FARR: CYTOKINESIS OF POLLEN-MOTHER-CELLS 257 



onion, which are attenuated at the equator. The fibers of the 

 larch then become of uniform diameter, and lastly thicken at the 

 equator, giving rise to swellings. The process continues until 

 the fibers begin to disappear near the nuclei, that is, shorten. 

 "All of the fibers that form cell-plate elements are completely 

 used up in the growth of the cell-plate." The swellings enlarge, 

 come in contact, and fuse. They do not split before fusing. The 

 central spindle grows peripherally, that is, centrifugally by the 

 addition of fibers. The cell-plate splits in the center and the new 

 wall is secreted (fig. 21). There are thus three processes going 

 on: cell-plate formation, plasma membrane formation, and wall 

 formation. In the larch they are all shown to take place centrif- 

 ugally, and Strasburger (67(2) figured the same condition in 

 Anthericum after the heterotypic division. Timberlake believes 

 that the wall formation may occur by the secretion of an unstain- 

 able solution, perhaps a carbohydrate, between the two plasma 

 membranes. The phenomenon of the separation of the two 

 plasma membranes is discussed in some detail. Of the process 

 he says: "It is hard to conceive of a layer of protoplasm becoming 

 differentiated into two separate layers similar in all apparent 

 respects to each other." 



Studies in physical chemistry since the date of Timberlake's 

 paper, especially in the behavior of colloids, should throw con- 

 siderable light upon these processes of cell-division. The well- 

 known fact of the crystalloid nature of cell-walls and starch grains 

 made it seem likely that the cell-plate is not of this nature, but 

 is more probably colloidal; and its being visible, both in living 

 and fixed material, would indicate that it is probably a suspensoid. 

 Though probably colloidal, it is not very different either physi- 

 cally or chemically from the rest of the cytoplasm, or a surface 

 boundary would form between them, such as delimits the nucleus; 

 in other words, the cell-plate differs from the plasma membranes 

 to which it gives rise, in that it is apparently permeable to sub- 

 stances in the cytoplasm indiscriminately. 



A number of interesting variations from the typical procedure 

 are noted by Timberlake in larch. "Whether the mother-cell 

 divides into four cells which form the pollen grains by successive 

 or simultaneous division depends upon the number of spindle- 

 fibers existing in connection with the first nuclear division. If 

 18 



