138 BOTANICAL GAZETTE [august 



The publication of these results has been delayed from time to 

 time, partly owing to the press of other work and partly through 

 the hope of including other members of the group in this report. 



I have reviewed in another connection (18) the scanty litera- 

 ture on the development of the embryo sac among members of 

 the Convallariaceae. Two accounts of the development of the 

 embryo sac of Convallaria majalis differ radically as to the fate of 

 the first four nuclei resulting from the division of the nucleus of the 

 mother cell. According to Wiegand's account (24), all 4 of these 

 nuclei enter into the structure of the mature embryo sac, while 

 Schniewind-Thies (21) reports that but one of 4 megaspores enters 

 into the structure of the embryo sac. 



Ernst (13) reports that in Trillium gr audi fl or tint the inner 

 of the 2 cells resulting from the heterotypic division develops into 

 the embryo sac, while Chamberlain (9) finds that in Trillium 

 recurvatum the inner of 4 megaspores develops into the embryo sac. 

 Ernst (13) also found that in Paris quadrifolia the inner of the 

 first 2 cells arising from the first division of the mother cell formed 

 the mature embryo sac by three divisions of its nucleus, as was the 

 case in Trillium grandiflorum. 



In the paper referred to above (18) I have shown that in Smila- 

 cina stellata the megaspore mother cell divides to form 4 fully 

 separated megaspores, and that the separating membranes later 

 disappear, forming thus a tetranucleate cell, from which, by one 

 division of its 4 nuclei, the mature embryo sac is formed. 



Since the publication of the above-mentioned paper, one other 

 member of the family has been reported upon. According to 

 Smith (23), the reduction divisions of the megaspore mother cell 

 of Clintonia borealis result in the formation of a tetranucleate cell. 

 the outermost nucleus of which is plump and normal, while the 

 remaining 3 inner nuclei are shrunken and disorganized. The 

 outermost nucleus divides twice, and 3 of the resulting 4 nuclei 

 organize an apparently normal egg apparatus, while the remaining 

 nucleus serves as the single polar nucleus. I have been able to 

 confirm Smith's results from material collected in the vicinity of 

 Iron Mountain. Michigan, a few sections of which I had already 

 prepared at the time of the publication of his results. 



