MITOSIS IN POLLEN MOTHER-CELLS. 13, 
but of a continuous linin thread in which are held the chromatin disks 
or granules, 
In an early stage the nuclear thread shows a marked tendency to con- 
tract into a ball or mass about the nucleolus. The contraction into a 
dense ball is regarded by some observers as a perfectly normal occur- 
rence, to which the name syzafszs has been given. My own investiga- 
tions have convinced me that the contraction of the nuclear thread into 
a ball is in a large measure due to the reagents, and that synapsis has 
little or no significance. It indicates probably a very sensitive con- 
dition of the nuclear thread or net at the stage in which the contraction 
occurs. 
Soon after the nuclear net has developed into the spirem, as men- 
tioned, the chromatin and linin elements split longitudinally (Fig. 5, 
D, a, 6, E). The daughter spirems remain either closely applied to 
each other, or, as sometimes happens, they may separate for longer or 
shorter intervals. They are always twisted upon each other, and, as 
a consequence, the two parallel rows of disks are not easily seen, 
especially where the chromatin thread makes short turns. The twist- 
ing of the daughter spirems upon each other persists after the trans- 
verse segmentation of the spirem into chromosomes, and in very many 
cases it is still to be seen during metakinesis (Figs. 6, 7). 
Very frequently portions of the spirem which run parallel with 
each other are connected by very fine threads, and, in some cases, as 
in the pollen mother-cells of Podophyllum, very delicate cytoplasmic 
threads seem to penetrate the nuclear membrane and fasten themselves 
to the chromatin spirem. At this stage also one or more nucleoli, of 
varying sizes and with a homogeneous or vacuolate structure, are pres- 
ent. The nuclear membrane, especially in Podophyllum, does not 
present from now on the sharp contour of the resting nucleus. It seems 
to consist merely of a cytoplasmic boundary (Fig. 5, F), and as will 
be pointed out in a later paragraph, we may conclude that the nuclear 
membrane consists of an extremely delicate kinoplasmic network, 
whose meshes in the resting nucleus are so closely arranged that only 
a sharp line is seen when observed in optical section. As soon, 
however, as the meshes widen with the increase in size of the nucleus 
the nuclear membrane loses its sharp contour. It cannot be asserted 
with absolute certainty that the fine threads extending from the nuclear 
membrane to the chromatin thread penetrate the membrane and con- 
tinue into the cytoplasm, but in Podophyllum the evidence seems to 
be in favor of such a view. At any rate there seems to be an intimate 
connection maintained between chromatin and cytoplasm. 
As karyokinesis progresses, the chromatin thread contracts, becom- 
