316 BOTANICAL GAZETTE [ocTOBER 
features of this small shrubby group, restricted to the southwestern region of Cape 
Colony, are fairly before us. Three of the five genera were investigated (Sarco- 
colla, Penaea, and Brachysiphon), suitable material of the other two (Endonema 
and Glischrocolla) not being available. _ 
{he morphological characters of the three genera examined are the same, 
so that one account can serve for all: The megaspore mother cell produces four 
nuclei, usually tetrahedrally arranged, and these migrate to the periphery of the 
mbryo sac, where each gives rise to a group of four nuclei. Three of the nuclei 
of each group are organized into cells which resemble an egg-apparatus, while 
the four remaining free nuclei fuse in the center of the sac to form the primary 
endosperm nucleus, which after fertilization forms a parietal layer of nuclei, walls 
appearing much later. The embryo has no suspensor, appearing first as a spher- — 
ical mass of cells, which elongates as the tissues are differentiated and the growing 
points are organized. 
This seems clearly an illustration of the formation of an embryo sac by the 
cooperation of four megaspores, in this case the product of each megaspore 
remaining remarkably distinct.—J. M. C. 
Embryo sac of Pandanus.—A preliminary note?? under this title has already 
been referred to in this journal.?3 The fuller account, with plates, has now been 
published.2+ Pandanus has long been regarded as a promising primitive mono- 
cotyledon, and its investigation is most timely. The general results are as follows: 
the archesporial cell (presumably solitary) cuts off a parietal cell which gives rise 
to several layers of cells separating the epidermis from the megaspore mother cell; 
the mother cell divides transversely into two daughter cells, the inner one of 
which directly produces the embryo sac, while the outer one divides anticlinally; 
¢ first division within the sac (the second reduction division) results in two polar 
nuclei; the micropylar nucleus divides, and there is no division of the daughter 
nuclei, nor is there usually any differentiation into egg and synergid; the antipodal 
nucleus gives rise to twelve nuclei, whether by simultaneous division or not was 
not determined; in the most advanced stages secured no nuclear fusion was 
observed, all fourteen nuclei remaining quite separate. 
The author still maintains that the embryo sac of Pandanus is a more ancient 
type than the ordinary eight-nucleate sac of angiosperms, and that it represents @ 
new type, “‘with its nearest analogue in Peperomia.”? It remains to investigate the 
fertilization stages of this interesting embryo sac, to determine whether the four- 
teen-nucleate condition really is the fertilization stage.—J. M. C. 
22 CAMPBELL, D. H., The embryo sac of Pandanus. Preliminary note. Annals 
of Botany 22:330. 1908. 
23 Bot. GAZETTE 45:364. 1908. 
*4 CAMPBELL, D. H., The embryo sac of Pandanus. Bull. Torr. Bot. Club 
36: 205-220. pls. 16, 17. 1909. 
