Form and genes often tell different stories about the evolution of animals, with molecular data generally considered to be more objective than morphological data. However, form provides the basis for the description of organisms, and the study of fossils crucially depends on morphology. Complex organisms tend to evolve as ‘mosaics’, in which parts may be modified at varying rates and in response to different selective pressures. Thus, individual anatomical regions may contain different phylogenetic signals. In the present study, we used computerized methods to ‘dissect’ the skulls of a primate clade, the guenons, into functional and developmental modules (FDM). The potential of different modules as proxies for phylogenetic divergence in modern lineages was investigated. We found that the chondrocranium was the only FDM in which shape consistently had a strong and significant phylogenetic signal. This region might be less susceptible to epigenetic factors and thus more informative about phylogeny. The examination of the topology of trees from the chondrocranium suggested that the main differences evolved at the time of the radiation of terrestrial and arboreal guenons. However, phylogenetic reconstructions were found to be strongly affected by sampling error, with more localized anatomical regions (i.e. smaller/less complex FDMs) generally producing less reproducible tree topologies. This finding, if confirmed in other groups, implies that the utility of specific FDMs for phylogenetic inference could, in many cases, be hampered by the low reproducibility of results. The study also suggested that uncertainties due to sampling error may be larger than those from character sampling. This might have implications for phylogenetic analyses, which typically provide estimates of support of tree nodes based on characters but do not generally take into account the effect of sampling error on the tree topology. Nonetheless, studies of the potential of different FDMs as proxies for phylogenetic divergence in modern lineages, such as the present study, provide a framework that may help in modelling the morphological evolution of present and fossil species. © 2008 The Linnean Society of London, Biological Journal of the Linnean Society, 2008, 93, 813–834.