SUMMARY Recent molecular research has provided a consistent estimate of phylogeny for the living papionin monkeys (Cercocebus, Lophocebus, Macaca, Mandrillus, Papio, and Theropithecus). This phylogeny differs from morphological phylogenies regarding the relationships of the mangabeys (Cercocebus and Lophocebus) and baboons (Mandrillus, Papio, and Theropithecus). Under the likely assumption that the molecular estimate is correct, the incongruence between the molecular and morphological data sets indicates that the latter include numerous homoplasies. Knowledge of how these homoplasies emerge through development is important for understanding the morphological evolution of the living papionins, and also for reconstructing the phylogenetic relationships and adaptations of their fossil relatives. Accordingly, we have used geometric morphometric techniques and the molecular phylogeny to investigate the ontogeny of a key area of morphological homoplasy in papionins, the face. Two analyses were carried out. The first compared allometric vectors of Cercocebus, Lophocebus, Macaca, Mandrillus, and Papio to determine which of the facial resemblances among the genera are homoplasic and which are plesiomorphic. The second analysis focused on early post-natal facial form in order to establish whether the facial homoplasies exhibited by the adult papionins are to some degree present early in the post-natal period or whether they develop only later in ontogeny. The results of our analyses go some way to resolving the debate over which papionin genera display homoplasic facial similarities. They strongly suggest that the homoplasic facial similarities are exhibited by Mandrillus and Papio and not by Cercocebus and Lophocebus, which share the putative primitive state with Macaca. Our results also indicate that Mandrillus and Papio achieve their homoplasic similarities in facial form not through simple extension of the ancestral allometric trajectory but through a combination of an extension of allometry into larger size ranges and a change in direction of allometry away from the ancestral trajectory. Thus, the face of Mandrillus is not simply a hypermorphic version of the face of its sister taxon, Cercocebus, and the face of Papio is not merely a scaled-up version of the face of its sister taxon, Lophocebus. Lastly, our results show that facial homoplasy is not restricted to adult papionins; it is also manifest in infant and juvenile papionins. This suggests that the homoplasic facial similarities between Mandrillus and Papio are unlikely to be a result of sexual selection.