Division in Mnhim hornum . 
145 
division, although simultaneous in the cells of each group, does not take 
place at the same time in the different groups. The process in all these 
divisions is identical, and the following description applies to any premeiotic 
mitosis. Divisions of the vegetative cells of the sporogonium have been 
examined, but differ in no respect from those of the archesporial cells. 
In some of the nuclei the extremely homogeneous structure of the 
nuclear reticulum is replaced by a more granular appearance ; the alveoli 
are larger and not so regular, and small granules of darkly-staining material 
make their appearance at the intersections of the network. Nuclei in this 
condition are constantly found in preparations fixed with acetic alcohol and 
the above-described homogeneous structure is only obtained when Flem- 
ming’s Mixture has been used. At this stage the superiority of the latter is 
clearly demonstrated. But even in material thus fixed some nuclei still 
exhibit the more granular appearance, and this condition, although frequently 
due to imperfect fixation, is in some cases the first sign of approaching 
mitosis. A little later a clear space appears in the middle of the nucleus 
(Fig. 3). This space is to be distinguished from the light area surrounding 
the nucleolus mentioned by Strasburger ( 23 , p. 10), and considered by him 
to be due to contraction at the time of fixing. Such an area is often seen 
in the nuclei of the resting cells (Fig. 1). The chromatin increases in 
amount and broad band-like areas appear in the more peripheral parts of the 
nucleus (Fig. 4). As the result of further contraction, these become gradu- 
ally narrower and more elongated (Fig. 5), and ultimately give rise to the 
spireme (Fig. 6). During this time the nucleolus has been losing its stain- 
ing capacity, and frequently shows lighter and darker areas within it. At 
the differentiation of the spireme it has completely disappeared. At this 
stage only a few small irregular masses of deeply-staining material can be 
seen, and these are in close connexion with the chromatic thread. 
Comparison of these stages with those described by Farmer and Shove 
( 9 ) in divisions of the cells of the root of Tradescantia reveal a close similarity, 
particularly with regard to the formation of the band-like chromatin masses. 
(Compare Figs. 4 and 5 with Figs. 3 and 4 of above work.) 
It could not be definitely ascertained whether the spireme consists of 
a continuous thread. In several cases it appears that free ends are present 
(Fig. 6), and that these are not due to removal of part of the nucleus in 
sectioning. Although careful search has been made, no traces of longi- 
tudinal fission have been discovered in the thread at this or at a later stage. 
The nuclear membrane, present in the earlier prophases, now completely 
disappears, and the spireme lies free in the protoplasm. Free ends can now 
be seen with certainty in the thread, and portions of it are arranged about 
a central plate, although, at this time, no sign of the spindle can be discovered 
(Fig. 7). Partition of the thread is not yet complete, and loops are found 
between the free ends ; the number of chromosomes cannot yet be ascer- 
