XII.] CONJUGATION AND SEXUAL REPRODUCTION. 1 6 



f 



the variety B^ a reversion which would be possible in an egg 

 in which the 'reducing division' takes place so that the com- 

 bination of idants, a aa a, is removed in the polar body, while 

 the combination, a abb, remains in the germ-nucleus of the 

 ovum. The primitive germ-cells of the next generation con- 

 tain the same combination, a abb, which is doubled in the 

 mother-germ-cells to a aaab b b b, 2in6. it is now clear that a 

 'reducing division' might occur, which would brmg the four 

 idants, bbbb, together into the germ-nucleus of an ovum, and 

 from an egg containing germ-plasm with this constitution there 

 must undoubtedly arise an individual of the variety B. 



In this illustration, which is of course far too simple, rever- 

 sion to the other variety might happen in the third generation. 

 In those cases, however, — and they are the usual ones,— in 

 which the number of idants is larger, and the proportion of 

 variety b much smaller, the exclusive predominance of the 

 latter can only take place far more slowly, and, as a rule, in 

 much fewer cases; for it depends upon the chance of a com- 

 bination of several idants b arising in certain ova, and of the 

 survival to maturity of the individuals which develope from 

 such eggs, — and these naturally must be far rarer than those 

 with a largely predominating number of idants a. Further- 

 more, there is no certainty that, among the eggs produced 

 by such individuals, any with an increased proportion of idants 

 b would find a place. 



These theoretical considerations harmonize well with the 

 results of experiment. Variety A can give rise to descendants 

 belonging to variety B, but this does not happen in all broods, 

 and often only after the lapse of numerous generations. And 

 the same is true of variety B in relation to the production of 

 variety A. In both cases, relatively few individuals change 

 into the other variety, and never all the descendants of one 

 mother. In the aquarium in which such a transformation has 

 occurred numerous individuals of the original form were in- 

 variably present, — a proof that it is always a rare exception for 

 such extreme combinations of germ-plasm to be formed. When, 

 however, this combination had once arisen, then such an in- 

 dividual gave rise, in all the cases observed, to offspring of her 

 own type. Thus a mother which arose from variety A, but has 

 passed over to variety By behaves as though her ancestors had 



