In recent years, most systematics studies have focused on phylogenetic analyses of molecular data sets. The latest trend has been to add molecular dating to these phylogenies utilizing methods such as nonparametric rate smoothing (NPRS) and penalized likelihood (PL) and calibrating these analyses using (often only one or very few) fossils. The success of such approaches is dependent on several assumptions, including a local clocklike behavior of evolution, the accuracy of the phylogeny, the correct phylogenetic placement of fossils, and the consistency of particular fossils in extrapolating rates throughout a given phylogenetic tree. An example of such an analysis of the Nymphaeales is provided to illustrate inappropriate use of fossils in this context and faulty results based on inadequate and/or inappropriate analyses. Neither fossil identifications nor a particular method of molecular dating should be called into question based on the disparity of a single analysis. Indeed, fossil observations and molecular dating are often at odds due to failure of the data to meet minimum assumptions of a clocklike behavior and poor or inadequate sampling of extant taxa, molecular sequence data, and/or fossils. Rejection or acceptance of either the fossils or the molecular dates resulting from their use should be considered in light of direct analysis of the fossils and compared to other analyses using other fossils and/or other extant data sets. Rejection of fossils based on unexpected results is merely verificationism.
Annals of the Missouri Botanical Garden
Vol. 95 • No. 1
Vol. 95 • No. 1
nonparametric rate smoothing