1 66 



NA TURE 



{yan. I, 1874 



from the flowers previously visited. Sometimes, also, 

 the three inner anthers are developed, and, completely 

 filled with pollen, closely surround and spontaneously 

 self-fertilise the two (in rarer cases three) stigmas, 

 cross-fertilisation being thus almost prevented. But 

 most of the flowers show an intermediate condition, having 

 only one or two of the three inner anthers developed. 



species the flowers vary and have always varied 

 in size, colour, the quantity of secreted honey, and con- 

 sequently in their attractiveness for insects. Whenever 

 in such a varying species the one variety possesses such a 

 degree of attractiveness for insects as to receive sufficiently 

 frequent visits from them, those variations which afford 



Fic. 2-;.— Sexual organs of Malvn roiundifolia, in their first period, longi- 

 tudinally bisected, seven times natural size, a, anthers : br, branches 

 of the style (sf) ; pr, pet.ils ; «, nectary ; fr, protecting hairs ; sr, sepals ; 

 cz', ovary : /5, filament-cylinder. 



Fig. 24. — Side view of the same organs in their second period. 



Without referring to many other genera which I have 

 ascertained to contain species quite analogous to those just 

 described, * we may, I think, admit as a summary of the 

 recorded facts, the following propositions : — In many 



Fig. 25.— Side view of the 



sexual organs of J/, sylvcstris, seven tin 

 natural size. 



1 



facility for cross-fertiUsation by insects have always been 

 preserved and accumulated by natural selection, whereas 

 the possibihty of self-fertiHsation has at the same time 

 frequently been lost. Hence we may infer that cross-fer- 

 tilisation is more advantageous to a plant than self- 



Fig. 26. — Side view of the long-styled flower of Polygonum Fagopytum, two leaves of the perianth having heen removed. «. nectaries ; a, anthers ; 

 j/, stigmas. Fig. 27. — Side view of the short-styled flower. FlG, 28. — Side view of the flower ai Polygonum Bistoria in its first period. Fig. 29. — Side 

 view in its second period. 



fertilisation. Whenever, on the contrary, another variety 

 of the same species presents so little attraction for in- 

 sects as to remain commonly overlooked by them, only 





Fig. 30. — Flo 



vUular 

 St, stign 



viewed laterally, two leaves of the perianth having been 



such individual peculiarities as induce self-fertilisation 



have been preserved and accumulated by natural selec- 



* Gemnium, Stcllaria, Cerastium, Rubus, Veronica, Carduus, Hieracium, 



tion, whereas cross-fertilisation by insects has frequently 

 become very difficult, although perhaps never quite im- 

 possible. Hence we may infer that self-fertilisation is by 

 no means absolutely disadvantageous to a plant, but only 

 when the offspring of self-fertilisation has to struggle for 

 existence with the offspring of cross-fertilisation. 



There is another curious point about the recorded facts. 

 We have seen that more and less attractive flowers 

 adapted to cross- or to self-fertilisation sometimes occur 

 in slightly differing, sometimes in well-marked varieties, 

 sometimes in doubtful, sometimes in good and distinct 

 species. 



If we believe the principle of evolution, and view 

 species as originated from varieties, varieties as originated 

 from slight individual differences, we may consider the 

 recorded facts as presenting and explaining one of the 

 many ways in which previously varying forms have been 

 transformed by naturalselectioninto different and diverging 

 species. Hermann Muller 



