It seems impossible, with such a state of affairs, that a race in which alcohol was widely used for a long period of time, could avoid extinction. At any rate, the races which have used alcohol longest ought to show great degeneracy--unless there be some regenerative process at work constantly counteracting this cumulative effect of the racial poison in impairing the germ-plasm. Such a proposition at once demands an appeal to history. What is found in examination of the races that have used alcohol the longest? Have they undergone a progressive physical degeneracy, as should be expected? By no means. In this particular respect they seem to have become stronger rather than weaker, as time went on; that is, they have been less and less injured by alcohol in each century, as far as can be told. Examination of the history of nations which are now comparatively sober, although having access to unlimited quantities of alcohol, shows that at an earlier period in their history, they were notoriously drunken; and the sobriety of a race seems to be proportioned to the length of time in which it has had experience of alcohol. The Mediterranean peoples, who have had abundance of it from the earliest period recorded, are now relatively temperate. One rarely sees a drunkard among them, although many individuals in them would never think of drinking water or any other non-alcoholic beverage. In the northern nations, where the experience of alcohol has been less prolonged, there is still a good deal of drunkenness, although not so much as formerly. But among nations to whom strong alcohol has only recently been made available--the American Indian, for instance, or the Eskimo--drunkenness is frequent wherever the protecting arm of government does not interfere. What bearing does this have on the theory of racial poisons? Surely a consideration of the principle of natural selection will make it clear that alcohol is acting as an instrument of racial purification through the elimination of weak stocks. It is a drastic sort of purification, which one can hardly view with complacency; but the effect, nevertheless, seems clear cut. To demonstrate the action of natural selection, we must first demonstrate the existence of variations on which it can act. This is not difficult in the character under consideration--namely, the greater or less capacity of individuals to be attracted by alcohol, to an injurious degree. As G. Archdall Reid has pointed out,[18] men drink for at least three different reasons: (1) to satisfy thirst. This leads to the use of a light wine or a malt liquor. (2) To gratify the palate. This again usually results in the use of drinks of low alcohol content, in which the flavor is the main consideration. (3) Finally, men drink "to induce those peculiar feelings, those peculiar frames of mind" caused by alcohol.