392 Mr. Griffith on the Impregnation o/'Dischidia. 



This is the first rudiment of the nucleus, or of the cavity within which the 

 future embryo is to be developed (Tab. XVIII. fig. 5). Just before the ex- 

 pansion of the corolla this mass becomes distinct, and frequently assumes a 

 rounded form. The globules I consider to be dislocated tissue. It consists 

 of an aggregation of brownish globules and smaller granules, which in fact 

 closely resemble the granules of the boyaux. The perfect ovule is compressed 

 laterally : the umbilical cord is attached close to its base. The upper margin 

 is convex ; the lower or dorsal generally straight, and often more or less 

 concave. The fissure is now very large, extending longitudinally from the 

 base of the umbilical cord for about one-third of the length of the convex upper 

 margin ; its lips or edges are gaping; it is of considerable depth. It appears 

 to narrow gradually towards its fundus. The grumous globular mass is now 

 very distinct, and the first indications of an excavation around this mass are 

 appreciable (Tab. XVIII. fig. 6). In flowers which have passed their meridian 

 (no impregnation however having taken place) the excavation is enlarged : 

 the grumous mass is now more irregular, and frequently appears to be broken 

 up, the component parts being then irregularly grouped together. 



In this genus the corolla is always closed to a considerable degree by the 

 connivence of its divisions. In addition to this, the divisions of the corolla in 

 D. Rafflesiana are furnished with short hairs, which project inwards. We 

 may therefore subtract the aid of foreign agency in determining the escape of 

 the poUinia from the anthers ; and we may likewise state, that when impreg- 

 nation takes place in a given flower, it is by the action of its proper pollinia. 



The pollinia are erect, oblong, somewhat compressed ; they have no dia- 

 phanous margin. The dehiscence takes place along the inner margin of each 

 mass, that is, as regards the cell of the anther in which it was previously 

 lodged. This corresponds with the margin of dehiscence of the pendulous 

 pollinia, first noticed by Mr. Brown. This line presents no appreciable struc- 

 tural difference, — nothing to indicate why the boyaux pass out here in pre- 

 ference to any other point. The fissui'es of communication are open in 

 D. Rafflesiana ; in the other species their edges are closely approximated as 

 far as their most prominent angles. 



The base of the stigma is stigmatic {i. e. papillose) in both species ; slightly 

 so in D. Rafflesiana, considerably more so in the other species. The cellular 



