68 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
original spireme segment, and its middle point to the distal pole. The 
rod-shaped tetrad becomes oriented in the spindle of the first matura- 
tion division with its long axis parallel to the spindle-axis, and at 
metaphase separates in the middle. In other words, the plane of the 
secondary split becomes the plane of the first maturation division, 
which is therefore equational. If now we may assume that the longi- 
tudinally split spireme segment has represented a pair of chromatin- 
threads which had conjugated side-by-side throughout their length, 
the plane of the primary split must be the plane of the reductional 
division, which becomes effective in the second spermatocyte mitosis. 
The tetrad A also forms rings, as shown in figure 62, 7, k, / (Plate 6). 
I have not been able to trace these rings into the metaphase to de- 
termine their orientation on the spindle, and furthermore I am quite 
uncertain whether the ring shape persists as far as the metaphase. 
Most of the metaphase figures show one tetrad in the form of a rod 
with its axis parallel to the spindle-axis, and with a constriction in the 
middle, as shown in figure 62, 7 and figure 79, A (Plate 7). Sometimes 
two or more rod-shaped tetrads are to be seen in the same spindle 
and with the same orientation. However, one of them is always in a 
more advanced stage of division than the others, and I have been in- 
clined to identify this precocious one with tetrad A. Figure 62, c-, 
indicates that such a conclusion is justified. Since the straight-rod 
condition is so characteristic of the metaphase, it may be that the 
rings also become transformed into straight rods by the time the 
metaphase is reached. | i 
The rings seem to have been formed either by a failure of the proxi- 
mal ends to separate during the formation of the secondary longi- 
tudinal split, or by a secondary union of these ends, 2. e. after the split 
had begun. For example, if a tetrad in the condition of figure 62, c, 
has the secondary split completed without the separation of the proxi- 
mal ends, a ring would result. So also would a ring be formed by a 
secondary union of the two proximal ends of a stage such as is seen in 
figure 62, d or e. In either event the region within the ring would 
represent the space formed as a result of the secondary longitudinal 
split. If the chromatids should now begin to separate at the proxi- 
mal end along the plane of the primary split, as seems to be indicated 
in figure 62, & and J, and if this process should be continued until a 
metaphase chromosome such as that shown in 7 is produced, there is 
every reason to believe that it would result in a separation of the 
original conjugants of the pair, and therefore constitute a reductional 
division. On the other hand, it is possible that the separation along 
