Abstrakt
In this study, we evaluate the impact of fossil assignments and different models of calibration on divergence time estimates carried out as Bayesian analyses. Estimated ages from preceding studies and liverwort inclusions from Baltic amber are used as constraints on a molecular phylogeny of Cephaloziineae (Jungermanniopsida) obtained from sequences of two chloroplast coding regions: rbcL and psbA.
In total, the comparison of 12 different analyses demonstrates that an increased reliability of the chronograms is linked to the number of fossils assigned and to the accuracy of their assignments. Inclusion of fossil constraints leads to older ages of most crown groups, but has no influence on lineage through time plots suggesting a nearly constant accumulation of diversity since the origin of Cephaloziineae in the early to Middle Jurassic.
Our results provide a note of caution regarding the interpretation of chronograms derived from DNA sequence variation of extant species based on a single calibration point and/or low accuracy of the assignment of fossils to nodes in the phylogeny.