Data from: Ecological causes of decelerating diversification in carnivoran mammals

Clade diversification is a central topic in macroevolutionary studies. Recently, it has been shown that diversification rates appear to decelerate over time in many clades. What causes this deceleration remains unclear, but it has been proposed that competition for limited resources between sympatric, ecologically similar species slows diversification. Employing carnivoran mammals as a model system, we test this hypothesis using a comprehensive time-calibrated phylogeny. We also explore several conceptually related explanations including limited geographic area and limited rates of niche evolution. We find that diversification slowdowns are strong in carnivorans. Surprisingly, these slowdowns are independent of geographic range overlap between related species and are also decoupled from rates of niche evolution, suggesting that slowdowns are unrelated to competition and niche filling. When controlling for the effects of clade diversity, diversification slowdowns appear independent of geographic area. There is a significant effect of clade diversity on diversification slowdowns but simulations show that this relationship may arise as a statistical artifact (i.e., greater clade diversity increases the ability of the gamma statistic to refute constant diversification). Overall, our results emphasize the need to test hypotheses about the causes of diversification slowdowns with ecological data, rather than assuming ecological processes from phylogenies alone.

支系分化（clade diversification）是宏观进化研究（macroevolutionary studies）领域的核心议题。近期研究表明，诸多类群的分化速率随时间推移呈现减速趋势。目前该减速现象的成因尚未明确，已有假说提出，同域分布、生态位相似的物种间对有限资源的竞争会延缓分化进程。本研究以食肉目哺乳动物为模式系统，依托一套经过全面时间校准的系统发育树（time-calibrated phylogeny）对该假说进行检验；同时还探讨了若干概念相关的备选解释，包括有限地理范围与受限的生态位演化（niche evolution）速率。研究结果显示，食肉目类群的分化减速现象显著。出乎意料的是，此类减速与近缘物种间的地理分布重叠并无关联，同时也与生态位演化速率脱钩，这表明分化减速与物种竞争及生态位填充无关。在控制支系多样性的影响后，分化减速与地理范围的相关性亦不复存在。支系多样性对分化减速存在显著影响，但模拟实验表明，该关联可能源于统计假象——即支系多样性越高，伽马统计量（gamma statistic）越能拒绝恒定分化的假设。综上，本研究结果强调，检验分化减速的成因假说需依托生态数据，而非仅通过系统发育数据推断生态过程。