Table 5.

Probabilities of detecting effects of climatic variables, and their variation among continents and between classes of amphibians and reptiles. The probabilities and their 95% confidence limits were calculated from a logistic regression in which continent, class and a continent × class interaction were the predictors. The response was binary with ‘1’ meaning that an effect of climate on some aspect of the species’ biology was reported. Each species within each study was treated as an independent data point (n = 412). Only data from Europe, North America, Central America, and South America were used because of small sample sizes for other geographical areas. Even within the set of regions with larger sample sizes, insufficient data were available to estimate probabilities of reporting effects of climate change for either amphibians or reptiles in South America.

North AmericaAmphibia0.520.410.61
North AmericaReptilia0.570.220.86
Central AmericaAmphibia0.620.420.78
Central AmericaReptilia0.670.250.92