mony of divorce, as to the cohesiveness of the domestic institution. So far as Savoy, Alsace-Lorraine, and Auvergne, our principal areas occupied by the Alpine or Celtic race, are concerned, the parallel with the map of divorce is quite close. The Mediterranean coast strip, nay, even the intrusive zone up the Rhône Valley, are indicated as areas where the family is less cohesive than in the upland areas of isolation. But what shall we say about Brittany? Racially, and in stability of the family as well, it belongs with Savoy
and Auvergne as an area of isolation, characterized by comparatively backward social phenomena. Nevertheless, inspection of our map shows it to be the region where such "home intermixture" is exceedingly prevalent. Less than one half the families live under entirely separate roofs, whereas in the other Celtic areas the proportion of independent families is often above ninety per cent.
This peculiar anomaly in the case of Brittany is all the more notable, as this region is one of the most conservative in all France, judged by the character of its social phenomena. Some disturbing factor is evidently at work. It seems to be purely environmental. Surprising as it may appear, this exaggerated "home intermixture" in the Armorican peninsula is apparently to a large degree referable to its geological and climatic peculiarities. Levasseur makes some