On this view, these plants are, in fact, merely fully fertile, and not fertile to an abnormal degree. Nevertheless, as characters of all kinds are liable to variation, especially with organisms unnaturally treated, and as in the four first and more sterile classes, the plants derived from the same parents and treated in the same manner, certainly did vary much in sterility, it is possible that certain plants in the latter and more fertile classes may have varied so as to have acquired an abnormal degree of fertility. But it should be noticed that, if my standards err in being too low, the sterility of all the many sterile plants in the several classes will have to be estimated by so much the higher. Finally, we see that the illegitimate plants in the four first classes are all more or less sterile, some being absolutely barren, with one alone almost completely fertile; in the three latter classes, some of the plants are moderately sterile, whilst others are fully fertile, or possibly fertile in excess.
The last point which need here be noticed is that, as far as the means of comparison serve, some degree of relationship generally exists between the infertility of the illegitimate union of the several parent-forms and that of their illegitimate offspring. Thus the two illegitimate unions, from which the plants in Classes 6 and 7 were derived, yielded a fair amount of seed, and only a few of these plants are in any degree sterile. On the other hand, the illegitimate unions between plants of the same form always yield very few seeds, and their seedlings are very sterile. Long-styled parent-plants when fertilised with pollen from their own-form shortest stamens, appear to be rather more sterile than when fertilised with their own-form mid-length stamens; and the seedlings from the former union were much more sterile than those from the latter union. In opposition to this relationship, short-styled plants illegitimately fertilised with pollen from the mid-length stamens of the long- styled form (Class 5) are very sterile; whereas some of the offspring raised from this union were far from being highly sterile. It may be added that there is a tolerably close parallelism in all the classes between the degree of sterility of the plants and their dwarfed stature. As previously stated, an illegitimate plant fertilised with pollen from a legitimate plant has its fertility slightly increased. The importance of the several foregoing conclusions will be apparent at the close of this chapter, when the illegitimate unions between the forms of the same species and their illegitimate offspring, are compared with the hybrid unions of distinct species and their hybrid offspring.
OXALIS.
No one has compared the legitimate and illegitimate offspring of any trimorphic species in this genus. Hildebrand sowed illegitimately fertilised seeds of Oxalis Valdiviana, but they did not germinate (5/4. 'Botanische Zeitung' 1871 page 433 footnote.); and this fact, as he remarks, supports my view that an illegitimate union resembles a hybrid one between two distinct species, for the seeds in this latter case are often incapable of germination.
[The following observations relate to the nature of the forms which appear among the legitimate seedlings of Oxalis Valdiviana. Hildebrand raised, as described in the paper just referred to, 211 seedlings from all six legitimate unions, and the three forms appeared among the offspring from each union. For instance, long-styled plants were legitimately fertilised with pollen from the longest stamens of the mid-styled form, and the seedlings consisted of 15 long-styled, 18 mid-styled, and 6 short-styled. We here see that a few short-styled plants were produced, though neither parent was short-styled; and so it was with the other legitimate unions. Out of the above 211 seedlings, 173 belonged to the same two forms as their parents, and only 38 belonged to the third form distinct from either parent. In the case of O. Regnelli, the result, as observed by Hildebrand, was nearly the same, but more striking: all the offspring from four of the legitimate unions consisted of the two parent-forms, whilst amongst the seedlings from the other two legitimate unions the third form appeared.