Everyone knows there are species, although no one has yet proposed a definition acceptable to all biologists, nor is anyone likely to. Nevertheless, there seems to be no disagreement as to what the process of speciation involves. Since species differences, no matter how species are defined, can be maintained only by barriers of some sort to gene exchange, speciation obviously involves the establishment of such barriers. Speciation can be discussed, therefore, without a precise definition of the end product. Most biologists, however, are likely to ask what there is to discuss, for not only has the process been described repeatedly in recent years, but nearly all treatments of the subject have been in essential agreement since Goldschmidt's arguments for speciation by "hopeful monsters" (1) were refuted a quarter of a century ago.Evolution leading to speciation, except speciation by polyploidy, is generally considered to be a gradual process that includes a succession of stages, from the formation of barely detectable ecological or geographical races to ecogeographic differentiation readily recognizable at the taxonomic level of subspecies. Subsequent development of reproductive barriers carries differentiation to the level of species. Speciation in this manner undoubtedly occurs in flowering plants, for in some genera all stages of the sequence can be found. On the other hand, the normal mode of speciation in many genera appears to have no direct relation to the formation of ecogeographic races or subspecies. It was also Goldschmidt's thesis (1) that speciation is a process distinct from subspeciation. He postulated that speciation results from the establishment of systemic mutations that in-volve chromosome reorganization and change the reaction norm of the entire genotype, but he was unable to substantiate this mechanism convincingly. I agree with him, however, that evolution involving chromosome reorganization is very different from that resulting from gradual accumulation of genic differences. This is not to say that chromosome rearrangements never become established in the course of ecogeographic differentiation, but rather that chromosome reorganization frequently is not a gradual one-by-one process. In the case of rapid chromosome reorganization, speciation is not an extension of ecogeographic race formation.Although I wish primarily to discuss very rapid, saltational speciation, which in recent years has generally received little attention, it is useful, for comparison, to consider first some of the general characteristics of species and speciation in groups of flowering plants when little or no chromosome reorganization is involved. My purpose is to emphasize that there are two distinct modes of speciation, that the products are not equivalent in their relationship to one another, and that they set the stage for different patterns of subsequent evolution. I realize that some of my statements are oversimplifications and that other examples would show that the alternatives are not always as sharply defined ...
