366 KANSAS UNIVERSITY SCIENCE BULLETIN. 



Care was taken to study both daughter groups, in one of which 

 it invariably occurred while the other offered no evidence of 

 the presence of a similar element. The constancy in size of 

 the accessory and the relation of its size here to the size of the 

 accessory as determined by figures 4 and 19, plate LXIX, 

 help to establish its identity. This condition of the chromatin 

 in the spermatid nucleus lasts for only a short time. The 

 large chromatin mass soon begins to break up into smaller 

 masses which eventually distribute themselves against the 

 nuclear wall. In an early stage of this disintegration the ac- 

 cessory is still easily recognized, as it lies on one side of the 

 nucleus apart from the rest of the chromatin. It still keeps 

 the elongated form it displayed in the second spermatocyte 

 anaphase. 



A whole cyst of young spermatid nuclei was counted in order 

 to determine in just what proportion of the spermatids of one 

 cyst the accessory occurred. The importance of this observation 

 is obvious, although it seems that the result could be predicted 

 safely from the evidence obtained from previous observations. 

 An account of the method used in counting may help to verify 

 the results. Serial sections of an entire cyst were obtained. 

 The cyst chosen for counting ran through twenty-four sections. 

 The sections were five micra in thickness. The exact condition 

 of the chromatin was found to vary in almost every cyst of this 

 generation. In the cyst counted were found the stages shown 

 in figures 28, 29 and 31, plate LXIX. From the conditions ex- 

 isting in such nuclei we would expect to be able to identify the 

 accessory from its shape and its position with respect to the 

 rest of the chromatin in the nucleus. Evidently the number of 

 nuclei in such a cyst and the number of spermatids formed 

 would be equal. All whole nuclei or almost whole nuclei and 

 fragments containing the accessory were counted. It was im- 

 possible to avoid the error of counting one nucleus as two where 

 a nucleus happened to be cut exactly in half; one half appearing 

 in two adjacent sections. The number of nuclei containing the 

 accessory was 484. The number of fragments of nuclei which 

 contained the accessory was 40. Thus the entire number of 

 spermatids in which the accessory occurred was 524. The 

 number of nuclei counted which contained no accessory was 

 577. These, however, would include the nuclei from which 

 small fragments containing the accessory had been severed. 



