology 4: 653-662. Allendorf FW. 1986. Genetic drift and the loss of alleles versus heterozygosity. Zoo Biology 5: 181-190. Allendorf FW. 1983. In Schonewald-Cox CM et al. (eds). 1983. Genetics and conservation : a reference for managing wild animal and plant populations. Benjamin/Cummings, California. Altschul SF, Gish W, Miller W, Myers EW, and DJ Lipman. 1990. alignment search tool. J. Mol. Biol. 215(3):403-410.  Basic local  Angers B and Bernatchez L. 1997. Complex evolution of a salmonid microsatellite locus and its consequences in inferring allelic divergence from size information. Mol. Biol. Evol. 14(3): 230-238. Anonymous. 1995. Forest Practices Code of British Columbia: Forest Practices Code Regulations. Province of British Columbia. Armbruster P, Bradshaw WE, and Holzapfel CM. 1998. Effects of postglacial range expansion on allozyme and quantitative genetic variation of the pitcher-plant mosquito, Wyeomyia smithii. Evolution 52: 1697-1704. Ash AN. 1997. Disappearance and return of salamanders to clearcut plots in the southern Blue Ridge mountains. Conservation Biology 11:983-989. Avise JC. 1994. Molecular Markers, Natural History and Evolution. Chapman and Hall, New York. BC Ministry of Environment, Lands and Parks. 1993. Pacific Giant Salamander. Wildlife in British Columbia at Risk, Brochure. Berger J. 1990. Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn sheep. Conservation Biology 4: 91-98. Berven KA and Grudzien TA. 1990. Dispersal in the wood frog (Rana sylvatica): implications for genetic population structure. Evolution 44(8): 2047-2056. Beschta RL, Bibly RE, Grown GW, Holtby LB, and TD Hofstra. 1987. Stream temperature and aquatic habitat: fisheries and forestry interactions. Pages 191-232 in Salo EO and TW Cundy (eds.). Streamside Management:  105  Forestry and Fishery Interactions. Contrib. 57, Inst. For. Resour., University of Washington, Seattle, WA. Blaustein AR and Wake DB. 1995. The puzzle of declining amphibian populations. Scientific American. April 1995: 522-527. Blaustein AR, Beatty JJ, Olson DH, and Storm RM. 1995. The biology of amphibians and reptiles in old-growth forests in the Pacific Northwest. Gen. Tech. Rep. PNW-GTR-337. Portland, OR: US Department of Agriculture, Forest Service, Pacific Northwest Research Station. Blaustein AR, Wake DB, and Sousa WP. 1994. Amphibian declines: judging stability, persistence, and susceptibility of population to local and global extinctions. Cons. Biol. 8: 60-71. Blears MJ, De Grandis SA, Lee H, and Trevors JT. 1998. Amplified fragment length polymorphism (AFLP): a review of the procedure and its applications. Journal of Industrial Microbiology and Technology 21:99-114. Bohonak AJ. 1999. Dispersal, gene flow, and population structure. The Quarterly Review of Biology 74(1): 21-45. Bohonak AJ, Davies N, Roderick GK, and FX Villablanca. Is population genetics mired in the past? Trends in Ecology and Evolution 13: 360. Bossart JL and Prowell DP. 1998. Genetic estimates of population structure and gene flow: limitations, lessons, and new directions. Trends in Evolution and Ecology 13: 202-206. Bowcock AM, Ruiz-Linares A, Tomforde J, Minch E, Kidd JR, Cavalli-Sforza LL. 1994. High resolution of human evolutionary trees with polymorphic microsatellites. Nature. 368: 455-457. Bren LJ. 1995. Aspects of the geometry of riparian buffer strips and its significance to forestry operations. Forest Ecology and Management. 75: 1-10. Brooker AL, Cook D, Bentzen P, Wright JM and RW Doyle. 