When we compare two series of this class which are grouped around different values, they may overlap each other. For instance, in a people of tall stature and another one of lower stature, the following theoretical distribution of numerical values of stature would be possible:— I │ II │ 1415-1455 mm. │1425-1455 mm. │ 5 cases 1455-1495 mm. │1455-1485 mm. │ 11 cases 1495-1535 mm. │1485-1515 mm. │ 44 cases 1535-1575 mm. │1515-1545 mm. │ 135 cases 1575-1615 mm. │1545-1575 mm. │ 325 cases │ │ 1615-1655 mm. │1575-1605 mm. │ 607 cases 1655-1695 mm. │1605-1635 mm. │ 882 cases 1695-1735 mm. │1635-1665 mm. │ 1000 cases 1735-1775 mm. │1665-1695 mm. │ 882 cases 1775-1815 mm. │1695-1725 mm. │ 607 cases │ │ 1815-1855 mm. │1725-1755 mm. │ 325 cases 1855-1895 mm. │1755-1785 mm. │ 135 cases 1895-1935 mm. │1785-1815 mm. │ 44 cases 1935-1975 mm. │1815-1845 mm. │ 11 cases 1975-2015 mm. │1845-1875 mm. │ 5 cases [Illustration: FIG. 2.] In these two series the group of statures from 1575 mm. to 1695 mm. occurs 1814 times in the first series, 3371 times in the second; that is to say, 1814 individuals are found in both classes, and 1557 (i.e., 3371-1814) are found only in the class of people of tall stature. In Fig. 2 I have represented these two series in one system of coordinates. Evidently all the individuals who belong to both series, and who are shown in the surface enclosed between both curves, are found in both series; and only the others, who are outside of this surface, and who belong to one of the groups, are not found in the other one. Bearing in mind these facts relating to types and variability, we are in a position to turn to a consideration of the characteristics of primitive man and of civilized man, and of members of distinct races.