The influence of breeding phenology on the genetic structure of four pond‐breeding salamanders

Abstract

Understanding metapopulation dynamics requires knowledge about local population dynamics and movement in both space and time. Most genetic metapopulation studies use one or two study species across the same landscape to infer population dynamics; however, using multiple co‐occurring species allows for testing of hypotheses related to different life history strategies. We used genetic data to study dispersal, as measured by gene flow, in three ambystomatid salamanders (Ambystoma annulatum, A. maculatum, and A. opacum) and the Central Newt (Notophthalmus viridescens louisianensis) on the same landscape in Missouri, USA. While all four salamander species are forest dependent organisms that require fishless ponds to reproduce, they differ in breeding phenology and spatial distribution on the landscape. We use these differences in life history and distribution to address the following questions: (1) Are there species‐level differences in the observed patterns of genetic diversity and genetic structure? and (2) Is dispersal influenced by landscape resistance? We detected two genetic clusters in A. annulatum and A. opacum on our landscape; both species breed in the fall and larvae overwinter in ponds. In contrast, no structure was evident in A. maculatum and N. v. louisianensis, species that breed during the spring. Tests for isolation by distance were significant for the three ambystomatids but not for N. v. louisianensis. Landscape resistance also contributed to genetic differentiation for all four species. Our results suggest species‐level differences in dispersal ability and breeding phenology are driving observed patterns of genetic differentiation. From an evolutionary standpoint, the observed differences in dispersal distances and genetic structure between fall breeding and spring breeding species may be a result of the trade‐off between larval period length and size at metamorphosis which in turn may influence the long‐term viability of the metapopulation. Thus, it is important to consider life history differences among closely related and ecologically similar species when making management decisions.

Landscape genetics reveals unique and shared effects of urbanization for two sympatric pool‐breeding amphibians

“… Broadscale patterns of genetic structure : We hypothesized broadscale patterns of genetic structure will be much stronger for wood frogs than spotted salamanders. Previous work to characterize the isolation‐by‐distance (IBD) relationships for these two species has supported this hypothesis through observations of clear positive correlations between geographic and genetic distances for wood frogs (Crosby et al, ; Peterman, Feist, Semlitsch, & Eggert, ; Richardson, ; Squire & Newman, ) and either high variance positive correlations (Burkhart et al, ; Peterman et al, ; Richardson, ; Zamudio & Wieczorek, ) or nonsignificant relationships (Purrenhage, Niewiarowski, & Moore, ; Whiteley, McGarigal, & Schwartz, ) for spotted salamanders. Additionally, we hypothesize that IBD relationships within species are not absolute, but are instead scale‐dependent, such that the strongest correlations between genetic isolation and distance will occur at some intermediate geographic scales of analysis.…”

Post‐Pleistocene differentiation in a Central Interior Highlands endemic salamander

“…These regions are dominated by temperate broadleaf and mixed forests (Ouachita Ecoregional Assessment Team, 2003;Ozarks Ecoregional Assessment Team, 2003). salamanders shows that effective population size and genetic diversity does not significantly differ among life stages (Peterman, Brocato, Semlitsch, & Eggert, 2016) and genetic inference is largely concordant among breeding seasons across the same landscape (Burkhart et al, 2017;Peterman et al, 2015). In total, we in Ambystoma spp.…”

Conventional methods for enhancing connectivity in conservation planning do not always maintain gene flow

“…Since it was not computationally feasible to use cross‐validation to limit over‐fitting, we used genetic algorithms to generate five preliminary single‐surface resistance maps for each species to identify potential variables for generating higher quality multisurface resistance maps (Supporting Information Table ; similar to Burkhart et al., ). Specifically, the MLPE models were refitted using restricted maximum likelihood, R 2 β statistics were calculated (Edwards, Muller, Wolfinger, Qaqish, & Schabenberger, ; using the r2 glmm r package; Jaeger, ), and variables were identified which adequately explained genetic variation for each species ( R 2 β ≥ 0.35).…”