19 14] SHA RP— SPERM A TOGENESIS IN MA RSIL fA 42 1 



Material and methods 



Sporocarps of Marsilia quadrifolia were cut open and placed 

 in water at room temperature. Sori were fixed at short intervals 

 until sperms were seen swimming in large numbers, a period of 10 

 or 12 hours. 



Several fixing reagents were employed, including those of Flem- 

 ming and Benda. The best results were obtained in preparations 

 stained with Haidenhain's iron alum-hematoxylin after the fol- 

 lowing fixing fluid: i per cent chromic acid 25 cc, water 75 cc, 

 glacial acetic acid 1 cc, 2 per cent osmic acid 14 drops. 



Description 



THE MALE GAMETOPHYTE 



A comparison of figs. 1-7 with fig. 8 will do more than a written 

 description to make clear the development of the male gametophyte. 



At the time when the gelatinous ring bearing the sori escapes 

 from the sporocarp, the microspore has in section the appearance 

 shown in fig. 1. The nucleus occupies a central position, and 

 large starch grains lie scattered throughout the cytoplasm. These 

 very soon move to the periphery, leaving the nucleus surrounded by 

 a zone of granular cytoplasm. The nucleus then passes to the side 

 of the spore, usually the side opposite the point where the spore 

 met the others of the tetrad, and cuts off a small prothallial cell 

 (wall 1, figs. 2, 3). No " basal cell," such as Belajeff (2) figures 

 above the prothallial cell in M. quadrifolia, was found in our 

 material. 



The next wall (wall 2, fig. 4) passes through the center of the 

 spore. The two hemispherical cells so formed behave alike, each 

 producing an antheridium in the following manner. A curved wall 



(wall 3, fij 

 formed in 



e> 



small 



S 



com 



theridium and defines the limits 



nous cell. There are no centrosomes in any of these mitoses. 



Each primary spermatogenous cell by four successive divisions 

 gives rise to a group of 16 spermatids (figs. 9-12). Since the two 

 spermatogenous masses formed in the two hemispheres are entirely 



