Parting ways: Pan-Homo divergence revisited

The timing of divergence between hominins and the bonobo-chimpanzee clade has been at the core of palaeoanthropological debate for over a century. The earliest molecular studies indicated divergence times ranging from 5 Ma to as recently as 1.3 Ma. This study critically reviews the trends of time estimates published between 1967 and 2023, and analyses how these are supported or rejected by the current molecular and fossil records. We compiled 202 divergence estimates and defined three distinct thresholds based on fossil evidence at 4.4 Ma (Australopithecus anamensis and Ardipithecus ramidus), 6.2 Ma (Orrorin tugenensis and Ardipithecus kadabba), and 7.2 Ma (Sahelanthropus tchadensis). We then used these thresholds to filter out molecular estimates that are too young to fit the fossil record. Overall, the data suggests a divergence event within the late Miocene, with each threshold pushing it further back, 8.63-6.38, 10.33-7.81, and 10.95-8.81 Ma, respectively. We use a quadratic regression to demonstrate that estimates have been slowly shifting from ~ 6 Ma to ~ 8.5 Ma over the past 56 years. A Bayesian meta-analysis of genomic estimates filtered by our most consensual threshold (i.e., assuming Australopithecus belongs to Hominini) indicates that the split must have occurred early in the late Miocene, most likely before 7 Ma (~ 99.5% posterior probability) with a pooled effect of 8.69-7.28 Ma. We conclude that, despite an initial bias towards younger estimates, the molecular timing for the last common ancestor (LCA) of Pan-Homo has been progressively approaching the intervals suggested by the current fossil record.