1994. Organization of Microsatellites Differs Between Mammals and Cold-Water Teleost Fishes. Can. J. Fish. Aquat. Sci. 51: 1959-1966. Brown GW and Krygier JT. 1970. Effects of clear-cutting on stream temperature. Water Resource. Res. 6:1133-1139. Bruford MW, Cheesman DJ, Coote T et al. (1996) Microsatellites and their application to conservation genetics In; Molecular Genetic Approaches in  106  (eds. Smith TB and Wayne RK), pp. 278-297. Oxford University Press, New York. Conservation  Bruford, MW and RK Wayne. 1993. Microsatellites and Their Application to Population Genetic Studies. Curr. Opin. Genet. Devel. 3: 939-943. Budd WW, Cohen PL, Saunders PR, and FR Steiner. 1987. Stream corridor management in the Pacific Northwest: I. Determination of stream-corridor widths. Environ. Manage. 11: 587-597. Burbidge AA and McKenzie NL. 1989. Patterns in the modern decline of western Australia's vertebrate fauna: Causes and conservation implications. Biological Conservation 50: 143-198. Burton RS. 1983. Protein polymorphisms and genetic differentiation of marine invertebrate populations. Marine Biology Letters 4: 193-206. Bury RB and Corn PS. 1988. Douglas fir forests in the Oregon and Washington Cascades: Relation of herpetofauna to stand age and moisture. Pages 11-22 in Szaro RE, Severson KE, and Pattons DR (eds.). Management of Amphibians, Reptiles and Small Mammals in North America. Symposium Proceedings, 19-21 July 1988, Flagstaff, Arizona. Bury RB. 1983. Differences in amphibian populations in logged and old growth redwood forest. Northwest Science 57:167-178. Caughley G . 1994. Directions in conservation biology. Journal of Animal Ecology 63(2): 215-244. Chakraborty R, Kimmel M, Stivers DN, Davison LJ, Deka R. 1997. Relative mutation rates at di-, tri-, and tetranucleotide microsatellite loci. Proc. Nat. Acad. Sci. 94(3): 1041-1046. Chakraborty R and Nei M. 1977. Bottleneck effects on average heterozygosity and genetic distance with the stepwise mutation model. Evolution 31:347 - 356. Cole EC, McComb WC, Newton M, Chambers CL and Leeming JP. 1997. Response of amphibians to clear-cutting, burning and glyphosate application in the Oregon Coast Range. Journal of Wildlife Management 61: 656-664. Conner EJ, Trush WJ, and Knight AW. 1988. Effects of logging on Pacific Giant Salamanders: influence of age-class composition and habitat complexity. Bull. Ecol. Soc. of Am. 69(suppl.): 104-105. Corn PS and Bury RB. 1989. Logging in western Oregon: responses of headwater habitats and stream amphibians. For, Ecol. and Manage. 29: 39-57.  107  Cornuet JM and Luikart G. 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144: 2001-2014. Crawford AM, Kappes SM, Patterson KA, Degotari MJ, Dodds KG, Freking BA, Stone RT, and CW Beattie. 1998. Microsatellite evolution: Testing the ascertainment bias hypothesis. Journal of Molecular Evolution. 46(2) 256260. Cross SP. 1985. Responses of small mammals to forest riparian perturbations. Pages 269-275 in Riparian Ecosystems and Their Management: Reconciling Conflicts in Use, Tuscon AR. Daugherty CH, Allendorf FW, Dunlap WW, and KL Knudsen. 1983. Systematic implications of geographic patterns of genetic variation in the genus Dicamptodon. Copeia. 1983: 679-691. Davies PE and M Nelson. 1994. Relationships between riparian buffer widths and the effects of logging on stream habitat, invertebrate community composition and fish abundance. Aust. J. Mar. Freshwater Res. 45: 1289-1305. Degani G and MR Warburg. 1978. Population structure and seasonal activity of the adult Salamandra salamandra (L) (Amphibia, Urodela, Salamandridae) in Israel, J. Herpetol., 12: 437-444. Dow BD and MV Ashley. 