I9i A SHARP— ORCHID EMBRYO SAC 381 



various species, while the ovules themselves are almost exactly 

 alike in structure , varying only in the matter of dimension. Thu 

 since the archesporial cell in the different species has the same 



general form and initiates a series of stages develonin^ under nrar- 



same 



expected. 



Whether a row of megaspores is produced or not seems, as 

 already pointed out, to be largely dependent upon the position of 



spindles 



he first two divisions. But megaspore mother 

 megaspore just before division are very much a 



in size, shape, and surroundings, and are acted upon by similar 

 external factors, so that whichever gives rise to the embryo sac 

 the same course is followed and the same end is reached. 



The 6-nucleate embryo sacs of Phajus, Corallorhiza, and 

 Broughtonia seems to show a tendency toward a further reduction 

 of the. vegetative portion of the gametophyte. 



In all of the species examined the endosperm nucleus, whether 

 arising from the fusion of two or more nuclei, disorganizes without 

 dividing, so that Calopogon palchellns (Pace 8), in which it may 

 give rise to as many as four free nuclei, remains as the only known 

 case where endosperm is developed in orchids. 



Summary and conclusions 



i 



. The archesporial cell in all of the species examined is hypo- 



becoming 



mother 



megaspore mo 



form 



two megaspores. The innermost megaspore gives rise to the 

 embryo sac. 



3- In Epidendrum variegatum and Bletia Shepherdii the mega- 

 spore mother cell often gives rise directly to the embryo sac; in 

 such cases four megaspores take part in the formation of the sac. 



dendrum variegatum, E. cochleatum, E. vcmicosum, 

 E- globosum, Coelogyne massangeana, and Pogonia macrophylla the 

 embryo sac is of the ordinary 8-nucleate type. In Bletia Shep- 



Ep 



