Caveats on the Use of Fossil Calibrations for Molecular Dating: A Reply to Parham and Irmis

In Parham and Irmis’s (2008) commentary on our article “Assessing Concordance of Fossil Calibration Points in Molecular Clock Studies: An Example Using Turtles” (Near et al. 2005), these authors address three areas of concern with respect to fossil calibration analyses in general and our article in particular: (1) absolute age estimates for fossil calibration points, including the accuracy of the reporting of age information and the inclusion of stratigraphic error; (2) the use of fixed calibration points in clock calibration studies; and (3) the phylogenetic placement of fossil taxa in our article. Before we discuss the points raised by Parham and Irmis, it is important to emphasize that the primary goal of our article was to provide a new methodological approach for incorporating uncertainty and error into molecular clock calibration estimates. In that article, we developed a new method, based on cross-validation with multiple calibration points, to identify outlier fossils that may have been inaccurately dated or phylogenetically misplaced or that had otherwise been providing inaccurate dates for clade divergences. We applied that method to turtles because they have an excellent fossil record and have been