This principle leads, De Vries says, in the domain of the descent theory to the conception that species have arisen from each other, not continuously, but by steps. Each new step results from a new combination as compared with the old one, and the new forms are thereby completely and sharply separated from the species from which they have come. The new species is all at once there; it has arisen from the parent form without visible preparation and without transitional steps. The mutation theory stands in sharp contrast to the selection theory. The latter uses as its starting-point the common form of variability known as individual or fluctuating variation; but according to the mutation theory there are two kinds of variation that are entirely different from each other. “The fluctuating variation can, as I hope to show, not overstep the bounds of the species, even after the most prolonged selection,—much less can this kind of variation lead to the production of new, constant characters.” Each peculiarity of the organism has arisen from a preceding one, not through the common form of variation, but through a sudden change that may be quite small but is perfectly definite. This kind of variability that produces new species, De Vries calls mutability; the change itself he calls a mutation. The best-known examples of mutations are those which Darwin called “single variations” or “sports.” De Vries recognizes the following kinds of variation:— First, the polymorphic forms of the systematists. The ordinary groups which, following Linnæus, we call species, are according to De Vries collective groups, which are the outcome of mutations. Many such Linnæan species include small series of related forms, and sometimes even large numbers of such forms. These are as distinctly and completely separated from each other as are the best species. Generally these small groups are called varieties, or subspecies,—varieties when they are separated by a single striking character, subspecies when they differ in the totality of their characters, in the so-called habitus. These groups have already been recognized by some investigators as elementary species, and have been given corresponding binary names. Thus there are recognized two hundred elementary species of the form formerly called _Draba verna_. When brought under cultivation these elementary species are constant in character and transmit their peculiarities truly. They are not local races in the sense that they are the outcome in each generation of special external conditions. Many other Linnæan species are in this respect like _Draba verna_, and most varieties, De Vries thinks, are really elementary species.