June 1891.] BOTANICAL SOCIETY OF EDINBURGH. 141 



We have next to consider the significance of this occa- 

 sional partition separating the embryo-sac into a micropyLar 

 and an antipodal half. Does it correspond to the first wall 

 of a developing prothallus, as generally believed, or to a wall 

 separating two spores, each of which will develop into a pro- 

 thallus (Ward's view), or are we dealing with a special 

 mother-cell dividing into two sporocytes (Warming- Vesque's 

 view)? 1 am inclined to hold the latter view, for the divi- 

 sion of the embryo-sac makes the impression, as Ward points 

 out distinctly, as being similar in nature to those preceding 

 it ; we know further that the walls of spore-mother-cells 

 break down whenever spores are reaching their maturity, 

 that they are, in other words, very transient ; in the embryo- 

 sac this process of dissolution of the wall seems to take place 

 even before spore-formation, or rather no definite wall can 

 be laid down, partly due to the vacuole and partly due to the 

 rapid increase in breadth of the embryo-sac taking place at 

 this time. If this wall is really a wall separating two 

 sporocytes, then the sporocytes ought to give rise to four 

 spores each, which they do — namely, the eight structures 

 within the sac. 



That we are really dealing with eight structures corre- 

 sponding to eight modified spores, becomes evident on studying 

 their development. For 1 find that after division of the 

 primary nucleus into a micropylar and an antipodal one, the 

 latter divides into two nuclei, one of which, the one next the 

 plerome-elements, precedes its sister-cell a short time in 

 division ; the two nuclei resulting thus begin to act at once 

 on the surrounding protoplasm, and to shut themselves off 

 from the central part of the embryo-sac by delicate walls. 

 Before the completion of this process, however, the remaining 

 nucleus lying usually to one side of the two nuclei abutting 

 on the plerome-elements, also divides, its antipodal nucleus 

 comes in close contact with the two other antipodal nuclei 

 and the embryo-sac-wall, and now the process of cell-wall 

 formation becomes very evident ; three nuclei shut themselves 

 off from the rest of the embryo-sac by highly refractive 

 walls, being fixed by their base to the wall of the embryo-sac, 

 while their free convex ends project into the cavity of the 

 embryo-sac. We find, therefore, if the occasional appearance 

 of a cell- wall separating the embryo- sac into a micropylar 