1996. Microsatellite Analysis of Seed Dispersal and Parentage of Saplings in Bur Oak, Quercus macrocarpa. Molecular Ecology 5: 615-627. Duellman WE, Trueb L. 1994. Biology  of Amphibians.  McGraw-Hill, New York.  Dupuis LA, Smith JNM, and F Bunnell. 1995. Relation of terrestrial-breeding amphibian abundance to tree-stand age. Cons. Biol. 9: 645-653. Edwards A, Hammond HA, Li J, Caskey CK, and R Chakraborty. 1992. Genetic variation at five trimeric and tetrameric tandem repeat loci in four human population groups. Genomics. 12:241-253. Ellegren H, Primmer CR, and Sheldon BC. 1995. directionality or bias? Nat. Genet. 11: 360-362. Estoup A and J Turgeon. 1996. probes  and amplification.  Microsatellite 'evolution':  Microsatellite markers: isolation with Version of 12/1996. Laboratoire de  Poissons, INRA 78352 Jouy-en-Josas, France.  non-radioactive  Genetique des  108  Excoffier, L , Smouse, P., and Quattro, J. 1992 Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data. Genetics 131:479-491. Excoffier L. 1993. WINAMOVA Version 1.55. Available at http://anthropologie.unige.ch/arlequin. Farr ACM. 1989. Status report on the Pacific Giant Salamander Dicamptodon tenebrosus in Canada. Prepared for the Committee on the Status of Endangered Wildlife in Canada (COSEWIC). Farr, ACM. 1985. Status report on the Pacific Giant Salamander (Dicamptodon ensatus) in Canada -1985. Unpublished report to the BC Ministry of Environment, Wildlife Branch. Ferguson HM and B Johnston. 1999. Status report on the Pacific Giant Salamander (Dicamptodon tenebrosus) in Canada. Prepared for the Committee on the Status of Endangered Wildlife in Canada (COSEWIC). Ferguson HM. 1998. Demography, Dispersal and Colonisation of Larvae of Pacific Giant Salamanders (Dicamptodon tenebrosus, Good) at the Northern Extent of their Range. MSc Thesis. University of British Columbia, Vancouver. Fitzsimmons NN, Moritz C, Limpus CJ, Pope L and R Prince. 1997. Geographic Variation of Mitochondrial and Nuclear Gene Polymorphisms in Australian Green Turtle Populations and Male-Biased Gene Flow. Genetics 147:18431854. Fitzsimmons NN, Moritz C, and Moore SS. 1995. Conservation and dynamics of microsatellite loci over 300 million years of marine turtle evolution. Mol. Biol. Evol. 12:432-440. Frankel OH and Soule ME. 1981. Conservation and evolution. Cambridge University Press. 327p. Frankham R. 1995a. Inbreeding and extinction: a threshold effect. Conservation Biology 9: 792-799. Frankham R. 1995b. Conservation genetics. Annual Review of Genetics 29: 305327. Frankham R. 1995c. Effective population size/adult population size ratios in wildlife: a review. Genet. Res. Camb. 66: 95-106.  109  Garza JC, Slatkin M, and Freimer NB. 1995. Microsatellite allele frequencies in humans and chimpanzees, with implications for constraints on allele size. Mol. Biol. Evol. 12: 594-603. Gibco BRL. 1999. AFLP Analysis System Manual. Life Technologies.  I and AFLP  Starter Primer Kit  Instruction  Gibbs JP. 1997. Genetic structure of redback salamander Plethodon cinereus populations in continuous and fragmented forests. Biological Conservation 86: 77-81. Glenn TC .1995. Microsatellite Manual, Version 5. Molecular Systematics Laboratory, Smithsonian Institution, Washington, DC. Goldstein DB, Linares AR, Cavalli-Sforza LL, and MW Feldman. 1995. An Evaluation of Genetic Distances for Use With Microsatellite Loci. Genetics 139:463-471. Goldstein DB, and Pollock DD. 1997. Launching microsatellites: A review of mutation processes and methods of phylogenetic inference. Heredity 88: 335-342. Good DA. 1989. Hybridization and cryptic species in Dicamptodon (Caudata Dicamptodontidae). Evolution 43: 728-744. Good DA and Wake DB. 1997. Phylogenetic and taxonomic implications of protein variation in the mesoamerican salamander genus Oedipina (Caudata: Plethodontidae). Revistade Biologica Tropical. 