Now, I do not contend that this fact in itself raises any difficulty against the theory of special creation. But I do think that a very serious difficulty is raised when to this fact we add another--namely, that on every biological region we encounter species related to other species in genera, and usually also genera related to other genera in families. For if each of all the constituent species of a genus, and even of a family, were separately created, we must hence conclude that in depositing them there was an unaccountable design manifested to make areas of distribution correspond to the natural affinities of their inhabitants. For example, the humming-birds are geographically restricted to America, and number 120 genera, comprising over 400 species. Hence, if this betokens 400 separate acts of creation, it cannot possibly have been due to chance that they were all performed on the same continent: it must have been design which led to every species of this large family of birds having been deposited in one geographical area. Or, to take a case where only the species of a single genus are concerned. The rats and mice proper constitute a genus which comprises altogether more than 100 species, and they are all exclusively restricted to the Old World. In the New World they are represented by another genus comprising about 70 species, which resemble their Old World cousins in form and habits; but differ from them in dentition and other such minor points. Now, the question is,--Why should all the 100 species have been separately created on one side of the Atlantic with one pattern of dentition, and all the 70 species on the other side with another pattern? What has the Atlantic Ocean got to do with any "archetypal plan" of rats' teeth? Or again, to recur to Australia, why should all the mammalian forms of life be restricted to the one group of Marsupials, when we know that not only the Rodents, such as the rabbit, but all other orders of mammals, would thrive there equally well. And similarly, of course, in countless other instances. Everywhere we meet with this same correlation between areas of distribution and affinities of classification.