Who painted the Flowers ? 39 



most important of all. As Dr. Asa Gray well says, 1 

 "The origination is the essential thing. . . . To be a 

 scientific explanation [the theory] should show, or enable 

 us to conceive, how insect-visitation operates or in any 

 way tends to develop colours, and originate apparatus. 

 . . . Thus far it does not appear how the visits of bees 

 to a blossom can make one hair white or black. For all 

 that yet appears, we may be indebted to bees for the 

 beauty of our gardens and the sweetness of our fields, 

 much as we are indebted to the postman for our letters. 

 Correspondence would flag and fail without him ; but 

 the instrument is not the author of the correspondence." 



It seems obvious, then, that if flowers had been 

 developed by bees, it is because it was their nature to be 

 so developed : and that nature was theirs before the 

 bees came. What development there has been must 

 have been along lines already laid down when the flowers 

 were made. The beauty which has resulted cannot be 

 attributed to the labourers who educed it, unless we are 

 prepared to credit the masons and carpenters with the 

 artistic merit of a cathedral. 



Another question which suggests itself refers to the 

 doctrine of development itself upon which the whole 

 argument depends. With regard to that doctrine, I must 

 for my own part say that in the observation of facts 

 within reach I meet with more apparently insoluble 

 difficulties than with fragments of proof. It is generally 

 assumed that the alternative to the development theory, 

 the supposition, namely, that all members of one species 

 are descended from one common ancestor originally 

 created in that form, is too violent to be entertained, and 

 that on development principles the difficulty disappears. 

 But, I would ask, must not developists suppose that all 

 these individuals are descended from one common 

 ancestor originally developed to this form ? Otherwise, if 

 there had been independent developments, how account 

 for the marvellous identity of results ? How, at least 



1 Contemporary Review, ut supra^ p. 606. 



