The Spermatogenesis of the Opossum (Didelphys virginiana) etc. 
.45 
A similar condition of polarization of the synaptic loops is parti- 
cidarly well illustrated in the spermatocytes of certain cockroaches descri- 
bed by Morse (1909). The reason why the relationship between centro- 
sphere and point of polarization can be so readily discerned in this material 
is due to the conspicuous character of the former. Possibly in all cases the 
polar orientation of the synizesis-mass is determined by the same in- 
fluence, the latter remaining obscure on account of the inconspicuous 
character of the centrosphere. 
Fig. 13 illustrates the important points respeeting this stage: 1. the 
loops are numerous (approximately twice the number characteristic of 
postsynaptic stages, fig. 16), delicate and intensely chromatic; 2. their 
lieight is approximately one half the diameter of the nucleus; 3. two of 
the loops have opened up with their free ends in apposition. Fig. 14 illu- 
strates a polar view of a similar stage showing more of the loops extended. 
The centrosphere here is located under the accessory chromosome in the 
center of the “bouquet”. In fig. 15 two of the extended loops have 
united end-to-end (telosynapsis). All of the loops have become slightly 
coarser and stain more intensely. 
A slightly later stage in this process is illustrated in fig. 16. Here 
most of the loops (about 9 — not all shown in Illustration; other loops 
in other sections) are nearly the height of the nuclear diameter, and the 
summits of the loops are marked by more compact chromatic knobs: 
the point of union of two threads (extended loops). The inference appears 
justified that the numerical reduction here takes place by a process of 
end-to-end union, as first suggested by Stevens (1905, 1906). There is 
as yet no sign of a duplicity of these final loops such as Morse (1909) 
describes in Blatta germanica, True, the loops are coarser just as Morse 
describes them in Blatta in Support of his position in favor of parasynapsis 
(Wilson), but this faet alone — and in the face of appearances indicating 
the opposite condition — does not necessarily indicate side by side union. 
The very slight indication of a split character of the chromosomes in later 
postsynapsis (diplotenic nucleus, fig. 20) need have no significance at all 
from the standpoint of reduction, but may simply be of the same nature 
as similar appearances in the early prophase of many dividing somatic 
cells (e. g. amphibia). 
Moreover, the nuclear Organization of the early postsynaptic phase 
leaves little doubt, I believe, as to the character of the reduction (i. e. 
by telosynapsis, Wilson) in this case. Here the Scattering loops are 
again united into an apparently continuous moniliform spireme (fig. 17). 
In later stages the original loops draw apart (figs. 18, 19 and 20) but 
