888 THE ORIGIN OF SPECIES 



the other hand, the trnnsportal of the lower eye of a flat- 

 fish to the upper side of tlie head, and the formation of 

 a }ivehensile tail, may be attributed almost wholly to con- 

 tinued use, together with inheritance. With respect to 

 the mammae of the higher animals, the most probable 

 conjecture is that primordially the cutaneous glands over 

 the whole surface of a marsupial sack secreted a nutri- 

 tious fluid; and that these glands were improved in func- 

 tion through natural selection, and concentrated into a 

 confined area, in which case they would have formed 

 a mamma. There is no more difficulty in understanding 

 how the branched spines of some ancient Echinoderm, 

 which served as a defence, became developed through 

 natural selection into tridactyle pedicellariaj, than in un- 

 derstanding the development of the pincers of crustaceans, 

 through slight, serviceable modifications in the ultimate 

 and penultimate segments of a limb, which was at first 

 used solely for locomotion. In the avicularia and vibrac- 

 ula of the Pol3'Zoa we have organs widely different in 

 appearance developed from the same source; and with 

 the vibracula we can understand how the successive gra- 

 dations might have been of service. With the poUinia 

 of orchids, the threads which originally served to tie 

 together the pollen-grains can be traced cohering into 

 caudicles; and the steps can likewise be followed by 

 which viscid matter, such as that secreted by the stigmas 

 of ordinary flowers, and still subserving nearly but not 

 quite the same purpose, became attached to the free ends 

 of the caudicles; — all these gradations being of manifest 

 benefit to the plants in question. With respect to climb- 

 ing plants, I need not repeat what has been so lately 

 said. 



