Conserv Genet (2011) 12:213–222 DOI 10.1007/s10592-010-0134-4
RESEARCH ARTICLE
Micro-spatial genetic structure in song sparrows (Melospiza melodia) Amy G. Wilson • Peter Arcese • Yvonne L. Chan Michael A. Patten
•
Received: 3 December 2009 / Accepted: 4 September 2010 / Published online: 7 October 2010 Ó Springer Science+Business Media B.V. 2010
Abstract The spatial genetic structure of populations is strongly influenced by current and historical patterns of gene flow and drift, which in the simplest case, is limited by geographic distance. We examined the microspatial genetic structure within 33 populations of song sparrows (Melospiza melodia) which included eight subspecies located across coastal areas in southern British Columbia (BC) and California. We also examined the effect of water barriers and local density estimates on genetic structuring. Across both regions, positive genetic structure was detectable at distances of less than 10 km. Genetic divergence was highest in Californian subspecies, perhaps due to reduced gene flow across sub-specific contact zones. In BC, populations distributed across islands displayed greater genetic structuring over similar spatial scales than those across mainland sites, supporting the prediction that water barriers reduce gene flow in this species. Our results confirm both the expectation for fine-scale genetic structure in these generally sedentary subspecies, and the role of
A. G. Wilson (&) P. Arcese Center for Applied Conservation Research, University of British Columbia, 2424 Main Mall, Vancouver, BC V6T 1Z4, Canada e-mail:
[email protected] A. G. Wilson Smithsonian Migratory Bird Center, National Zoological Park, P.O. Box 37012-MRC 5503, Washington, DC 20013, USA Y. L. Chan Department of Biological Sciences, Stanford University, Stanford, CA 94305, USA M. A. Patten Oklahoma Biological Survey, University of Oklahoma, Norman, OK 73019, USA
landscape features in generating geographic variation in genetic structure. Keywords Spatial autocorrelation Genetic structure Isolation by distance Islands
Introduction Natural populations typically have restricted dispersal across certain landscapes and this can lead to the nonrandom spatial distribution of individuals with respect to their genetic similarity, a situation referred to as spatial genetic structure (Wright 1943). In the absence of selection, more related individuals will occur in closer proximity to each other, leading to positive genetic structure. The spatial extent over which positive genetic structure is detectable has been referred to in the literature as the ‘patch size’ (Sokal 1979; Sokal and Wartenberg 1983; Smouse and Peakall 1999) and the ‘genetic patch width’ (Epperson 2003), the latter of which will be adopted in this paper. The genetic patch width should vary among species that differ in vagility (Bohonak 1999), as well as among populations within a species due to the influence of physical or ecological barriers on dispersal patterns. The relative influence of landscape features on gene flow can be determined by examining the correlations between genetic distance and landscape features. Isolation by distance (IBD), is the most commonly tested model, and is equivalent to a null model, where the genetic distance between populations is simply correlated with the geographic distance between them (Wright 1943). Cases where IBD patterns are absent or weak often indicate that a more complex suite of ecological or physical factors are structuring populations. In support of this, several recent studies
123
214
show that genetic structure is more closely correlated to ecological gradients (Cooper 2000; Ruegg et al. 2006), dispersal barriers (Lee et al. 2001; Funk et al. 2005; Crispo et al. 2006; Pe´rez-Espona et al. 2008) and reproductive barriers (Irwin et al. 2005), than to a simple estimate of the geographic distance. For small populations, the level of genetic drift could also be more influential in determining the genetic similarity between populations than geographic proximity (Hutchison and Templeton 1999). Alternatively, the absence of an IBD pattern may be an artefact of sampling beyond the geographic scale of genetic structuring (Smouse and Peakall 1999), and many spatial genetic studies are conducted across broad spatial scales. Avian populations generally show low genetic structure (Crochet 2000), leading researchers to design sampling schemes at a broad spatial scale to increase the chance of detecting differences. Therefore, we have limited data regarding the fine-scale genetic structuring within avian populations (but see Friesen et al. 1996, Woxvold et al. 2006). The inferences that can be made from fine-scale spatial genetic structure have significant relevance for conservation. Understanding the scale at which gene flow predominates can indicate the approximate scale of demographic independence (Diniz and Telles 2002; Scribner et al. 2005). Spatial genetic structure is also indicative of the spatial scale over which populations diverge genetically, which can be crucial information for genetic management programs and establishing translocation criteria (Scribner et al. 2005). Spatial genetic structure has informed invasive species control (Hampton et al. 2004; Darling and FolinoRorem 2009), wildlife disease management (Deyoung et al. 2009), climate change modeling (Scoble and Lowe 2010) and connectivity of both common (Vignieri 2005) and vulnerable populations (Watts et al. 2004). We examined spatial genetic structure in song sparrows (Melospiza melodia) across a broad geographical range along the pacific coast of North America. Song sparrows are an ideal study species for genetic studies because they are widely distributed across North America in a range of landscapes, and show considerable subspecific divergence across their range (Aldrich 1984; Arcese et al. 2002). Song sparrows have often served as a focal study species, and therefore, we have a wealth of data on their population ecology, inbreeding patterns, behavior and phenotypic variation. This knowledge enables us to interpret genetic patterns within a well-informed framework that can then be utilized to understand patterns in other insular species with similar life histories, but for which detailed study is logistically difficult. Recent analyses of long-term dispersal data showed that insular populations of non-migratory song sparrows are relatively sedentary, with low immigration rates among islands (Wilson and Arcese 2008). At a large geographic
123
Conserv Genet (2011) 12:213–222
scale, song sparrow populations on islands tend to diverge genetically (Pruett et al. 2008), have reduced genetic diversity (Wilson et al. 2009) and are phenotypically divergent from the mainland populations (Patten and Pruett 2009). One exception to the divergence pattern is Vancouver Island, which has the same song sparrow subspecies (M.m.morphna) as the nearby mainland. Despite the fact that Vancouver Island has several vertebrate endemics (Campbell et al. 1990; Cook and MacDonald 2001; Darling 2003), the effect of the Georgia Strait on population genetic divergence has never explicitly been examined. In this study, we measure the spatial genetic patch width and genetic structuring of song sparrow populations on islands and in continuous mainland landscapes. Our sampling design explicitly tests the effect of large and small water barriers, sampling over similar spatial scales, and controlling for covariates that affect genetic structuring such as population size and subspecific boundaries.