45(3): 1185-1205. Gotelli D, Sillero-Zubiri C, Applebaum GD, Roy MS, Girman DJ, Garcia-Moreno J, Ostranders EA, and RK Wayne. 1994. Molecular Genetics of the Most Endangered Canid: the Ethiopian Wolf Canis simensis. Molecular Ecology 3: 301-312. Grimaldi MC, CrouauRoy B. 1997. Microsatellite allelic homoplasy due to variable flanking sequences J. Mol. Evol. 44(3): 336-340. Harris RB and Allendorf FW. 1989. Genetically effective population size of large mammals: an assessment of estimators. Conservation Biology 3:181-191. Hartl DL and Clark AG. 1997. Principles of Population Associates, Inc. Sunderland, Massachusetts.  Genetics,  3  rd  Ed. Sinauer  Hawkins CP, Murphy ML, Anderson NH, and Wilzbach MA. 1983. Density of fish and salamanders in relation to riparian canopy and physical habitat in streams of  no  the north-western United States. Canadian Journal of Fisheries and Aquatic Sciences 40: 1173-1185. Hay JM, Ruvinsky I, Hedges SB and LR Maxson. 1995. Phylogenetic relationships of amphibian families inferred from DNA sequences of mitochondrial 12S and 16S ribosomal RNA genes. Mol. Biol. Evol. 12(5): 928-937. Haycock RD. 1991. Pacific Giant Salamander Dicamptodon tenebrosus Status Report. Unpublished report to the BC Ministry of Environment, Wildlife Branch. Hedgecock D. 1986. Is gene flow from pelagic larval dispersal important in the adaptation and evolution of marine invertebrates? Bulletin of Marine Science 39: 550-564. Hedrick PW. 1999. Perspective: Highly variable loci and their interpretation in evolution and conservation. Evolution 53: 313-318. Hedrick PW, Lacy RC, Allendorf FW, and ME Soule. 1996. Directions in conservation biology: Comments on Caughley. Conservation Biology 10: 1312-1320. Hewitt GM. 1996. Some genetic consequences of ice ages, and their role in divergence and speciation. Biol. J. Linn. Soc. London. 58: 247-276. Hillier L and P Green .1991. sequencing  primers.  OSP: A computer program for choosing PCR and PCR Methods and Applications 1, 124-128.  DNA  Huang WG, Cipriani G, Morgante M, and R Testolin. 1998. Microsatellite DNA in Actinidia chinensis: isolation, characterisation, and homology in related species. Theoret. Appl. Genet 97(8): 1269-1278. Hutter CM, Schug MD, and Aquadro CF. 1998. Microsatellite variation in Drosophila melanogaster and Drosophila simulans: A reciprocal test of the ascertainment bias hypothesis. Mol. Biol. Evol. 15(12): 1620-1636. Jaeger RG. 1980. Microhabitats of a terrestrial forest salamander. 265-268.  Copeia 1980:  Johnson MS and R Black. 1998. Increased genetic divergence and reduced genetic variation in populations of the snail Bembicium vittatum in isolated tidal ponds. Heredity 80: 163-172. Johnston B. 1998. Terrestrial Pacific Giant Salamanders (Dicamptodon tenebrosus Good): Natural History and Their Response to Forest Practices. Msc. Thesis. University of British Columbia, Vancouver.  in  Barbara Johnston and Leonardo Frid. Impact of clearcut logging on movements by terrestrial pacific giant salamanders (Dicamptodon tenebrosus). Unpublished manuscript. Kelsey KA. 1995. western  Responses Washington.  WA.  of headwater  stream amphibians  to forest practices  in  PhD Dissertation, University of Washington, Seattle,  Kimura M and T Ohta. 1971. Theoretical Aspects of Population Genetics. Princeton University Press, Princeton NJ. Kimura M and JF Crow. 1964. The number of alleles that can be maintained in a finite population. Genetics 49:725-738. Kleeberger SR and JK Werner. 