An Introduction to Phylogenetically Based Statistical Methods, with a New Method for Confidence Intervals on Ancestral Values

Interspecific comparisons have played a prominent role in evolutionarybiology at least since the time of Charles Darwin. Since 1985,the "comparative method" has been revitalized by new analyticaltechniques that use phylogenetic information and by increasedavailability of phytogenies (often from molecular data sets).Because species descend from common ancestors in a hierarchicalfashion, related species tend to resemble each other (elephantslook like elephants); therefore, cross-species data sets generallydo not comprise independent and identically distributed datapoints. Phylogenetically based statistical methods attempt toaccount for this fact. Phylogenetic methods allow traditionaltopics in comparative and ecological physiology to be addressedwith greater rigor, including the form of allometric relationshipsand whether physiological phenotypes vary predictably in relationto behavior, ecology or environmental characteristics, whichprovides evidence about adaptation. They can also address newtopics, such as whether rates of physiological evolution havediffered among lineages (clades), and where and when a phenotypefirst evolved. We present brief overviews of three phylogeneticallybased statistical methods: phylogenetically independent contrasts,Monte Carlo computer simulations to obtain null distributionsof test statistics, and phylogenetic autocorrelation. In a newresult, we show analytically how to use independent contraststo estimate ancestral values and confidence intervals aboutthem. These confidence intervals often exceed the range of variationobserved among extant species, which points out the relativelygreat uncertainty inherent in such inferences. The use of phytogeniesshould become as common as the use of body size and scalingrelationships in the analysis of physiological diversity.