Between the highest mean observed among the Southern Patagonians, and the lowest mean among the Bosjesmans, we find a difference of 0^m·554 (1ft 9·8 inch). The difference between individuals will be 0^m·930 (3ft 0·6 inch). But I think that it ought to be reduced to 0^m·790 (2ft 7·1 inch), adopting as a mean the height 1^m·14 (3ft 8·88 inch) given by Barrow as the height of a Bosjesman woman who had had several children. We are thus certain that we are not taking a case of teratological dwarfishness as a possible normal state. Travellers have not often measured separately the height of men and women. Uniting the facts of this nature which I have been able to collect, we find 0^m·141 (5·57 inch) as the mean difference between the heights of the sexes, and 0·973 as the mean ratio, the woman being everywhere shorter than the man. Among the Lapps, according to Capel Brooke and Campbell, the mean difference is as high as 0^m·278 (10·94 inches); in Austria, it is as low as 0^m·037 (145 inch) according to Liharzik. V. _Proportion of the body and of the limbs._ In all the races of our domestic animals, the relative development of the different parts of the body, the _proportions_, have a characteristic value, which is equal and frequently superior to that of height. No one would think of separating the greyhound from the harrier. It ought to be exactly the same with man. With the animal, races are formed by a selection more or less open, and undertaken for a fixed purpose. The proportions of the different parts of the body thus acquire a fixity, which cannot be found in human races on account of the absence of selection. This variability is found even in the simplest relations, and in those which might be considered fundamental. Such is the relation of the height of the head to the total height. Gerdy, who has taken up this question in a special manner, has found that the height of Frenchmen is rarely beyond 7½ heads, most frequently a little more than 8 heads, and sometimes 9. The artistic ideal is no more fixed than the reality, in spite of the mathematical rules laid down, from Vitruvius to Liharzik and Silberman. The table drawn up by Audran shows the variation from 7-19/48 heads (the Egyptian Termes) to 7-43/48 (the Farnese Hercules). The difference between these two extremes is exactly half a head. Painters have taken still more liberty. Raphael has only given a height of 6 heads to some of his figures, and Michael Angelo has given them 8 or more. The Pythian Apollo (7-42/48 heads), the Laocoon (7-27/48), are nevertheless _chefs-d’œuvre_, and we rightly bestow an equal amount of admiration upon the two Italian masters. The reason is just the same as with the rest of organised beings: man’s organism is not subject to absolute laws, nor to a rigorously fixed development.