208 BOTANICAL GAZETTE [SEPTEMBER 
occurring in Crucianella by Lroyp (15). He reports that in three 
species of Crucianella investigated, four megaspores were formed 
which were not separated by cell walls, and that “only occasional 
exceptions could be found to this.” The upper of these four nuclei 
develops into the embryo sac, and the three lower finally degenerate. 
Not uncommonly these three lower nuclei undergo division. Many 
cases were seen, Lioyp reports, in which the four megaspores had 
each divided once, thus forming eight nuclei in a common cavity. 
He says: “If these divisions are regarded as the first mitoses of an 
embryo sac we have four embryo sacs lying tandem.” Only the 
one lying adjacent to the micropyle attains full development. Al- 
though he reports the disintegration of the lower megaspores, his 
figures admit other interpretations. As late as the third division of 
the functional megaspore, what he interprets to be one of the three 
lower megaspore nuclei is shown dividing in the same cavity with 
the four dividing nuclei of the embryo sac. The dividing nuclei 
seem to have the same appearance in every way. This would cer- 
tainly show a prolonged activity on the part of these three inner 
megaspores. This delayed germination of the lower megaspores and 
the similarity of the dividing nuclei in every way, open up many 
possibilities. The exact fate of the eight free nuclei formed by the 
division of all four of the megaspores is left unsettled. LLoyD com 
cludes that the outer megaspore attains full development, and that 
the eight nuclei therefore never organize to form an embryo sac jointly. 
This is a most interesting case, and a complete history of the embryo 
sac from the mother cell may throw light on the interpretation of 
multinucleate embryo sacs. 
This omission of the division walls between the megaspores of a 
tetrad is found by Luovp also in Asperula, which is related to Cru- 
cianella. Any one of four spores may germinate to form the a 
sac, and the/other three finally degenerate. The three spores wit 
do not germinate are at times difficult to distinguish from antipod 
nuclei. 
An important conclusion to be derived from the behavior : of 
Fichhornia, Avena, Crucianella, and Asperula is that the - 
walls are not essential to the individualization of a spore, and tha ; 
the failure to form division walls between the megaspores does N° 
> 
SESE EET SD ea eae 
