4 



ROYAL HORTICULTURAL SOCIETY. 



place to the clear green tint ; but it remains during the whole 

 existence of the plant on the part which it occupies, and where it 

 traces a coloured circle. From this moment all was clear : if the 

 hybrids of D.ferox, allied to another species of the white group, 

 have brown stems and violet blossoms, it is because D.ferox itself 

 possesses the germ of this colouring. In the pure species, the 

 colouring remains in a rudimentary state, occupying only the 

 small interval which extends from the commencement of the root 

 to the cotyledons ; in the hybrid it acquires an enormous increase, 

 extending over all parts of the plant, and manifesting more espe- 

 cially its action on the flower. Here then is a first mode of varia- 

 tion induced by the crossing of two species, and which produces 

 its effects on the first hybrid generation. The second generation 

 is about to offer us variations of another kind, and still more 

 remarkable. 



All these hybrids, though sterile at the first seven or eight 

 dichotomies, were very fertile in those which were developed 

 later. Their seeds, sown last spring, gave me in the second 

 generation nineteen plants of D. feroci-lcevis, and twenty-six of 

 Iwvi-ferox. The two sets still resembled each other, but by a cha- 

 racter diametrically opposite to that which was the prominent 

 trait of the preceding generation. The most astonishing diversity 

 of feature succeeded the former great uniformity, a diversity of 

 such an extent, that out of the forty-five plants of which the two 

 sets were composed, no two were found which precisely resembled 

 each other. They differed in stature (in the proportion of one to 

 four), in habit, in the form of the leaves, the colouring of the 

 stem and flowers, the degree of fertility, the size of the fruit, and 

 the degree of aculeation. "With the exception of a single indi- 

 vidual of the Icevi-ferox set, which had completely reverted to 

 D. Icevis, with this slight difference that there was still at the base 

 of the stem a circle of purplish violet, not one of these plants had 

 sensibly approached this species, and there was only a very small 

 number in which one could recognize faint resemblances to D. 

 ferox ; the greater part even more closely resembled D. Stra- 

 monium and D. quercifolia, with which they had no relationship, 

 than the species from which they descended. Some had white 

 flowers and green stems, either self-coloured or tinted with purple 

 at the base, while others had violet flowers of various shades, and 

 stems more or less brown, sometimes even of a purplish black as 

 deep as that of D. Tatula, which is the most perfect type of the 

 violet group ; the fruit was of all sizes, from that of a filbert to 



