No. 3.] THALASSEMA AND ZIRPHAEA. 607 



is probably the beginning of a process that will be treated 

 in detail after the description of the morphology of the 

 chromosomes in metaphase. From the complete failure to find 

 forms indicating a direct transformation of the open rings into 

 the metaphase figures, it seems not improbable, that the thick 

 rods may represent a later and more concentrated and com- 

 pressed stage of the rings. This also appears to be indicated 

 by measurements which show the rings to be of greater length 

 than the rods. The open ring, however, is not to be consid- 

 ered a necessary stage, for it is quite evident that many of the 

 rods arise directly from the similar structures present in the 

 germinal vesicle. 



More difficult to determine are the occasional two short rods 

 lying side by side {k, I, m), as well as crosses or ophiurid forms 

 («), the arms of which appear perfectly solid without the slight- 

 est indication of a division into halves. A few may possibly 

 be explained by supposing an accidental adherence of two 

 distinct chromosomal rods. Others may have resulted from a 

 premature division, either longitudinal or transverse, of a single 

 rod, and the subsequent rotation of the halves ; or the cross 

 might be considered a ring compressed along two perpendicular 

 diameters, whereby the four included quadrants were converted 

 into loops. The close adhering of the halves of these loops 

 might well obscure the line of separation and cause the arms 

 to appear solid. Analogy with Zirphaea would strongly sug- 

 gest the last-mentioned view. 



Despite the varied forms presented during prophase, the 

 chromosomes of the equatorial jjlate exhibit considerable uni- 

 formity. Hence the various prophase forms must in some man- 

 ner be convertible into a uniform type of metaphase figure. 

 The commonest and, as we shall consider it, the typical form 

 assumed by the chromosomes in the latter stage is that of a 

 cross, with a pair of broad arms in the equatorial plane, and 

 the narrower perpendicular arms directed toward the poles 

 of the division figure (PI. XXXI, Fig. 12, Text-fig. II, 0, p). 

 The latter pair or polar arms are of varying length, and invari- 

 ably end distally in a solid knob, to which a single traction- 

 fiber is attached. The more proximal portion is often seen 



