June, 1905.] Embryo Sac and Embryo of Batrachium. 3.55 



The primary sporogenous cells continue to divide and appar- 

 ently give rise to the axial layer of tapetal cells. The origin of 

 the peripheral layer from the wall cells and the axial layer from 

 the sporogenous, seems to accord with their origin in R. delphi- 

 nifolius, both from examination of the slides and from the 

 observations recorded in Coulter's Life History of Ranunculus 

 (3). This refers the origin of the tapetum to the primary arche- 

 sporium instead of referring the axial layer to the inner tissue of 

 the androecium. Frequently a splitting was observed between 

 the sterile wall layers and the tapetum but quite as often it could 

 be seen between the tapetum and the sporogenous tissue (Fig. 9), 

 and sometimes seemed separated from both. As the stamen 

 matures the cells are forced past each other and misplaced, 

 making it extremely difhcult to determine the origin of the 

 tapettnn unless a careful study of a series of stages has been made. 



The primary sporogenous cells then divide a number of times 

 so that a central cross section shows sometimes as manv as 

 twelve microsporocytes (Fig. 8), while a longitudinal section 

 shows from three to four rows (Fig. 11). The tapetal layer does 

 not disintegrate early but is still quite well organized after the 

 separation of the tetrads. 



The microsporocyte divides to form four microspores (Fig. 

 12-13). No cases of more were found as has been reported in 

 Ficaria (4) and other Ranunculaceae but in some cases the sep- 

 aration is incomplete. This is shown in one of the pollen grains 

 in Fig. 25. In many cases the microspore never germinates (Fig. 

 14), in fact scarcely one to four. The tube nucleus and the gen- 

 erative nucleus lie close together. Just before pollination the 

 generative cell becomes lenticular and divides to form the sperm 

 nuclei (Fig. 15). These are not readily seen because of the 

 abundant starch granules, the deep color which the pollen grain 

 takes, and the crowding of the three nuclei. In the slides of R. 

 delphinifolius there were found similar cases of two male nuclei 

 before the germination of the pollen ttibe. 



Before the lamina has entirely enclosed the nucellus, the 

 archesporium can be distinguished (Fig. 16-17). The occurrence 

 of two or more archesporial cells is not at all unusual and in many 

 cases the struggle for supremacy results disastrously for all con- 

 cerned. The remains of other archesporial cells can almost alwavs 

 be seen around the megasporocyte. There is no evidence of the cut- 

 ting off of any primary parietal cell but the reduction division 

 occurs at once. The lower of the two cells divides first and in 

 many cases the division of the upper seemed never to pass bevond 

 the formation of the spindle (Fig. 18-19). This is not unlike the 

 development of the megaspores as reported by Mottier (8). In 

 a few cases there seemed to be two complete sets of megaspores 

 but the writer did not observe any twin embryo sacs though it is 



