Statistical Methods in Anthropology

Abstract

IN their interesting letter in Nature¹ on statistical methods in anthropology, Dr. J. Bronowski and Mr. W. M. Long suggest that the reason why Prof. Zuckerman's findings were at odds with those based on classical comparisons was that he had compared each measurement separately instead of using the comprehensive test provided by the discriminant function. They took as an example the discriminant function calculated from measurements on forty human and forty-four chimpanzee lower milk canines. This gave a wide separation in the ranges of the two groups and conclusively indicated (a) that the Kromdraai and Taungs milk canines are not those of chimpanzees and (b) that they fall within the human range.