Methods Sampling protocol We collected spatially extensive data from georeferenced M. m. morphna individuals in populations across coastal BC, Vancouver Island and the Southern Gulf Islands at scales ranging from 0.2 to 450 km (Fig. 1). In 2005, 150 genetic samples across nine sites were collected in a pairwise manner along coastal southern BC and Vancouver Island (Fig. 1a). Population sites and sample sizes are as following: Campbell River (n = 19), 2) Powell River (n = 18), 3) Texada Island (n = 17), 4) Qualicum (n = 18), 5) Sechelt (n = 16), 6) Alaksen (n = 20), 7) Duncan (n = 19) 8) Sooke (n = 16) and 9) Padilla Bay (n = 7). Within the Southern Gulf Islands, 211 individuals were sampled across ten islands and islets (Fig. 1b) and sample sizes are as following: Shell islands, (n = 26), Dock Islands (n = 36), Mandarte (n = 78), Halibut (n = 13) and Sidney (n = 52) Islands. During the breeding season, adult birds were captured in mist nets using playbacks of male song. All individuals were banded with a uniquely numbered band, capture coordinates were recorded and blood samples were taken. Blood samples were taken from the brachial vein using a sterile 30 gauge needle. The 20–50 ll blood sample was collected in plain glass capillary tubes and immediately placed in 1 ml of Queen’s Lysis buffer (Seutin et al. 1991). Bleeding was stopped with pressure before birds were released at the capture site. Individuals were genotyped at eight microsatellite markers on a LI-COR 4200 DNA analyzer. DNA extraction, polymerase chain reaction conditions and individual genotyping are described in detail elsewhere (Temple 2000; Chan and Arcese 2002; Wilson et al. 2008).
Conserv Genet (2011) 12:213–222
215
Fig. 1 Map of sampling sites of song sparrow populations. a depicts sampled M.m.morphna populations along the coastal mainland and Vancouver Island (southern BC): 1 Campbell River, 2 Powell River, 3 Texada Island, 4 Qualicum, 5 Sechelt, 6 Alaksen, 7 Duncan, 8 Sooke and 9 Padilla Bay. Inset shows sampled M.m.morphna populations in the Southern Gulf Islands: 10, 11 Shell islands, 12–14 Dock Islands, 15 Reay, 16 Sidney, 17 Mandarte and 18 Halibut. b depicts sampled
sites in the San Francisco Bay region where sampling including populations of M.m.gouldii (1,2), M. m. samuelis (3,4), M. m. maxillaris (5,6), M. m. heermanni (7) and M. m. pusillula (8,9). c depicts sampling sites in the Salton Sea population, where the black areas are the sampling range of M. m. heermanni individuals and the white areas are the sampling range of M.m.fallax individuals
Microsatellite data were available from previously published work for San Francisco Bay (214 individuals across nine sites and five subspecies: M. m. samuelis, M. m. maxillaris, M. m. pusillula, M. m. gouldii and M. m. heermanni, Chan and Arcese 2002, 2003, Fig. 1c), and the Salton Sea (58 individuals across two subspecies: M. m. heermanni and M. m. fallax, Fig. 1d).
