Response : The Body Temperature of Tyrannosaurus rex

Abstract

Response: Millard concludes that T. rex is an ectotherm from an analysis of the data in our (R.E.B. and W.J.S.) report (1). We disagree with his conclusion because his application of a statistical model to Furthermore, body temperatures in endotherms are controlled within narrow limits ( ± 20C), resulting in a restricted range of bone isotopic values. Measurements of intrabone isotopic heterogeneity ranged from 0.4 to 0.8 per mil, respectively, from a modem cow and deer (3), using sample sizes similar to those taken from T. rex. Millard's statistical reasoning would suggest that these endotherms underwent body temperature variations of 100 to 200C. With these modem animals we know that this is not the case, which also suggests that the Haversian system independence assumption is invalid. We agree that the spread of the isotopic data may be a minimum estimate of variation.Statistical tests of means and variances from data in the report (1 ) may be used to evaluate a hypothesis of endothermy in T.rex. These tests were completed independently (by those of us at the University of Alabama, Birmingham) and the results suggest the following. (i) If T. rex were an endotherm, isotopic variance of extremity bones (tibia, metatarsal, phalange, and mid and distal caudal vertebrae) ought not to differ significantly from isotopic variance of core bones (rib, gastralia, and dorsal vertebrae). A two-tailed F test yields a test statistic of 1.010, the probability of which is 0.49; thus the null hypothesis of endothermy cannot be rejected. (ii) If T. rex were an endotherm, the mean isotopic value of the bone recording the coldest temperature (distal caudal vertebra) ought to differ by less than 4CC from the mean isotopic value of the bone recording the warmest temperature (dorsal vertebra). A one-tailed, twosample t test on these two bones yields a test statistic of 0.659, the probability of which is 0.26; thus once again providing no evidence to reject the null hypothesis of endothermy. These additional statistical analyses are consistent with the hypothesis of endothermy in T. rex. Our point still stands that the lack of significant intrabone or interbone 8P variation in T. rex suggests it was homeothermic.