Further studies on the Gametogenesis of Pandanis sinuatus, Say. 231 
numbers of these festes, yet in uo case has it been possible to count 
the number of chromosoines in the early part of synezesis. The paired 
arrangement of them, however, is clearly seen. Xear the entl of this 
Stage all of the filaments can sonietimes be distingnished, and are eight 
in number, thicker than in the leptotane stage, and double at least in 
some parts, fig. 6. 
The eight double filaments now become dispersed toward the per- 
iphery of the nucleus, giving the diplotane stage, figs. 7 and 8. The double 
filaments offen interlace so that they cannot be counted, although where 
counting is possible eight are found. The double chromosomes now shor- 
ten, fig. 9, and I once thought that a transverse constriction was present, 
yet this might possibly be an accidental notch in the ragged outline of 
the double chromosome, and further study has not cleared up this point. 
Leptotane Synezesis>Synapsis Diplotane 
Shortening Ist. Mat. Spindle 
At a little later stage, however, a distinct second division is present, trans- 
forming the double chromosome into a cetrad, fig. 10. The shape of the 
tetrad is such that its longitudinal axis cannot be determined, and there- 
fore it is impossible to decide whether the second division (constriction) 
is longitudinal or transverse. 
The two maturation mitoses then rapidly follow, distributing the 
parts of each tetrad into the four spermatids, figs. 11 — 13. Owing to the 
form of the tetrads it would be difficult to decide whether the reduction 
in number of the chromosomes occured in the first maturation mitosis 
(pre-reduction) or in the second (post-reduction), but before going into 
this question we may briefly consider whether pre- and post-reduction 
have different effects in heredity. 
Obviously in ordinary sexual reproduction if reduction occurs at all, 
it is immaterial whether it takes place during the first or second matiu’a- 
tion mitosis. Weismaxx found that there is no second maturation divi- 
sion in certain parthenogenetic eggs, and concluded that reduction is thus 
omitted. Recent investigations show, however, that if synapsis occurs 
