Bayesian transmogrification of clade divergence dates: a critique


Correspondence: Michael Heads, Buffalo Museum of Science, 1020 Humboldt Parkway, Buffalo, NY 14211-1293, USA.



The crucial step in Bayesian dating of phylogenies is the selection of prior probability curves for clade ages. In studies on regions derived from Gondwana, many authors have used steep priors, stipulating that clades can only be a little older than their oldest known fossil. These studies have ruled out vicariance associated with Gondwana breakup, but only because of the particular priors that were adopted. The use of non-flat priors for fossil-based ages is not justified and is unnecessary. Tectonic calibrations can be integrated with fossil calibrations that are used to give minimum clade ages only.