tort] BROWN & SHARP—EPIPACTIS 445 
and in others into those of only a portion of one. This may be due, 
however, to some condition such as nutrition, which is external to 
the megaspores, and is probably not due to potentialities inherent 
in the various megaspore nuclei, for it would seem that the nucleus 
of each megaspore, if placed under proper conditions, would have 
the potentialities for producing the nuclei of a complete sac. This 
conclusion is supported by the large number of cases in which the 
development of more than one megaspore in a tetrad has been 
described (Coutrer and CHAMBERLAIN 4). Differences in the 
potentialities of the megaspore nuclei, moreover, could not explain 
the differences in development, for the course can be predicted at 
metaphase of the reducing division. Different potentialities if 
they existed would, therefore, have to be in the nuclei of the differ- 
€nt megaspore mother cells; but according to present theories of 
heredity all mother-cell nuclei possess equal potentialities. The 
most reasonable conclusion would seem to be that the different 
courses of development are due to conditions external to the nuclei, 
and that the fate of a nucleus will depend on its position. It would 
seem probable, moreover, that the conditions which determine the 
fate of a nucleus, when four megaspores combine to form a normal 
Sac, must be the same as those which determine the fate of the 
nuclei of a sac formed from a single megaspore. The formation of 
a normal sac from four megaspores in Lilium (CouLTER and CHAM- 
BERLAIN 4), Smilacina (MCALLISTER 9), and also in the large 
number of cases in which a row of megaspores is not formed, as 
well as from the aposporous outgrowths into the cavity of the 
degenerated embryo sac of Hieracium (ROSENBERG 13), and in 
Alchemilla (MurBecxk 10) from the megaspore mother cell without 
a reduction in the number of chromosomes, would seem to in- 
dicate that the formation of a sac is not due to the nature of the 
cell from which it is produced, but that a normal sac will be formed 
from any cell subjected to the conditions under which a megaspore 
would produce one. The determining conditions in all of these 
cases, or at least most of them, are probably the same as in Epi- 
pactis, and since these conditions appear to be widely distributed 
among the angiosperms, they may have been the original cause of 
the evolution of the eight-nucleate sac. This could be true even 
