Diversity-dependent speciation and extinction in hominins

Abstract

The search for drivers of hominin speciation and extinction has tended to focus on the impact of climate change. Far less attention has been paid to the role of interspecific competition. However, research across vertebrates more broadly has shown that both processes are often correlated with species diversity, suggesting an important role for interspecific competition. Here we ask whether hominin speciation and extinction conform to the expected patterns of negative and positive diversity dependence, respectively. We estimate speciation and extinction rates from fossil occurrence data with preservation variability priors in a validated Bayesian framework and test whether these rates are correlated with species diversity. We supplement these analyses with calculations of speciation rate across a phylogeny, again testing whether these are correlated with diversity. Our results are consistent with clade-wide diversity limits that governed speciation in hominins overall but that were not quite reached by the Australopithecus and Paranthropus subclade before its extinction. Extinction was not correlated with species diversity within the Australopithecus and Paranthropus subclade or within hominins overall; this is concordant with climate playing a greater part in hominin extinction than speciation. By contrast, Homo is characterized by positively diversity-dependent speciation and negatively diversity-dependent extinction-both exceedingly rare patterns across all forms of life. The genus Homo expands the set of reported associations between diversity and macroevolution in vertebrates, underscoring that the relationship between diversity and macroevolution is complex. These results indicate an important, previously underappreciated and comparatively unusual role of biotic interactions in Homo macroevolution, and speciation in particular. The unusual and unexpected patterns of diversity dependence in Homo speciation and extinction may be a consequence of repeated Homo range expansions driven by interspecific competition and made possible by recurrent innovations in ecological strategies. Exploring how hominin macroevolution fits into the general vertebrate macroevolutionary landscape has the potential to offer new perspectives on longstanding questions in vertebrate evolution and shed new light on evolutionary processes within our own lineage.

Fig. 1. Speciation and extinction dates and phylogeny used in subsequent analyses.

a, Species lifespans, comprising the time between speciation and extinction dates based on three datasets. Orange: published fossil FADs and LADs estimated without taking fossil preservation into account. Light blue: speciation and extinction dates estimated in a Bayesian framework incorporating time-based variability in fossil preservation rates. Dark blue: speciation and extinction dates estimated in a Bayesian framework incorporating within-lineage variability in fossil preservation rates. Note that these taxa are those the published dates and our new database have in common; actual analyses incorporated Homo ergaster in the no-preservation-prior dataset and Homo rudolfensis in the preservation prior datasets. Homo erectus s.l. refers to Homo erectus sensu lato. b, The Parins-Fukuchi et al. phylogeny used in this study, with species coloured by taxonomic grouping (yellow: Homo; green: non-Homo).

Fig. 2. A divergent relationship between macroevolutionary rates and species diversity in Homo.

a, Results from PyRate birth–death models, run across three datasets with different fossil preservation priors (no preservation priors: published FADs and LADs; time-based preservation variability, where preservation is allowed to vary every 1 million years—prior applied to fossil occurrence data from three databases; and within-lifetime preservation variability, where preservation rate is allowed to vary across a species’ lifespan—prior applied to fossil occurrence data from three databases). In the latter two models, the preservation rate was also allowed to vary between lineages. The posterior distribution of the correlation parameter is shown, with the 50% credible interval shaded and the 95% credible interval indicated by the outline. The mean correlation parameter is indicated by a thick line. Results for speciation are indicated in blue; those for extinction are shown in red. b, The relationship between diversity 500,000 years before tip height and speciation rate (tip DR) across the Parins-Fukuchi et al. phylogeny. Shaded area indicates 95% confidence interval.