THE THREE EOKMB OF LTTnnUji SALICAEIA. 193 



This result would appear to be one of high importance, for with 

 dimorphic plants it is ensured at the risk of occasional sterility ; 

 not only is the pollen of each plant useless or nearly useless to 

 that individual, but so is the pollen of all the plants of the same 

 form, that is, of half the total number of individual plants. In 

 that extensive class of plants called by C. K. Sprengel dieho- 

 gams, in which the jjollen of each flower is shed before its own 

 stigma is ready, or in which the stigma (though this ease occurs 

 more rarely) is mature before the flower's own pollen is ready 

 sterility can hardly fail to be the occasional result ; and it would 

 be the inevitable result with both dichogamous and reciprocally 

 dimorjjliic flowers unless pollen were carried by insects (and in 

 some few species by the wind) from one flower or plant to the 

 other. As with reciprocal dimorphism so with dichogamy, 

 within the same genus some of the species are and some are not 

 thus characterized. Again, in the same genus, as in that of 

 Trifolium, some species absolutely require insect-aid to produce 

 seed, others are fertile without any such aid ; now when insects are 

 requisite for fertilization, pollen will generally be carried from one 

 flower to the other. We thus see, by means of reciprocal dimor- 

 phism, of dichogamy, and of insect-aid, that some species require, 

 or at least receive, incessant crosses with other individuals of the 

 same species ; whereas other species of the same genera can be, 

 and probably are often fertilized during long periods by the 

 pollen of their own flowers. Why this wide difference in the 

 frequency of crosses should occur we are profoundly ignorant. I 

 will only further remark on this head, that it would be a great 

 mistake to suppose that many flowers, which are neither reci- 

 procally dimorphic nor dichogamous, nor require insect-aid for 

 their fertilization, nor show any particular adaptation in their 

 structure for the visits of insects, are not habitually crossed with 

 the pollen of other individuals ; this occurs, for instance, habitually 

 with cabbages, radishes, and onions, which nevertheless are per- 

 fectly fertile (as I know by trial) with their own pollen without 

 aid of any kind. 



But it may be further asked, granting that reciprocal dimor- 

 phism is of service by ensuring at each generation a cross (but I 

 am far from pretending that it may not have some additional 

 unknown signification), why did not dimorphism suffice for 

 L. salicaria and Grafferi ? why were they rendered reciprocally 

 trimorphic, entailing such complicated sexual relations P Wo 

 cannot answer, except perhaps so far : — if we suppose two xflants of 



«J2 



