SPERMATOGENESIS OF THE DRAGON-FLY. 2J3 



can be separately distinguished opens up the possibility of 

 experimentation upon cell division when a solution isotonic with 

 the body fluid of the dragon-fly nymphs can be determined. 



IV. SPERMATOGENESIS OF SYMPETRUM SEMICINCTUM (SAY). 



(a) The Spermatogonial Period. 



The length of larval life in this form is unknown and it prob- 

 ably varies with temperature and food supply. Nymphs ob- 

 tained early in May this year w r hich measured i cm. in length 

 were at least one year old for the adults had not begun to emerge. 

 These nymphs possessed no visible gonads. In the youngest 

 larvae bearing reproductive organs, it was exceedingly difficult 

 to find spermatogonial divisions. Until the life cycle is known 

 no explanation can be offered for this, but it may be that the 

 gonads develop during the winter of the second year of larval 

 life after the insects are in the mud at the bottom of the streams 

 where it is hard to find them. The material of this particular 

 species was difficult to work with on account of: (i) the small 

 size of the cells and their closely crowded condition ; (2) the 

 considerable number of chromosomes; (3) the irregular arrange- 

 ment of the cysts as to age, referred to under III. 



The spermatogonial cells could be easily distinguished by 

 their large nucleoli and their large-mesh nuclear network. The 

 chromatin granules, sparsely scattered along the linin network 

 in the center of the cell, were collected in small clumps close to 

 the nuclear membrane. This arrangement of the chromatin gave 

 to the nucleus a clear appearance. Frequently small chromatin 

 bodies appeared in the network, but these were inconstant in 

 number and apparently of no importance in the later develop- 

 ment. The nucleolus appeared more often as composed of a 

 clear ground substance, probably linin, in which masses of 

 chromatin were imbedded. It sometimes looked like a solid 

 chromatin mass. Lewis and Lewis ('15) found that in a living 

 cell, the nucleolus was never a compact body, but was coarsely 

 granular and large in proportion to the nucleus. The degree of 

 contraction of the ground substance inclosing the granules, 

 depending upon the fixation of the material, would account for 



