108 BOTANICAL GAZETTE [ FEBRUARY 
and is frequently to be found crowded close into the angle 
between the egg and the synergids, as in fig. 38. It is much 
larger than the egg nucleus, but in other respects resembles 
it quite closely. Its nucleolus usually contains one large and 
several smaller vesicles. The body of the sac is occupied by a 
large vacuole. 
The antipodal cells are generally three in number, separated 
by walls and arranged in a row, the innermost being much the 
largest. The Compositae have long been known to exhibit great 
variability in the antipodal region, a multiplication of cells and 
nuclei being reported in Senecio aureus by Mottier (20), in Aster 
and Solidago by Martin (19), and in Aster Nove-Anghe by 
Chamberlain (23). The same variability has been seen, to a cer- 
tain extent, in Silphium, in all the species studied. The wall 
may be lacking between two of the three nuclei, giving but two 
cells (fig. 39), either or both of which may show strong indica- 
tions of direct nuclear division. Again, the nucleus may divide 
in any one of the three cells (figs. 40,41). Finally,in one case, 
seven antipodal cells were found, containing eight nuclei, with 
indications of amitotic division (jig. 42). 
In a recent paper by Mlle. Goldflus (26), already referred to, 
it has been shown that the antipodal cells in the Compositae 
possess at times a digestive function, and in other forms serve to 
conduct food from the ovule to the growing sac with its embryo. 
All the evidence goes to show that in Silphium these cells alt 
conductive, rather than digestive. It is equally clear that 1” 
those forms where the antipodals burrow back into the chalazal 
region, a fact which may perhaps be correlated with the reduced 
size of the nucellus, the surrounding cells must be destroyed by 
the action of a ferment secreted by the cells of the encroaching 
tissue. The possession of either of these functions is claimed by 
the author to be inconsistent with the interpretation of the antip- 
odal cells as a vegetative or prothallial region of the gameto- 
phyte. But it must be admitted that the most natural function 
of the vegetative, as distinguished from the reproductive, region 
of a prothallus, is that of nutrition, of which digestion and 