autocorrelation coefficient r(h) indicates that individuals occurring at a particular distance class h, have a tendency to covary in the same direction from the sample genetic mean. If gene flow is limited by distance at the spatial scale of the study, a positive correlation (r(h)) is expected at the shorter distance classes, but this correlation will become zero at distance classes where drift predominates over gene flow. The smallest distance class at which the correlation coefficient intersects the axis of r = 0, is taken as an estimate of the genetic patch size (Sokal 1979; Peakall et al. 2003). All individuals were georeferenced, enabling us to calculate detailed inter-individual spatial distances. Spatial autocorrelation was calculated separately for San Francisco Bay, southern BC and Salton Sea providing three estimates of spatial autocorrelation over the same distance classes. We divided the sampling range into four distance classes: 0.5, 2, 10 and 25 km. The Salton Sea region only had three distance classes: 2, 10 and 25 km. Each distance class included all pair-wise distance values greater than the preceding class value and smaller than the focal distance class value. For the Southern Gulf Islands we used a distance class set of 0.5, 2, 5 and 10 km. Different distance class choices were evaluated, but did not alter the resulting interpretation. Statistical significance for the multivariate
Individual spatial autocorrelation analyses—within population genetic structuring Spatial autocorrelation coefficients can be used in population genetic studies to provide a measure of the genetic similarity of individuals as a function of the geographic distance between them. We analyzed the scale of spatial genetic structure using multivariate spatial autocorrelation analysis (Smouse and Peakall 1999) as implemented in GENALEX version 6.01 (Peakall and Smouse 2006). This method calculates an autocorrelation coefficient (r(h)) among individuals which are separated by specified distance class intervals (h). The autocorrelation coefficient reflects the tendency of individuals in particular distance classes to covary in the same, opposite or in random directions from the genetic mean. Therefore, a positive
123
216
spatial autocorrelation is based on 95% confidence intervals generated by bootstrapping around the estimated r for each distance class and a 95% null interval is generated around the value of r = 0. A statistically significant autocorrelation is one whose 95% confidence intervals do not overlap with the 95% confidence intervals of the null. For all analyses, we used 1000 permutations and bootstrap iterations. Spatial genetic relationships were visualized using the genetic landscape interpolation analysis available in ALLELES IN SPACE v. 1.0 (Miller 2005). The genetic surface is based on inter-individual distances of sampled individuals and interpolated distances in areas where individuals were not sampled. Across the genetic landscape, the peaks and troughs indicate high and low genetic distances between individuals respectively. Population level genetic structure Population-level genetic structure was examined using partial Mantel tests, generalized linear mixed models and Bayesian clustering analyses. Mantel tests have already been utilized for the San Francisco Bay and Salton Sea populations (Chan and Arcese 2003; Pruett et al. 2008) so those particular analyses were not repeated for those two areas. We calculated genetic differentiation using a standardized F’ST, which was particularly valuable for this study as our microsatellite loci are highly variable, which deflates the theoretical range of FST values (Hedrick 1999, 2005). We standardized our FST values against the FSTMAX value (Hedrick 2005), which was calculated with the program RECODEPC v. 0.1 (Meirmans 2006). Pairwise FST values (Weir and Cockerham 1984) and significance were calculated based on 1000 permutations using FSTAT v 2.9.3.2 (Goudet 2001). Standardized divergence measures (F0ST ) were used in partial Mantel tests as implemented in IBDWS v. 3.15 (Jensen et al. 2005). Statistical significance was calculated based on 20,000 permutations. For the BC and San Francisco Bay, the geographic distance between populations was calculated both as the shortest straight path between populations (straight distance) and the shortest distance path around water barriers (effective distance). Within southern BC, we controlled for the water barrier imposed by the Georgia Strait by calculating effective distance using either a northern stepping stone pattern over the Broughton Archipelago or a southern route over the San Juan/Gulf Islands. Among the BC Southern Gulf Island populations, we calculated inter-individual distances at the distance from island shore to island shore. In addition to geographical distance, we examined the influence of population size on genetic structure. Local population densities as estimated from smoothed abundance grid data from the
123
Conserv Genet (2011) 12:213–222
Breeding Bird Survey (BBS, Sauer et al. 2008) were included in models for BC and San Francisco Bay. Exact estimates of Southern Gulf Island population sizes were available from ongoing demographic studies (Wilson and Arcese 2008). As a complement to the partial Mantel tests, we also used generalized linear mixed models to examine how population size may influence genetic structure. We combined the San Francisco Bay and Vancouver data and restricted our model testing to three factors: relative density, intra- or inter-specific comparison and region. All models included the focal population as a random factor. These five models were evaluated against an intercept-only model. Model selection was based on the Akaike’s Information Criterion (AICc) where we considered models with DAICc B 2 to be well supported by the data (Burnham and Anderson 2002). All statistical analyses were conducted using the Program R (version 2.8.1). We also implemented the Bayesian clustering program STRUCTURE v 2.3.3 (Pritchard et al. 2000, Falush et al. 2003) to estimate the number of genetic clusters (K) within the British Columbian samples (Fig. 1a, b). Pruett et al. (2008) determined that there were five major genetic clusters within San Francisco Bay and Salton Sea populations, so these analyses were not repeated here. To analyze the British Columbian samples, we used the admixture model with correlated allele frequencies without using sampling locality as a prior. We ran STRUCTURE for 20 replicates at each K value ranging from 1 to 18, the latter of which was the number of sampling localities. Each STRUCTURE run consisted of an initial burn-in of 105, with 5 9 106 iterations. The most supported value of K was inferred from the posterior probabilities (Pritchard and Wen 2003) and the DK method of Evanno et al. (2005). We also calculated a ‘clusteredness’ index (G) sensu Rosenberg et al. (2005), which reflects the overall partitioning of membership coefficients of individuals (q) within each population across the inferred clusters. The G index ranges from equal assignment across clusters (G = 0) to high proportionate assignment to a one or a few clusters (G = 1) (Rosenberg et al. 2005). The program STRUCTURE HARVESTER v0.3 was used to process the STRUCTURE results files (Earl 2009) and CLUMMP V1.2.2 (Jakobsson and Rosenberg 2007) was used to summarize across the runs for the most probable K value.
