Physical anthropologists have long recognized the extent of morphological variation present in Homo sapiens, yet few studies have directly compared the extent to which human variability differs from the variability found in other species.  Theoretical considerations indicate that genotypic (and therefore phenotypic) variation should decrease under directional selection.  Since Australopithecus afarensis is much more similar cranially to the modern African apes than to modern humans, it is likely that humans have diverged to the greatest extent from the common human-chimpanzee-gorilla ancestor.  Certainly there has been strong selection for increasing brain size (and hence cranial size) in hominids during the last 2 million years.  One might therefore expect humans to be less variable cranially than the African apes.  Experimental selection experiments in mice and Drosophila, however, indicate that phenotypic variation does not necessarily decrease with directional selection, and may even increase.

In order to investigate the question of comparative variability in African apes and humans, craniometric data for Pan paniscus, Pan troglodytes, Gorilla gorilla, and Homo sapiens, was gleaned from the literature.  Data for males only was used in order to avoid the effects of sexual dimorphism.  The coefficient of variation (CV) was used for these interspecific comparisons to minimize the effect of size.  Because extant African apes are limited geographically to only a part of Sub-Saharan Africa, two comparisons were run: one using human CV's derived from a pooled estimate of variance including all geographical regions, and another using human CV's derived from a within-geographical-region estimate of variance.

Comparisons utilizing the pooled human data indicate that, as a species, humans do not appear to be less variable craniofacially than the other African apes.  Further, humans in fact appear to be more variable in measures of cranial breadth.  Comparisons utilizing within-geographical-region estimates indicate that human geographical groups are somewhat less variable in facial-masticatory measurements, but entirely comparable variability exists in the measures of the cranial vault.  This analysis indicates that, to the extent these measurements are genetically controlled, genotypic variance has not decreased (with respect to the African apes) in the face of directional selection.

This analysis raises at least 2 questions of significance concerning human and African ape evolution.  1) Why are measures of cranial breadth so variable in humans, considering that this dimension is quite heritable?  This may indicate that some balancing mechanism is maintaining variation.  2) Mitochondrial DNA studies have indicated the possibility of a bottleneck in human evolution with respect to the African apes.  If the rates of mtDNA evolution are constant in these species, the bottleneck must have been both recent and transient (since variation does not appear to have been greatly reduced in modern humans). Alternatively, mtDNA rates may not be constant in humans and African apes.