io Campbell.—Studies on the Araceae. 
the neighbouring sections the synergidae could be distin¬ 
guished (Fig. 23). The endosperm-nucleus could not be 
found, nor was there any trace of a third antipodal cell. 
Whether this peculiar condition of things is normal in 
Aglaonema remains to be seen, as no other specimens were 
found in which the egg-apparatus or early stages of the anti¬ 
podal cells could be made out. 
The Endosperm . 
Very soon after fertilization has been effected the primary 
endosperm-nucleus divides. Whether the first division is 
accompanied by the formation of a division-wall could not 
be determined ; but it is not impossible that such is the case, 
as Hofmeister described such a division for Pothos longifolia. 
While the number is still very small (Fig. 26), walls have been 
formed between the large nuclei, and the embryo-sac is thus 
filled with the endosperm, almost from the beginning. This 
early filling up of the embryo-sac is very characteristic of the 
Araceae, and Hofmeister called attention to it in a number 
of genera, although his work seems to have been largely 
neglected in recent studies on the embryo-sac. There is not 
found in these plants—at least in the forms hitherto examined— 
the development of a peripheral protoplasmic layer in which 
are embedded the free nuclei. 
The endosperm-nuclei are very large and stain readily. 
They possess one large nucleolus (Fig. 25) or sometimes two 
or more. They are separated by delicate cell-walls, which 
are, however, perfectly apparent. The endosperm-cells are 
usually largest at the apex of the embryo-sac, and smallest at 
the antipodal end. They contain little granular contents. 
The embryo-sac is at first broadly oval in outline (Figs. 22, 
26), but later elongates a good deal with the increase in size 
of the ovule, and there is a marked bend produced at the base 
of the sac, by its extension on the side next the funiculus. 
This bending of the embryo-sac is a very constant feature in 
the species under consideration (Figs. 27, 31). 
