326 BOTANICAL GAZETTE [October 



Fig. 13. — Four megaspores, apparently all increased in size but the first 

 disintegrating, and the second and fourth larger than the third. 



Fig. 14. — A different arrangement of the four megaspores, probably 

 occurring more often than that shown in fig. 13. 



Fig. 15. — Approaches still more nearly the usual tetrad arrangement. 



Figs. 16-22. — The variations in the early stages of the megaspores. 



Fig. 2^. — Apparently five megaspores; one nucleus is very small and seems 

 to have only one chromosome, probably resulting from an abnormal division 

 like that in fig. 12. 



Fig. 24. — Third and fourth megaspores developing; the nucellar cells 

 surrounding these are already showing signs of disintegration, and only the 

 three at the base are still normal in appearance. 



Fig. 25. — In this case also the epidermal cells are becoming pale, especially 

 over the upper vigorous megaspore, which appears to be more active than 

 the fourth. 



Fig. 26. — The third megaspore forming the embryo sac and in mitosis for 

 the first division; the epidermal layer partly disorganized, leaving the sac in 

 contact with the inner integument and an open micropyle. 



Fig. 27. — Two-celled sac formed from second megaspore; the third 

 megaspore also developing; the epidermal layer still further disorganized. 



Fig. 28. — Embryo sac formed from the fourth megaspore; first division 

 showing 10 chromosomes, the other megaspores represented by a formless mass 

 above; epidermal layer still perfect; integuments well advanced. 



Fig. 29.— All the epidermal layer, especially the two upper cells of the 

 nucellus, lighter than the adjacent integument cells; two-celled sac from the 

 fourth megaspore; the second and third megaspores have persisted longer 

 than usual. 



Fig. 30. — The entire ovule with two-celled embryo sac; ovule with loose 

 spongy tissue at base; nucellus entirely lacking over greater part of sac. 



Fig. 31. — Two-celled sac with nuclei in mitosis for the second division; 

 the 10 chromosomes may be seen in the upper nucleus; the upper half of the 

 sac is in contact with the integument, the nucellus having entirely disappeared 

 from this region. 



Fig. 32. — Similar to fig. 31, but the lower nucleus somewhat in advance 01 

 the upper; in the lower the chromosomes are short and thick, while in the 

 upper the spirem is segmenting. 



Fig. 33.— Spindles for the second mitosis in the embryo sac; this sac 

 developed from the second megaspore; the third megaspore also developing. 



Fig. 34. — The four-nucleate embryo sac with the spirems more or less 

 completely segmented for the third division. 



Fig. 35.— Spindles for the third division in the embryo sac; in the upper 

 part of the sac one spindle is almost at right angles to the paper. 



Fig. 36.— Eight-nucleate sac soon after the third division. 



• • 



