MOTTIER AND NOTHNAGEL: CHROMOSOMES OF ALLIUM 557 
the nucleus in a pollen mother-cell soon after the last somatic 
division. The structure of the whole cell is the same as that of 
any somatic cell from any meristematic region. The growth 
period of both cell and nucleus now begins, and the very marked 
increase in the size of the nucleus as compared with that of the 
cell is very conspicuous in this as well as in other species of Allium 
(Fic. 2, 7, 8, 9). In Fic. 2 the nucleus is almost if not quite as 
large as it ever becomes. The nuclear reticulum is uniform, and 
the nucleoli may or may not be evenly spaced in the cavity of the 
nucleus. They do not lie in the same plane, and in making the 
drawing the focus was necessarily changed. Fic. 2 and all others 
represent rather thick sections of cells. Sometimes the chromatin 
granules form larger and smaller aggregates, which may be grouped 
about the nucleoli or removed from the latter, but we do not find 
large fused masses of chromatin such as Bonnevie has figured 
and described as ‘‘Chromatinknoten.”’ A glance at FIG. 2 shows 
further that there may be a tendency to form a thread, that is, 
‘there will be seen stretches of linin in which the granules are 
arranged in lineal series. As pointed out by one of us (Mottier, 
07) for Lilium Martagon, there is a tendency in Allium cernuum 
to form a delicate thread or spirem just before or as the nucleus 
passes into the synaptic contraction. Fic. 2 is about ready to 
begin the contraction of its net into the compact mass. FIG. 3 
is a faithful attempt to illustrate the nuclear structure passing 
into synapsis, and FIG. 4 is a similar stage but includes the whole 
nucleus. These two figures were found in the same section of the 
loculus, in which were to be seen variously different stages of the 
early contraction. The writers wish to state most emphatically 
that there is no evidence of the fusion of two spirems during the 
contracting process. The spirem is formed directly from the nuclear 
network in the only way possible for a net to make a continuous 
thread, namely, by the breaking or dissolving of threads of certain 
meshes and the fusion of others. In the fusion of meshes several 
threads are seen to unite just as frequently and as certainly as 
one may find the fusion of only two threads. The appearance of 
the lateral union of two threads of certain meshes in several parts 
of the nucleus previous to synapsis is the strongest evidence, in 
the opinion of the writers, that those observers can bring forward 