1983. movement of Ambystoma maculatum,  Post-breeding migration and summer J. Herpetol., 17:176-177.  Kleeberger SR and JK Werner. 1982. Home range and homing behavior of Plethodon cinereus in northern Michigan, Copeia 1983: 409-415. KruglyakS, Durrett RT, Schug MD, and CF Aquadro. 1998. Equilibrium distributions of microsatellite repeat length resulting from a balance between slippage events and point mutations. Proc. Nat. Acad. Sci. 96(18) 10774-10778. Lamberson RH, McElvey R, Noon BR and C Voss. 1992. A dynamic analysis of northern spotted owl viability in a fragmented forest landscape. Cons. Biol. 6: 505-512. Lande R. 1994. Risk of population extinction from fixation of new deleterious mutations. Evolution 48: 1460-1469. Lande R. 1988. Genetics and demography in biological conservation. Science 241: 1455-1459. Laurance WF. 1991. Ecological correlates of extinction proneness in Australian tropical rain forest mammals. Conservation Biology 5: 79-89. Luikart G, Allendorf FW, Cornuet JM, Sherwin WB. 1998. Distortion of allele frequency distributions provides a test for recent population bottlenecks. Journal of Heredity 89: 238-247. Luikart GH. 1997. bottlenecks  Montana.  Usefulness of molecular markers for detecting and monitoring genetic change. PhD Dissertation.  population  University of  112  Lagercrantz U, Ellegren H, and L Andersson. 1993. The abundance of various polymorphic microsatellite motifs differs between plants and vertebrates. Nucl. Acid. Res. 21 (5): 1111 -1115. Leonard WP, Brown HA, Jones LLC, McAllister DR, and RM Storm. 1993. Amphibians of Washington and Oregon. Audubon Society, Seattle WA. Levinson G, and Gutman GA. 1987. Mol. Biol. Evol. 4:203-221 Lomolino MV and R Channell. 1995. Splendid isolation: patterns of geographic range collapse in endangered mammals. Journal of Mammology 76: 335-347. Mace GM, Smith TB, Bruford MW and RK Wayne. 1996. An overview of the issues. In: Smith TB, and Wayne RK (eds.). Molecular genetic approaches in Conservation. Oxford University Press. New York. Maiorana VC. 1978. Difference in diet as an epiphenomenon: space regulates salamanders. Can. J. Zool. 56: 1017-1025. Mantel N, 1967. The detection of disease clustering and a generalized regression approach. Cancer Research 27: 209-220. McCombs A and W Chittenden. 1990. account  of logging and sawmilling  The Fraser Valley Challenge: An illustrated in the Fraser Valley. Treeline Publishing.  McPhail JD and CC Lindsay. 1970. Introduction: Freshwater Fishes of North-western Canada and Alaska. In: Stevenson (ed.) Freshwater Fishes of North-western Canada and Alaska. Queen's Printer for Canada, Ottawa. Miller LM and AR Kapuscinski. 1997. Historical analysis of genetic variation reveals low effective population size in a northern pike (Esox lucius) population. Genetics 147: (3) 1249-1258. Miller MP. 1998. AMOV-PREP 1.01: A program for the preparation of AMOVA input files from dominant-marker raw data. Available at: http://herb.bio.nau.edu/~miller/tfpga.htm Miller MP. 1997. program  Tools for population genetic analyses (TFPGA) 1.3: A windows for the analysis of allozyme and molecular population genetic data.  Available at:  http://herb.bio.nau.edu/~miller/tfpga.htm.  Miller Jr. RG. 1991. Simultaneous  Statistical  Inference.  Springer-Verlag, New York.  Milligan BG, Leebans-Mack J, and AE Strand. 1994. Conservation genetics: beyond the maintenance of marker diversity. Molecular Ecology 3: 423-435.  