DAVIS: SPERMATOGENESIS. 93 
4. Tae MATURATION PERIOD. 
A. Dissosteira carolina. 
1. Auwtosomes. 
At the end of the growth period the loops of the spireme, though still 
showing their polar arrangement, have taken a peripheral position, 
while the longitudinal split has become indistinct. In the next stage 
(g), which marks the beginning of the maturation period (Plate 4, 
Figs. 57, 58), the spireme loops have become detached from the nuclear 
membrane and are irregularly distributed through the nucleus. ‘The 
chromatin becomes aggregated into larger granules, while the longi- 
tudinal split is more distinct than at any previous stage, the space 
between the two series of chromatin granules being wider than at any 
other time. 
In stage h (Fig. 59) the loops have become converted into definitive 
tetrads. These vary greatly in shape, but may be roughly grouped 
into three types: (1) straight or curved rods, (2) crosses, and (8) rings, 
or loops with their free ends crossed. ‘These types are all, I believe, 
modifications of a common fundamental form composed of two longi- 
tudinally split threads or rods of equal length joined end to end. Each 
tetrad is apparently formed from a single loop of the spireme in the 
following manner: When the loops first become free they are appar- 
ently of uniform appearance throughout their entire length, but a little 
later the longitudinal split begins to widen at the middle of the loop 
forming a diamond shaped opening (Plate 7, Figs. 160, 161). This 
opening I believe to be at the point of union of two univalent autosomes 
of the spermatogonia. ‘These rod-shaped autosomes are usually 
more or less curved, rarely straight. ‘This is the simplest type of tet- 
rad which I have found in this species, but in Hippiscus a still simpler 
form occurs (Fig. 159). In this genus there is no widening of the 
longitudinal split at the middle, but there is at this point a distinct 
break in the chromatin, the intervening space being bridged over by 
linin. 
The second type of tetrad has the form of a cross (Figs. 165-167) 
in which the four arms may be approximately equal in length, or one 
pair may be longer than the other. A further complication is caused 
by the arms being usually more or less bent, the free ends of each pair 
tending to approach each other. However, both pairs of arms never 
bend in the same direction. Figure 169 shows a case where this 
