On Some Stages in the Life History of Gnetum. 
259 
In these cases the number of sac-nuclei taking part in the endosperm- 
fusion varies from 0 to 14 (see Table I). Further, while the fusing-nuclei 
are always of the same generation* as the oosphere-nucleus, their number 
is in no way dependent upon the generation to which they belong, e.g. 
there are 
14 in the 7th and 5th generations (Cols, i and iii). 
2 in the 6th and 4th generations (Cols, ii, ix, x). 
14, 8, 7, 6,f 5,t 4, or 0 in the 5th generation (Cols, iii, iv, v, vi, viii). 
These facts establish at least a probability that the essential character 
which qualifies a sac-nucleus to participate in this fusion is its freedom. 
In all cases included in the table this is the only obvious character 
possessed by these nuclei which is not shown by others which do 
not join in the fusion. The inconstancy of the number of fusing-nuclei 
in certain cases — 4, 5, or 6 in SarcocoUa squamosaX — shows that, in 
these at least, the number is determined by circumstances, not by 
fixed characters inherent in the nuclei themselves. In Cypripedium a 
synergid§ which becomes free from the egg-apparatus and wanders down 
the sac, fuses with the single antipodal nucleus. It appears from the 
description that the synergrid-cell as a whole undergoes this change of 
position " as if it were being pushed from its actual position by the inrush 
of the contents of the (pollen) tube." It therefore enjoys the same 
freedom as a free nucleus. The opportunist character of its participation 
is suggested by the possibility that it sometimes fails. || In Went's 
Podostemaceae, the primary antipodal nucleus develops no further, and 
there is therefore only one polar nucleus; no fusion occurs and no endosperm 
is formed. In Pejjeromia hispidula, the sac contains only two cells, viz. 
one synergid and the oosphere ; all the other nuclei are free and these all 
fuse.^ Similar conditions and the same results are seen in Eujjhorhia 
virgata.** Feperomia pellucida with a two-celled egg-apparatus and six 
peripheral-cells in the antipodal region leaves only seven nuclei free 
for the endosperm fusion. ft The difference between this species and 
^ In Pandanus, in which the antipodal-nuclei are of a later generation than the 
oosphere and the upper polar, a varying number of antipodal-nuclei — one to six or 
more, usually two or three, fuse with the single upper polar-nucleus to form one, 
sometimes two primary endosperm-nuclei. Campbell, 1911. 
t Occasionally five or six sac-nuclei take part in the fusion in the Penaeaceae 
(Stephens, 1909, p. 367, figs. 20 b, 23). 
X Stephens, loc. cit. 
§ Pace, - 907, p. 359. 
II Loc. cit., p. 360. 
% Johnson, 1914, fig. 84. (The details of certain abnormal cases were not fully 
Interpreted, loc. cit., p. 369, fig. 87). 
** Vide Samuels, 1912, p. 104. 
ft Campbell. 
