XII.] CONJUGATION AND SEXUAL REPRODUCTION, 211 



otherwise inevitable death, we ought to find it as a fiandamental 

 process, occurring without a single exception. It is hardly 

 necessary to say that this is not the case. Least of all ought its 

 appearance to depend obviously upon external conditions of life. 

 But this is certainly the case ; the periodicity of its appearance can 

 be proved to depend upon adaptation. 



In many thousands of species of the higher animals amphi- 

 mixis invariably makes its appearance at the outset of every 

 generation, for no egg can develope without fertilization. This 

 is true of the whole Vertebrate sub-kingdom. Isolated excep- 

 tions to this general law suddenly begin to appear in the group 

 of the Arthropods. Certain eggs, in which we should have 

 thought fertilization would be the necessary preliminary to 

 development, have gained the power of developing un- 

 aided, — viz. the power of producing males alone (bees), 

 while the same eggs, if fertilized, would produce females. 

 In plant-lice, on the other hand, females emerge from unferti- 

 lized ova, and not one generation only, but two, three, and 

 even many, succeed each other before a sexual generation 

 occurs and, with it, amphimixis. How far this latter is from 

 being a process of multiplication, and how superficial is the 

 connexion which usually obtains between amphimixis and 

 multiplication, are shown in the bark-lice, e. g. Phylloxera. In 

 these it has already been mentioned that the sexual generation 

 consists of minute animals devoid of mouth and of the power 

 of taking food. The female lays a single egg, so that, as in the 

 primitive form of conjugation, the number of individuals is not 

 increased by reproduction, but diminished by half Nature 

 could hardly express with greater clearness the stress which she 

 lays on amphimixis ; nor could she argue in a more convinc- 

 ing way that increase is distinct from amphimixis, and that the 

 quickening of new germs need not be dependent upon the latter. 



If amphimixis were a process of rejuvenescence we could 

 hardly believe that its occurrence in the life of a species 

 would be so excessively fluctuating, — sometimes taking place 

 in each generation, sometimes recurring after a lapse of two, 

 three, or even as many as ten generations, sometimes being 

 absent for forty generations, as I have proved to be the case in 

 Cypris reptans. It might be suggested that the recurrence of 

 amphimixis does not depend on the number of generations of 



p 2 



