Primitive A ngiosperms . 1 3 3 
period. The fact that an empty cell is commonly found in the neighbour- 
hood of that in which such mitosis is taking place confirms the conclusion 
naturally drawn from the number of chromosomes that the dividing nucleus 
is a fusion product. Apogamous proembryos are probably formed from the 
daughter nuclei of such mitoses. 
The apogamous proembryos may be formed in situ ; that is, within the 
jacket-cells to which their primary nuclei belong. In such cases their 
development rarely proceeds far. But nuclei from the jacket-cells some- 
times penetrate the wall which divides them from the central cell of the 
archegonium, and there is good reason to think that the proembryos found 
within it are produced by the activity of such escaped nuclei. Such pro- 
embryos often develop suspensors, which may even bear rudimentary em- 
bryos at their tips ( 11 , p. 282). Since they are always formed after fertiliza- 
tion, their production may depend on some stimulus transmitted from the 
pollen-tube. Miss Berridge even thinks it possible that an escaped jacket- 
nucleus may sometimes fuse with a nucleus from the pollen-tube (ll,p. 283). 
In Gnetum the archegonium has disappeared altogether. Karsten ( 50 ) 
and Lotsy ( 55 ) have shown that a number of equivalent nuclei lie near the 
apex of the embryo-sac. Each of these seems to be a female nucleus, and in 
Gn . Gnemon at any rate each pollen-tube which discharges its contents into 
the embryo-sac fertilizes two of them (Lotsy ( 55 ), p. 96 and Fig. 45). Before 
fertilization there is no appearance of cell-formation round these nuclei. 
The prothallus fills the lower part of the embryo-sac. In Gn. Gnemon 
it is practically complete before fertilization (Lotsy) ; but in the species de- 
scribed by Karsten its development is arrested at an early stage, and is not 
resumed until fertilization of one or more of the apical nuclei has taken place. 
Details of the fertilization of Welwitschia are still lacking, but Pearson 
( 64 ) has described the development of the young embryo-sac very fully, and 
it is clear that cell-formation is complete in every part of it some time before 
fertilization. Each archegonium must therefore be represented by one cell 
at least, not, as in Gnetum , by a single nucleus. As in Gnetum , the apical 
end of the embryo-sac is clearly differentiated from the lower part. 
Miss Berridge suggests that the endosperm in Angiosperms may be 
comparable to the apogamous- proembryos produced within the arche- 
gonium of Ephedra distachya by division of nuclei escaped from jacket-cells. 
Her interpretation of the Angiospermous embryo-sac is best given in her 
own words : — 
‘ Here therefore we have a process occurring in the embryo-sac of a 
Gymnospermic genus, which shows a remarkable likeness to the develop- 
ment of the Angiospermic endosperm after triple fusion ; that is to say, we 
have cell-formation resulting from the fusion of nuclei, one of which at least 
is allied to the egg-nucleus, and taking place under the stimulus due to the 
entry of the pollen-tube. 
