[What's wrong with "development of isolating mechanisms" (and its special case in form of "reinforcement")?].

Examined are the conceptual grounds of the "reinforcement" model that play an important role in explanatory schemes of the theory of evolution. According to the current views, reinforcement (of obstacles for crossing under natural selection when there occurs a repeated contact with hybridization between two substantially divergent forms, up to the level where started hybridization is terminated and the forms become unable to cross) should be a common or even frequent phenomenon. In fact, reinforcement turns out to be a rare event needed to be specially searched for, all such events are dubious, etc. Here, empirical data on reinforcement pro and contra are analyzed, using mostly ornithological (and some other zoological) materials. It is shown that in most cases where the theory predicts reinforcement to occur, it does not happen. On the contrary, something quite opposite is observed--selection for non-crossing enhancement does not stop up the leaks in gene pools of the forms that are due to their hybridization within the zone of a repeated contact, but instead the stable hybrid zone is established which serve as a conduit for gene exchange between the forms. Through hybridization, those forms can adopt strange genes, remaining, at that, their isolation at the level of separate species. This is achieved through the system of self-non-self recognition in population system of both forms being in contact. Within the zone of a repeated contact, conditions are usually such that individuals repeatedly make mistakes forming mixed pairs and producing hybrids, i.e., at the level of individuals, self-non-self recognition is often inefficient, especially at the range edges, in zone of repeated contact and such. Efficient recognition, that ensures species separateness even under conditions of continuing hybridization, takes place upon inclusion of individuals of different origin--phenotypically 'pure' ones, hybrids, and backcrosses--into spatio-ethologic population structure of both forms, upon acquisition of territories by established groups, their inclusion in flocks and other intra-population associations of resident individuals. At that, groups of both forms predominantly incorporate their 'own' ('pure') phenotypes and reject 'strange' ones, including hybrids of intermediate phenotype. If the forms have attained the species level of separateness, hybrids do not form their own stable groups, but try to individually inset themselves into the habitation of both parents, which turns out to be not so successful as compared with phenotypically 'pure' individuals of the given form. Thereby, hybrids happen to be directly excluded from the population system of both forms even if hybridization goes on and produces them in sufficient quantity while hybrids and backcrosses themselves are not inferior to 'pure' individuals either in survivorship or reproductive success (those ones who became residents have occupied a territory and are reproducing). It is shown that usual examples of reinforcement are not only sparse but could be explained differently. Particularly, 'splitting up' of a population system which gives rise to two forms of different ecology with following strengthening/enhancement of the divergence through sexual selection is better explained by our model of self-non-self recognition than by reinforcement. Relying on this concept, the conclusion is drawn that the model of non-crossing 'reinforcement' in zones of a repeated contact with hybridization is not supported by the data, as well as the model of character displacement that follows from it. Apparently, the reasons behind its popularity consist in conformity with the 'ideology' of a more general theory that is agreed upon by most researchers, and corresponds, in whole, to known facts (in this case, the theory of modern evolutionary synthesis).