ied the same men from different points of view. How much more confusing if each chanced to hit upon an entirely different set of ten thousand men! This, be it noted, is generally the case in practice. Nevertheless, although there is always danger in such inferences, we are fortunate in possessing so many parallel investigations that they check one another, and the tendencies all point in one direction.
These tendencies we may discover by means of curves drawn as we have indicated above in our diagram. By them we may analyze each group in detail. Every turn of the lines has a meaning. Thus, the most noticeable feature of the Sardinian curve of statures is its narrowness and height; the Ligurian one is broader at the base, with sloping sides; and the Scotch one looks as if pressure had been applied at the apex to flatten it out still farther. The interpretation is clear. In Sardinia we have a relatively unmixed population. Nearly all of the people are characterized by statures between five feet one inch (1·56 metres) and five feet five inches (1·65 metres). They are homogeneous, in other words: and they are homogeneous at the lower limit of human variation in stature. The curve is steepest on the left side. This means that the stature has been depressed to a point where neither misery nor chance variation can stunt still further; so that suddenly from seven per cent of the men of a height of five feet one and a half inches [more frequent than any given stature in Scotland] we drop to two per cent at a half inch shorter stature. A moment's consideration will show that the narrower the pyramid, the higher it must be. One hundred per cent of the people must be accounted for somewhere. If they do not scatter sidewise, their aggregation near the center will elevate the apex, or the shoulders of the curve at least. So that a sharp pyramid points to a homogeneous people. If they were all precisely alike, a single vertical line one hundred per cent high would result. On the other hand, a flattened curve indicates the introduction of some disturbing factor, be it an immigrant race, environment or what not. In this case the purity of the Sardinians is readily explicable. They have lived in the greatest isolation, set apart in the Mediterranean. A curve drawn for the Irish shows the same phenomenon. Islands demographically tend in the main to one or the other of the extremes. If unattractive, they offer examples of the purest isolation, as in Corsica and Sardinia. If inviting or on the cross-paths of navigation, like Sicily, their people speedily degenerate into mixed types. For if incentive to immigration be offered, they are approachable alike from all sides. The Scotch, as we have observed, are more or less mixed in type, and unequally subjected to the influences of environment; so that their curve shows evidence of heterogeneity.
