Our analysis of molecular phylogenies provides compelling evidence that vertebrate diversification events occurred much more recently in the northern hemisphere than in the southern hemisphere. This pattern was remarkably consistent across different latitudinal subsets of species, suggesting that the geographic variation in species’ ages that we observed here was not substantially influenced by the exclusion of species that were distributed solely at low or high latitudes.
Why are vertebrates from the northern hemisphere younger than those from the southern hemisphere? One plausible explanation, as outlined in the Introduction, is that the observed disparity in diversification histories reflects geographic differences in the extent of climatically-driven extinction events that occurred throughout the Quaternary. Some regions of the southern hemisphere (e.g., inland Australia) were also severely affected by dry conditions during the Quaternary
, but palynological data indicate that forested and partially forested areas still persisted in coastal regions, even during the Last Glacial Maximum
. Although lineages of some of the species used in our analyses diverged prior to the Quaternary, this does not mean that climatic oscillations during the Quaternary could not be responsible for the observed patterns. Many species lost genetic lineages during the Quaternary
, and thus a species may have (randomly) lost its oldest lineage while undergoing range contraction in response to Quaternary climate change. For example, a species that had similar levels of genetic differentiation in both hemispheres prior to the Quaternary would have had a higher probability of losing its oldest genetic lineage in the northern hemisphere because the magnitude of Quaternary climate change was greater there.
Our findings accord with the results of previous analyses that focused exclusively on amphibians and reptiles
[12, 13], and thus demonstrate that the patterns revealed by these earlier studies apply to vertebrates as a whole. Indeed, we found no evidence for an interaction between taxonomic class and hemisphere of origin, suggesting that the relationship between geographic origin and species age was broadly consistent across the five classes of vertebrates included in our analyses. Although differences between hemispheres were only statistically significant for mammals and reptiles when each taxonomic group was analyzed separately, partitioning our dataset by taxonomic group entailed a considerable loss of statistical power. In addition, the regression coefficients describing the effects of hemisphere of origin were in the same direction for four out of five taxa, the exception being birds. Nevertheless, similar results were obtained for birds when a single northern hemisphere species that had exceptionally divergent populations was removed from our analyses.
At first glance, this consistency in the effects of geographic location on diversification histories is somewhat surprising given the marked differences in dispersal abilities between the taxa included in our analyses
[17, 18]. One potential reason for this seemingly counterintuitive result is that differences in dispersal abilities between taxa are trivial in comparison to the disparity in Quaternary climate changes between the northern and southern hemispheres. However, it should be noted that the mean age of birds, a group that has the greatest dispersal abilities of the taxonomic classes that we considered, showed the least disparity between hemispheres.
Interestingly, studies of antbirds and passerines have revealed greater genetic distances (based on allozymes) within neotropical species than in species from the Nearctic
[19, 20]. Similar patterns have also been documented in a suite of birds and mammals from the New World
, as well as in plants
 and vertebrates
 distributed across the globe. Collectively, the results of these studies suggest greater intraspecific genetic variation in regions that have experienced stable long-term climatic conditions. Thus, our finding that diversification events occurred earlier in the southern hemisphere, where climatic oscillations have also been less severe in the past, is in broad accordance with earlier studies on a wide range of taxa.
We also found that the relationship between hemisphere of origin and species age did not differ between ectotherms and endotherms, and this pattern was also evident when we considered each of these two groups in isolation. Future research could usefully explore whether species-level traits that do not show a strong taxonomic signal (e.g., habitat-use) explain additional geographic variation in species ages
However, we cannot dismiss the possibility that our findings may be an artifact of differing levels of taxonomic knowledge about the vertebrate faunas of the northern and southern hemispheres
[13, 20]. Species from the southern hemisphere have generally been less intensively studied than species from the northern hemisphere, and thus some southern hemisphere ‘species’ included in our analyses may actually represent composite taxa. This taxonomic ignorance would, in turn, result in over-estimations of species’ ages in the southern hemisphere. Dubey & Shine
 demonstrated that geographic disparity in taxonomic ignorance is a relatively poor explanation for differences in the ages of amphibian species between hemispheres, but that this mechanism could potentially explain the different diversification histories of the two hemispheres for reptiles. However, we consider this interpretation less likely than climatically-driven extinction events given that our findings were relatively consistent across such broadly divergent taxa (see above). We further consider it unlikely that our results are an artifact of using intraspecific rather than interspecific divergence times, given that Dubey & Shine
 found very similar geographic patterns in species ages regardless of whether the analysis was performed with intraspecific divergences (across orders and suborders) or with the cladogenic events responsible for species formation
. Thus, the same parameters appear to influence these two types of measures
. Our finding that diversification events occurred earlier in the southern hemisphere may also have been influenced by the restricted distributions of some taxonomic groups throughout the globe. Some taxa are strictly or more widely represented in one hemisphere, which could impact species age estimations (considering that each taxa exhibits different life-history traits and characteristics).
Our analyses also revealed differences in the timing of intraspecific diversification events between taxonomic classes, and between ectotherms and endotherms. In particular, the mean ages of ectotherms (fishes, amphibians, and reptiles) were older than those of endotherms (mammals and birds). This pattern could be the result of several different mechanisms, and these mechanisms are not mutually exclusive:
The differences we observed between endothermic and ectothermic vertebrates could have been due to differences in the dispersal abilities of these groups. Endothermic vertebrates are typically capable of dispersing greater distances than ectothermic vertebrates (e.g.,
[17, 18, 24]), and taxa that are capable of dispersing greater distances generally have more homogeneous population structures compared to less vagile taxa. Such homogenization would result in fewer divergent lineages within endothermic vertebrates, and ultimately more recent intraspecific diversification events.
Alternatively, the observed differences between endothermic and ectothermic vertebrates could have been due to differences in energetic demands during hibernation. When abiotic conditions become unfavorable, ectothermic vertebrates are capable of lowering their rates of resource consumption to a far greater degree than are endothermic vertebrates
[25–27], and thus ectotherms may have been less adversely affected by harsh winter conditions during the Pleistocene. In addition, many ectothermic vertebrates possess efficient mechanisms to deal with sub-zero temperatures (e.g., freeze-avoidance or freeze tolerance strategies:
[28–30]). Hence, ectothermic vertebrates may have been more likely to survive climatic fluctuations during the Quaternary compared to endotherms, at least in locations where the minimum temperature required for ectotherm activity was reached during part of the year.
Finally, differences in the timing of intraspecific diversification events between groups may have been caused by variation in life-history traits. For example, reproductive isolation is faster in particular groups such as mammals (compared to birds and amphibians) due to a higher possible rate of regulatory evolution, causing higher probabilities of developmental incompatibilities between mammal species
. Consequently, for the same level of genetic divergence, two lineages could be considered sister species in mammals, and intraspecific lineages in birds or amphibians. The vertebrate classes we examined also differ markedly in a wide array of additional traits (e.g., body size), and such traits are known to impact rates of molecular evolution