362 BOTANICAL GAZETTE [june 



the mother cell of Lilium is said to ''function directly as a mega- 

 spore/' and the resulting eight-nucleate sac seems to be quite normal. 

 The inference is that the formation of megaspores has been omitted, 

 and that the nucleus of the mother cell holds the same relation to the 



^ 



eight nuclei of the embryo sac that the megaspore nucleus does in 

 ordinary cases. Analysis of the situation, however, shows that this 

 is not true, for megaspores, at least their nuclei, cannot be omitted. 

 The first two divisions from the mother cell are reduction divisions, 

 and therefore the first four nuclei of the embryo sac of Lilium are meg- 

 aspore nuclei, to be recognized as such by their cytological history and 

 structure. This means that the nuclei of four megaspores have 

 entered into the structure of the female gamctophyte of Lilium; and 

 that to complete its ante-fertilization development each of these 

 nuclei divides once. Therefore, in the case of Lilium, there is only 

 one division after the reduction divisions, instead of the usual three; 

 and the total number of divisions is reduced from the customary five 

 to three. In this case the history of an embryo sac containing the 

 usual structures shows a very unusual reduction in the number of 

 successive divisions, and this reduction seems to have been attained 

 by many more monocotyledons than dicotyledons. 



The case of Cypripedium' is an illustration of a similar reduction; 

 but with the old methods of interpretatiofi it would stand as remark- 

 ably different from LiHum and most angiosperms, for the completed 

 embryo sac, ready for fertihzation, contains only four nuclei. But 

 in this case the mother cell divides, and one of the daughter cells forms 

 the embr}'0 sac. The first division of the nucleus of this dai 

 is the second reduction division, resulting in two megaspore nuclei. 

 It is these two megaspore nuclei that are involved in the development 

 of the female gametophyte, each dividing once as in the case of LiHum. 

 Therefore, in Cypripedium also the usual five successive divisions 

 are reduced to three; the only difference being that in Lilium four 

 megaspore nuclei are involved, while in Cypripedium only two are 

 used. In her paper Miss Pace calls attention to this comparison, and 

 also to the fact that the eHmination of one more division would result 

 in an oogenesis similar to that of animals. 



I Pace, Lltlu, Fertilization in Cypripedium. BoT. Gazette 44:353-374- 

 pis. 2^2-/. 1907. 



