158 AMPHIMIXIS OR ESSENTIAL MEANING OF [XII. 



generations the number of different idants would have been 

 diminished from twenty to ten, and the whole group would con- 

 sist often pairs, for instance ««, bb, cq dd, ee,ff, gg, hh, it, kk, the 

 idants in each pair being identical. In the course of later 

 generations the number of different idants might be diminished 

 still further, although more gradually. 



We are thus led to believe that, in persistent parthenogenesis 

 unbroken by bisexual reproduction, a great uniformity of germ- 

 plasm will at length arise, and, as a result, a great uniformity 

 of individuals. We cannot doubt this if we consider that each 

 fresh simplification of the germ-plasm, when it has once ap- 

 peared, is unable to revert towards complexity because fertili- 

 zation, i. e. the introduction of foreign idants, is excluded. As 

 soon as the 'reducing division' causes a single one out of 

 the twenty maternal idants in the segmentation nucleus of 

 the egg to become double, it has been shown above that 

 one of the other idants must be irretrievably lost not onl}'- 

 to the maternal germ-plasm and to the daughter, but also to the 

 descendants of every generation. Among all the numerous 

 possible combinations there is only one which leads to no 

 diminution in the number of different idants, viz. the above- 

 mentioned arrangement a, b, c, d, e, /, and this is an exact 



repetition of the maternal combination. Hence the diminution 

 in the number of different idants is far more probable than the 

 maintenance of the complete series, and this probability will be 

 repeated in each successive generation, until only two kinds of 

 idants remain in the germ-plasm. When, however, this point 

 is reached ^, the reverse becomes true ; for the probability that 

 idants a alone, or b alone, would be left in the egg-nucleus by 

 the ' reducing division' is much less than that both kinds would 

 exist side by side. 



This becomes clear if we consider a definite case. Instead 

 of the twenty idants which have been assumed hitherto, let us 

 take only half as many, viz. ten, and let us suppose that they 

 have been already reduced to two different kinds, a and b. 

 These double themselves in the mother-egg-c-ell to twenty — ten 

 a and ten b. The following combinations are then possible for 



^ Even before this point is reached the probabihty begins to change. — 

 A. W. 1892. 



