But in attempting to divide mankind into such groups, an ethnologist is met by a serious and apparently insurmountable difficulty. The gradation of color is so imperceptible from the clearest white to the jettest black; and even anatomical peculiarities, normal in one branch, are found to exist, albeit in exceptional cases, in many others; so that the ethnographers scarce know where to stop in their classification, and while some recognize but three grand varieties, others contend for five, for eleven, or even for a much greater number. This difficulty arises, in my estimation, mainly from the attempt to class mankind into different species, that is, groups who have a separate origin; and also, from the proneness to draw deductions from individual instances, by which almost any absurdity can be sustained, or truth refuted. As we have already inveighed against the latter error, and shall therefore try to avoid falling into it; and as we have no desire to enter the field of discussion about unity or plurality of species, we hope, in a great measure, to obviate the difficulties that beset the path of so many inquirers. By the word _race_[20] we mean, both here and in the body of the work, such branches of the human family as are distinguished in the aggregate by certain well-defined physical or mental peculiarities, independent of the question whether they be of identical or diverse origin. For the sake of simplicity, these races are arranged in several principal classes, according to their relative affinities and resemblances. The most popular system of arrangement is that of Blumenbach, who recognizes five grand divisions, distinguished by appellations descriptive either of color or geographical position, viz: the White, Circassian, or European; the Yellow, Altaic, Asiatic, or Mongolian; the Red, American, or Indian; the Brown, or Malay; and, lastly, the Black, African, or negro. This division, though the most commonly adopted, has no superior claims above any other. Not only are its designations liable to very serious objections, but it is, in itself, entirely arbitrary. The Hottentot differs as much from the negro as the latter does from the Malay; and the Polynesian from the Malay more than the American from the Mongolian. Upon the same principle, then, the number of classes might be indefinitely extended. Mr. Gobineau thought three classes sufficient to answer every purpose, and these he calls respectively the white, yellow, and black. Mr. Latham,[21] the great ethnographer, adopts a system almost precisely similar to our author's, and upon grounds entirely different. Though, for my own part, I should prefer a greater number of primary divisions, I confess that this coincidence of opinion in two men, pursuing, independent of, and unknown to each other, different paths of investigation, is a strong evidence of the correctness of their system, which, moreover, has the merit of great simplicity and clearness.