8 STUDIES IN SPERMATOGENESIS. 



was also conspicuous in all of the cells of this group. The spireme 

 here, as also in figure 99, is fine and closely interwound. In figure 

 99 and again in figure 100 the element x is joined to the spireme as it 

 is throughout the spireme stage. In the ' ' bouquet " or " polarized ' ' 

 stage the combined nucleolus and element x are always at one side of 

 the group of loops and down very close to the base of the figure (figs, 

 loi, 103). In figure 102 most of the loops are cut across. The stage 

 shown in figures 104 and 105 is a later one than that just described. 

 Here we have again a continuous spireme connected with the element 

 X, making it seem improbable that the bivalent chromosomes are 

 really separated in the bouquet stage. Figure 106 gives some of the 

 variations in form of the combined nucleolus and element x. The last 

 of the five figures was taken from a giant cell containing at least twice 

 the usual amount of chromatin. In one giant cell four unusually 

 large combinations of this kind were found, and a corresponding 

 amount of chromatin in the spireme. In figure 107 one sees the 

 spireme divided into segments still joined by linin bridges. In figure 

 108 similar segments may be seen, one of them showing a longitudinal 

 split. The element x is present, but the nucleolus has disappeared. 

 In many cases the split, if it appears at all, closes quickly and the 

 chromosome bends in U-shape, as in figure 109, plate iv. This figure 

 also shows two centrosomes (t). In other cases the split persists as in 

 figure 1 10 and leads to the formation of crosses of a tetrad character 

 (figs. Ill, 112, 113), as in Stenopelmatus and many other insects. 

 Figures 114 to 117 show later stages of the U-shaped chromosomes. 

 Perfect rings are rare. All sorts of variations are seen, broad and 

 narrow U -shapes, rings split at one point or the opposite points, a 

 U split at the bottom (fig. 114), pairs of parallel rods (fig. 115), and 

 occasionally rods constricted in the middle and showing a longitudinal 

 split in each half, as in figure 116. Figure 117 shows different views 

 of the split rings. Apparently all of these forms straighten out so that 

 the two components of the bivalent chromosome stand end to end as 

 dumbbells or compressed crosses in the metaphase of the first matura- 

 tion spindle (figs. 123-125). The element x remains concentrated and 

 more or less spherical in form. Figures 1 18-122 are equatorial plates, 

 with X absent in figure 120, in the same plane as the 11 other chromo- 

 somes in figure 119, far to one side in figure 118, and near one pole of 

 the forming spindle in figure 122. It is also shown in various positions 

 with regard to the spindle in figures 123 to 126 and 128 to 132. In 

 figure 125 it is apparently double, and again in figure 129. In figure 

 130 one lagging chromosome shows the dyad nature of the products 



