Beyond Janzen's Hypothesis: How Amphibians That Climb Tropical Mountains Respond to Climate Variation.

Janzen's hypothesis (JH) posits that low thermal variation selects for narrow physiological tolerances, and thus small species distributional ranges and high species turnover along tropical elevational gradients. Although this hypothesis has been intensely revisited, it does not explain how many tropical species may exhibit broad distributions, encompassing altitudinal gradients. Moreover, the physiological responses of tropical species remain largely unknown, limiting our understanding on how they respond to climate variation. To fill these knowledge gaps, we tested a major component of JH, the climate variability hypothesis (CVH), which predicts broader thermal tolerance breadth (Tbr = CTmax - CTmin) with broader temperature variation. Specifically, we sampled populations of five amphibian species distributed in two mountain ranges in Brazil's Atlantic Forest to test how CTmin and CTmax vary along elevational gradients. Since both thermal and water balance traits are pivotal to the evolutionary history of amphibians, we also measured rates of dehydration and rehydration and their relations with thermal tolerances. We found that broader temperature variation with increasing altitude did not always lead to broader Tbr, since changes in CTmin and CTmax were species-specific. In addition, we found that water balance did not show consistent variation with altitude, also with low correlations between hydric and thermal traits. While we also found that highland populations are at lower risk of thermal stress than lowland counterparts, both are living far from their upper thermal limits. As a consequence of intraspecific variation in physiological traits and spatial variation in climate along altitude, responses to climate variation in tropical amphibian species were context-dependent and heterogeneous. Together with recent studies showing thermal tolerances of some tropical amphibians comparable to temperate taxa, our findings highlight that several responses to climate variation in tropical species may not conform to predictions made by either the CVH or other important hypotheses concerning physiological variation. This reinforces the need to overcome geographical bias in physiological data to improve predictions of climate change impacts on biodiversity.