II.] INCIPIENT STRUCTURES. 69 



ries away the pollen-mass glued to his back, and then when 

 he has his next involuntary bath in another flower, as he 

 crawls out the pollen-mass attached to him comes in con- 

 tact with the stigma of that second flower and fertilizes it. 

 In the other example (Catasetum), when a bee gnaws a 

 certain part of the flower, he inevitably touches a long deli- 

 cate projection, which Mr. Darwin calls the antenna. " This 

 antenna transmits a vibration to a certain membrane, which 

 is instantly ruptured ; this sets free a spring by which the 

 pollen-mass is shot forth like an arrow in the right direc- 

 tion, and adheres by its viscid extremity to the back of the 

 bee!" 



Another difficulty, and one of some importance, is pre- 

 sented by those communities of ants which have not only a 

 population of sterile females, or workers, but two distinct 

 and very different castes of such. Mr. Darwin believes that 

 he has got over this difficulty by having found individuals 

 intermediate in form and structure 34 between the two work- 

 ing castes ; others may think that we have in this belief 

 of Mr. Darwin, an example of the unconscious action of vo- 

 lition upon credence. A vast number of difficulties similar 

 to those which have been mentioned might easily be cited 

 those given, however, may suffice. 



There remains, however, to be noticed a very important 

 consideration, which was brought forward in the North 

 British Review for June, 1867, p. 286, namely, the neces- 

 sity for the simultaneous modification of many individuals. 

 This consideration seems to have escaped Mr. Darwin, for 

 at p. 104 of his last (fifth) edition of " Natural Selection," 

 he admits, with great candor, that until reading this arti- 



34 Mr. Smith, of the Entomological department of the British Mi^eum, 

 has kindly informed me that the individuals intermediate in structure are 

 very few in number not more than five per cent. compared with the 

 number of distinctly differentiated individuals. Besides, in the Brazilian 

 kinds these intermediate forms are wanting. 



