432 BOTANICAL GAZETTE [JUNE 
nucleolus was not in contact with the nuclear membrane, but was 
lying free in the nucleus. The “sickle stage,’”’.was seen only in 
poor preparations, and hence I am inclined to regard it as an 
artificial product. In many nucleoli at the same stage there are 
a number of smaller vacuoles, instead of a large central one 
(fig. 7). There is no doubt but that the term vacuole is a mis- 
nomer, but for lack of a better one this name will be used for 
the larger clearer areas in the nucleolus. 
After the division of the chromatin granules the entire chro- 
matin band or spirem undergoes longitudinal splitting, producing 
a double linin thread, each thread containing a single row of 
chromatin granules (figs. 8-ga). The double number of chro- 
matin bands makes a very characteristic appearance when com- 
pared with the earlier stages before splitting. There does not 
appear to be any substance connecting the two chromatin bands, 
the longitudinal fission appearing complete. 
At this stage there often appear peculiar radiations or tan- 
gential filaments in the cytoplasm. These generally stretch 
from one side of the cell to the other, passing the nucleus as 
tangents (fig. §). Whether this appearance was an artificial 
production or not I could not determine, but it is probable that it 
isa natural condition, as the threads appeared in numerous sections 
which did not seem to be otherwise disturbed. At this stage 
the two centrospheres, which were sometimes seen, still lie ncn 
together beside the nucleus. 
After the splitting of the chromatin band the two resulting 
bands now begin to twist on one another, the twisted spirem 
being in marked contrast with the former parallel arangee 
(figs. ro, roa). In the meantime the nucleus has enlar 
: considerably. After the two threads have twisted quite closely 
together the resulting twisted chromatin band arranges itself sO 
; as to form twelve loops, the heads of the loops being close to. : 
_ the nuclear membrane. Each loop contains from one to three 2 
twists. At first the double nature of the chromatin band is still a 
. ie evident (gs. Z 1-116), but later the two linin threads are — 
. ely associated, (caimpcete oF the a ae ~ nce 
