SKELETAL REMAINS FROM PERU STEWART 177 



The comparison carried out in table 9 is by means of the simple 

 and probably crude device known as the average difference of the 

 means. This method has been employed by Shapiro and others to call 

 attention to metrical similarities among peoples of the far north (see 

 Stewart, 1939). According to the current interpretation of the aver- 

 age difference, a figure that approaches 2 suggests a doubtful iden- 

 tity of type. From this point of view there is little justification for 

 grouping together the Pueblo and Spoon River groups. On the other 

 hand, the Spoon River and Maples Mills groups are perhaps properly 

 separated as different types. However, if an average difference of 

 2.06 is sufficient to separate Centralid from Sylvid, it is debatable 

 whether a difference of 1.8 is sufficiently low to warrant the union of 

 the Pueblos and Peruvians. 



Viewed from another angle, the first two groups in table 9 repre- 

 sent peoples of small build, whereas the last two groups are of large 

 build. This difference in build is reflected in the high average differ- 

 ence between groups 2 and 4. Now, if instead of comparing sizes we 

 compare shapes, we find that the average differences between the in- 

 dices are 2.06 (groups 1-2), 2.30 (3-4), and 2.28 (2-4). Thus, the 

 Peruvians are distinguished from the Pueblos by having a distinctly 

 rounder head, broader face, and lower orbits. Both of these groups 

 in turn are distinguished from the remaining two groups chiefly by 

 the difference in relative head height. 



The data on stature are also of interest in the present connection. 

 We have seen that male stature in the Chicama Valley, as computed 

 by Pearson's formula e, is 159.4 cm. Hooton's data on the Pecos 

 Pueblo (1930) when handled in the same way yield a figure of 162.2. 

 Also, Hooton's data on Madisonville (1920), which Neumann (1941a) 

 has identified as Centralid, give a stature near 167 cm. The Sylvids 

 are probably just as tall as the Madisonville population. Although 

 these figures may be regarded as comparable within the limitations 

 of the series, because they are all calculated in a like manner, they do 

 not entirely accord with the data on the living. Thus, the modern 

 highland male population through Ecuador, Peru, and Bolivia has a 

 stature close to 160 cm. (cf. Gillin, 1911; Hurtado, 1932) ; the recent 

 Pueblos average 164 cm. (Hrdlicka, 1935). 



These facts all go to show that the generalization we have been 

 considering is based upon too little knowledge of the physical composi- 

 tion of the American Indians. Although this criticism naturally does 

 not invalidate the general classificatory scheme, nevertheless it casts 

 doubts upon the scale of the differences that distinguish the individ- 

 ual types, as well as their distribution and number. 



Everyone will admit that the American Indian is variable in all 

 his physical characters and, furthermore, that some temporal and 



