Page:The American Indian.djvu/388

322

accompanying tables. If, for instance, there were no correspondences between the statures of social groups within the same area, we should expect the distributions in the successive columns to be similar, or to have a chance relation. This is not what we observe, but, on the contrary, in almost every case there is a tendency to cluster around an area average. Thus, our initial assumption seems justified and we may expect a somatic grouping of natives at least roughly coincident with the culture grouping.

If, however, we disregard all such assumptions and proceed empirically with the compilation of a distribution map we reach a result shown in Fig. 100. First, we may be reminded