Parallel and non-parallel phenotypic responses to environmental variation across Lesser Antillean anoles.

Species whose ranges encompass substantial environmental variation should experience heterogeneous selection, potentially resulting in local adaptation. Repeated covariation between phenotype and environment across ecologically similar species inhabiting similar environments provides strong evidence for adaptation. Lesser Antillean anoles present an excellent system in which to study repeated local adaptation because most species are widespread generalists occurring throughout environmentally heterogenous island landscapes. We leveraged this natural replication to test the hypothesis that intraspecific variation in phenotype (coloration and morphology) is consistently associated with environment across 9 species of bimaculatus series anoles. We measured dorsal coloration from 173 individuals from 6 species and 16 morphological traits from 883 individuals from 9 species, spanning their island ranges. We identified striking, but incomplete, parallelism in dorsal coloration associated with annual precipitation in our study species. By contrast, we observed significant patterns of morphological isolation-by-environment in only 2 species and no signal of parallel morphological evolution. Collectively, our results reveal strong divergent natural selection by environment on dorsal coloration but not morphology.

Reservoir frogs: seasonality of Batrachochytrium dendrobatidis infection in robber frogs in Dominica and Montserrat.

Emerging infectious diseases are an increasingly important threat to wildlife conservation, with amphibian chytridiomycosis, caused by Batrachochytrium dendrobatidis, the disease most commonly associated with species declines and extinctions. However, some amphibians can be infected with B. dendrobatidis in the absence of disease and can act as reservoirs of the pathogen. We surveyed robber frogs (Eleutherodactylus spp.), potential B. dendrobatidis reservoir species, at three sites on Montserrat, 2011-2013, and on Dominica in 2014, to identify seasonal patterns in B. dendrobatidis infection prevalence and load (B. dendrobatidis genomic equivalents). On Montserrat there was significant seasonality in B. dendrobatidis prevalence and B. dendrobatidis load, both of which were correlated with temperature but not rainfall. B. dendrobatidis prevalence reached 35% in the cooler, drier months but was repeatedly undetectable during the warmer, wetter months. Also, B. dendrobatidis prevalence significantly decreased from 53.2% when the pathogen emerged on Montserrat in 2009 to a maximum 34.8% by 2011, after which it remained stable. On Dominica, where B. dendrobatidis emerged seven years prior to Montserrat, the same seasonal pattern was recorded but at lower prevalence, possibly indicating long-term decline. Understanding the dynamics of disease threats such as chytridiomycosis is key to planning conservation measures. For example, reintroductions of chytridiomycosis-threatened species could be timed to coincide with periods of low B. dendrobatidis infection risk, increasing potential for reintroduction success.

Assessing population viability while accounting for demographic and environmental uncertainty.

Predicting the future trend and viability of populations is an essential task in ecology. Because many populations respond to changing environments, uncertainty surrounding environmental responses must be incorporated into population assessments. However, understanding the effects of environmental variation on population dynamics requires information on several important demographic parameters that are often difficult to estimate. Integrated population models facilitate the integration of time series data on population size and all existing demographic information from a species, allowing the estimation of demographic parameters for which limited or no empirical data exist. Although these models are ideal for assessments of population viability, they have so far not included environmental uncertainty. We incorporated environmental variation in an integrated population model to account for both demographic and environmental uncertainty in an assessment of population viability. In addition, we used this model to estimate true juvenile survival, an important demographic parameter for population dynamics that is difficult to estimate empirically. We applied this model to assess the past and future population trend of a rare island endemic songbird, the Montserrat Oriole Icterus oberi, which is threatened by volcanic activity. Montserrat Orioles experienced lower survival in years with volcanic ashfall, causing periodic population declines that were compensated by higher seasonal fecundity in years with high pre-breeding season rainfall. Due to the inclusion of both demographic and environmental uncertainty in the model, the estimated population growth rate in the immediate future was highly imprecise (95% credible interval 0.844-1.105), and the probability of extinction after three generations (in the year 2028) was low (2.1%). This projection demonstrates that accounting for both demographic and environmental sources of uncertainty provides a more realistic assessment of the viability of populations under unknown future environmental conditions.