Frozen Evolution. Or, that’s not the way it is, Mr. Darwin. A Farewell to Selfish Gene.

The phenotypes of existing species are more or less adapted to the conditions of the environment in which they live. If they are to start to use a new niche and enable the formation of a new species by branching speciation, they must first adapt to this new niche. This evolutionary change can only rarely occur in a single step, i.e. can occur as a result of a single mutation. Mostly a number of gradual evolutionary changes are necessary for exploitation of a new niche; these change the phenotype corresponding to the requirements of the original niche to a phenotype corresponding to the requirements of the new niche. There are a number of intermediate stages between the terminal states of the particular evolutionary pathway, where the particular intermediate links, transition forms, will probably exhibit suboptimal phenotypes from the viewpoint of both the old and new niches. Thus, their carriers will be at a disadvantage compared to the carriers of the original phenotype, which will substantially retard the progress of speciation. The entire phenomenon can be elegantly described by the model of the adaptive landscape (see I.13). There are a number of peaks in the adaptive landscape to which the individual species can gradually climb through the action of natural selection. Each of the peaks corresponds to a potential niche. Some of the peaks are occupied (exploited niches), while others are empty. In order for the members of a certain species to move from one peak to another, i.e. in order for them to form a new species, they must be able to first “climb down” from their adaptive peak and overcome the valley in the adaptive landscape, i.e. they must survive with a suboptimal phenotype from the viewpoint of fitness for a great many generations (Fig. XXI.4).

Fig. XXI.4 Speciation and overcoming valleys in the adaptive landscape. The scheme depicts a one-dimensional adaptive landscape with two adaptive peaks. In order for a species to be able to occupy the higher peak, some of its members would first have to “climb down” into the valley of the adaptive landscape, i.e. gain a phenotype that is disadvantageous from the standpoint of fitness, and only then, as a consequence of further mutations, “climb up” to a new unoccupied peak, i.e. gain the phenotype that is optimal in the particular adaptive landscape.

            In some types of speciation, suboptimality of the transition phenotypes is not a great drawback. For example, if part of the population finds itself in an isolated territory with very different conditions than those to which the phenotype of the original species was optimally adapted, the phenoptype of the population can begin to adapt to the new conditions even if the changes lead to temporary or permanent loss of fitness under the original conditions. Because the members of the population will not find themselves in the original conditions for a long time, their potentially worse competitiveness will not be a detriment. However, in some types of speciations, the newly formed species is in constant contact with the original species and, in this case, the reduced competitiveness of its members can be detrimental. Prevention of speciation that is based on worse adaptability of the individuals with transition phenotypes can be an important factor in those types of speciation where the new species evolves in close contact with the parent species. There are two basic factors that can facilitate the formation of a new species, to be more precise phenotype differentiation, under these conditions. The first factor is the multidimensionality of the actual adaptive landscape. While the landscape to which we are accustomed is three-dimensional, the adaptive landscape is actually multidimensional and each of these dimensions corresponds to one phenotype trait of a living organism. There are valleys between the individual peaks in a three-dimensional landscape that correspond to an area with suboptimal phenotypes in the three-dimensional model of the adaptive landscape. In contrast, in a multidimensional adaptive landscape, there are sorts of ridges or upland plains between the individual, often very distant peaks, along which the population can move, through the action of suitable selection pressure, from one peak to another without first climbing down to the valley.

            Another factor that allows species to pass from one adaptive peak to another is genetic drift and the related possibility of neutral evolution. In small populations, to be more exact in populations with small effective size, selection has rather low efficacy. This means that, in small populations, individuals with suboptimal phenotypes, located in deep valleys of the adaptive landscape, can survive for long times and even predominate through the effect of genetic drift. Then, when the effective size of the population again increases, the progeny of these individuals can climb back to the surrounding peaks in the adaptive landscape, including formerly unoccupied peaks. Sewall Wright explained adaptive evolution and certain forms of speciations by his shifting balance hypothesis, which is based on the principle of alternation of the effect of genetic drift with that of natural selection (VII.3).s