1907] COKER— DEVELOPMENT OF SEED IN PONTEDERIACEAE 299 



wall, contains a large nucleus which has the same origin as in Ponte- 

 deria. This is also just what is found in the case of Saiirurus cernuus 

 (Johnson ii), but In none of these cases does the lower endosperm 

 nucleus divide.^ 



In Limnocharis, according to Hall (8), the lower of the two nuclei 

 formed by the division of the megaspore nucleus passes to the lower 

 end of the sac, but does not divide to form antipodals and a lower 

 polar nucleus. The upper polar nucleus, formed in the usual way, 

 sinks to the lower end of the sac, and there divides without fusion 

 (there being no lower polar). A cell wall resulting from this division 

 cuts the sac in two. The nucleus left In the upper cell divides to 

 form the endosperm, but the lower does not divide. 



In both Naias and Zannichellia, Campbell (3) has described 

 a large nucleus near the antipodal end of the sac, which increases 

 greatly In size, but does not divide. He did not determine the origin 

 of this nucleus, but offers two suggestions, either that the large lower 

 nucleus Is the lower polar, or that it is the product of the first division 

 of the definitive endosperm nucleus. The latter explanation seems 

 most probable. 



In Elodea canadensis ^ Wylle (15) finds that there is generally 

 present a large extra nucleus in the antipodal end of the sac. He 

 says: "The sudden appearance of the extra nucleus in the antipodal 

 group, when one has in mind the behavior of the polars, might suggest 

 that these nuclei do not always fuse, and that one of them passes 

 down to the lower end of the embryo sac and joins those in the antipo- 

 dal pouch. The general evidence, however, is against this view. 

 The irregular number of nuclei displayed here and the general arrange- 

 ment of the cytoplasm about them would indicate that any increase 

 in number has come from division among the antipodals, an activity 

 in this region that often results from fertilization. In all embryo 



sacs studied at earHer stages the lower polar had passed out of tho 

 tip, and its return to the antipodal group seems improbable.'' Is it 

 not more probable that this nucleus is formed as in Pontederia ? 



2 Cook mentions a single division occurring once in Nymphaea odorata. 



This discussion might be indefinitely extended by referring to the numerous cases 

 of chambered sacs and other peculiarities of the endosperm of dicotyledons, but 

 homologies are so improbable in such details that it is not necessary to cite them 

 further. A good exposition of this subject may be found in Coulter and Chamber- 

 lain's Morphology 0} the angiosperms. 



'-} 



1 



■1 



