XII.] CONJUGATION AND SEXUAL REPRODUCTION. 1 35 



possibility of a repetition (two or more times) of the same 

 combination, and the further possibihty of the development of 

 those very germ-cells which contain identical germ-plasm — 

 the probability of such an occurrence is so excessively remote 

 that it may be considered practically non-existent, and we 

 have no reason for wondering that identical individuals have 

 never been observed among the children successively born 

 in any human family. 



To my mind the doubUng of the idants before the ' reducing 

 division ' possesses this very significance : — it renders possible 

 an almost infinite number of different kinds of germ-plasm, 

 so that every individual must be different from all the rest. 

 And the meaning of this endless variety is to afford the material 

 for the operation of natural selection. 



It might perhaps be objected that sufficient variety could 

 have been attained without the doubling of the rods, and that, 

 although the difference between the numbers of combinations 

 produced with and without doubling is certainly very con- 

 siderable, the number of rods would have been large enough 

 without increase, since, as a matter of fact, the number of 

 descendants developed is always smaller than the number 

 of possible combinations. Eight idants without doubling 

 give seventy combinations ; these multiplied by the seventy 

 combinations of the other parent yield 4900 varieties of 

 germ-plasm in the fertilized ova, and potentially an equal 

 number of different offspring. We might suppose that this 

 number would suffice in all cases ; for when the germ-cells are 

 far more numerous (many animals producing 100,000 or even 

 upwards of 1,000,000 ova, not to mention spermatozoa) only 

 a very small percentage can enter upon development, and 

 of these but very few can arrive at maturity and reproduce 

 themselves. It would be sufficient, we might think, if there 

 were only a few more thousand combinations of germ-plasm 

 than of individuals which attain maturity. 



There is, however, much to be said on the other side. If 

 we are not able to determine by calculation the number of 

 differing individuals which are necessary in order that the 

 development of the species may be guided by natural selec- 

 tion, we can scarcely fail to recognize that it is only by the 

 widest possible choice that the best possible adaptation of 



