40 



T.W. HOLLIDAY AND S.E. CHURCHILL 



Body Linearity Relative to Mass 



Another body shape feature known to covary with climate is relative 

 body linearity. In living populations, the weight: height, orponderal, 

 index is used as a measure of this relationship (e.g., Newman, 1961 ; 

 Schreider, 1964, 1975; Eveleth, 1966; Hiernaux etal, 1975). This 

 relationship is most easily quantified skeletally via relative femoral 

 head size (i.e., antero-posterior femoral head diameter/femoral 

 bicondylar length x 1 00). This index should reflect relative linearity, 

 since the femoral head is highly correlated with body mass, while 

 femoral length is highly correlated with stature. This skeletal index 

 was (not surprisingly) found to vary significantly between males and 

 females, with males possessing relatively larger femoral heads than 

 females (two-tailed / test, p < 0.0001 ), and thus Gough's Cave 1 is 

 compared only to other males for this trait. 



Table 6 reports the summary statistics for this trait among the 

 comparative samples and the Cheddar specimen. Within the recent 

 humans, there is a clear clinal pattern from Sub-Saharan Africa 

 through North Africa and into Europe, such that the femoral head 

 becomes relatively larger with increasing latitude (see also Ruff, 

 1994). With regard to fossil humans, note the extremely high indices 

 exhibited by the male Neandertals. For this index, both Neandertal 

 males (La Chapelle-aux-Saints 1 and La Ferrassie 1 ) fall beyond the 

 99th percentile of recent European males (n = 134). The other 

 European fossils, including the Mesolithic males and the Gough's 

 Cave 1 specimen himself are virtually identical to recent Europeans 

 for this trait. Only the EUP sample slightly deviates from the 

 European pattern of relatively large femoral heads; they are more 

 similar to recent North Africans in that their femoral heads are 

 somewhat smaller (although not as small as those of the Sub-Saharan 

 Africans). 



Body Breadth Relative to Stature 



Bi-iliac breadth, or bi-cristal breadth, as it is sometimes called, is 

 measured as the transverse diameter of the superior margin of the 

 pelvic girdle. This raw measurement is correlated with climatic 

 variables (Crognier, 1981 ; Ruff, 1994), but its fit with climate and/or 

 geography significantly improves when it is scaled to a linear dimen- 

 sion of the body such as stature (Roberts. 1978; Ruff, 1991. 1993, 

 1994). For the samples presented here, stature is unknown, and 

 therefore must be predicted from long bone length, e.g. femoral 

 length. In such cases, then, predicted stature is each individual's 

 femoral length subsequent to an arithmetic manipulation, (i.e., femo- 

 ral length x slope, +Y-intercept). Such prediction formulae inevitably 

 introduce error into the analysis, however, since biologically speak- 

 ing, many individuals are expected to fall well above or well below 

 the predictive line. Thus, to avoid the introduction of further error, 

 stature is not predicted for this analysis, but rather, femoral length 

 (which is highly correlated with stature) is used in its stead. 



The first means by which the body breadth to height relationship 

 can be investigated is via the computation of ratios - in this case, bi- 

 iliac breadth / femoral bicondylar length x 100. Due to the fact that 

 females have wider trunks relative to stature than do males, the 

 values for this index are significantly different between the sexes 

 (two-tailed t test, p < 0.000 1 ), and therefore the Cheddar specimen is 

 compared solely to males for this variable. Table 6 reports the 

 summary statistics for the males in the comparative sample and the 

 Cheddar specimen. Gough's Cave 1 lies well within 1 standard 

 deviation of the Mesolithic, LUP and recent European male means. 

 Likewise, his value is only 1 .4 standard deviations above the North 

 African mean. However, he falls over 3 standard deviations above the 

 recent Sub-Saharan African mean; as discussed below, this group is 

 characterized by some of the longest limbs and narrowest trunks of 



Table 6 Summary statistics for Gough's Cave 1 , fossil and recent human 

 males - femoral head/femoral length ratios (FHAP/FL) and bi-iliac 

 breadth/femoral length ratios (BIB/FL). 



FHAP/FL 



BIB/FL 



Gough's Cave 1 10.7 63.3 



European Mesolithic 10.7,0.5,6 62.1,3.2,6 



European Late Upper Palaeolithic 10.8,0.7,15 61.2.5.1.10 



European Early Upper Palaeolithic 10.1, 0.4, 1 56.6, 3.2, 6 



European Neandertals 12.3,0.4,4 69.8.-, 1 



Recent Europeans 10.6,0.5.134 61.2.3.4.126 



Recent North Africans 9.9,0.6,72 57.3,4.4,60 



Recent Sub-Saharans 9.5, 0.6, 53 52.6. 3.0, 49 



any humans. Interestingly, while based on extremely small samples, 

 the earlier European fossil samples stand in marked contrast to each 

 other and to recent Europeans. The Neandertals (albeit solely repre- 

 sented by the La Chapelle-aux-Saints 1 specimen) are characterized 

 by an extremely high index, indicative of their broad body breadth 

 relative to stature (Ruff, 1991, 1993, 1994;Trinkause/a/., 1994).By 

 way of contrast, the earliest modern European males (represented by 

 6 individuals) have low indices; in fact, their mean index falls 

 between those of the North and Sub-Saharan Africans. 



A second means of evaluating relative body breadth has been used 

 extensively by Ruff (1991, 1993, 1994), and involves plotting rela- 

 tive bi-iliac breadth indices, like those calculated above, against 

 stature in bivariate space. Using this method, one can evaluate the 

 relationship between the 'size-corrected' index and a measure of 

 overall size (in Ruff's case, stature; here again, femoral length is 

 used in its stead). Ruff has shown that among recent humans, there is 

 little overlap among broad geographically circumscribed samples 

 for this bivariate relationship, and thus this method could provide 

 some insight into the relative position of the Cheddar specimen. 

 Figure 1 is a scatter plot of the bi-iliac breadth/femoral length ratios 

 regressed on femoral length for the recent Sub-Saharan Africans 

 (squares), the recent Europeans (crosses) and Gough's Cave 1 (star). 

 The lines fitted to the recent samples are least-squares regression 



350 



417 483 



Femoral Length 



Fig. 1 Scatter plot of bi-iliac breadth/femoral length index on femoral 

 length. Recent Europeans are indicated by crosses; recent Sub-Saharan 

 Africans by squares. Gough's Cave 1 is indicated by a star. The lines for 

 the two recent human samples are least-squares regression lines. 



