When a species is not adapted to cross-pollination, but is self-pollinated, and produces a larger range of forms, they are not the result of natural selection, but rather, as De Vries explains in regard to such mutations, or species in the making, due to the production of mutants in the early history of the plant, differing in one or more characters from the type, and perpetuated by the addiction of the plant to self-pollination. By degrees the creation of small variations from the original will bring about great divergence between the parent stock and the latest mutants, until the extreme forms may rightly be considered distinct species, especially if the connecting forms have failed to survive. Much then depends, in our conception of species, upon the survival of intermediates, or the reverse. The existence of natural selection, indeed, largely hinges upon the occurrence of such a phenomenon as survival, or the reverse, or in other words of adaptations to meet special conditions.

Different from continuous variations are those produced by a change in the environment. Such small continuous variations were emphasized by Darwin as being advantageous and in the struggle for existence liable to be selected, other plants which do not possess them not surviving. The extremes survive, hence the difficulty experienced in tracing the evolution of species owing to the disappearance of connecting forms.

So far as the results of artificial selection go, in the cultivation of species, Jol annsen has not found that divergent forms arise; the usual result is the production of a pure line with a higher average weight of seed (the special character experimented with), and beyond this no great variation occurs.

But there are other variations of kind rather than degree, and one obtains sports (as in the case of white flowers), discontinuous variations, or mutations. According to Darwin such sports or mutations are rare. His experience is in direct opposition to that of De Vries, who worked upon Cenothera, the parent plant of his mutants being, however, of doubtful nature. Whether mutations are frequent or not, species may arise suddenly. Owing to the influence of cross-pollination, indeed, it is natural that mutants should, once originated, be quickly altered, and their origin obscured, by the influence of crossing, which is, as a rule, the prevalent type of pollination. Mutations would also become crossed with typical forms, and rare mutants would be eliminated.

Self-pollination would, on the contrary, help to perpetuate mutations, and the rarer, more extreme form of pollination, or cleistogamy, whereby outside influence of insect visitors is entirely excluded, would have the effect of establishing any mutations that could arise and in due course of producing plants entitled to rank as sub-species, or even new species.

There is in general a tendency in plants to preserve specific identity, given uniform conditions of growth. The plant itself has a tendency to avoid variation as a rule, a fact reflected in the prepotency of pollen or the readiness with which one individual of a species is crossed with another of the same species, rather than with that of another species, or even genus. Also like individuals of the same species are more likely to be crossed than dissimilar individuals. This tendency is, however, discounted by the erratic conduct of the agents that effect cross-pollination, namely insects, though it has been shown by Lord Avebury that, as a rule, insects have a marked preference for visiting plants of the same species in their wanderings to and fro in search of pollen or honey. But, apart from colour and form of flower, insects are not, we must assume, to be credited with the power of discrimination between forms differing in other respects, where those features do not affect their search for the above items of food.

The relative scarcity of impure crosses or hybrids shows how unerring is the instinct of insects. Hybrids, as distinct from species and varieties, are not the main links in the chain of evolution of species. They are exceptional illustrations of the influence of cross-pollination in the production of new forms. Incidentally they may be the origin of new species, but they illustrate the manner in which species or varieties may occur, since they are the results of the same process which brings about evolution of forms in one direction, by cross-pollination. Mutations, etc, produced in self-fertile plants are illustrations of a similar process of evolution brought about by other means, and due presumably to the inherent tendency to vary (where variation occurs), which is obscured somewhat in crossed plants by the influence of the different characters of the parents in a cross-pollinated plant, or, at any rate, the possibility of their difference, since distance may in such cases be responsible for variation in the two individuals crossed, whereas in self-pollination no such difference exists, and the plant is grown under similar conditions.

At one time it was considered hybrids were rare, and generally sterile. A Frenchman, Naudin, disproved this, and found many hybrids capable of fertilization.

It was reserved for the Austrian monk, Gregor Mendel, to explain and express arithmetically the principles of hybridity. He found the result of crossing two different forms was not the production of intermediates, but of plants with the characters of one of the parents. Wrinkled and smooth-seeded peas when crossed produced hybrids with round peas in the first generation, and he termed the round-pea type a dominant, whilst the other type was recessive. The hybrid, when pollinated with its own pollen, in the second or F2 generation, produced some plants with round, some with wrinkled peas, in the proportion of 3 to 1; i.e. a quarter of the descendants of a generation revert back to the original with a recessive character. The actual result in the first generation may be of an intermediate type, and in the F2 generation some of both characters of the parents, and some intermediates, will occur. The resultant hybrid may appear to be intermediate, when really the result is a combination of two dominant characters giving the semblance only of an intermediate. These principles can be readily worked out where the parent plants do not present a complex of characters each of which is likely to influence the resulting hybrid. In such cases the elimination of each character separately is the only means of determining the nature of the parents in a first cross, and it becomes more difficult in a second or third cross.