The Morphology of Ruppia Maritima. 141 



As to thie number of chromosomes distributed to the daughter- 

 nuclei in this division, it is impossible to state it exactly, after a care- 

 ful study. Murbeck, however, succeeded in finding a very clear 

 case where eight appeared in the nuclear plate, and judging from 

 the count in the two preceding mitoses, and also from general 

 appearances in the division, it appeared to me as if eight was the 

 most probable number in Ruppia maritima. The chromosomes in this 

 division, as in the much later one when two male cells are formed 

 (p. 142), are extremely small — much smaller than in the first and 

 second divisions in the pollen mother -cell^so that even with the 

 highest powers of the microscope they are very difficult to identify. 

 A few words should be added regarding the generative cell at 

 this period. As Murbeck has noted, its nucleus is smaller than 

 the tube-nucleus, but, judging from the staining reactions of the two, 

 it contains a comparatively larger amount of chromatin. Moreover, 

 a definite aggregation of cytoplasm surrounds the generative nucleus. 

 As Murbeck has announced in Ruppia rostellata, so here no membrane 

 separates this cell from the rest of the microspore, and it is evident 

 that one sees here a naked cell, such as occurs in the case 

 of the egg or synergidae of the embryo-sac. On the other hand, 

 Wiegand (1899, pp. 352, 353, figs. 41, 42) finds the generative 

 cell in Potamogeton — which, by the way, is not formed until the 

 " spores reach their full size " — enclosed in a definite membrane. 

 A distinct wall is also found in Typha by Schaffiier (1897. I.) and in 

 Naias and Sparganium by Campbell (1897, 1899). Rosenberg (1901. I.) 

 in Zostera mentions only a very indistinct wall at a later stage, 

 when the two male cells are formed, separating them from the 

 surrounding pollen -plasm. 



In most cases, though by no means always, a narrow space separ- 

 ates this cell from the rest of the microspore, which a study of a 

 large number of sections shows is undoubtedly due to a contraction. 

 (PI. XII, fig. 92). . 



The microspore now enters upon a long period of growth, during 

 which it increases in length from the 175—225 // of the above des- 

 cribed stages to the 500—550 /.i of the mature condition. During 

 this period it gradually assumes its curious bow-shaped form, with 

 the characteristic dumbbell-like expansions at the ends, and also a 

 considerable dilatation centrally at the convex side of the bow 

 (PI. XII, fig. 95). 



A pollen -grain so remarkable in shape is not unique among the 

 Potamogetonaceae; for the pollen-grain of Cymodocea is also un- 

 usual, being a long filamentous grain measuring, according to Bornet 



