DETECTING SHIFTS IN DIVERSIFICATION RATES WITHOUT FOSSILS

Studies of shifts in diversification rates and adaptive radiations are difficult when there are no fossils because past events cannot be inf erred. The phylogenies of recent species, however, allow one to infer the patterns of past diversifications. I present a new method for esti mating the diversification rate of a Lineage, provided that a phylogen y of recent species, constructed, for instance, with molecular data, i s available. This method was inspired by survival models and takes int o account species that are not included in detailed phylogenetic data, provided that approximate dates of origin of these species are known. Likelihood ratio tests and Akaike Information Criterion make it possi ble to test for differences in diversification among lineages or group s of lineages and, thus, to evaluate adaptive radiation hypotheses. Th e present modeling approach can easily be extended to include temporal variations in diversification rates. A simulation study showed that t he method is statistically consistent, avoiding Type I and Type II err ors, and that it is robust to periodic or random fluctuations in the s peciation rate. An example is presented with a composite phylogeny of primates.