A Comparative Study of the Clu'omosomes etc. 
257 
(Plate XIX, Figs. 51, 52). The iinion never takes place by a twisting of 
two threads about each other, but merely by the parallel approximation 
of two single threads, which show the plane of union as a light straight 
line when the thi’ead is viewed at a favorable angle. The zygotene threads, 
showing their doubleness well in osmic acid material, soon enter the 
bouquet stage, with the free ends of the loops pointing toward one side 
of the nucleus, that is, toward the positive pole (Figs. 53, 54, 55). The 
presence of a centrosome at this pole could not be determined in my 
material. Stages where one (or more) of the zygotene threads had not 
yet formed a complete loop, as in Figure 53, were often encountered. 
I could not detect at this stage an exact correspondence of the granules 
of the two components of the double threads such as has been maintained 
by adherents both of metasyndesis and of parasyndesis. In E. curvata 
the granules are more distinct, and I shall refer to theni in the description 
of that species. 
As the polar arrangement of the loops is assumed, there is a gradual 
shortening and thickening of the threads to the pachytene condition. 
It is here that synizesis occurs, which, however is not of long duration. 
In properly fixed material, especially that fixed in Hermann’s fluid, the 
chromatin never clumps together into a shapeless, homogeneous mass, 
but the shortencd loops, which crowd together at one side of the nucleus, 
maintain theh- general form and position, exhibiting their doubleness 
when cut in cross section. The threads soon become longer and less 
deeply stained, and show the light longitudinal clef t more clearly (Plate XIX, 
Fig. 55). This is the stage most favorable for a study of the double threads. 
Side views show that there is always one loop (Fig. 55 M.) which is longer 
than any of the others, and this doubtless corresponds to the pair of 
macrochromosomes seen in the spermatogonia. Another characteristic is 
the presence of a single pair of deeply staining threads which do not form 
a loop (Plate XIX, Fig. 55). This pair is peculiar in several other ways. 
The halves are often united by one end merely (Fig. 54), and only later 
become approxünated tlu’oughout their entire length. But by far the 
most strikhig peculiarity is the difference in structure exhibited by the 
components of this pair, one {x) being compact and homogeneous, the 
other {y) more like the members of the autosome loops; however, the 
latter {y) stains more deeply than the autosome threads and is made up 
of larger and more distinct granules (Plate XIX, Fig. 55 xy). This pair 
can be foUowed through the entire growth period, appearing with re- 
markable clearness in the stages foUowing the polar arrangement of the 
loops. Its cüstinctness depends, however, a great deal on the degroe 
17 * 
