THE SPERMATOGENFSIS OF THE SPIDER. 1 73 



but I think such to be the case, and also that the nineteen loops 

 are connected by a band of linin only. These loops grow thicker 

 and show a longitudinal split gradually increasing in width, while 

 the nucleus becomes more swollen with nuclear sap (PI. I., Figs. 

 10-13). At this time a definite polarity can be noted, the blind 

 ends of the loops being directed away from that portion of the 

 cell which contains the greater mass of cytoplasm and the cen- 

 trosome. The accessory chromosomes lie in the embrace of the 

 free ends of a loop and near the centrosome. In testes Hearing 

 maturity, the majority of the cells are in this stage. This is fol- 

 lowed by a rather rapid shortening of the loops when they draw 

 down toward what Montgomery has called the distal pole (PI. I., 

 Figs. 14, 15). In his description of this stage Wagner says: 

 " Der Linin faden (resp. die Reihen der chromatin-korner) bildet 

 Schleifen, die alle gleich lang sind und die gleich Richtung 

 haben : in dieser Weise theilt sich der Linin faden in Stiicke von 

 gleichen Lange. Gleichzeitig bildet sich der Nucleolus." While 

 I agree with him in regard to the same general direction of the 

 loops I do not agree with him in his statement that they are equal 

 in length. At this stage the difference in the size of the loops 

 can be plainly seen. This condition favors Montgomery's view 

 of the end-to-end conjugation of chromosomes of like size 

 during synapsis and which has been confirmed by Sutton in his 

 work on Brachystola. It seems highly probable, in the light 

 of recent research, that Montgomery's theory in regard to the 

 pairing of paternal and maternal chromosomes during synapsis 

 is the true explanation of the numerical reduction occurring at 

 this time. 



In the prophase of the primary spermatocytes, the bend of the 

 loops becomes more acute, while the arms shorten and thicken 

 and the now V-shaped chromosomes, varying in size, are scat- 

 tered through the nuclear cavity. In doubly-stained material 

 the accessory chromosomes appear as two red rods lying side by 

 side among the violet V-shaped chromosomes. The latter now 

 split from apex to base, opening out to form double V's and, 

 sometimes after, sometimes during this process are drawn into 

 the equator of the first maturation spindle (PI. I., Figs. 16-20). 

 In the metaphase the accessory chromosomes always take a per- 