113  Mills SL and Smouse PE. 1994. Demographic consequences of inbreeding in remnant populations. American Naturalist 144: 412-431. Moore SS, Sargeant LL, King TJ, Mattick JS, Georges M, and DJS Hetzel. 1991. The conservation of dinucleotide microsatellites among mammalian genomes allows the use of heterologous PCR primer pairs in closely related species. Genomics. 10: 654-660. Murphy ML, Hawkins CP, and Anderson NH. 1981. Effects of canopy modification and accumulated sediment on stream communities. Transactions of the American Fisheries Society 110: 469-478. Murphy ML and Hall JD. 1981. Varied effects of clear-cut logging on predators and their habitat in small streams of the Cascade Mountains, Oregon. Can. J. Fish. Aquat. Sci. 38: 137-145. Nathan R, Safriel UN, and H Shrihai. 1996. Extinction and vulnerability to extinction at distribution peripheries: an analysis of Israeli breeding avifauna. Israel Journal of Zoology 42: 361-383. Nei M.1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583-590. Neill WE. 1998. Recovery of Pacific Giant Salamander populations threatened by clearcut logging, Final Report. Submitted to World Wildlife Fund Canada, Endangered Species Recovery Fund (Jan. 1998). Neill WE and JS Richardson. 1997. Persistence of threatened Pacific Giant Salamander populations under forest harvesting. Final Report submitted to Forest Renewal British Columbia (April 1997). Newbold JD, Erman DC and KB Roby. 1980. Effects of logging on macroinvertebrates in streams with and without buffer strips. CJFAS 37: 10761085. Nussbaum RA, Brodie Jr. ED, and Storm RM. 1983. Amphibians and reptiles of the Pacific Northwest. The University of Idaho Press. Nussbaum RA. 1976. Geographic variation and systematics of salamanders of the genus Dicamptodon Strauch (Ambystomatidae). Miscellaneous Publications of the Museum of Zoology, University of Michigan 149: 1-94. Nussbaum RA and Clothier GW. 1973. Population structure, growth, and size of larval Dicamptodon ensatus (Eschsholtz). Herpetologica 25: 257-262.  114  Nussbaum RA. 1969. Nests and eggs of the Pacific Giant Salamander, Dicamptodon ensatus (Eschscholtz). Northwest Science 47:218-227. Orti G, Pearse DE, and JC Avise. 1997. Phylogenetic assessment of length variation at a microsatellite locus. Proc. Natl. Acad. Sci. 94: 10745-10749. Palumbi SR. 1994. Genetic divergence, reproductive isolation, and marine speciation. Annual Review of Ecology and Systematics 25: 547-572. Paetkau D, Waits LP, Clarkson PL, Craighead L, and Strobeck C. 1997. An empirical evaluation of genetic distance statistics using microsatellite data from bear (Ursidae) populations. Genetics 147: 1943- 1957. Peakall R, Gilmore S, Keys W, Morgante M, and Rafalski A. 1998. Cross-species amplification of soybean (Glycine max) simple sequence repeats (SSRs) within the genus and other genera: Implications for the transferability of SSRs in plants. Mol. Biol. Evol. 15(10): 1275-1287. Pechmann JHK, Scott DE, Semlitsch RD, Caldwell JP, Vitt LJ, and Gibbons JW. 1991. Declining amphibian populations: the problem of separating human impacts from natural fluctuations. Science 253: 892-895. Petranka JW. 1998. Salamanders of the United States and Canada. Smithsonian Institution Press. WA. Petranka JW, Eldridge ME, and KE Haley. 1993. Effects of timber harvesting on Southern Appalachian Salamanders. Cons. Biol. 7: 363-370. Phillips K. 1990. Where have all the frogs and toads gone? Bioscience 40: 422: 424. Pimm SL, Jones HL, and Diamond J. 1988. On the risk of extinction. American Naturalist 132: 757-785. Primack RB. 1993. Essentials of Sunderland, Massachusetts.  conservation  biology.  Sinauer Associates.  