Size and Scaling in Human Evolution

Science  06 Dec 1974:
Vol. 186, Issue 4167, pp. 892-901
DOI: 10.1126/science.186.4167.892


Our general conclusion is simply stated: many lineages display phyletic size increase; allometric changes almost always accompany increase in body size. We cannot judge adaptation until we separate such changes into those required by increasing size and those serving as special adaptations to changing environments.

In our view, the three australopithecines are, in a number of features, scaled variants of the "same" animal. In these characters, A. africanus is no more "advanced" than the larger, more robust forms. The one early hominid to show a significant departure from this adaptive pattern toward later hominids—cranially, dentally, and postcranially—is H. habilis from East Africa. The australopithecines, one of which was probably a precursor of the Homolineage, were apparently a successful group of basically vegetarian hominids, more advanced behaviorally than apes (87), but not hunter-gatherers.

The fossil hominids of Africa fall into two major groupings. One probable lineage, the australopithecines, apparently became extinct without issue; the other evolved to modern man. Both groups displayed steady increase in body size. We consider quantitatively two key characters of the hominid skull: cranial capacity and cheek tooth size. The variables are allometrically related to body size in both lineages. In australopithecines, the manner of relative growth neatly meets the predictions for functional equivalence over a wide range of sizes (negative allometry of cranial capacity with a slope against body weight of 0.2 to 0.4 and positive allometry of postcanine area with a slope near 0.75). In the A. africanus to H. sapiens lineage, cranial capacity increases with positive allometry (slope 1.73) while cheek teeth decrease absolutely (slope — 0.725). Clearly, these are special adaptations unrelated to the physical requirements of increasing body size. We examined qualitatively other features, which also seem to vary allometrically. Of course, many characters should be studied quantitatively, but we think that the scheme outlined here should be treated as the null hypothesis to be disproved.