1904] MERRIMAN: VEGETATIVE CELL DIVISION 195 
or if the linin has broken down as in later stages, the split is the 
result of looking down upon the vacuoles which were formerly 
inclosed by the tetrads. That the split is not sharply defined is 
due. to the persistence of the less deeply stained linin substance, 
which still connects in an uneven manner the masses of chro- 
matin. 
The growth and extension of the chromatic figure imply a 
correlated extension of the nucleus until the nuclear membrane 
breaks as a result of the interior pressure. Previous to the dis- 
ruption of the nuclear membrane, the coils of the spireme have 
been quite regularly arranged within the nucleus, as shown in 
Jigs. 1g and 15, following the same arrangement that is to be 
seen in the early spireme, fig. ¢. With the weakening of the 
membrane or closely following its dissolution, and with the con- 
sequent changes in pressure, the coils of the spireme break 
transversely at the places where they were bent. This breaking 
does not imply that the segments of the spireme have become 
completely separated, for, as fig. 30 shows, the linin substance 
may still connect the segments of the spireme into a more or 
less continuous thread. 
The number of segments thus arising seems to be inconstant, 
apparently varying from ten to thirty or more. There isa great 
deal of uncertainty, however, as to the number of segments, 
because many of them are cut in sectioning, and thus the appar- 
ent number is increased. A study of fig. 30, which shows 
thirty-eight such segments, would seem to indicate that none of 
the chromosomes there has been cut by the knife, because there 
exists linin between the successive segments of the bands. Ifa 
chromosome had been cut off square by the knife, its end would 
appear sharply truncated, with no linin substance connecting it 
with an adjacent chromosome. Such evidence as this and that 
presented by the cells shown in figs. 37, a and 4, leads to the 
conclusion that in vegetative cells of Allium the spireme is not 
invariably broken into sixteen chromosomes, as has been main- 
tained by other investigators, but that the number of segmetits 
is dependent upon the size and course of the spireme in the 
nucleus. This transverse breaking of the spireme into segments 