Queller DC, Strassmann JE, Hughes CR. 1993. Microsatellites and kinship. Trends Ecol. Evol. 8: 285-288. Randall MG. 1982. The dynamics of an insect population throughout its altitudinal distribution: Coleophora alticollela (Lepidoptera) in Northern England. Journal of Animal Ecology 51: 993-1016. Raymond M, and Rousset F. 1995. Genepop (version 3.1d): population genetics software for exact tests and ecumenicism. J. Heredity 86: 248-249.  115  Redenbach Z, Taylor EB. 1999. Zoogeographical implications of variation in mitochondrial DNA of Arctic grayling (Thymallus arcticus). MOLECULAR Ecology 8: 23-35. Richardson JS and WE Neill. 1998. Headwater amphibians and forestry in British Columbia: Pacific Giant Salamanders and Tailed Frogs. Northwest Science 72(Special Issue 2): 122-123. Richardson JS and WE Neill. 1995. Distribution patterns of two montane stream amphibians and the effects of forest harvest: The Pacific Giant Salamander and Tailed Frog in south-western British Columbia. Unpubl. Report to the British Columbia Ministry of Environment, Lands and Parks. Rico C, Rico I, Hewitt G. 1996. 470 million years of conservation of microsatellite loci among fish species. Proceeding of the Royal Society of London Series BBiological Sciences 263: 549-557. Rogers DJ and SE Randolph. 1986. Distribution and abundance of tsetse flies (Glossina species). Journal of Animal Ecology 55:1007-1025. Rozen S and H Skaletsky. 1998. Primer3. Code available at http./wwwgenome.wi.mit.edu/genome software/other/primer3.html. Routman E. 1993. Population structure and genetic diversity of metamorphic and paedomorphic populations of the tiger salamander, Ambystoma barbouri. J. Evol. Biol. 6:329-357. Rubinsztein DC, Amos W, Leggo J, Goodburn S, Jain S, Li SH, Margolis RL, Ross CA, and MA Fergusonsmith. 1995. Microsatellite evolution - evidence for directionality and variation in rate between species. Nature Genetics 10: 337343. Sage RD and JO Wolff. 1986. Pleistocene glaciations, fluctuating ranges, and low genetic variability in a large mammal (Ovis dalli). Evolution 40: 1092-1095. Schlotterer C and D Tautz. 1992. Slippage Synthesis of Simple Sequence DNA. Nucl. Acid. Res. 20 (2): 211-215. Schneiders, Kueffer J-M, Roessli D, and L Excoffier. 1999. Arlequin Version 1.1. Available at: //anthropologie.unige.ch/arlequin. Schug MD, Hutter CM, Wetterstrand KA, Gaudette MS, Mackay TFC, and CF Aquadro. 1998. The mutation rates of di-, tri- and tetranucleotide repeats in Drosophila melanogaster. Mol. Biol. Evol. 15(12): 1751-1760.  116  Scribner KT, Arntzen JW, and Burke T. 1994. Comparative analysis of intra- and inter-population genetic diversity in Bufo bufo, using allozyme, single-locus microsatellite, minisatellite and multilocus minisatellite data. Mol. Biol. Evol. 11(5): 737- 748. Shaffer HB and McKnight ML. 1996. The polytypic species revisited: Genetic differentiation and molecular phylogenetics of the tiger salamander Ambystoma tigrinum (Amphibia: Caudata) complex. Evolution. 50(1): 417433. Slatkin M. 1993. Isolation by distance in equilibrium and non-equilibrium populations. Evolution 47: 264-279. Slatkin M. 1985. Gene flow in natural populations. Ann. Rev. Ecol. Syst. 18: 393430. Smith TB, and Wayne RK. 1996. Molecular genetic (eds). Oxford University Press. New York.  approaches  in  Conservation.  Sokal RR and FJ Rohlf. 1995 Biometry, 3 Edition. WH Freeman & Co. New York. rd  Spotila JR. 1972. Role of temperature and water in the ecology of lungless salamanders. Ecol. Monogr. 42: 95-125. Stallings RL, Ford AF, Nelson D, Torney DC, Hildebrand E, and RK Moyzis. 1991. Evolution and distribution of (GT)n repetitive sequences in mammalian genomes. Genomics. 