STUDIES ON INSECT SPERMATOGENESIS. 343 



In the case of the Golgi apparatus, the evidence is as yet very 

 meager, but none the less positive. Gatenby ('18) has recently 

 given some account of dictyokinesis in several genera of pul- 

 monates, and in the L'nnacidcc at least he found evidence of a 

 very exact distribution of the Golgi elements, corroborating the 

 similar observations of Platner ('89) made many years ago. 

 Although his observations were not made on the germ cells, 

 Deineka ('12) has given one of the clearest accounts of dictyo- 

 kinesis so far published, and he points out the apparently equal 

 distribution of the dictyosomes to the daughter cells. The ob- 

 servations on Hemiptera which I have already given furnish, it 

 seems to me, the best ground on which to base an opinion ; for 

 the essentials of dictyokinesis are so clear that it is possible to 

 arrive at a reliable conclusion. And indeed it may be affirmed 

 that the Golgi apparatus of the spermatocytes is divided very 

 equally among the spermatids, so that each receives a fourth of 

 the original amount. Accordingly, in the case of the Golgi ap- 

 paratus there is also a definite regularity in the quantities dis- 

 tributed in division to the daughter cells. 



Concerning the mass of cytoplasm proper, an approximately 

 equal division often occurs, at other times a very unequal one, 

 but in; any case the same regularity is a characteristic feature. 

 Summing up the evidence, therefore, our first question concern- 

 ing the existence of regularity in the mass division of the cyto- 

 plasm may be answered affirmatively. There is conclusive evi- 

 dence that the cytoplasm and its principal formed elements are 

 distributed with great regularity by the processes of cell division, 

 the most usual result being that each daughter cell receives a 

 rather accurate half of the total quantity of cytoplasmic materials. 



Concerning the second question, namely, as to the mechanism 

 by which this regularity in mass distribution is achieved, two 

 answers at once suggest themselves : ( I ) The division may be a 

 purely chance separation of elements distributed at random in 

 the cytoplasm; or (2) It may be accomplished by definite activi- 

 ties related to those which produce the spindle and the resultant 

 phenomena of karyokinesis. Thus far the relation of the cyto- 

 plasmic elements to the centrioles and the spindle in general has 

 received scant attention, and the idea of chance distribution has 



