The Embryo-Sac of Angiosperms. 383 
the archesporial cell may without dividing form one megaspore 
mother-cell, it does not seem reasonable to suppose that the division 
of the mother-cell to four megaspores may not also he left out, and 
the mother-cell function directly as a megaspore. 1 ” The reduction- 
division would then he shifted so as to take place at the germination 
of this spore. 2 
2. On Pearson’s hypothesis, at the end of free nuclear divisions 
all the nuclei in the embryo-sac—whether one, two, or four 
spores germinate within it—are potential or reduced gametes. If 
this view is accepted, it will be evident that the embryo-sac is much 
more plastic in its ontogeny than was formerly supposed. The 
usual grouping of the nuclei at the two ends of the sac (seen in the 
eight-nucleate sac and also in Gunnerd) is a natural consequence 
of the elongated shape it assumes and the formation of a central 
vacuole. A striking parallel is seen in Hyncinthus oriental is, 3 where 
the microspores sometimes germinate like megaspores, with three 
successive mitoses, giving rise to four nuclei at each end of the 
pollen-grain, with a large central vacuole. Moreover, in various 
apogamous species, where the reduction-division does not occur, 
and the normal row of four megaspores may be suppressed, 
development of the embryo-sac follows the usual course. The fact 
that there is no grouping in Peperoniia, and that in the Penaeacese 
the sac is quadri-polar, may be a consequence of its spherical or 
nearly spherical outline in these forms. 4 
On the other hand it may he urged that the embryo-sac varies 
so very little in its ontogeny, and its contents are usually so 
markedly differentiated, that it is hard to see, in the large number 
of cases where a megaspore row is not formed, how four megaspores 
can have germinated within it, without the normal number and 
arrangement of the nuclei being disturbed. This is especially so if 
we regard the sixteen-nucleate sacs mentioned above as cases 
where this has happened, for in Pcperomia the nuclei are irregularly 
distributed, and in the Penaeaceae the sac is quadri-polar. It is difficult 
to see how a similar result has been avoided in cases such as that 
of the Liliaceae, where tetrad-formation may be partially or entirely 
suppressed within the limits of a single order without apparently 
influencing the development of the embryo-sac. 
3. There are cases in which any one of the megaspore row 
may function 5 ; cases in which more than one megaspore may 
develop 5 ; cases in which no walls are formed between the four 
megaspore nuclei, so that they are enclosed within a single cell 
though only one develops 5 ; and one case (Crucianella) 0 in which 
the megaspore nuclei thus enclosed in the mother-cell each divide 
once, the nuclei derived from the micropylar megaspore-nucleus 
then forming the embryo-sac, while the others degenerate. All 
these indicate the possibility of four megaspores becoming enclosed 
and germinating equally within the one mother-cell. 
Against this it may be pointed out that whenever a tetrad of 
1 McAllister, 1909, p. 212. Brown, 1908, p. 456. 
2 Strasburger, 1894 ; Farmer and Moore, 1907, p. 196 ; Ernst, 
1908 (B), pp. 26-29. 
s Nemec, ex Coulter & Chamberlain (03), p. 74. 
4 Brown, 1908. 
5 Coulter and Chamberlain, 1903, p. 76-87. 
c Lloyd, 1902, ex Coulter and Chamberlain, 1903, p. 86. 
