tified it in northern Africa. From all these places it has now disappeared more or less completely. Only in two or three other localities does it still form an appreciable element in the living population. There is one outcrop of it in a small spot in Landes, farther to the southwest, and another away up in Brittany, in that peculiar population at Lannion which we left in our last article with a promise to return to it. On the island of Oloron off the west coast there seems to be a third survival. A very ancient type has also been described by Virchow in the islands off northern Holland, which is quite likely of similar descent.
In all these cases of survival above mentioned, geographical isolation readily accounts for the phenomenon. Is that also a competent explanation for this clearest case of all in our population in Dordogne? Why should these peasants be of such direct prehistoric descent as to put every ruling house in Europe to shame? Has the population persisted simply by virtue of numbers, this having been the main center of its dispersion in prehistoric times? Or is it because of peculiarly favorable circumstances of environment? It certainly is not due to isolation alone, for this region has been overrun with all sorts of invaders, during historic times at least, from the Romans to the Saracens and the English. Nor is it due to economic unattractiveness; for, be it firmly fixed in mind, the Cro-Magnon type is not localized in the sterile Limousin hills, with their miserable stunted population. It is found to-day just to the southwest of them in a fairly open, fertile country, especially in the vicinity of Bordeaux. These peasants are not degenerate; they are, in fact, of goodly height, as indeed they should be to conform to the Cro-Magnon type. In order to determine the particular cause of this persistence of an ancient race, we must broaden our horizon once more, after this detailed analysis of Dordogne, and consider the whole southwest from the Mediterranean to Brittany as a unit. It is not impossible that the explanation for the peculiar anomalies in the distribution of the Alpine stock hereabouts may at the same time offer a clew to the problem of the Cro-Magnon type beside it.
The main question before us, postponed until the conclusion of our study of the Dordogne population, is this: Why has the Alpine race in the southwest of France, in direct opposition to the rule for all the rest of Gaul, spread itself out in such a peculiar way clear across the Garonne Valley and up to the Pyrenees? It lies at right angles with the river valley instead of along it. In other words, why is not the Alpine type isolated in the unattractive area of Auvergne instead of overflowing the fertile plains of Aquitaine? The answer is, I think, simple. Here in this uttermost part of France is a last outlet for expansion of the Alpine race repressed on every side by an aggressive alien population.
