for the quotient mean observed range/ a for various sizes of n. The ap- 

 proximate S.D.s obtained in this way are marked in Table 9 by a super- 

 script b. 



In several instances the S.D. has been computed both directly and from 

 the range for a given sample, in order to compare the results. For fairly large 

 samples the 2 values approximate moderately well: for example, for 95 male 

 H. lar, the values obtained are 7.51 (from the deviations) and 7.23 (from the 

 range). For 85 female H. lar the 2 estimates are 7.81 and 6.94, respectively. 

 With smaller samples, however, the discrepancy is larger: for 34 male chim- 

 panzees, the estimates are 40.61 and 41.77 c.c, whilst for 27 female chimpan- 

 zees, the estimates are 3370 and 41.25 c.c. In the case of these 4 series, only 

 the S.D. derived from the actual deviations has been used in the later esti- 

 mates of skewing (Table 12). 



The standard deviation estimated by either method is then expressed 

 as a percentage of the sample mean to give the coefficient of variation (V). 



Variability: results on living hominoids. Among the different sam- 

 ples of hominoid cranial capacities cited, V varies widely. For single sex 

 samples, where n is reasonably large, the lowest value of V is 7.22 per cent 

 (for 95 male H. lar), while the highest value, excluding modern man, is 

 13.30 per cent (for 72 male gorilla). 



In the chimpanzee series, V for males varies from 7.94 to 1 1.63 per cent, 

 the figure for 163 total pooled males being 9.76 per cent. Among female 

 chimpanzees, V varies from 8.24 to 9.24 per cent, the figure for 200 pooled 

 females being 8.74 per cent. Thus, while considerable variability of variabil- 

 ity obtains among various samples of chimpanzees, the overall impression is 

 that male capacities are somewhat more variable than female. 



For male orangutan, V ranges from 7.70 to 1 1.72 per cent, the value for 

 203 pooled males being 9.19 per cent. Among females, the values of V stretch 

 from 9.31 to 10.24 P er cent, whilst the value for 199 pooled females is 10.60 



NOTES TO TABLE 9: 



" Determined from actual deviations by Ashton ( 1950) alter Zuckerman (1928). Ashton and Spence 

 (1958), Randall (1943-1944), and by my research assistants, Miss C. J. Orkin and Mr. C. Block 



6 Estimated by Miss C. J. Orkin and myself from the sample range, bv the use of a table ol relation- 

 ships between the sample frequency, the mean observed range, and the standard deviation (Simpson, 

 Roe. and Lewontin 1960, p. 141, Table 1). 



2£ 44 



