On Some Stages in the Life History of Gyiet^im. 
241 
proved to occur in G. Gnemon* (Plate XLVIII, fig. 21a, b). We liave^ 
found no indication that a primary uninucleate endosperm cell is ever 
formed except in this manner. These processes commence at the chalazal 
end of the sac and proceed gradually, though sometimes a little irregularly, 
upwards. Sometimes a compartment may be formed above a number of 
nuclei which are not yet enclosed (Plate XL VII, fig. 18), and frequently 
the nuclei in lower compartments may still remain free when fusion has 
occurred in those above them (Plate XLYII, fig. 19). If the enclosed 
nuclei do not fuse — as occurs occasionally when an unduly small number is 
enclosed — the compartment increases in size f and its protoplasm becomes 
highly vacuolated (Plate XLYII, fig. 19 e). This is also true of similar 
compartments in the endosperm of Welwitschia.+ Nuclear fusion in an 
endosperm compartment is frequently followed immediately by cell-division 
(Plate XLYII, fig. 19 d) ; whether occurring immediately or not, the cells 
resulting from the division are smaller than those which gave rise to 
them.§ 
It should be emphasised that, with the exception of such cases as that 
shown (Plate XLYII, fig. 19), in which the enclosed nuclei are always few 
in number, every cell of the primary endosperm is uninucleate, the nucleus 
being formed by gamete-fusion. When this nucleus divides a persistent 
cell-plate appears and uninucleate cells result ; so far as is known this also 
is always true for the primary cells. In later stages, however, the endo- 
sperm cells may become multinucleate (Plate XLYIII, fig. 20), and 
frequently in the seed the endosperm appears to be mainly composed of 
multinucleate cells. There are some indications that this condition results 
from nuclear fragmentation ; the appearance and reactions of the free 
nuclei are always consonant with the view that they are in process of 
degeneration. This multinucleate condition of the old endosperm cells of 
G-netum is probably to be ascribed to the causes which produce similar 
results in the prothalli of Ephedra, || Taxus,^ and other of the lower 
gymnosperms. 
While the origin of the endosperm of the lower half of the sac is now 
known, that of the micropylar region is still in doubt. Lotsy ** describes 
it (in G. Gnemon) in the following terms : About the time of the 
formation of zygotes it is seen that a greater or smaller number of the 
* Cf. Pearson, loc. ext., p. 321. 
t This is probably the explanation of the curious cell figured by Lotsy (1899 
fig. 62), and described by him as an " Archegonium-like formation." 
X Pearson, 1909, p. 353, fig. 38 b. 
§ Pearson, loc. ext., p. 352. Cf. also Pearson, 1915 a, fig. 23. 
II Strasburger, 1880, pp. 106, 109 ; Jaccard, 1894, p. 34. 
IF Jaeger, 1899. 
** Lotsy, 1899, p. 97. 
