Sperm-Cells in A sclepias Cornuti,. Decaisne. 131 
condition of the nucleolus as shown in Figs. 10 and 12-16 
When there is no fragmentation there are usually no vacuoles 
present, though this is not always the case. 
Fragmentation of the nucleolus has been described by 
Farmer (1893) for Lilium Mart agon , and (1895) for Pellia 
epiphylla , where the particles into which it breaks up are 
found usually on young chromosomes or close to them. 
Farmer suggests that this proximity is due to ‘ mechanical ’ 
causes. From the staining reactions in A sclepias, the sugges- 
tion is very strong that it functions as a reservoir for chromatin 
or for some allied substance, as Farmer (1895) and Pfitzner 
have suggested. 
Miss Sargant (1896), in her studies on Lilium Mar tag on, 
thinks that the nucleolar substance serves as material for the 
spindle-fibres, or at least to solder them to the chromosomes. 
This may be its function in A sclepias, as the spindle-fibres 
seem to be of nuclear origin. Other facts, however, argue 
against this interpretation. 
Zimmermann (1893) believes that the nucleoli are thrust out 
into the cytoplasm, and then in the metaphase taken up again 
into the daughter-nuclei, £ omnis nucleolus e nucleolol But if 
this were true in A sclepias, then it becomes more difficult than 
ever to explain the loss of nucleolar substance, as such, after 
the first division of the primary pollinium-cell, for Stevens 
(1898), by careful measurements, has shown that the volume 
of the nucleoli of the daughter-nuclei of the first division is 
only about one-tenth of the volume of the nucleoli of the 
primary pollinium-cell nucleus. 
After the pollinium-cells are formed they contain prominent 
rod-shaped bodies which will be referred to later. These 
bodies stain very deeply with the gentian-violet, and it is 
possible that this fact may be associated with the disappear- 
ance of the nucleolar substance, as such. Stevens 1 argues for 
the nutritive function of the nucleolus, but thinks that its 
substance goes to form c the kinoplasmic spindle and the cell- 
plate? On this supposition, however, large nucleoli would be 
1 Loc. cit, p. 83. 
K 2 