10: 807-815. Stebbins RC. 1951. Amphibians California Press.  of Western  North America.  Berkeley: University of  Stevens James. 1996. Applied Multivariate Statistics for the Social Sciences 3 Ed. Lawrence Erlbaum Associates Inc. New Jersey. rd  Storfer A. 1999. Gene flow and population subdivision in the salamander, Ambystoma barbouri. Copeia 1: 174-181.  Streamside  Tautz, D, and M Renz. 1984. Simple Sequences Are Ubiquitous Components of Eukaryotic Genomes. Nucleic Acids Res. 12: 4127-4138. Taylor AC, Sherwin WB, and RK Wayne. 1994. Genetic Variation of Microsatellite Loci in a Hairy-Nosed Wombat Lasiorhinus krefftii. Mol. Ecol. 3: 277-290. Taylor EB, Pollard S, and D Louie. 1999. Mitochondrial DNA variation in bull trout (Salvelinus confluentus) from north-western North America: implications for zoogeography and conservation. Molecular Ecology 8: 1155-1170.  117  Taylor JS, Durkin MH, and F Breden. unpublished. The death of a microsatellite: a phylogenetic perspective on microsatellite interruptions. Dept. of Biological Sciences, Simon Fraser University. Taggart JB, Hynes RA, Prodohl PA et al. 1992. A simplified protocol for routine total DNA isolation from salmonid fishes. J. Fish. Biol. 40(6): 963-965. Templeton AR, Routman E and CA Phillips. 1995. Separating population structure from population history: A cladistic analysis of the geographical distribution of mitochondrial DNA haplotypes in the Tiger salamander, Ambystoma tigrinum. Genetics 140: 767-782. Valdes AM, Slatkin M, and NB Freimer. 1993. Allele Frequencies at Microsatellite Loci: The Stepwise Mutation Model Revisited. Genetics 133: 737-749. Valsecchi E and W Amos. 1996. Microsatellite markers for the study of cetacean populations. Mol. Ecol. 5(1): 151-156. Vesely DG. 1996. headwater  Terrestrial amphibian abundance and riparian buffer strips, Oregon Coast Range.  State University, Corvallis.  species  richness  in  MSc Thesis. Oregon  Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Homes M, Freijters A, Pot, J, Peleman J, Kuiper M, and Zabeau M. 1995. AFLP: a new technique for DNA fingerprinting. Nucl. Acids. Res. 23: 4407-4414. Waples RS. 1998. Separating the wheat from the chaff: patterns of genetic differentiation in high gene flow species. Journal of Heredity 89(5): 438-450. Waples RS. 1987. A multispecies approach to the analysis of gene flow in marine shore fishes. Evolution 41:385-400. Weber JL, and C Wong. 1993. Mutation of human short tandem repeats. Human. Mol. Genet. 2: 1123-1128. Weber JL. 1990. Informativeness of human (CA) polymorphisms. Genomics. 7: 524-530. n  Weir BS. 1990. Genetic data analysis. Sinauer Associates, Inc. Sunderland, Massachusetts. Weir BS and Cockerham CC. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358-1370.  118  Welsh HH. Jr. 1990. Relictual amphibians and old-growth forests. Cons. Biol. 4: 309-319. Whitlock MC and McCauley DE. 1998. Indirect measures of gene flow and migration: Fst * 1/(4Nm +1). Heredity 82: 117-125. Williams MAJ, Dunkerly DL, De Decker P, Kershaw AP, and T Stokes. 1993. Quaternery Environments. Edward Arnold, New York. Wright S. 1978. Evolution and the genetics of Populations. Vol. 4. Variability Within and Among Natural Populations. University of Chicago Press, Chicago. Wright S. 1931. Evolution in mendelian populations. Genetics 16: 97-159. Wright S. 1921. Systems of mating. Genetics. 6:111-178. Zar JH. 1984. Biostatistical Analysis 2 edition. Prentice Hall, Englewood Cliffs, New nd  Jersey.  119  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0074848/manifest

Comment

Related Items