308 BOTANICAL GAZETTE [October 



the cell in synapsis is somewhat shorter and broader than the other. 

 The walls of one cell are dotted in, but the walls of the other and 

 both nuclei are drawn. 



The mother cell divides in the usual manner (figs. 4, 8, 9). A 

 full series was studied, but only a few drawings will be shown. In 

 fig. 5 the synapsis stage is shown and in fig. 8 the spirem. The 

 chromosomes are quite short and thick in fig. 9, and the haploid 

 number (10) can be counted. The daughter cells sometimes differ 

 slightly in size, but as a rule the difference is not marked (fig. 10). 

 In fig. 11 the lower daughter cell has the spindle already formed for 

 the second division, while the upper daughter cell has only formed 

 the chromosomes, the nuclear membrane being still complete and 

 very distinct. But in fig. 12 both daughter nuclei are in the early 

 telophase of this division. In the upper cell one chromosome did 

 not reach the pole and was left out of the megaspore nucleus. 

 This condition was seen only a few times; and, as the upper mega- 

 spore always disintegrates in the material studied, it does not seem 

 to be of any importance in the life-history of this plant. One 

 example of the same condition was found in the first division. Here 

 it might affect the life-history; for very often the second megaspore 

 develops. If the nuclei continued to divide as usual, this might give 

 an egg with one less than the usual number of chromosomes, in this 

 case 9 instead of 10. 



Figs. 13-15 show the different positions of the megaspores, a 

 straight row in fig. 13, the two lower in a row and the two upper 

 side by side in fig. 14, and approaching the tetrad form in fig. 15- 

 Chodat (3) gives in his fig. 661 a straight row of four megaspores 

 and in fig. 662 a row of three cells; in both cases the lower mega- 

 spore has enlarged to produce the embryo sac. 



Embryo sac. — While in angiosperms it is probably true that the 

 lowest of the four megaspores usually produces the embryo sac, the 

 others disintegrating, cases showing that any of the four may 

 function have been reported, and in some instances all four show 

 sac tendencies. Coulter and Chamberlain (41 p. 84) give a 

 summary of the literature on this subject. In Parnassia apparently 

 the second or the third as frequently develops as the fourth (figs- 

 16-29), but no case was found in which the first developed. These 



