No abstract

A species’ thermal sensitivity and its exposure to climate variation are key components in the prediction of its vulnerability to climate change. We tested the thermal sensitivity of a tropical amphibian that lives in a mild constant climate in which the thermal tolerance range is expected to closely match the experienced environmental temperature. The air temperature that this species is exposed to varies between 21.9 and 31.6°C with an annual mean of 27.2°C. We estimated the microhabitat water temperature variation under vegetation shade, which buffers the temperature by 1.8°C in relation to that of the air, and with open canopy, where the water was 1.9°C warmer than the air temperature. With broods of tadpoles split into five treatments (15°C, 21°C, 28°C, 31°C, and 33°C), we estimated the critical thermal maximum (CTMax) and critical thermal minimum (CTMin) after at least 7 days of acclimation. Both CTMax (42.3°C) and CTMin (11.8°C) were more extreme than the temperature range estimated for the field. We estimated the optimum temperature (To = 28.8°C) and the thermal performance breadth (range: 23.3–34.1°C) based on growth rate (g/day). The animals were able to acclimate more extensively to cold than to warm temperatures. These performance curve traits closely matched the air temperature. The estimated vulnerability varied according to the microhabitat prediction model used. The combination of tadpole data on thermal sensitivity and macro‐ and microhabitat variation provides a necessary framework to understand the effects of climate change on tropical amphibians.

The advertisement call of Dermatonotus muelleri was originally described by Nelson (1973) in a brief section of a review on the mating calls of the Microhylinae. He used two calls from São Leopoldo, state of Minas Gerais, in Brazil to determine that they have i) dominant frequency between 1.500-2.200 kHz (mean 1.854 + 0.216 kHz), and ii) harmonic intervals between 0.140 and 0.150 kHz (0.146 +/- 0.005 kHz). Nelson (1973) based his description on an audiospectrogram produced with high frequency resolution and did not quantify the pulse structure of the calls. More recently, Giaretta and colleagues (2013) expanded on the original description using a larger set of calls recorded from Gurinhat, state of Minas Gerais, in Brazil. They quantified the temporal structure of the call and confirmed that the dominant frequency is around 1.8 kHz. In addition, they identified a secondary low frequency band at 667 Hz.

No abstract