Absolute or Proportional brain size: that is the question. A reply to Leigh's (2006) comments

In his comments, Leigh (2006) argues that when absolutebrain size rather than proportional size is plotted againstage, the juvenile Homo erectus specimen Mojokerto fallsinto the extant Homo sapiens range of variation. On the otherhand, we showed that, when proportional endocranial volume(PEV, i.e., the ratio between the endocranial volume of an in-dividual at a given age and the average adult endocranial vol-ume) is used as the means of comparison, Mojokerto fallswithin the chimpanzee range rather than the human range(Coqueugniot et al., 2004). Although these observations mayseem contradictory, there is, in fact, no major disagreementbetween them. Absolute and proportional brain sizes addressdifferent questions. Additionally, that H. erectus and H. sapi-ens may have had similar absolute brain sizes at an earlyage does not necessarily imply similar developmental curves.As the adult brain size is quite different between the two spe-cies, this may actually imply different developmental trajecto-ries. Finally, with respect to the variability of absolute andproportional brain size in humans and apes, our data yield dif-ferent results than do the data used by Leigh. This discrepancymay result from several methodological issues that we alsofind important to clarify.Age determination for MojokertoIt is certainly possible to debate our age determination forMojokerto of 0.5e1.5 years. However, previous age estimates(from one to eight years old) have been based on vague argu-ments, including some based on the braincase size of the spec-imen, which led to circular reasoning about brain development.We utilized the most accurate available age indicators. Thefossa subarcuata closure and the fontanelle were most reliable.Our study also showed that the external tympanic morphology,mentioned by Leigh, is a poor age indicator. A fused or unfusedstate of the tympanic plate is observed at variable frequenciesin all the age classes between birth and adulthood. It shouldbe noted that, if, based on this feature, an older individualage (ca. 4e6 years old) is favored for Mojokerto, then thePEVof this juvenile H. erectus would be closer to the modernhuman distribution (Fig. 1), but its absolute brain size would bepushed farther away from the modern human distribution.Problems with material and methodologyThe study of brain development in fossil hominins is pri-marily based on measurement of the endocranial volume(EV). In contrast, the data utilized by Leigh represent post-autopsy brain weights, whereby EV is estimated by applyinga coefficient of specific gravity. The autopsy data introduceimportant limitations for precisely inferring EV and for evalu-ating its variability in humans and in apes. The cerebrospinalfluid surrounding the brain represents about 12% of the endo-cranial cavity (Peters et al., 1998). During an autopsy, a vari-able quantity of the cerebrospinal fluid escapes from thesubarachnoid space and from the ventricles, making it difficultto accurately control for this factor a posteriori. Another im-portant factor that artificially increases variability is the cause