110 BOTANICAL GAZETTE [ AUGUST 
tubes were found in the funiculus before the division that results 
in the four megaspores, and at various times tubes had passed | 
the chalaza while there was no sign of endosperm in the sacs. 
Usually when a tube gets close to the sac its penetration is 
rapid, and Treub says its growth within the nucellus is rapid. 
It seems, then, that there would not be time for a great develop- 
ment of endosperm between the time of the entry of the tube 
into the chalaza and its entry into the sac. 
3. Inferring the exact time of fertilization from the presence 
of pollen tubes is not conclusive. Anyone who has tried to 
distinguish a pollen tube in the confused strand of tubular antip- 
odal prolongations will readily understand the difficulty of 
determining without a doubt what is tube and what is sac. 
Even if tubes are seen, it is not an easy matter to tell whether 
they are young or old, and it must have been much more diffi- 
cult with teased preparations. 
4. Very little was said or known twelve years ago about the 
retarded division of the egg. Now it is known that the angio- 
sperm egg often rests for a time after its fertilization, while 
the formation of endosperm begins at once. It was the writer's 
good fortune to study such a phenomenon in the embryo sac 
of Asclepias (2), in which the exact time of fertilization was 
observed, and the egg rested after fertilization until the endosperm 
had passed its 16-celled stage. If the exact time of fertilization 
had not been seen, one would have been apt to judge the resting 
fertilized egg to be an unfertilized one, and would have been 
inclined to say that endosperm division took place before fertil- 
ization. Johnson’s study of the Piperaceae (3) has brought to 
light the same resting period in the egg of Piper medium. He 
figures a sac with twenty-two endosperm nuclei in cross-section, 
already walled off, and the egg still undivided. In the same 
paper he says of Heckeria umbellata that the embryo sac becomes 
filled with cellular endosperm before the egg divides. It is only 
fair to add, however, that fertilization was not observed in these 
species, but the appearance of the egg leaves little doubt of its 
fertilization long before it divides. 
5. The fact that Treub found a definite wall about the egg in 
