260 KANSAS UNIVERSITY SCIENCE BULLETIN. 



the plate, and the daughter groups are therefore inclined to 

 each other instead of being parallel. No other peculiarities in 

 the division of the ordinary chromosomes presented them- 

 selves. (Figs. 17, 18, 19, 20.) 



GENERAL CONSIDERATIONS. 



The points of interest in the spermatogenesis of Xiphidium 

 relate to the accessory chromosome, and I shall therefore first 

 briefly summarize its history and then speak of the accounts of 

 other investigators. 



It is to be noted that the accessory chromosome is a single, 

 very large spermatogonial chromosome that enters the pro- 

 phase of the first spermatocyte without undergoing any ex- 

 tensive separation of its chromomeres. While it forms a 

 thread, this is always denser and more compact than the gen- 

 eral spireme. The result is that the element always stains 

 strongly, after the manner of metaphase chromosomes. It is 

 thus made very conspicuous in its position on the periphery of 

 the nucleus. It passes through a variety of forms, but finally 

 enters the metaphase of the first spermatocyte as a bent rod, 

 and in this form passes undivided to one pole of the spindle. 

 The result is to produce two types of second spermatocytes. 



In the metaphase of the second spermatocyte the accessory 

 chromosome takes its place in the equatorial plate and is there 

 divided along the plane of its early longitudinal cleavage — an 

 equation division. From each first spermatocyte there are ac- 

 cordingly produced four spermatids, two of which possess an 

 accessory chromosome while the other two are without it. 

 These results are in accordance with my more extensive work 

 ('02) upon other members of the same family, and vary at 

 points from the preliminary account upon Xiphidium. 



The only other investigators that have concerned themselves 

 with the spermatogenesis of the locustids are de Sinety ('01), 

 Stevens ('05), and Otte ('06). The work of Sabatier ('90) 

 dealt only with the transformation stages, and so does not con- 

 cern the present discussion. The results of de Sinety and 

 Stevens in the main agree with mine; those of Otte are rad- 

 ically different in essentials. I wish here briefly to note the 

 work of the latter author upon the accessory chromosome only. 

 I hope to consider the maturation divisions of insects in detail 

 in a subsequent paper. 



