Rate of Weight Loss of Tree Frogs at Various Temperatures and Humidities

Sixteen Hyla regilla, four each from four localities, were desiccated at each of four temperatures (10, 15, 20, 25°C); each animal was weighed five time in 23 hr. Three of the frogs maintained at 25°C died after a loss of approximately 36 to 52% of body weight during 23 hr; none died at lower temperatures. The rate of weight loss was significantly higher at high temperatures than at low temperatures and increase logarithmically with temperature. The total difference between localities was not statistically significant. Fifteen combinations of temperature (5,10, 15, 20, 25°C) and humidity (35 to 40, 50 to 55, 60 to 80% R.H.) were successively established in a plant growth chamber. The mean rate of weight loss of 6 Hyla regilla from each of 11 localities and 6 H. californiae from 1 locality was obtained at each set of conditions. The effect on rate of weight loss of a given increase in temperature is greater at low humidities than at high humidities. The rate of weight loss was adjusted by initial weight and by eigenvector values from principal component of morphological values. The adjusted rates were than analyzed by split—plot design analysis of variance. The sample means were found to differ significantly with respect to temperature, humidity and locality. The interaction of locality and temperature of humidity was not significant; temperature and humidity did have a significant interaction. Multivariate discriminant analysis separated the population into three groups. One group consisted of frogs from southern California desert and Baja California localities. The second group consisted of frogs from southern Carolina coast and mountains, and frogs from Yosemite. The third consisted of frogs from the northern part of the range. The rate of weight loss of different localities demonstrates a relation between latitude and some linear function of the first two discriminant axes. The data were reanalyzed after removal of the 25°C conditions. The split—plot design analysis of variance then indicated that the rate of weight loss of the frogs from the different localities were not significantly different. The discriminant analysis did not show a grouping as discrete as before, but the general pattern was about the same. The grand means for each locality of all frogs for all conditions were obtained and compared. A single sample (Lee's Camp) differed greatly from all samples, more than Hyla regilla differs from H. californiae. No relation between available environmental moisture and rate of weight loss was discerned. The linear arrangement of the discriminant values of the northern populations suggests some relation to the "ecogegrophical rules." This relationship is not shown in the discriminant values of the rates from southern localities.