440 
George Arnold 
not been üble to observe tbe transverse achromatic bridge in the 
middle of eacb Segment. 
Schockaekt, in bis work on Thysanoxoon (’OO and '01) has descri- 
bed a similar state of atfairs, which he thinks fully Support the 
theory of an end-to-end juuction of tbe somatic cbromosomes. 
For the later stages, up to the formation of the gernini, about 
to be dcscribcd, my results do not differ largely from those of Schleie, 
but bis interpretation does differ of course, owiug to bis contention 
that the tliick spireme is formed bv the longitudinal pairing of uni- 
valent threads. But, like Schleip, I liave also failed to find anything 
in these later stages of the prophase similar to the pseudo-tetrad 
formation described by Mattiesen. 
After the synapsis, which as has already been pointed out, 
represents a period of extreme condensation of botli the linin and 
chromatin, the spireme Segments open out again and lie loosely in 
the nucleus uear its periphery (Fig. 9). 
The chromatin now gradually leaves the ends of and masses up 
towards the middle of each linin band. This takes some time (Figs. 10, 
11, 12 aud 13), but eventually the microsome granules which make 
up the chromatin are no longer separately distinguishable, so thickly 
does the chromatin coudense towards the middle. In this way, the 
ends of each spireme segment consist of linin almost devoid of any but 
the smallest chromatin particles (see Fig. 13 lin). These free ends 
stain more and more faintly and apparently dissolve in the nuclear 
sap. leaving behind thin slireds, dotted here and there with the 
miautest of chromatin granules (Figs. 14, 15 and 16). 
In sections stained with Heidenhain’s Iron Alum-Haematoxylin, 
the linin is not stained, or only very indistinctly , so that in a cell 
in which the chromatin has not completely Condensed (Fig. 14), the 
chromatin masses appear entirely separated from each otlier. Probably 
this explains Mattiesen ’s pseudo-tetrads, as he used that stain. But 
wheu stained with the triple stain described above, the linin shows 
up most clearly, and the continuity of each segment is apparent. The 
chromatin now becomes more evently distributed in the remains of 
the linin segments. The ends of these segmeuts now beud over and form 
the characteristic loops, and eight-shaped figures (Figs. 15 and 16). 
The triple stain used does not seem to stain the centrosomes, as 
judging by the condition of the nucleus tliey ought to be seen in 
Figs. 15 and 16. However, Fig. 13 stained with Thionin and Orange G, 
shows the aster and libres verv clearlv. 
