certain Penaeaceae. 
375 
it forms an interesting transition stage between that type and a possible 
ancestor possessing a larger number of free nuclei, sufficient to form several 
primary endosperm nuclei where here only one is present. On the other 
hand, if, as seems more probable, it is a derived form, the same hypothesis 
is equally helpful, especially in explaining how it is possible for a normal 
endosperm to arise from a four-nucleate instead of the usual bi-nucleate 
fusion. 
Summary. 
Species of Sarcocolla , Penaea , and Brachysiphon , three of the five 
genera of the rather isolated order Penaeaceae, have been examined. The 
development and structure of the embryo-sac and embryo in all three 
genera is the same. The megaspore mother-cell forms four nuclei by 
a reduction division. These are usually tetrahedrally arranged. They 
migrate to the periphery of the embryo-sac, and there as a rule lie cross- 
wise, one at each end and the other two at the sides ; their position, 
however, often varies. Each gives rise to a group of four nuclei. Cells are 
formed round three of these nuclei in each group, and these cells in each 
case assume more or less of the appearance of an egg-apparatus ; whether 
they can all function thus is uncertain. The four remaining free nuclei 
fuse in the centre of the sac to form the primary endosperm nucleus. After 
fertilization this forms a parietal layer of endosperm nuclei, cell-walls 
appearing only at a late stage. 
The embryo has no suspensor. It divides to form a sphere of cells, 
which elongates as the plerome is formed. It has a broad root-tip, no 
root-cap, a massive hypocotyl, and rudimentary cotyledons. 
This embryo-sac is probably to be regarded as a derived form, with all 
four megaspores included in its development, rather than as a primitive one ; 
but in our present state of knowledge of the embryo-sac of the Angio- 
sperms, this cannot be considered as definitely proved. It is suggested that 
in either case its endosperm is formed, like that of W elwitschia , by a fusion 
of potential gametes, all the nuclei in the sac being looked upon as 
potential or reduced gametes. 
This investigation was begun in the Botanical Laboratory of the 
South African College, at the suggestion of Professor Pearson, to whom 
I am indebted for much help and advice in its early stages, and for assist- 
ance in procuring material. I wish also to thank Professor Seward for 
kindly allowing me to continue it in the Cambridge Botany School. 
