Fragmentation of habitat may be one of the greatest negative impacts on species and has received a great deal of attention among conservation biologists (e.g. Saunders et al. 1991; Keyghobadi 2007). Among the variety of biological processes negatively affected by habitat fragmentation, one of the most important may be the geographic structure of genetic diversity and reduction in gene flow among populations. Populations in these habitats may be both reduced in size—resulting in genetic drift—and isolated—limiting gene flow (Nei 1973; Keyghobadi et al. 2005). Both of these processes erode individual heterozygosity and local genetic diversity over time, which may increase both regional and range-wide differentiation and inbreeding depression. This may ultimately reduce individual fitness, disease resistance and the ability of populations to adapt to local environmental changes (Keller & Waller 2002).
Although these impacts on populations are clearly causes for concern, it is not certain that the imprint of reduced gene flow was caused by anthropogenic habitat modification. Rather, species may have existed as a collection of isolated populations for thousands of generations. Several studies have shown that some taxa naturally occur as fragmented populations because of the effects of Pleistocene glaciations or changes in sea level (Sumner et al. 2004; Jordan & Snell 2008). Several studies have attempted to tease apart historical and contemporary aspects of gene flow using markers that evolve at various rates, acquiring collections made prior to agricultural modification, or comparing populations in fragmented and unfragmented landscapes (Johnson et al. 2003; Pertoldi et al. 2006; Martınez-Cruz et al. 2007). Finding markers suitable for investigating processes occurring at different timescales is not guaranteed, and for many species, particularly those preserved in formalin, most pre-agricultural material is not suitable for range-wide analyses with historical samples. Finally, some species simply no longer occur in unfragmented areas, making comparisons of population structure impossible.
Using the eastern massasauga rattlesnake (Sistrurus catenatus) as an exemplary taxon living in a fragmented landscape, Chiucchi & Gibbs (2010) ask whether contemporary population isolation, drift and interpopulational gene flow at regional and range-wide scales is incidental—stemming from historical isolation prior to colonization of North America by Europeans. The eastern massasauga (Fig. 1) is a medium-sized venomous snake that occurs throughout central North America and generally lives in unmodified habitat composed of prairie, swamp, bogs, grassy meadows or rolling plains (Campbell & Lamar 2004). As a federal candidate to be classified as threatened or endangered by the US Fish and Wildlife Service and as a federal Species at Risk in Canada, this taxon no longer occurs in any broadly continuous, unfragmented habitat, and few or no appropriate pre-agrarian historical genetic samples exist. With 19 species-specific microsatellite loci taken from 388 individuals across 19 contemporary localities, Chiucchi & Gibbs (2010) compared both contemporary and historical gene flow and population structure. Using STRUCTURE (Pritchard et al. 2000) and several other tests, the authors demonstrated that populations do exhibit significant structure locally and regionally. In fact, noticeable population structure was detected between populations separated by only 7 km.
Given the signature of population isolation, the paper tests the two hypotheses (e.g., contemporary vs. historical migration) using BAYESASS (Wilson & Rannala 2003) to examine gene flow over the last few generations and MIGRATE (Beerli 2002) to assess gene flow averaged over a period from the present to a time prior to European colonization (approximately 10 000 years). Therefore, it is expected that large differences between the estimates of migration using these two methods will result if human habitat modification was primarily responsible for contemporary structure. Surprisingly, differences between both estimates of migration were centred on zero (Fig. 2) and were significantly correlated using a Mantel test. This suggests that these snakes have existed in isolated populations characterized by low gene flow since the Pleistocene. Consistent with reduced gene flow over the long periods of time, Chiucchi & Gibbs (2010) demonstrated with another coalescent method (2mod; Ciofi et al. 1999) that populations have existed in a drift–migration equilibrium over a long time span. Moreover, even though populations show a > 10 × difference in effective population size on a regional scale, they do not show declines over time.
While reduced, local levels of gene flow are an interesting result in the massasaugas, and this may not be atypical among snakes (King 2009), particularly rattlesnakes, which often occur in small population sizes and have long generation times. Few studies have been conducted on par with the one presented in this issue, but interestingly, a paper on the isolated populations of the threatened timber rattlesnake (Crotalus horridus) in Upstate NY demonstrated that while philopatry and communal hibernacula are common among populations, gene flow among these sites is actually quite high as evidence from microsatellite paternity studies revealed (Clark et al. 2007). Therefore, trends among related species cannot be generalized, particularly in the case of S. catenatus where low gene flow is a historical trend.
For massasaugas, limited dispersal and population isolation even at close distances may be the historical norm. While extreme philopatry may be one explanation (Gibbs & Weatherhead 2001), regional isolation, fragmentation of habitat and local extinction caused by the Wisconsin Glacial Episode may also be an alternative explanation to account for historically low migration. With respect to conservation, it remains to be seen whether these currently isolated populations are suffering from reduced fitness owing to moderate levels of inbreeding. However, if they have persisted in this isolated state historically, it is possible that genetics relative to ecological factors is less important for long-term survival.
In the current biodiversity crisis, it seems reasonable to understand if and how species survive in fragmented populations over long periods of time despite the negative impacts on population size and genetic structure predicted to occur in isolation. Chiucchi & Gibbs (2010) add to an important body of evidence that demands historical factors that influenced modern population structure be addressed when examining at-risk species. Luckily, these authors have provided a strong method with which to do this.