Results Individual spatial autocorrelation analyses Within both San Francisco Bay and southern BC, genetic and geographic distances were correlated at distance
Conserv Genet (2011) 12:213–222
classes less than 10 km. In San Francisco Bay, the correlation at the shortest distance class (B0.5 km) was considerably higher (r = 0.039, 95% CI: 0.045, 0.033, Fig. 2a) than the correlation at the same distance class within the southern BC populations (r = 0.020, 95% CI: 0.026, 0.015, Fig. 2b) suggesting reduced gene flow in San Francisco Bay. In southern BC and San Francisco Bay regions, however, the autocorrelation coefficient was notsignificant at a distance class of B10 km, providing an estimate of a genetic patch size that falls within a range of 2–10 km for both regions. Within the Salton Sea, the spatial correlation at the 0.5 km distance class was relatively high (r = 0.067, 95% CI: 0.173–0.015, Fig. 2c), but due to limited sample size, the 95% confidence interval of the estimate included zero, precluding a robust estimate of genetic patch size in the region. Within the Southern Gulf Islands, the correlation at the first distance class of B0.5 km was high (r = 0.077, 95% CI: 0.082–0.072, Fig. 2d), but was not statistically significant at distance classes beyond 2 km. The STRUCTURE cluster analyses identified the most probable model as K = 3. Under this model, cluster membership distributions (G) were similar across individuals on Vancouver Island (G = 0.51), the coastal mainland (G = 0.48) and Texada Island (G = 0.48) (Fig. 3), where membership coefficients were skewed towards two common clusters. However, cluster membership assignment was more skewed among individuals in the Southern Gulf Island populations on Dock Islands (G = 0.58), Halibut (G = 0.42), Sidney (G = 0.45), and most notably on Mandarte Island (G = 0.70), where individuals have disproportionate assignments to one of the three clusters (Fig. 3). The genetic landscape within the coastal mainland of BC also has particular areas of divergence, indicated by the higher peaks occuring in the Campbell River and Qualicum populations. The grid size and weighting parameters had no effect on the relative shape of the landscape surface (Fig. 4a). The genetic landscape surface within the San Francisco Bay also depicted the strong divergence of the M.m.pusillula (Mmp1 and Mmp2) and M.m.hermanni (Mmh1) population (Fig. 4b). Population level genetic structure The Mantel test revealed no correlation between the straight-line geographic and genetic distance (F0ST ) between populations (r = -0.036, P = 0.56), even after accounting for populations separated by the Strait of Georgia (r = -0.011, P = 0.54). In contrast, we found strong IBD in the Southern Gulf Islands, after correcting for variation in island size (r = 0.62, P = 0.0003) or population size (r = 0.62, P = 0.0003).
217
In our generalized linear mixed model analysis, the most supported model of factors affecting genetic structure was a model including subspecies and focal population. Genetic divergence between populations of different subspecies was greater than within subspecies (b = -0.089 SE: 0.016).
Discussion This study showed that fine-scale genetic structuring occurs within song sparrow populations, and that the scale of this structuring (termed as genetic patch size) is less than 10 km. The genetic patch size was comparable between BC and San Francisco Bay populations, although the latter had higher genetic distances over smaller geographic distances. The higher structuring among San Francisco Bay populations is likely due to restricted gene flow across subspecific boundaries (Chan and Arcese 2003). However, small island populations in BC also showed considerable structuring, with genetic patch sizes of less than 2 km, demonstrating that landscape barriers can lead to differences in genetic structure between insular and continuous populations even within a subspecies. Multivariate spatial autocorrelation is infrequently used within vertebrate genetic surveys (but see Peakall et al. 2003), so there are few studies available to compare the genetic patch size found in this study. However, when this method was applied within populations of white-breasted thrashers (Ramphocinclus brachyurus, Temple et al. 2006) and superb fairy wrens (Malurus cyaneus, Double et al. 2005), positive genetic structure was detected at scales of less than 200 m. Therefore, passerine populations can have detectable fine-scale genetic structuring that can be of importance for correctly interpreting population genetic structure (Smouse and Peakall 1999; Schwartz and McKelvey 2008). However, the difference in the strength of the spatial autocorrelation among song sparrow populations, demonstrates that spatial genetic structure shows geographical variation as a result of localized factors such as water barriers or subspecific boundaries. Water barriers have been shown to restrict movement in other avian taxa over distances as short as a few kilometers (Paine 1985; Smith et al. 2005; Hayes and Sewlal 2004), and our genetic data suggests song sparrows are similarly limited. The shorter genetic patch width and IBD genetic structuring found among the Southern Gulf Island populations provides further evidence that small water barriers (i.e.\7 km) can have detectable genetic effects. Our markrecapture and genetic data were concordant, as Mandarte Island, which has the lowest immigration rate (Wilson and Arcese 2008), showed the highest genetic divergence and skewed clustering. The genetic structure among islands
123
218
Conserv Genet (2011) 12:213–222
Fig. 2 Correlogram of the correlation coefficient (r) between genetic and geographic distance at four distance classes. Correlation coefficients were calculated at the individual level for song sparrow populations within: a San Francisco Bay b Salton Sea, c BC, d Southern Gulf Islands. The permuted 95% confidence interval (dashed lines) around the null of r = 0, and the bootstrapped 95% confidence intervals around the correlation for each distance class are also shown
was likely mediated by dispersal patterns and not genetic drift because distance, and not population size, was most strongly correlated with genetic structure. Despite fine-scale structuring, there was an absence of IBD among the continuous, non-insular populations. The absence of IBD may be because the interval between sampled populations exceeded the 10 km genetic patch size (Smouse and Peakall 1999). If populations had been sampled along a finer gradient of distances, perhaps an IBD pattern would have been detected. At the broader scale, we were not able to demonstrate a genetic discontinuity that spatially coincided with the Georgia Strait. The only indication of a barrier signal was that the two sampled populations with the largest genetic distance (Qualicum and Sechelt) are situated directly across the Strait from one another. There is a gradual increase in divergence among populations located along the northeastern edge of Vancouver Island (Fig. 4) from Sooke to Campbell River, which likely corresponds to
123
reductions in suitable habitat for M.m.morphna as BBS densities were lower in Campbell River. To our knowledge, no other genetic study in the Pacific Northwest has adopted a comparable sampling design to evaluate water barriers. The majority of genetic studies in this area have sampled at a broad geographic scale, and so evaluate insular divergence based on a single island population. Examples from Vancouver Island are grey wolves (Canis lupus, Geffen et al. 2004), wapiti (Cervus elaphus, Polziehn et al. 2000) and American marten (Martes americana,) that were genetically divergent from the mainland, supporting the prediction that the Georgia Strait may limit dispersal for some taxa. In the case of grey wolves, low dispersal rates across the Georgia Strait led to a near 20 year absence from Vancouver Island, following extirpation in the 1950s (Scott and Shackleton 1982). Low recolonization and small population size led to allee effects in the recovering populations (Mun˜oz-Fuentes et al. 2010) (Fig. 5).
