The variety thus reproduced itself purely in the eldest, while the normal type reproduced itself purely in the third, and almost purely in the second and last; so that it would seem, at first, as if the normal type were more powerful than the variety. But all these children grew up and intermarried with normal wives and husbands, and then note what took place—Salvator had four children, three of whom exhibited the hexadactyle members of their grandfather and father, while the youngest had the pentadactyle limbs of the mother and grandmother; so that here, notwithstanding a double pentadactyle dilution of the blood the hexadactyle variety had the best of it. The same prepotency of the variety was still more markedly exemplified in the progeny of two of the other children, Marie and George. Marie (whose thumbs only were deformed) gave birth to a boy with six toes, and three other normally formed children; but George, who was not quite so pure a pentadactyle, begot, first, two girls, each of whom had six fingers and toes; then a girl with six fingers on each hand, and six toes on the right foot, but only five toes on the left; and lastly, a boy with only five fingers and toes. In these instances, therefore, the variety, as it were, leaped over one generation to reproduce itself in full force in the next. Finally, the purely pentadactyle André was the father of many children, not one of whom departed from the normal parental type.’ 169. The instances now quoted illustrate two things. Both tell us how varieties arise, we may say spontaneously, or in other words we cannot tell how; and the former instance, that of the Ancon breed, shows us moreover that such varieties when they do occur may be rendered permanent by means of artificial selection. If the six-fingered descendants of Gratio Kelleia had been forced to intermarry amongst themselves it is highly probable that we should have had a permanent hexadactyle variety of the human race. It has likewise been shown by Charles Darwin that the pouter, the fan-tail, the carrier, and the tumbler are all varieties of the common rock-pigeon. 170. It thus appears that permanent varieties may be produced by artificial selection. Now Darwin and Wallace have brought before us the very great fact that similar changes can also be produced by natural selection.