XII.] CONJUGATION AND SEXUAL REPRODUCTION. 1 29 



and mingling of the parent nuclear substances presupposed by 

 him, 'similar predispositions would arrange themselves closer 

 to one another than dissimilar ones, and from the similar but 

 varying paternal and maternal predispositions an intermediate 

 form might arise by mutual influence.' I have printed the 

 words ' intermediate form ' in italics because it appears that so 

 much depends upon it ; for obviously the intermediate form of 

 predisposition must be looked upon as one and no longer as 

 two separate predispositions. Hence, according to Hertwig, 

 the fusion of two parental germ-plasms produces an inter- 

 mediate form of germ-plasm in ivhich each predisposition is not 

 doubled, but remains single. Furthermore, this germ-plasm 

 could grow, and could be represented by a larger or smaller 

 mass, but it is impossible that it could be halved without losing 

 its character as germ-plasm, except it were first doubled 

 in size, and all its predispositions were doubled and sym- 

 metrically arranged on each side of the plane of division like 

 the antimeres in a bilaterally symmetrical animal. But even in 

 this last case a ' reducing division,' that is a putting on one side 

 of half the number of the corresponding but individually distinct 

 chromatin elements, is impossible because both halves would 

 contam precisely similar predispositions. O. Hertwig deceives 

 himself in believing that he can assume a halving of the number 

 of chromatin elements while his conception of the composition 

 of the germ-plasm only admits of a halving of mass. In his 

 germ-plasm, made by the fusion of paternal and maternal 

 predispositions, there are no differing predispositions of one 

 and the same part or organ : the parental differences have, 

 according to his view, neutralized each other, and each predis- 

 position is present as a single intermediate variety. Whence 

 comes the necessity or the possibility of any reduction ^ What 

 are the units which are to be reduced in number ? 



It is clear that we cannot avoid the assumption of higher 

 units of germ-plasm, each one of which contains, collected 

 together, the varied predispositions of the species, whether called 

 by my term ancestral plasm, or by any other name. I shall 

 attempt to explain elsewhere that this conception is not only in- 

 dispensable for our understanding of the ' reducing division,' but 

 that it is even rendered necessary by the phenomena of heredity. 

 At present I do not propose to do more than show that the 



VOL. II. K 



