374 METAMORPHOSIS 



arisen. In reality, however (Weismann, 1892 a, p. 421), the phenomenon may be 

 otherwise explained. The white-flowered species may have arisen from a violet- 

 flowered stock and may, so to speak, have the capacity for forming this colour in a 

 latent state (compare p. 376), but which in the hybrid becomes actual. Whether 

 all the unexpected colours which appear in hybrids are to be explained in the 

 same way must remain at present an open question. A much more frequent 

 variation of the hybrid from the parents lies in this, that the hybrid, as a rule, 

 differs in its growth- energy. This energy may be feebler than that of the parent 

 plants when the parents are not closely related forms. In this case the seeds ger- 

 minate badly and the seedlings are difficult to rear. Or — and this is particularly 

 applicable to hybridization between nearly-related races — ' they are remarkable 

 for their size, rapidity of growth, early blooming, free flowering, longer period of 

 life, great capacity for multiplication, abnormal size of individual organs, and 

 similar characters ' (Focke, 1881, p. 475). If, for instance, the hybrid Datura 

 tatula 5 X jD. stramonium a^ attains a height of two metres, while the parents only 

 attain a height of about one metre, we may say that the hybrid has acquired 

 a new character, nevertheless it is only a quantitative and not a qualitative 

 variation such as we might obtain otherwise, e.g., by over-nutrition in seed- 

 formation or good manuring in germination. One may look just as little 

 upon the increased growth-energy of the hybrid as on the other, at all events, 

 frequent characteristic, its diminished fertility, as a serious objection to the 

 view that hybrids show no new characters. This diminished fertility is usually 

 manifested in partially or completely unfertile pollen, more rarely in immature 

 ovules. For this reason it is often possible to rear fruits and seeds only by 

 pollinating from the parental line, although there is a class of hybrids whose 

 ovules are quite fertile with their own pollen {Salix, Hieracium). At the ex- 

 treme limit of sterility are many species of Rhododendron, Epilobium, &c., 

 hybrids of which, in general, do not even form flowers (compare p. 376). 



Let us now inquire what the hybrids of the second generation look like, 

 that is to say, the plants arising from the seeds produced by the first hybrid 

 generation. In this case, even less than the other, no general rule can be laid 

 down, even leaving out of consideration self-sterile hybrids. Among fertile 

 hybrids there are doubtless those in which the offspring after self-fertilization 

 are quite similar to the parents (Mendel's Hieracium hybrids, Correns, 1901 a, 

 pp. 75, 80) and, in contrast to these forms, those in which the offspring are entirely 

 different. [Owing to the discovery of parthenogenesis in Hieracium the uni- 

 formity of their offspring appears in an entirely different light ; there are hybrids, 

 however, which are perfectly constant (De Vries, 1903, 66).] Much work has 

 recently been carried out on the latter class of hybrids, and the numerical relation- 

 ships of the individual variations have been worked out on Mendel's principles. 

 In experiments of this kind, all, or at least very many, of the seeds produced 

 must be sown, but in the earlier investigations on this subject only a few plants 

 were raised, and hence no conclusions as to the constancy or inconstancy of 

 hybrid progenies could be arrived at. We cannot do more than give one 

 example from the very voluminous literature on the subject. Let us consider 

 a pea-hybrid that has arisen from two races differing in one character only, say 

 the colour of the flowers, which in race A are red, in race B white. As we have 

 already seen red is the dominant character. All the hybrids of the first genera- 

 tion have red flowers. If the seeds produced by self-pollination from these 

 hybrids be raised, the majority of the offspring will be found to produce red 

 flowers, but a certain proportion will be white-flowered. An enumeration shows 

 that 25 per cent, are white-flowered and 75 per cent, red-flowered. All the off- 

 spring of these white-flowered plants remain white-flowered, while of those of the 

 red-flowered forms one-third remain unaltered in colour (red) while of the 

 other two-thirds 25 per cent, are white and 75 per cent. red. In order to 

 explain this extremely pecuHar result Mendel (1866) assumed that the 



