of the same species, and the study of crosses and hybrids is therefore a means of separating, to some extent, the influence of one parent from the influence of the other. This is true, however, only with reference to characteristics which are of recent acquisition, for the greater part of the history of two allied species has been the same, and they show in common everything except what has been acquired by each one since they diverged from their common ancestor.

Crossing gives no way of showing whether these common characteristics are or are not transmitted by one parent or the other or by both, but it does give us this information regarding characteristics which appear in one species but not in the other, and it is therefore the best means at our disposal for studying the influence of each parent upon the offspring.

Crossing as a Cause of Variation.

According to our theory of heredity, we can easily see how the crossing of two species or varieties should lead to variability, for when two species or varieties are crossed certain cells of the body will be hybrids between the gemmules of the male parent and the ovarian particles. inherited through the female from the egg of the preceding generation. Now the ovarian particle transmits the properties of a cell like that of the female parent, while the gemmule transmits those of a corresponding cell in the father. It is plain that corresponding cells of a female of one species or variety and of a male of another species or variety must be more different from each other than corresponding cells in a male and female of the same species or variety. The hybrid cell formed by their union would, therefore, be expected to differ more from each of them, that is, to vary more than it

does in the offspring of parents of the same variety. It is well known that this is the case; that, in domesticated animals and plants at least, crossing is a great causeaccording to some older writers the only cause-of variation.

Darwin says that it is probable that the crossing of two forms when one or both have long been domesticated or cultivated, adds to the variability of the offspring, independently of the commingling of the characters derived. from the two parent forms. He believes that new characters arise in this way in hybrids between domesticated forms, forms which have been rendered variable through cultivation, but he doubts whether we have, at present, sufficient evidence to prove that the crossing of species which have never been cultivated leads to the appearance of new characters.

The following illustrations of this law are quoted from his Variation (Vol. ii. p. 319):

"Gärtner declares, and his experience is of the highest value on such a point, that when he crossed native plants which had not been cultivated, he never once saw in the offspring any new character; but that from the odd manner in which the characters derived from the parents were combined, they sometimes appeared as if new. When, on the other hand, he crossed cultivated plants, he admits that new characters occasionally appeared. .. According to Kölreuter, hybrids in the genus Mirabilis vary almost infinitely, and he describes new and singular characters in the form of the seeds, in the colors of the anthers, in the colyledons being of immense size, in new and highly peculiar odors, in the flowers expanding early in the season, and in their closing at night. With respect to one lot of these hybrids he remarks that they presented characters exactly the reverse

of what might have been expected from their parentage.

"Professor Lecoq speaks strongly to the same effect in regard to this same genus, and asserts that many of the hybrids from Mirabilis jalapa and multiflora might easily be mistaken for distinct species, and adds that they differed in a greater degree than the other species of the genus from M. jalapa. Herbert has also described the offspring from a hybrid Rhododendron as being as unlike all others in foliage as if they had been a separate species. The common experience of floriculturists proves that the crossing and recrossing of distinct but allied plants, such as the species of Petunia, Calceolaria, Fuchsia, Verbena, etc., induces excessive variability: hence the appearance of quite new characters is probable. M. Carriere has lately discussed this subject; he states that Erythrina cristagalli had been multiplied by seed for many years, but has not yielded any varieties; it was then crossed with the allied E. herbacia, and the resistance was now overcome, and varieties were produced with flowers of extremely different size, form, and color."

Darwin, therefore, concludes that crossing, like any other change in the conditions of life, seems to be an element, probably a potent one, in causing variability.

The variability of hybrids is quite as explicable by Darwin's Pangenesis hypothesis as it is by our theory of heredity, although I do not see why, on the hypothesis of pangenesis, the hybrid offspring of domesticated forms should be any more variable than those produced between wild species.

The Offspring of Hybrids more variable than the First Generation

There is another aspect of the variability of hybrids which is very remarkable, and which is in perfect agreement with our theory of heredity, but, so far as I am aware, absolutely inexplicable without it.

This is the law that although the offspring of the first generation are generally uniform when two species or races are crossed, the subsequent generations of children produced by these hybrids display an almost infinite diversity of character. (Darwin, Variation, ii. p. 321.)

Darwin also refers to this curious law in the Origin of Species, p. 260, and attempts an explanation of it. He says: "The slight variability of hybrids in the first generation, in contrast with that in the succeeding generations, is a curious fact, and deserves attention. For it bears on the view which I have taken of one of the causes of ordinary variability, namely, that the reproductive system from being eminently sensitive to changed conditions of life, fails under these circumstances to perform its proper function of producing offspring closely similar in all respects to the parent form. Now, hybrids in the first generation are descended from species (excluding those long cultivated) which have not had their reproductive systems in any way affected, and they are not variable; but hybrids themselves have their reproductive systems seriously affected, and their descendants are highly variable."

According to this view, the variability of the descendants of hybrids is a sort of monstrosity, due to the failure of the reproductive organs to perform their proper functions; ordinary variability is not monstrosity, but is perfectly normal, and as the variability of hybrids

has precisely the same character, I think we cannot regard it as due to unnatural disturbance.

According to our theory, variation is due to the action of changed or unnatural conditions upon certain cells of a preceding generation. Now, as characteristics of both parents are mingled in a hybrid, it must nearly always happen that certain cells with peculiarities of one parent will be in contact with, or will depend in some way upon, cells with peculiarities inherited from the other species. There will therefore be a lack of the perfect adjustment between each cell and its neighbors, which has been brought about in each parent by natural selection, and this imperfect adjustment will cause the cell which is unfavorably placed to throw off gemmules. The cells of the body of a hybrid will therefore be unusually prolific of gemmules, and will transmit variability to later generations.

According to our hypothesis, a hybrid is more likely to transmit variability than a pure species, because more of its cells are placed under circumstances favorable to the production of gemmules.

For the same reason a hybrid between two domesticated or cultivated forms must have more tendency to vary than one produced by crossing two wild species, for the domestic or cultivated parents live under unnatural conditions, and therefore have more tendency than wild species to transmit gemmules, and thus cause variability.

The Sex of the Parent affects the Variability of Hybrids.

I have shown that the body of a hybrid is peculiarly favorable for the production of gemmules, and that, for this reason, the descendants of hybrids are variable