386 ON THE NUMBER OF POLAR BODIES AND [VI. 



egg-cells, would not apply to an equal extent in the case of 

 the male germ-cells. Among the egg-cells it may be very 

 important that each one should have its special individual 

 character, produced by a somewhat different composition of 

 its germ-plasm, inasmuch as a considerable proportion of the 

 eggs frequently developes, although this is never the case with 

 all of them. But the production of sperm-cells is in most 

 animals so enormous that only a very small percentage can be 

 used for fertilization. If, therefore, e.g. ten or a hundred 

 spermatozoa contained germ-plasm with exactly the same 

 composition, so that, as far as the paternal influence is con- 

 cerned, ten or a hundred identical individuals would result if 

 they were all used in fertilization, such an arrangement would 

 be practically harmless, for only one spermatozoon out of an 

 immense number would be emplo3^ed for this purpose. From 

 this point of view we might expect that the ' reducing division ' 

 of the sperm-nucleus would not take place at the end of the 

 development of the sperm-cell, but at some earher period. 

 There is no necessary reason for the assumption that this 

 division must take place at the end of development, and with- 

 out some cause natural selection cannot operate. It is, of 

 course, conceivable that the causes of other events may also 

 involve the occurrence of this division at the end of develop- 

 ment ; but we do not at present know of any such causes. I 

 should not consider the influence of the specific histogenetic 

 nucleoplasm, i. e. the spermatogenetic nucleoplasm, to be such 

 a cause, because the quantitative proportions are very different 

 from those which obtain in the formation of egg-cells, and 

 because it is not inconceivable that the small quantity of true 

 germ-plasm which must be present in the nuclei of the sperm- 

 cells at every stage in their formation might enter upon a 

 'reducing division' with the spermatogenetic nucleoplasm, even 

 when the latter preponderated. 



As soon as we can recognize with certainty the forms of 

 nuclear division which are 'reducing divisions,' the question 

 will be settled as far as spermatogenesis is concerned. It has 

 been already established that various forms of nuclear division 

 occur at different periods of spermatogenesis. I make this 

 assertion, not only from my own observations, but also from 

 observations which have been made and insisted upon by 



