APPENDIX 347 



original parents equally, yet in this experiment the type of the seed 

 parent obtained so great a preponderance that out of each sixty- 

 four plants of the first generation fifty-four exactly resembled it, or 

 only differed in one character. It is seen how rash it must be under 

 such circumstances to draw from the external resemblances of 

 hybrids conclusions as to their internal nature. 



Gartner mentions that in those cases where the development was 

 regular, among the offspring of the hybrids, the two original species 

 were not reproduced, but only a few individuals which approached 

 them. With very extended developmental series it could not in 

 fact be otherwise. For seven differentiating characters, for 

 instance, among more than 16,000 individuals — offspring of the 

 hybrids — each of the two original species would occur only once. 

 It is therefore hardly possible that these should appear at all among 

 a small number of experimental plants; with some probability, 

 however, we might reckon upon the appearance in the series of a 

 few forms which approach them. 



We meet with an essential difference in those hybrids which re- 

 main constant in their progeny and propagate themselves as truly as 

 the pure species. According to Gartner, to this class belong the 

 remarkably fertile hybrids, Aquilegia atropurpurea canadensis, 

 Lavatera pseudolbia thuringiaca, Geum urbano-rivale, and some 

 Dianthus hybrids; and, according to Wichura, the hybrids of the 

 Willow family. For the history of the evolution of plants this cir- 

 cumstance is of special importance, since constant hybrids acquire 

 the status of new species. The correctness of the facts is guaran- 

 teed by eminent observers, and cannot be doubted. Gartner had 

 an opportunity of following up Dianthus Armeria deltoides to the 

 tenth generation, since it regularly propagated itself in the garden. 



With Pisum it was shown by experiment that the hybrids form 

 egg and pollen cells of different kinds, and that herein lies the reason 

 of the variability of their oft's[)ring. In other hybrids, likewise, 

 whose offspring behave similarly we may assume a like cause; for 

 those, on the other hand, which remain constant, the assumption 

 appears justifiable that their reproductive cells are all alike and 

 agree with the foundation-cell [fertilised ovum] of the hybrid. In 

 the opinion of renowned physiologists, for the purpose of propaga- 

 tion one poDen cell and one egg cell unite in Phanerogams ^ into a 



1 In Pisum it is placed beyond doubt that for the formation of the new embryo a 

 perfect imion of the elements of both reproductive cells must take place. How 



