1907] PACE—FERTILIZATION IN CYPRIPEDIUM 363 



Lilium condition); Campbell (7) two or three in Dicffenbachia ; 

 and JuEL (29) two in Taraxacum. The very few cases reported 

 show the rarity of this condition among the angiosperms investigated. 

 Four of the five families represented in this enumeration, to which 

 the Orchidaceae are now added, are all monocotyledons; and this 

 fact, taken together with the still greater reduction in the Lilium 

 type, is in harmony with the theory that monocotyledons are a 

 specialized offshoot from dicotyledons, as held by Jeffrey (28), 

 QuEVA (37), Miss Sargant (38), and others. 



However, ScHNIE^VIND-THIES (40) has shown that in Lilium the 

 heterotypic and homotypic divisions take place within the embryo 

 sac; that is, there is no permanent wall formed in either of the divi- 

 sions from mother cell to the four megaspore nuclei. It follows, there- 

 fore, that all four megaspore nuclei take part in forming such an 

 embryo sac. In Cypripedium, on the other hand, a wall accompanies 

 the first division of the mother cell, but no wall is formed in connec- 

 tion with the second division. In this case, therefore, two megaspore 

 nuclei enter into the formation of the embryo sac, instead of four as 

 in Lilium, or one as in most angiosperms. It is interesting to note 

 here that if the usual megaspore walls were formed in Lilium and in 

 Cypripedium, each sac would contain two nuclei — the egg and its 

 sister, which is the polar that usually unites with the antipodal polar 

 to form the primary endosperm nucleus. 



Embryo sac. — An embryo sac with only four nuclei has not been 

 reported before, but there is just the same number of divisions between 

 the mother cell and the egg in this case as in Lilium. After all, the 

 essential feature in the history of the egg is the distribution of the 

 chromosomes in the reduction di\isions. It is possible that the use 

 of more than one megaspore in organizing the sac holds some neces- 

 sary relation to "double fertihzation," for in both Lilium and Cypri- 

 pedium, if walls appeared in the usual way so that only one megaspore 

 was used in forming the sac, and there were no additional di\asions 

 between mother cell and egg, there would be only two nuclei in the 

 sac, and "double fertilization," at least "triple fusion," would be 

 impossible. 



It is interesting in this connection to compare this series with 

 that in the animal egg. There the primary oocyte (mother cell) 



