th] LIFE HISTORY OF THE PONTEDERIACEA 329 
hese cells there is no appearance of crowding; their walls break 
in and their protoplasm becomes continuous with that of 
te megaspore ( fig. 24). Before the complete absorption of 
te tapetal cells the nucleus of the megaspore, situated about 
tiemiddle of the cell, divides into two (jig. 24). The lower 
nif of the megaspore at this time ceases to grow, while the 
per half widens out and assumes the shape shown in jigs. 
5-27. 
Thedivision into micropylar and antipodal groups of four nuclei 
-bllows in the usual way without any regular time order. Fig. 25 
Sows an embryo sac with four nuclei, fig. 26 with eight nuclei, 
i. 27 fusion of the polar nuclei. The position of the definitive 
ucleus is always near the place where the embryo sac con- 
‘nets into the narrow basal portion. This nucleus is large, has 
_*very large, clear nucleolus, and is connected with the wall of 
te sac by strands of protoplasm. The fusion of the polar 
wuclei certainly occurs before pollination, for I have found the 
, fsion accomplished in a flower whose perianth was not yet 
| Mcosed. That this act is not dependent upon the entrance of 
| aN tube is proved very conclusively in Caltha palustris, 
€ 
| ave observed sterile ovules through whose thickened 
3 cae Pollen tube can pierce, yet whose embryo sacs always 
| = © seven-nuclear stage. In Eichhornia, also, the embryo 
Ways reaches the mature condition, in which, as we have 
ce t of the pollen tubes break down without penetrating 
: Micropyle. 
, fog a tibodals of Eichhornia are evanescent, staining feebly 
et di . frst, and rarely with definite walls. Their nuclei do 
CS of either directly or by mitosis. , 
ei “pparatus has the usual arrangement. The oat 
Provided with striated “ filiform” tips and a prominent 
>... he egg is relatively small and appears to be stung 
- bat om aga a pollen tube reaching to the embryo = " 
“Nelens . and that is represented in fig. 29- The see 
e € tip of the tube is probably one of the male nuciel, 
