162 



PROCEEDINGS OF THE NATIONAL MUSEUM 



increases when they are isolated from the skeleton. This factor 

 probably plays a part in the Peruvian series. As for deformity, no 

 one can be sure of entirely excluding it, because there is then the 

 danger of going too far and removing true brachycranic individuals. 

 As others have noted (see Stewart, 1940), the Pecos series includes 

 a good number of slightly deformed skulls, which fact is reflected 

 in the high sigmas for the vault diameters and indices. Then, for 

 our purposes, the Shell Mound and Arikara series, being largely 

 sexed from the skeleton and being completely free from deformity, 

 furnish the best check on variability. 



Table 3. — Comparative standard deviations: Males 



1 Cupisnique, Mochica, and Miscellaneous Chicama ('see table 2). 



' Hooton (1930). Total series A: 50 or less individuals represent the "undeformed" subseries; l.i her 



numbers include deformed skulls. 



3 Newman and Snow (1942). Total Shell Mound Series, (able 27. 



« Yon Bonin and Morant (1938), table 12. The figures for cranial module and upper facial index have 

 been supplied by the author. 



In examining table 3 we see that the Peruvians show about the same 

 variability as the Arikara and Shell Mound series, except chiefly in 

 the cranial index and the length and breadth of the skull. There 

 seems to be good reason to believe, therefore, that this unnatural 

 variability of the vault diameters, as in the case of the Pecos group, 

 is caused by the inclusion of a few slightly deformed individuals. 

 We may note that Howells (1941) reports a "mean sigma" for the 

 cranial index of 3.22 for 23 European series, and von Bonin 

 and Morant (1938') give 3.12 for 14 North American Indian series 

 comprising 1,073 skulls. 



