62 2 GRIFFIN. [Vol. XV. 



When once within the equatorial plate they again assume 

 this compacted condition (PL XXXIV, Fig. 50). Often they 

 are not dissimilar to the ring chromosomes in the first sperma- 

 tocyte division in Elasmobranchs, according to Moore's figures 

 ('95), and if we imagine the latter still further compressed lat- 

 erally so as to bring the halves into contact, the resemblance 

 would be perfect. Their arrangement within the plate agrees 

 essentially with that of Thalassema and is subject to but little, 

 if any, variation (PL XXXIV, Fig. 50). The loops are each 

 attached to a single traction-fiber, leaving the lateral ones quite 

 free — an arrangement that occurs as an exception in the first 

 spermatocyte division of Calopteiins (Wilcox, '90). Rarely one 

 of the tetrads has adjacent spheres attached to the traction- 

 fibers, as is the rule in Caloptctms. 



In Caloptetms and all forms with similar tetrads the spheres 

 retain their individuality and diverge in pairs. In Zirphaea, 

 however, the " spheres," being but contracted loops, the whole 

 chromosome becomes pulled out into a homogeneous ring 

 (PL XXXIV, Fig. 51). The quadruple appearance is thus 

 obliterated and is never regained ; so that henceforth we have 

 to deal with rings and daughter-V's. The details of the 

 process are easily followed ; the elongation of the polar loops 

 (PL XXXIV, Fig. 51) and the gradual shortening and disap- 

 pearance of the lateral loops, followed by division of the ring 

 into daughter-V's (PL XXXIV, Fig. 52), as well as the subse- 

 quent behavior of the latter, all take place exactly as in Tha- 

 lassema, but with greater clearness. 



It is highly probable that the rings and double rods present 

 throughout the growth period arise by a transverse segmenta- 

 tion and partial longitudinal splitting of an original spireme 

 thread, as shown in Thalassema and other forms characterized 

 by similar chromatic elements. This being granted, it follows 

 that the polar mitoses represent one " equation " (longitudinal) 

 and one "reducing" (transverse) division. Whether the reduc- 

 tion takes place in the first or the second polar mitosis, is, how- 

 ever, impossible to decide in this case, because of the close 

 similarity in appearance of the four loops. 



