244 University of California Publications in Agricultural Sciences [Vol. 2 



The two nuclear divisions precede cell division as in many of the 

 higher plants. It will be seen from the account which follows that the 

 reduction division is typical in all respects. 



Before the first, or so-called heterotypic division occurs, the plastids 

 lie between the wall and the nucleus (figs. 1 and 2). As the prophase 

 progresses they enlarge, and some of them divide. The division is 

 usually a simple bipartition, but some appearances suggestive of bud- 

 ding were seen. The cell wall is very thin at first but gradually 

 thickens during the first division. The plastids remain between the 

 nucleus and the wall until after the first spindle has disappeared, when 

 they gradually move, or more probably are moved, between the two 

 groups of late anaphase chromosomes, there forming a ring which 

 nearly fills the space between the two nuclei (fig. 4). This position 

 is retained until after the second division. In the late second ana- or 

 early telophase a portion of the ring of plastids is drawn between each 

 of the two sets of daughter nuclei (fig. 7). The ring then breaks in 

 four places, so that one quarter of it, and consequently about a fourth 

 of the plastids, come to surround the inner face of each nucleus (figs. 

 7 and 8). The outer wall now pushes inward between each of the four 

 nuclei and finally separates the pollen cells (fig. 9) . Before the inpush- 

 ing of the outer wall two cell plates form at right angles to each other. 

 The method of cell division in Ginkgo, then, is a combination of cell 

 plate formation and cytokinesis. The inner portion of the wall (which 

 lies next to the plastid-bearing area) thickens greatly while the outer 

 wall remains unchanged. As the inner wall attains its final thickness, 

 the plastids become generally distributed between it and the nucleus, 

 and a new wall, staining in gentian-violet, appears about each pollen 

 cell. The old wall stains in orange or in light green. As the pollen 

 cells grow they burst the thin outer walls of this case, leaving an empty 

 shell. The plastids now appear smaller, since the amount of starch in 

 them is considerably reduced. 



Juranyi (1872, pi. 31, fig. 13) figured a light band between the two 

 daughter nuclei of the pollen mother cells of Ceratozamia, which is 

 very like the position of the starch grains of Ginkgo at this stage. 

 Sprecher (1907, p. 155) figured the cell plate formation of the pollen 

 mother cells of Ginkgo, but neither figures nor mentions plastid or 

 starch grain distribution. Moore (1903) figured plastids like those of 

 Ginkgo in the pollen mother cells of PaJIavincinia. but does not discuss 

 their distribution. Smith (1907) shows starch grains in the pollen 



