Hi Christopher, and thanks for your question.
The short answer to your first question is...I don't know. But here are three possibilities: First, the difference in the observed slope across all species, and the slope predicted by the model, might be due to differences in metabolic rate among groups that I haven't addressed. For example, after accounting for the effects of body size and temperature, there are still differences between the resting metabolic rates of ectotherms and endotherms. Or second, it could be related to systematic differences in the temperature estimates I used and the actual temperatures of interest. Estimating mean or preferred body temperatures, particularly for ectotherms, will always be messy. And, it's hard to know exactly what temperature would be best to use. Is it max temp? Or average annual temp over some evolutionary time period? Does the latter equate to mean or preferred temp? Third, of course, it could be that the model prediction is wrong. The temperature dependence of brain mass shown here is certainly within the realm of those observed in studies of metabolic rate-but these too vary.
As for the 2nd question, the answer is much the same as the first. But, I'd be cautious in reading too much into a significant difference among a groups like amphibians. Data for groups like this are relatively scarce, and not necessarily representative of the group as a whole.
Hope this clarifies things a bit.
Thanks for your reply and really interesting paper. I'm working on this area in chondrichthyans specifically, and was particularly interested in your findings in ectotherms.- Christopher Mull •