I48 BOTANICAL GAZETTE [august 



doubt as to the individuality of the 4 cells, and that they are the 

 morphological equivalents of megaspores is beyond question. 



In Smilacina racemosa (19), although but the 2 outer reduction 

 nuclei enter into the structure of the embryo sac, they too are for a 

 time fully separated by a split cell plate. 



In Streptopus roseus, also, the cell plate which separates the 2 

 homoeotypic nuclei entering into the structure of the embryo sac 

 very often splits, thus causing a complete separation of these 2 

 nuclei. I have referred above to evidence which suggests that at 

 times this separation may be permanent. 



Xo other conclusion seems possible than that 4 fully separated 

 cells arising from a megaspore mother cell by the reduction divisions 

 must be megaspores. That the 2 cells which in Smilacina racemosa 

 and in Streptopus rosens fuse to form the 2-celled stage of the embryo 

 sac are also megaspores seems equally clear. 



The cell membranes separating the reduction nuclei in the mega- 

 sporangia of Smilacina sessifolia, S. amplexicaalis, and Maianthe- 

 mum canadense, although in my material showing no splitting into 

 two distinct cell walls, are nevertheless much more complete and 

 persistent than the evanescent cell plates such as have been reported 

 at times for the embryo sac of Lilium and in other genera. These 

 membranes extend out and join the lateral walls and all traces of a 

 central spindle disappear. It seems very possible that an examina- 

 tion of sufficient stages would show that these membranes also split 

 as is the case in Smilacina stellata. 



The transition from forms like the foregoing, in which the 

 division membranes persist for some time, to the lily type of 

 embryo sac is strongly suggested by those forms in which the mem- 

 branes are incomplete and evanescent, as is the case in Medeola 

 and Clinlonia. Additional evidence on this point has been con- 

 tributed by Brown and Sharp for Epipactis (5). According 

 to their account, although the embryo sac usually arises from 

 the inner of an axial row of 4 megaspores, it also frequently 

 arises from the megaspore mother cell, evanescent cell membranes 

 being formed, and it may possibly be formed from the 2 inner 

 megaspores. Evidence on this latter point, however, was not 

 conclusive. 



