146 
Dig by. — On the Arckesporial and 
It is only very rarely that association of a long length of spireme 
thread is to be seen proceeding in a nucleus, but such cases, when present, 
are most striking. The half univalent spireme (thread) is thrown into loops 
which are orientated towards one portion of the nuclear periphery, and there 
the sides of the loops (i. e. threads) are seen to be associating in pairs (Figs. 
47 and 48). This investigation has necessitated the examination of a very 
large amount of material, and only these two diagrammatic examples have 
been found. If the spireme loops are orientated solely towards the nuclear 
poles, then the chance of cutting the nucleus at the precise angle to demon- 
strate the phenomenon would be remote. In the majority of nuclei at this 
stage, association can be detected as taking place between lengths of spireme 
thread in the synaptic knot, but these cases are naturally not nearly so con- 
vincing as in the nuclei figured. Gregoire ( 10 ) has published a beautiful 
drawing (Planche I, Fig. 24) of this polarization showing the association 
in a most graphic manner. He writes: ‘ C’est surtout dans YOsmunda 
regalis que nous avons rencontre les plus beaux cas de dualites d’accole- 
ment. . . . Dans la Fig. 24 on voit des filaments 1 minces nettement orientes 
vers un pole du noyau, chose assez rare dans les sporocytes — nos figures en 
apportent le premier cas aussi clair, a notre connaissance * (p. 378). He 
interprets the association of threads in synapsis as the association of two 
lengths of entire univalent spiremes, each associating thread representing an 
entire somatic chromosome ; whereas the conclusion arrived at during the 
present investigation is that each length of associating spireme represents 
the longitudinal half of a univalent spireme or univalent chromosome. 
The association of threads in pairs to form the spireme filament is 
completed before the spireme begins to come out of synapsis ; the emerging 
spireme filament, therefore, is double throughout its length — no single unasso- 
ciated threads remain (Fig. 49). Accordingly from late synapsis onwards the 
nature of the spireme is that of an entire univalent filament, i. e. the product 
of the lateral association of the two longitudinal halves (threads) of univalent 
spireme, but separation between the halves (threads) may still remain more 
or less evident. This spireme filament is homologous with the continuous 
spireme filament of a somatic prophase, before it segments transversely into 
the somatic chromosomes. 
The nucleolus occasionally buds during synapsis. No typical formation 
of chromatic bodies has been observed, but von Derschau ( 2 ) has figured and 
described them in O. regalis. 
Early Hollow Spireme Stage. 
As the univalent spireme filament comes out of synapsis (Fig. 49) 
fission is, as a rule, conspicuous. Sometimes the separation of the two 
associated threads is considerable, sometimes it is shown merely by the 
1 ‘ Threads according to the nomenclature adopted in this paper (see foot-note, p. 159). 
