366 BOTANICAL GAZETTE [june 



that four megaspore nuclei are concerned, and that there is no "row" 

 developed by the mother cell at all. It is a very interesting per- 

 formance as described; for the original nucleus of the sac produces 

 four nuclei in tetrahedral arrangement; these four separate and 

 become parietally placed (one micropylar, one antipodal, two equa- 

 torial) ; each of the four produces a group of four nuclei ; three 

 nuclei of each group organize after the fashion of an egg apparatus; 

 and the four free nuclei pass to the center of the sac and fuse to 

 form the primary endosperm nucleus. We have here at least the 

 assurance of simultaneous successive divisions, which we do not 

 have in the case of all 16-nucleate sacs. If four megaspore nuclei 

 are involved in this organization, the reduction in the number of 

 successive divisions is just that which probably occurs in Peperomia. 



Some evident conclusions may be drawn from the above statement, 

 and they should be of use in investigations of the angiosperm embryo 

 sac. The nuclear divisions within the embryo sac vary in character; 

 they may include both reduction divisions, the second one, or neither 

 of them (the usual case). The appearance of more than eight 

 nuclei in an embryo sac is no evidence that there has been any more 

 than the usual number of successive divisions. As many as thirty- 

 two nuclei may appear in an embryo sac without increasing the 

 usual number of divisions. It is absolutely necessary to trace the 

 nuclear succession from mother cell to completed embryo sac before 

 any safe conclusions can be drawn as to the significance of the con- 

 ditions observed within the sac. The cases of an increased number 

 of nuclei, as in Peperomia, which have sometimes been cited as 

 showing a primitive condition of the sac, have actually turned out 

 to be cases of reduction from the ordinary condition. There is a 

 tendency among many groups to eliminate the divisions that follow 

 the two reduction divisions, but the tendency is not general among 

 angiosperms, since among the Sympetalae it does not seem to exist; 

 hence the forms which exhibit it may be regarded as more or less 

 specialized in this feature. 



If an angiosperm embryo sac should be found containing sixteen 

 or more nuclei derived from a single megaspore nucleus, it may be 

 regarded as relatively primitive; but such a sac has not yet been 

 demonstrated. 



The University of Chicago 



