528 MB. H. MARSHALL WAED OK 



In fig. 16 (Vesque) the two nuclei in the lower cell are pro- 

 bably from the first division of the enibryo-sac contents ; and here 

 he has figured the central row abutting on the embryo-sac (cf. 

 Plate XIX. figs. 1 and 2, &c). He also finds the perpendicular- 

 wall form in the uppermost of the cut-off cells (compare Vesque's 

 figs. 18-20 with our Plate XVIII. figs. 6-8, and Plate XIX. 

 fig. 1), again overlooking the presence of an axial row of cells 

 below the sac. The fate of this upper cell is described in accord- 

 ance with our views. 



But now appears a startling difference in Vesque's account. 

 He finds the lowermost of the cells resulting from division of the 

 embryo-sac mother cell now becoming divided once or twice by 

 cross walls ; but on referring to his figs. 19, 21, 23, and 25, it 

 appears probable that he here begins to recognize the presence 

 of the cells of the axial row, a series recognizable in median lon- 

 gitudinal sections throughout (Plate XVIII. fig. 7 and Plate XIX. 



In Plate XIX. I have carefully drawn the appearance pre- 

 sented by a section at two planes differing by the thickness of a 

 whole cell: the one (fig. 3) shows the embryo-sac and axial row; 

 the other (fig. 4) shows an apparent axial row of cells, as if one 

 cell had become cut up into three by transverse walls. Such ap- 

 pearances are very apt to mislead in the investigation, and are of 

 course due to the series of cells around the central row, one of 

 which, in the figure referred to, is radiating in a plane perpendi- 

 cular to that of the section. 



Before attempting any further explanation of what occurs in 

 JButomus, we may describe the facts so far obtained in other 

 ovules. 



Embryo-sac $c. o/Alisma plantago. (Plate XX.) 



The earlier stages in the development of the ovule and embryo- 

 sac offer nothing specially remarkable. So far as my observa- 

 tions go, the description given for JButomus covers the main 

 points ; and, overlooking some difierence in detail due to the rela- 

 tively smaller number and larger sizes of the cells, we may start 

 our description at the stage when the two integuments have al- 

 most ceased growing (Plate XX. fig. 1). 



The embryo-sac has driven up the two " cap-cells " in the form 

 of a hvaline cone to the anov nf Bia nnnoiin. . n ^A rt%a n^lls in the 



