The next taxonomic division below that of species is the subspecies. A subspecies is a regional population of a polytypic species (a species with a number of separate populations) which meets two tests: (1) it occupies a distinct geographical territory; (2) it differs from other subspecies o f the same species in measurable characteristics to a considerable degree (to be specified shortly).
Subspecies must by definition be allopatric: if several subspecies were to inhabit a single region, they would breed together and the differences between them would be obliterated. Within its own geographical territory, which has an environmental character of its own, the subspecies has achieved, or is in the process of achieving, an adjustment to its local food supply, to the local climate, and to the behavior patterns of other animal species with which it shares its domain. After each subspecies has worked out a balance with all other elements in its local environment, it is not likely to change very much until its situation changes: natural selection will prune off unfavorable mutations that arise locally and keep the favored gene ratio constant.
Over the border, which may be a natural barrier such as a range of mountains or a patch of desert, or even a critical isotherm, may be found another subspecies of the same species, equally well established in a state of equilibrium with its environment. As the two environments differ in certain details, so do the genetic structures of its occupants. What is good for A is less advantageous for B, and vice versa. In each territory, natural selection keeps the gene structure of the local subspecies constant by also eliminating unfavorable genes that flow over the border. However, genes which are unfavorable in both environments may be eliminated in both populations, so that A and B may evolve together into a new polytypic species that retains its original set of subspecies. This is what we think happened when a number of human sub> species passed the threshold from Homo erectus to Homo sapiens.
Taxonomists have set up an arbitrary procedure to determine whether two or more populations within a species are morphologically different enough to qualify as subspecies. It is called the overlap test and is applied both to visible criteria, such as tooth size, and to invisible ones, such as blood groups. If in any well defined, presumably heritable morphological character, a representative sample of population A differs from a representative sample of population B to or beyond a critical degree, then we are dealing with subspecies. The critical degree is 75 per cent. If 75 per cent or more individuals of A are different from 100 per cent of B, then the two are probably subspecies.'
This method was devised for use on large samples of living animal populations and it can be applied to modern anthropometric series, but it is rarely if ever useful in the study of fossil man because we have few samples large enough for analysis by probability statistics. When applied to modern human populations, this test shows that Homo sapiens is at present a polymorphic species divided into a number of clearly differentiated subspecies, each centered in its own territory.
The concept of subspecies is essentially zoological and is used almost entirely to describe regional variations in animal species. However, paleontologists also use it occasionally, to describe steps in a single evolutionary line which they consider too small to merit the rank of separate species. Such units may be called successional subspecies, or waagenons-named for a mid-nineteenth century paleontologist, W. Waagen.' In order to keep confusion to a minimum I shall not use the word subspecies in this book to designate such successive units. When successive species must be split, I shall do it in terms of the evolutionary levels or grades through which they have passed.