6 



110 VAL HOKTICULTUKAL SOCIETY. 



plants of low stature, more or less deformed, as different from 

 each other as from the hybrids of the first generation; besides, 

 like the two first, they were sterile, or at the most yielded only a 

 few fruits, in which the seeds were only incompletely developed. 

 Three new plants of the second generation, cultivated in 1864, 

 presented the same diversities of physiognomy; they no more 

 resembled those of the former year than the first hybrids. One 

 of these, which approached M.jalapa very sensibly, was extremely 

 fertile ; the two others flowered very unequally, and did not yield 

 a single seed. "What results from this second experiment is still 

 the irregular variation of the offspring of a hybrid plant, when it 

 does not resume the livery of the species from which it descends. 



It may be asked whether this propensity of hybrids to vary 

 continues to the third and following generations, when they pre- 

 serve their fertility. My answer to this question is as follows : — 



I observed in 18G3 and 186-1 the sixth and seventh generations 

 of a hybrid which I have kept for several years, Linaria purpureo- 

 vulgaris, both represented by some hundreds of individuals. A 

 good number of these last reverted, some completely, the others 

 partially, to the form of Linaria vulgaris with yellow flowers, a 

 small number to those of Linaria purpurea with purple flowers. 

 Others, still very numerous, inclined, so to speak, towards neither 

 the one nor the other, but nevertheless did not resemble the 

 hybrid of the first generation. There were all possible kinds of 

 variation ; tall or dwarf stature, broad or narrow leaves, the corolla 

 deformed in various ways, discoloured, or exhibiting unusual tints, 

 and out of all these combinations there did not result two indi- 

 viduals which were perfectly alike. It is very clear that we have 

 to do here with irregular variation, which engenders only indivi- 

 dualities, and that uniformity is not established between the de- 

 scendants of hybrids, except on the condition that it resumes the 

 normal livery of the parent species. 



Similar facts, to which all the attention which they merited has 

 not been paid, have been produced, and are produced daily in the 

 practice of Floral Horticulturists. Here is a well-known and 

 well-authenticated instance : there exist in gardens two species of 

 Petunia perfectly defined, the one (P. nyctaginiflora) has white 

 flowers, the other (P. violacea) has purple flowers, neither of which 

 at present has varied, but intercrossing easily and yielding hybrids 

 as fertile as themselves. In the first generation, all the hybrids 

 are alike ; in the second they vary in the most remarkable degree, 

 some reverting to the white species, others to the purple, and a 