Conserv Genet (2011) 12:213–222
219
Fig. 3 Results of STRUCTURE analyses for coastal mainland and Southern Gulf Islands. A model of K = 3 was most supported. Each column represents an individual where cluster membership assignment is on the y-axis. Population codes are 1 Campbell River, 2 Powell River, 3 Texada Island, 4 Qualicum, 5 Sechelt, 6 Alaksen, 7 Duncan, 8 Sooke and 9 Padilla Bay. Inset shows sampled M.m.morphna populations in the Southern Gulf Islands: 10, 11 Shell, 12–14 Docks, 15 Reay, 16 Sidney, 17 Mandarte and 18 Halibut
Fig. 4 Genetic landscape for a M.m.morphna in southern BC and b San Francisco Bay. Genetic landscapes are shown on the left and corresponding map of sampling area is on the right. Genetic surfaces were calculated by interpolation based on a distance weighting value of 1, with a 50 9 50 grid. Surface heights are proportionate to the genetic distance of the population. Population codes in southern BC correspond to 1 Campbell River, 2 Powell River, 3 Texada Island, 4 Qualicum, 5 Sechelt, 6 Alaksen, 7 Duncan, 8 Sooke and 9 Padilla Bay. Population codes in San Francisco Bay correspond to M.m.gouldii (1,2), M. m. samuelis (3,4), M. m. maxillaris (5,6), M. m. heermanni (7) and M. m. pusillula (8,9)
Haida Gwaii and the Alexander Archipelago, are located 48 and 16 km respectively off- shore, and are both recognized endemism hotspots (Cook and MacDonald 2001; Cook et al. 2006) with multiple examples of genetically
divergent populations (Bidlack and Cook 2002; Lucid and Cook 2004, Topp and Winker 2008). In Europe, strong divergence between Spanish and Moroccan populations of great buzzards (Otis tarda; Broderick et al. 2003) and
123
220
Fig. 5 Relationship between the pair-wise genetic distance ðF0ST =ð1 F0ST ÞÞ and log geographic distance (km) for San Francisco Bay (black circles and black triangles) and southern BC (grey circles) populations. Comparisons between populations of the same subspecies are shown with circles, while comparison between populations of different subspecies are shown with triangles
Greater Mouse-eared bat (Myotis myotis, Castella et al. 2000) has been shown across the 16 km expanse of the Strait of Gibraltar. The influence of a particular dispersal barrier can be surprising, for instance the 48 km land barrier of the Isthmus of Panama is a greater dispersal impediment for seabirds than expanses of hundreds of kilometers of the Pacific Ocean (Steeves et al. 2005). Dispersal barriers that are inconsequential over an evolutionary time-frame, may be relevant over ecological time, but it is the latter that can be the most difficult to detect because genetic divergence may be low (Waples and Gaggiotti 2006). Many commonly used analytical methods have difficulty detecting structure and inferring rates of gene flow in species or populations with low genetic divergence (Waples 1998, Latch et al. 2006). In our study, the absence of a barrier signal for the Georgia Strait in a sedentary songbird is an example where a biologically significant barrier did not leave a statistically significant signal (Waples 1998). Dispersal data from mainland and insular song sparrow populations (Nice 1937; Halliburton and Mewaldt 1976; Smith et al. 2006; Wilson and Arcese 2008) suggest that the Georgia Strait almost certainly restricts dispersal over an ecological time-frame. However, song sparrows in this region have low genetic divergence due to recent colonization or a large Ne, reducing the power to detect barriers using typical numbers of individuals and loci. Despite the absence of a genetic discontinuity associated with the Georgia Strait, there were detectable genetic consequences, as diversity declined in both Vancouver Island and Southern Gulf Island populations (Wilson et al. 2009). Through spatially extensive sampling within a single subspecies, we were able to show that microspatial genetic
123
Conserv Genet (2011) 12:213–222
structuring can exist in a vagile taxon such as a bird, particularly for insular populations. Dispersal occurring over this microspatial scale is of importance for short-term conservation (Sunnucks and Taylor 2008), and our study highlights the influence of water barriers in restricting dispersal among insular populations. These dispersal restrictions were correlated with reductions in genetic diversity in this species (Wilson et al. 2009) and could potentially lead to the establishment of local genetic genetic architecture (i.e. co-adapted gene complexes created through drift), resulting in outbreeding depression (Templeton 1986, Knowlton and Jackson 1993), which has been reported in our study population (Marr et al. 2002). At the microevolutionary level, our data on the divergence patterns within a single subspecies due to landscape variation is a biologically meaningful baseline that can be used to interpret divergence in populations of