This is a fact not generally recognised by naturalists and anthropologists who have touched upon these questions. For instance, it has been admitted as proved that domestic races, when they have returned to the feral state, reassume all the original characters of the species. This is a mistake. The fact is, that both with animals and plants, _escaped races_ lose a certain number of characters, and frequently the most apparent ones, which they owe to domestication; they reassume others which they had lost during their period of servitude, but the former are more frequently only diminished and masked by the latter. If fruit-trees escaped from our orchards, if our horses, dogs, cattle, and pigs, when they have become wild, had really reassumed the original type of the species, they ought to present in every area which they inhabit the marked uniformity characteristic of animals who were never subject to man. This is not the case. They ought in particular to preserve no trace of their acquired characters. Now, the latter are partly persistent. Vans Mons has found apple-trees and pear-trees of Belgium, in a wild state, in the forest of the Ardennes. The prickles had reappeared, the fruit had become small and bitter again, but the principal cultivated varieties were still to be recognised. I have established an analogous fact with regard to cling-stone and free-stone peach-trees in a valley of the Cévennes. Similarly Martin de Moussy has recognised in the troops of dogs which have become wild in America, all the chief races from which they had been derived, although they reassumed the general characters of the tan-coloured type. IX. From the number of observations which have been collected among plants and animals, and of which I can only notice a small number, it is easy to understand the appearance and multiplication of human races, and to account for certain general facts, some of which are closely connected with our history. Let us state at starting that with man, as with animals, _varieties_ have appeared at times which may be classed among hemitery. Individuals, exhibiting from their birth exceptional characters, are none the less healthy, and sometimes have very remarkable power of transmission. Edward Lambert, born in 1717 of perfectly healthy parents, had all his life a kind of carapace more than an inch thick, and irregularly fissured, which gave him the name of _the porcupine man_. All his children, to the number of six, and his two grandchildren, inherited this strange modification of the skin, although his wife and his daughter-in-law did not show the least trace of it. In the Colburn family, four generations were marked with polydactylism which was derived from the grandmother of the great calculator. At the fourth generation, four children out of eight still had supernumerary fingers, though at each generation normal blood was mixed with the teratological blood.