interesting observations upon this subject. Where peasant houses are closely aggregated or bunched in little villages, it is easy for each family to maintain its separate dwelling, and yet for them all to co-operate with one another in daily labor. On the other hand, the peasant whose house is quite apart from those of his neighbors, placed squarely, perhaps, in the center of his landed property, must of necessity take his farm laborers into his own household. Thus, where population is scattered evenly over a district, not in closely built hamlets, but in widely separated houses, it generally happens that there is considerable "home intermixture." Several families or parts of families live under the same roof. Applying these considerations to Brittany, it seems as if the very low percentage of separate "home families" were a result of just such a broadcast distribution of population. This absence of hamlets in turn is a direct result of geology and climate. In Brittany the rainfall is very heavy; water courses and springs abound on all sides. The soil is at the same time thin, overlying an impervious granite formation. This makes it possible to build houses wherever convenient, without anxiety concerning water supply. The exact opposite of this occurs along the dry Mediterranean coast, where water is a marketable commodity; and in those departments with a permeable chalk soil, where water disappears rapidly in subterranean streams. In these latter cases houses inevitably collect about the water courses and springs, and a high proportion of aggregated population at once is manifested, with all that is thereby implied, socially speaking. One of the first results would be that each family in such a hamlet might occupy its own dwelling exclusively.
Geographical factors have also operated in still another way in Brittany to discourage the growth of closely built villages. This region is so remote from any of the routes of military invasion from the east that no necessity has ever arisen for compacting the population in villages capable of ready defense. Levasseur gives this as an important element in producing the contrasts in the proportion of urban population between the different parts of France. In all of our areas of isolation, the Alps, Auvergne, or Brittany, protected by Nature against intrusion of enemies, the population can safely scatter as it will. In any case, as we have said, the effect upon the family, especially in all that concerns its separate existence under a roof by itself, is very patent.
If the geographical isolation peculiar to the areas occupied by the Alpine race is thus potent in the way we have indicated, why may it not appear in political as well as in social affairs? Conserva-
- Bulletin de l'Institut Internationale de Statistique, iii, 1888, pp. 70 et seq.