adaptively divergent song sparrows (Basham and Mewaldt 1987) that are also listed as Federal Special Concern Species. The genetic divergence between M.m.morpha populations separated by over 200 km is lower than the genetic differences found between subspecies that are less than 10 km apart in southern California. These differences suggest restricted dispersal occurs and/or populations are small and fragmented, which refute the assertion the genetic divergence between these Californian subspecies is inconsequential for conservation (Zink 2010). Spatial autcorrelation analyses can provide conservation biologists with crucial information regarding landscape permeability and dispersal processes (Diniz and Telles 2002; Sunnucks and Taylor 2008; Schwartz and McKelvey 2008). If populations are sampled across a variety of scales and potential barriers, a comprehensive insight can be gained regarding how landscape permeability is shaping the genetic structure of a species. Acknowledgments We thank S. Wilson, A. Marr, L.F. Keller, J.M.N Smith, C. Begus, C. Ritland, A. Miscampbell, H. Yueh, W. Easton, and J. Hope for assistance on various aspects of this study. Site access was provided by the Tsawout and Tseycum Bands, A. and H. Brumbaum, T. and M. Boyle, Parks Canada, Canadian Wildlife Service, Nature Trust, Powell River Municipality, and Sooke Municipal District. We gratefully acknowledge financial support from NSERC, American Ornithologists Union Bleitz Grant and the Friends of Ecological Reserves. All work was conducted under permits of the UBC Animal Care Committee and Environment Canada. This manuscript benefitted from comments made by V. Friesen, E. Taylor, D. Irwin, S. Aitken, K. Cheng, S. Wilson and two anonymous reviewers.
References Aldrich JW (1984) Ecogeographical variation in size and proportions of Song Sparrows (Melospiza melodia). Ornithological Monographs 35 Arcese P, Sogge MK, Marr AB, Patten MA (2002) Song Sparrow (Melospiza melodia), The Birds of North America Online. In: Poole A (ed) Cornell Lab of Ornithology, Ithaca
Conserv Genet (2011) 12:213–222 Basham MP, Mewaldt LR (1987) Salt-water tolerance and the distribution of south San Francisco Bay song sparrows. Condor 89:697–709 Bidlack AL, Cook JA (2002) A nuclear perspective on endemism in northern flying squirrels (Glaucomys sabrinus) of the Alexander Archipelago, Alaska. Conserv Genet 3:247–259 Bohonak AJ (1999) Dispersal, gene flow and population structure. Quart Rev Biol 74:21–45 Broderick D, Idaghdour Y, Korrida A, Hellmich J (2003) Gene flow in great bustard populations across the Strait of Gibraltar as elucidated from excremental PCR and mtDNA sequencing. Conserv Genet 4:793–800 Burnham KP, Anderson DR (2002) Model selection and inference: a practical information-theoretic approach. Springer, New York Campbell RW, Dawe NK, McTaggart-Cowan I, Cooper JM, Kaiser GW, McNall MC (1990) Birds of British Columbia, vol II. Royal British Columbia Museum, Victoria, British Columbia, Canada Castella V, Ruedi M, Excoffier L, Ibanez C, Arlettaz R, Hausser J (2000) Is the Gibraltar Strait a barrier to gene flow for the bat Myotis myotis (Chiroptera: Vespertilionidae)? Mol Ecol 9:1761–1772 Chan Y, Arcese P (2002) Subspecific differentiation and conservation of song sparrows (Melospiza melodia) in the San Francisco Bay region inferred by microsatellite loci analysis. Auk 119:641–657 Chan Y, Arcese P (2003) Morphological and microsatellite differentiation in Melospiza melodia (Aves) at a microgeographic scale. J Evol Biol 16:939–947 Cook JA, MacDonald SO (2001) Should endemism be a focus of conservation efforts along the North Pacific Coast of North America? Biol Conserv 97:207–213 Cook JA, Dawson NG, MacDonald SO (2006) Conservation of highly fragmented systems: the north temperate Alexander Archipelago. Biol Conserv 133:1–15 Cooper ML (2000) Random amplified polymorphic DNA analysis of southern brown bandicoot (Isoodon obesulus) populations in Western Australia reveals genetic differentiation related to environmental variables. Mol Ecol 9:469–479 Crispo E, Bentzen P, Reznick DN et al (2006) The relative influence of natural selection and geography on gene flow in guppies. Mol Ecol 15:49–62 Crochet PA (2000) Genetic structure of avian populations—allozymes revisited. Mol Ecol 9:1463–1469 Darling LM (2003) Status of the Vancouver Island Northern Pygmyowl (Glacidium gnoma swarthi) in British Columbia. Wildlife Bulletin B-113. B.C. Ministry of Sustainable Resource Management. Conservation Data Centre, and B.C. Ministry of Water, Land and Air Protection, Biodiversity Branch, Victoria, BC. 14 pp Darling JA, Folino-Rorem NC (2009) Genetic analysis across different spatial scales reveals multiple dispersal mechanisms for the invasive hydrozoan Cordylophora in the Great Lakes. Mol Ecol 18:4827–4840 Deyoung RW, Zamorano A, Mesenbrink BT et al (2009) Landscapegenetic analysis of population structure in the Texas Gray Fox Oral Rabies Vaccination Zone. J Wildl Manage 73:1292–1299 Diniz JAF, Telles MPD (2002) Spatial autocorrelation analysis and the identification of operational units for conservation in continuous populations. Conserv Biol 16:924–935 Double MC, Peakall R, Beck NR, Cockburn A (2005) Dispersal philopatry and infidelity: dissecting local genetic structure in superb fairy-wrens (Malurus cyaneus). Evolution 59:625–635 Earl DA (2009) STRUCTURE HARVESTER v0.3. http://users.soe.ucsc. edu/*dearl/software/struct_harvest/ Epperson BK (2003) Geographical genetics. Princeton University Press, Princeton New Jersey Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol 14:2611–2620
221 Falush D, Stephens M, Pritchard JK (2003) Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164:1567–1587 Friesen VL, Montevecchi WA, Gaston AJ et al (1996) Molecular evidence for kin groups in the absence of large-scale genetic differentiation in a migratory bird. Evolution 50:924–930 Funk WC, Blouin MS, Corn PS et al (2005) Population structure of Columbia spotted frogs (Rana luteiventris) is strongly affected by the landscape. Mol Ecol 14:483–496 Geffen E, Anderson MJ, Wayne RK (2004) Climate and habitat barriers to dispersal in the highly mobile grey wolf. Mol Ecol 13:2481–2490 Goudet J (2001) Fstat a program to estimate and test gene diversities and fixation indices Version 293 http://wwwunilch/izea/ softwares/fstathtml Halliburton R, Mewaldt LR (1976) Survival and mobility in a population of pacific coast song sparrows (Melospiza melodia gouldii). Condor 78:499–504 Hampton JO, Spencer PBS, Alpers DL et al (2004) Molecular techniques, wildlife management and the importance of genetic population structure and dispersal: a case study with feral pigs. J Appl Ecol 41:735–743 Hayes FE, Sewlal JAN (2004) The Amazon River as a dispersal barrier to passerine birds: effects of river width, habitat and taxonomy. J Biogeog 31:1809–1818 Hedrick PW (1999) Perspective: highly variable loci and their interpretation in evolution and conservation. Evolution 53: 313–318 Hedrick PW (2005) A standardized genetic differentiation measure. Evolution 59:1633–1638 Hutchison DW, Templeton AR (1999) Correlation of pairwise genetic and geographic distance measures: Inferring the relative influences of gene flow and drift on the distribution of genetic variability. Evolution 53:1898–1914 Irwin DE, Bensch S, Irwin JH, Price TD (2005) Speciation by distance in a ring species. Science 307:414–416 Jensen JL, Bohonak AJ, Kelley ST (2005) Isolation by distance web service. Bmc Genet 6:13 Knowlton N, Jackson JBC (1993) Inbreeding and outbreeding in marine invertebrates. In: Thornhill NW (ed) The natural history of inbreeding and outbreeding. University of Chicago Press, Chicago, IL, pp 200–249 Latch EK, Dharmarajan G, Glaubitz JC, Rhodes OEJ (2006) Relative performance of Bayesian clustering software for inferring population substructure and individual assignment at low levels of population differentiation. Conserv Genet 7:295–302 Lee PLM, Bradbury RB, Wilson JD et al (2001) Microsatellite variation in the yellowhammer Emberiza citrinella: population structure of a declining farmland bird. Mol Ecol 10:1633–1644 Lucid MK, Cook JA (2004) Phylogeography of Keen’s mouse (Peromyscus keeni) in a naturally fragmented landscape. J Mamm 85:1149–1159 Meirmans PG (2006) Using the AMOVA framework to estimate a standardized genetic differentiation measure. Evolution 60: 2399–2402 Miller MP (2005) Alleles in Space (AIS): computer software for the joint analysis of inter-individual spatial and genetic information. J Hered 96:722–724 Mun˜oz-Fuentes V, Darimont CT, Paquet PC, Leonard JA (2010) The genetic legacy of extirpation and re-colonization in Vancouver Island wolves. Conserv Genet 11:547–556 Nice MM (1937) Studies in the life history of the song sparrow I: a population study of the song sparrow. Trans Linnean Soc New York 6:1–328 Paine RT (1985) Reestablishment of an insular winter wren population following a severe freeze. Condor 87:558–559
123
222 Patten MA, Pruett CL (2009) The Song Sparrow, Melospiza melodia, as a ring species: patterns of geographic variation, a revision of subspecies, and implications for speciation. Syst Biodivers 7:33–62 Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in Excel Population genetic software for teaching and research. Mol Ecol Notes 6:288–295 Peakall R, Ruibal M, Lindenmayer DB (2003) Spatial autocorrelation analysis offers new insights into gene flow in the Australian bush rat Rattus fuscipes. Evolution 57:1182–1195 Pe´rez-Espona S, Pe´rez-Barberı´a FJ, Mcleod JE et al (2008) Landscape features affect gene flow of Scottish Highland red deer. Mol Ecol 4:981–996 Polziehn RO, Hamr J, Mallory FF, Strobeck C (2000) Microsatellite analysis of North American wapiti (Cervus elaphus) populations. Mol Ecol 9:1561–1576 Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155:945– 959 Pritchard JK, Wen W (2003) Documentation for STRUCTURE software: version 2. Available from http://pritch.bsd.uchicago. edu Pruett CL, Arcese P, Chan YL et al (2008) Concordant and discordant signals between genetic data and described subspecies of Pacific coast song sparrows. Condor 110:359–364 Rosenberg NA, Mahajan S, Ramachandran S et al (2005) Clines, clusters, and the effect of study design on the inference of human population structure. PLoS Genet 1:e70 Ruegg K, Slabbekoorn H, Clegg S, Smith TB (2006) Divergence in mating signals correlates with ecological variation in the migratory songbird Swainson’s thrush (Catharus ustulatus). Mol Ecol 15:3147–3156 Sauer JR, Hines JE, Fallon J (2008) The North American Breeding Bird Survey, Results and Analysis 1966–2007. Version 5.15.2008. Accessed 1 May 2010 Schwartz MK, McKelvey KS (2008) Why sampling scheme matters: the effect of sampling scheme on landscape genetic results. Conserv Genet 10:441–452 Scoble J, Lowe AJ (2010) A case for incorporating phylogeography and landscape genetics into species distribution modelling approaches to improve climate adaptation and conservation planning. Divers Distrib 16:343–353 Scott BMV, Shackleton DM (1982) A preliminary study of the social organization of the Vancouver Island wolf. In: Harrington FH, Paquet P (eds) Wolves of the world. Noyes, Park Ridge, pp 12–25 Scribner KT, Blanchong JA, Bruggeman DJ et al (2005) Geographical genetics: conceptual foundations and empirical applications of spatial genetic data in wildlife management. J Wildl Manage 69:1434–1453 Seutin G, White BN, Boag PT (1991) Preservation of avian blood and tissue samples for DNA analyses. Can J Zool 69:82–90 Smith TB, Calsbeek R, Wayne RK et al (2005) Testing alternative mechanisms of evolutionary divergence in an African rain forest passerine bird. J Evol Biol 18:257–268
123
Conserv Genet (2011) 12:213–222 Smith JNM, Keller LF, Marr AB, Arcese P (2006) Conservation and biology of small populations. Oxford University Press, New York Smouse PE, Peakall R (1999) Spatial autocorrelation analysis of individual multiallele and multilocus genetic structure. Heredity 82:561–573 Sokal RR (1979) Ecological parameters inferred from spatial correlograms. In: Patil GP, Rosenzweig ML (eds) Contemporary quantitative ecology and related ecometrics. International Cooperative Publishing House, Fairland, MD, pp 167–196 Sokal RR, Wartenberg DE (1983) A test of spatial autocorrelation analysis using an isolation-by-distance model. Genetics 105: 219–237 Steeves TE, Anderson DJ, Friesen VL (2005) The Isthmus of Panama: a major physical barrier to gene flow in a highly mobile pantropical seabird. J Evol Biol 18:1000–1008 Sunnucks P, Taylor AC (2008) The application of genetic markers to landscape management. In: Pettit W, Cartright I, Bishop K, Lowell D, Pullar D, Duncan D (eds) Landscape analysis and visualisation: spatial models for natural resource management and planning. Springer, Berlin, Germany, pp 292–317 Temple, M (2000) Microsatellite analysis of extra-pair fertilization in the Ipswich sparrow (Passerculus sandwichensis princeps). M.Sc thesis, Dalhousie University Templeton AR (1986) Coadaptation and outbreeding depression. In: Soule´ ME (ed) Conservation biology: the science of scarcity and diversity. Sinauer Associates, Sunderland, MA, pp 105–116 Vignieri SN (2005) Streams over mountains: influence of riparian connectivity on gene flow in the Pacific jumping mouse (Zapus trinotatus). Mol Ecol 14:1925–1937 Waples RS (1998) Separating the wheat from the chaff: Patterns of genetic differentiation in high gene flow species. J Hered 89: 438–450 Waples RS, Gaggiotti O (2006) What is a population? An empirical evaluation of some genetic methods for identifying the number of gene pools and their degree of connectivity. Mol Ecol 15:1419– 1439 Watts PC, Rouquette JR, Saccheri J et al (2004) Molecular and ecological evidence for small-scale isolation by distance in an endangered damselfly, Coenagrion mercurial. Mol Ecol 13: 2931–2945 Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis of population-structure. Evolution 38:1358–1370 Wilson AG, Arcese P (2008) Influential factors for natal dispersal in an avian island metapopulation. J Avian Biol 36:341–347 Wilson AG, Arcese P, Keller L et al (2009) The contribution of island populations to in situ genetic conservation. Conserv Genet 10:419–430 Woxvold IA, Adcock GJ, Mulder RA (2006) Fine-scale genetic structure and dispersal in cooperatively breeding apostlebirds. Mol Ecol 15:3139–3146 Wright S (1943) Isolation by distance. Genetics 28:114–138 Zink RM (2010) Drawbacks with the use of microsatellites in phylogeography: the song sparrow Melospiza melodia as a case study. J Avian Biol 41:1–7
