For more than a century systematists and anthropologists have been describing, Cataloguing, and classifying races in man, in animals, and in plants. Yet despite the hundreds or perhaps thousands of investigations which have been published on this subject, the concept of race remains ambiguous. This ambiguity is inherent in the method, based on description of morphological averages, by which the problem has been approached. A racial type or standard is arrived at, which is, in effect, a system of statistical averages. Once such a type is established it becomes a tool for comparing races with one another and for deciding whether any given individual does or does not belong to the race in question. In other words, the procedure has been the same as that employed in systematic zoology and botany to delimit species and to assign a given individual to one species or another.
An implied justification of this procedure lies in the principle, the validity of which is securely established, that species evolve from races; hence species and races are assumed to be consubstantial. Another, and apparently no less valid, justification can be claimed, in that the procedure has, on the whole, stood the test of experience: systematists and anthropologists not only have succeeded in reducing the chaos of diverse racial variants to an intelligible system, but also have produced a coherent picture of the distribution and migration of many strains over the earth's surface. Nevertheless, although the method serves well as a device for the delimitation of species, it proves to be totally inadequate for the analysis of intraspecific variability.
Although it is true that species arise from races, they differ from races by virtue of the reproductive isolation of the former and the absence of such isolation in the latter. A species is a genetically closed system; species do not regularly exchange genes. Because of this fact, any given individual (species hybrids excepted) always belongs to a given species, never to two or more. Races, on the other hand, are genetically open systems. Their populations are channels through which genes can and do flow from race to race. And, because genes may vary independently from one another, an individual may carry some genes which occur frequently in the representatives of one population or race and other genes more characteristic of another. Such an individual may, in this sense, belong to two or more races at the same time; it is in fact compounded of elements of both.
To analyze and describe races adequately it is necessary to analyze and describe the distribution of the variable genes one by one. A system of morphological averages may well serve as an exploratory device, but a basic understanding of principles of racial variation can come only from knowledge of . . .