Abstract
Divergence dating has become one of the most popular applications of molecular phylogenetics over the last decade. This technique combines models of molecular evolution with information about ages of select groups within a phylogenetic tree to infer divergence times for the entire tree in the form of a chronogram. These chronograms have in turn informed many areas of evolutionary biology including historical biogeography. evolution of adaptive characters, and rates of species diversification. The information used to calibrate divergence dating studies most frequently comes in the form of fossil evidence. This paper explores some potential impacts of fossil calibration uncertainty on divergence dating results, using work from myself and collaborators on ants and bees. Among Hymenoptera in general, divergence dating studies so far have been largely restricted to these two taxa (Brady et al. 2009). The poor preservation and vast incompleteness of the fossil record prevalent in most insect groups presents several concerns. In studies that are only able to employ a few fossil calibrations, fossils with incomplete taxonomic or temporal data may exercise a pathological influence on dating results. There are several potential avenues to assess this, but a simple and direct approach is to remove suspect calibrations and see how this changes the results. For example, Brady et al. (2006b) used divergence dating to examine the timing of eusocial evolution in sweat bees of the family Halictidae. The fossil record for halictid bees is quite poor, so only three fossil calibrations were available for this study. After conducting analyses using all three constraints, the authors removed the two recent Dominican amber calibrations and reran the analyses. The results were the same with and without this removal, indicating that eusociality evolved approximately 20-22 million years ago in three separate lineages of sweat bees. If one is particularly concerned about the age assigned to the fossils, one could investigate whether different age assignments influence results. For example, in a divergence dating study on ants, Moreau et al. (2006) observed that there was some uncertainty of about 5 million years over the ages of three fossil strata used to provide some of their fossil minimum age calibrations. They ran separate analyses using different age assignments for these three strata, and reported that this resulted in a 28-million-year age difference for the age of extant ants. However; when Brady et al. (2006a) repeated this same procedure on a similar data set, they found a difference of only 0-2 million years. It appears that factors other
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call