f'liAr. VIII. AND TRIMORPIIISM. 205 



while tlic sterility of the union from which they were derived 

 was hy no means f^n\it. ^Vith liybrids raised from the same 

 seed-capsule the deforce of sterility is innately variable, so it 

 is in a marked manner with illefritimate plants. Lastly, many 

 hybrids arc profuse and persistent flowcrers, while other and 

 more sterile hybrids produce few flowers, and arc weak, miser- 

 able dwarfs ; exactly similar cases occur with the illegitimate 

 oflspring of various dimordhic and trimorphic plants. 



Altogether there is the closest identity in character and be- 

 havior between illegitimate plants and hybrids. It is hardly 

 an exaggeration to maintain that the former are hybrids, but 

 jjroduced within the limits of the same species by the imjiroper 

 imion of certain forms, while ordinary hybrids are produced 

 from an improper union between so-called distinct species. 

 We have also already seen that there is the closest similarity in 

 all respects between first illegitimate unions and first crosses 

 l)e+wcen distinct species. This will perhaps be made more 

 fully ajiparcnt by an illustration : we may suppose that a bot- 

 anist found two well-marked varieties (and such occur) of the 

 loiig-stjled form of the trimorphic Lythrum salicaria, and that 

 he determined to try by crossing whether they Mere specifi- 

 cally distinct. He would find that they yielded only about one- 

 fifth of the proper number of seed, and that they behaved in all 

 the other above-specified respects as if they had been two dis- 

 tinct species. But to make the case sure, he would raise plants 

 from his supposed hybridized seed, and he would find that the 

 seedlings were miserably dwarfed and utterly sterile, and that 

 they behaved in all other respects like ordinary hybrids. He 

 might then maintain that he had actually proved, in accordance 

 with the common view, that his two varieties were as good 

 and as distinct species as any in the world ; but he would be 

 completely mistaken. 



The facts now given on dimorphic and trimorphic plants 

 an» important, because they show us, first, that the ])hysi()logi- 

 cal test of lessened fertility, both in first crosses and in hyl)ri(ls, 

 is no safe critericm of specific distinction ; secondly, because we 

 may conclude that there is some unknown bond which connects 

 (li(> infertility of illegitimate unions with that of their illegiti- 

 mate oflspring, and we an* led to extend the same view to lirst 

 crosses and hybrids ; thirdly, because we find, and this seems 

 to me of especial importance, that two or three forms of the 

 same species may exist and may differ in no respect, except in 

 their reproductive organs, and yet be sterile when united in 



