108 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
Fig. 82) is easily distinguishable in the spermatid for some time, as 
its disintegration takes place much later than that of the autosomes. 
G. Steiroxys trilineata. 
1. Autosome. 
At the close of the growth period the loops of the polar spireme 
become freed from the nuclear membrane, as in the Acrididae. More- 
over, the chromatin and linin become much concentrated so that in 
stage g (Plate 5, Fig. 70) the diameter of the threads is much less than 
during the preceding stage. At the same time, as a’ result of the 
aggregation of the chromatin, the longitudinal split reappears. Some- 
what later the loops have become converted into tetrads, which stain 
deeply and have such a ragged outline that it is almost impossible to 
determine their structure; but they have apparently (Figs. 71, 72) 
forms similar to those of the Acrididae. ‘There is always one tetrad 
which is much larger than the others and is undoubtedly formed from 
the two large autosomes of the spermatogonia. ‘The form of this 
tetrad is well shown in Figure 71. It is apparently formed by the 
arms of the loops becoming twisted around each other, and, as in the 
Acrididae, each of these arms no doubt represents a univalent auto- 
some. During the late prophase of the first maturation division this 
autosome becomes disposed on the spindle with its long axis at night 
angles to the axis of the spindle (Fig. 73). In addition to the large 
element there are thirteen smaller autosomes (Plate 8, Fig. 189), which 
have approximately the same size relations as the autosome pairs of 
the spermatogonia. In the case of the large autosome the following 
division is apparently longitudinal, but, Judging from analogy with 
the Acrididae, results in the separation of its univalent components. 
As the two components separate (Fig. 190) each is split longitudinally 
and is evidently composed of two curved rods lying side by side. In 
the case of the smaller autosomes I have found it impossible to deter- 
mine their orientation on the spindle with any accuracy, but, as in the 
large autosome, the first division is probably reducing. As the small 
autosomes separate (Fig. 190), their dyad structure is usually apparent, 
although the longitudinal split is often difficult to distinguish and can 
best be seen in polar views of the anaphase (Fig. V, p. 106). Figure 
191 shows the beginning of the telophase, the autosomes being collected 
in a mass near the poles of the spindle, while Figure 192 is a late telo- 
phase. During the ‘‘semiresting stage” of the secondary spermato- 
