One or two more cases of fertile hybrids may be given before we pass on to the corresponding experiments in plants. Professor Alfred Newton received from a friend a pair of hybrid ducks, bred from a common duck (Anas boschas), and a pintail (Dafila acuta). From these he obtained four ducklings, but these latter, when grown up, proved infertile, and did not breed again. In this case we have the results of close interbreeding, with too great a difference between the original species, combining to produce infertility, yet the fact of a hybrid from such a pair producing healthy offspring is itself noteworthy. Still more extraordinary is the following statement of Mr. Low: "It has been long known to shepherds, though questioned by naturalists, that the progeny of the cross between the sheep and goat is fertile. Breeds of this mixed race are numerous in the north of Europe."[55] Nothing appears to be known of such hybrids either in Scandinavia or in Italy; but Professor Giglioli of Florence has kindly given me some useful references to works in which they are described. The following extract from his letter is very interesting: "I need not tell you that there being such hybrids is now generally accepted as a fact. Buffon (_Supplements_, tom. iii. p. 7, 1756) obtained one such hybrid in 1751 and eight in 1752. Sanson (_La Culture_, vol. vi. p. 372, 1865) mentions a case observed in the Vosges, France. Geoff. St. Hilaire (_Hist. Nat. Gén. des reg. org._, vol. iii. p. 163) was the first to mention, I believe, that in different parts of South America the ram is more usually crossed with the she-goat than the sheep with the he-goat. The well-known 'pellones' of Chile are produced by the second and third generation of such hybrids (Gay, 'Hist, de Chile,' vol. i. p. 466, _Agriculture_, 1862). Hybrids bred from goat and sheep are called 'chabin' in French, and 'cabruno' in Spanish. In Chile such hybrids are called 'carneros lanudos'; their breeding _inter se_ appears to be not always successful, and often the original cross has to be recommenced to obtain the proportion of three-eighths of he-goat and five-eighths of sheep, or of three-eighths of ram and five-eighths of she-goat; such being the reputed best hybrids." With these numerous facts recorded by competent observers we can hardly doubt that races of hybrids between these very distinct species have been produced, and that such hybrids are fairly fertile _inter se_; and the analogous facts already given lead us to believe that whatever amount of infertility may at first exist could be eliminated by careful selection, if the crossed races were bred in large numbers and over a considerable area of country. This case is especially valuable, as showing how careful we should be in assuming the infertility of hybrids when experiments have been made with the progeny of a single pair, and have been continued only for one or two generations.