Changes also affected the skull of Malpaisomys: its skull is narrow at the level of the interorbital breadth, a Dinaciclib CDK inhibitor character unknown in Mus. Such a character is related to a peculiar position of the eyes that may be adaptive as it allows a better protection from predatory birds, the eastern Canary Islands being characterized by open environments and great expanses of recent lava fields covered by scarce vegetation. Presently, molecular data indicate that the lineage of Malpaisomys split at a similar time as the genus Mus radiation but the exact order of lineage origins remains impossible to ascertain. The chronograms derived from the BEAST and Multidivtime analyses are presented in Figure 6. Estimated ages and 95% credibility intervals of notable nodes are detailed in Table 4. Two independent runs of Multidivtime gave the same results. On the whole, the estimations of Multidivtime are slightly older that the ones obtained from the BEAST analysis, but in most cases, 95% CI are extensively overlapping and are of the same order. The estimations are particularly congruent among the Murini including Malpaisomys and for the deepest nodes of the tree while the most important discrepancies between the two methods are found among the Rattini. Different clock methods are known to lead to disparate results due to their inherent handling of the rate MLN4924 heterogeneity across lineages. When the lineage rate assumption is violated, relaxed-clock methods were proven to produce biased estimated credibility intervals. This would explain that a specific evolutionary rate could characterize the Rattini tribe and account for the slight disparities between the two dating methods. Nonetheless, estimations and the 95% CI for the three calibration points are congruent with the paleontological evidence. We re-ran the analyses, omitting each of the calibration points in turn to assess whether the molecular dates are the result of the fossil constraints and we obtained similar estimates. The split of the main lineages of Murinae took place between 12.3 and 9.8 Ma /12.7 and 10.8 Ma. These estimates are similar to the ones obtained by Lecompte et al., slightly older than the ones reported by Rowe et al., and much younger than the ones found in Jansa et al.. The main difference between our study and the one undertaken by Rowe et al. corresponds to the addition of a third calibration interval. When this constraint is removed from the Multidivtime analysis, our estimates converged to the divergence-dates obtained by Rowe et al.. With respect to the study of Jansa et al., the differences are probably not only due to sampling different taxa but also to the use of calibration points based on much older splits in the mammal lineage.