94 



OX HYBRIDIZATION AMONGST VEGETABLES. 



some other Hymenocallis in an impoverished and half-effete state, 

 probably produced the variety. The four seedlings were pre- 

 cisely similar to each other : they have not often flowered, and 

 have not borne seed, but they have been rather neglected. It is 

 desirable that further experiments should be tried on Hymenocallis 

 by pollen of P. Ulyricum, and vice versa; but I do not think 

 they will blend, though I am a little less confident on that point 

 than I should have been some years ago. If these four seedlings 

 had been raised by the accidental access of pollen nearly effete 

 from some other Hymenocallis that had flowered lately in the 

 same house, their flower should have been modified. Is it pos- 

 sible that a grain of its own pollen, nearly effete, had touched 

 the stigma, in spite of my precautions, and that its defect was 

 the cause of variety ? If so, an important clue would be ob- 

 tained. Is it possible that the pollen of the cognate genus Pan- 

 cratium, without being able to fertilize the ovules, could help 

 the defective grain of its own kind to some ingredient in which 

 it was defective, and so obtain some influence over the produce, 

 without being actually its parent ? If, as I believe, two grains 

 of pollen cannot act simultaneously in the same ovule, that could 

 not be ; but it is a point open to inquiry, and upon which I 

 merely say, that where I have carefully mixed the pollen of 

 twelve species of Calceolaria, that of one only took effect, and 

 that I have not succeeded in any attempt to effect a double or 

 mixed fertilization at the same time. 



In former publications I laid very great stress — and I now 

 believe too much stress — on the form of the fruit, for I think 

 that important part of vegetation is no less capable of modi- 

 fication and change than other parts thereof. The extensive 

 genus Iris, with great general similarity of aspect, exhibits some 

 anomalous diversities. In Iris setosa, otherwise very like I. 

 Virginica, the conspicuous erect petals of the genus have dis- 

 appeared, and given place to three slender bristles; in Iris 

 Sibirica, and the species closely connected with it, the solid 

 flower-stalk has become a fistulous pipe ; in the Gladdon, the 

 outer and second coats of the seed, which in the fistulous sorts is 

 hollow, becomes filled with pulp, and the seed assumes the colour 

 and semblance of a ripe berry. The bulbous races have their 

 own peculiarities. Theory and experience lead me to think that 

 the whole of the extensive bearded race that occupies the Mediter- 

 ranean formation and its skirts, creeping on as far as Nepal, from 

 the many very dwarf species to the large Germanica and pallida, 

 are easily convertible, and stand almost on the footing of local 

 varieties. I cannot blend them with the fistulous Siberian or 

 with the species belonging to the Virginian type. I think the 

 bearded Mediterranean, the fistulous Siberian, the Virginian 



