Ambiguous support for extra-tropical accelerated diversification of rosids

Abstract

Sun et al.1 used a comprehensive phylogenetic and a locality based climatic dataset to examine how past climates have driven diversification across 17 orders of rosids. They concluded that tropical rosids diversify slower than the (younger) non-tropical counterparts in part due to a strong negative relationship between paleo-temperature and diversification (i.e. higher diversification rates under lower temperatures). Their conclusions are based on tip rates (derived from BAMM2; DR3) and correlations with current temperature (e-SIM4) or binary tropicality data sets (BAMM/STRAPP5; FiSSE6) as well as tree-wide estimates of diversification with paleo-temperature (BAMM; RPANDA7) or tropicality (BiSSE8; HiSSE9). Here, I highlight several inconsistencies in their diversification analyses as well as a systematic error pertaining to the RPANDA model selection procedure, which, together with several minor technical issues, weaken the support for Sun et al.’s1 conclusions. A re-analysis of their BiSSE/HiSSE and RPANDA analyses are performed.