194 PROBLEMS OF EVOLUTION 



obtains. And thus insects have been gardeners with an isolated 

 garden for each species of flower, gardeners for whom new 

 varieties as they arose proceeded to isolate themselves. 

 Sterility I have as yet spoken only of complete intersterility between 

 and in- S p ec i es ant j varieties. But probably this is too strong a term to 

 use in many cases. When some plants are beginning to deviate 

 from the type to which the rest of the species conform, it may 

 well be that comparative infertility exists between them and 

 their kin, rather than complete sterility. It is well known that 

 when two pollens are put upon the same flower, though both 

 may be capable of fertilising it, yet when they have to compete, 

 one is prepotent. When the transferred pollen comes from 

 another flower of the same variety, it is believed to be prepotent 

 over the home-grown pollen. But when it has developed on a 

 flower of a different variety, it is likely to have less impregnating 

 power than that with which it has to compete. When we are 

 dealing, not with incipient varieties but, with species, absolute 

 sterility is the rule. 



In order to perpetuate a new variety that has developed some 

 helpful characteristic when the deviation from the type first 

 appears, it is enough if it be comparatively infertile with its 

 allies. For flowers that are fertilised with pollen from the same 

 stock will produce more seed and the variety will not disappear. 

 Wind- and Professor Henslow has maintained that insect fertilisation is 

 iHs^d ^ no a dvantage to a ptant, since those that are usually self- 

 flowers fertilised get on equally well, in many cases producing an 

 enormous amount of seed, and occupying more of the earth's 

 surface than flowers with elaborate contrivances to secure cross- 

 fertilisation. Some sedges (genus Carex) are very widely dis- 

 seminated, and some obscure composite such as the common 

 cudweed (Gnaphalium uliginosurn). 



These are facts which it is impossible to dispute, but the 

 inference from them is unsound. If some species of orchids 

 are comparatively scarce, it is not because they depend on 

 insect fertilisation. On the contrary, it is probable that insect 

 fertilisation has saved a comparatively feeble race from ex- 

 tinction. In such cases, as I have said already, we must con- 



