292 BOTANICAL GAZETTE [APRIL 
in Cryptomeria (29). The work of Lotsy (30) on Gnetum indicates 
the same thing, and occasionally slight evidences of this are found 
in Pinus and other conifers. 
Even angiosperms sometimes have this competition between 
female gametophytes, which seems to occur as a functional form of 
developmental selection. This is especially true of some of the 
lower Archichlamydeae and Monocotyledons. Casuarina (37) has 
been reported to have as high as twenty or more megaspores, of 
which several enlarge considerably, but only one is functional. 
Alchemilla (20) has been observed with five or six ripe megaspores; 
Fic. 27.—Selection between female gametophytes in Ranunculus septentrionalis: 
A, section of nucellus showing eight-celled archesporium; B, later stage showin, 
several female gametophytes and aborting megaspores in early stages of development, 
400; after CouLTER (12); cut lent by D. Appleton Company. 
in Arisaema (8) something similar has been reported; and in Ranun- 
culus COULTER (12) found as many as eight archesporial cells and 
three embryo sacs within the same ovule (fig. 27). Numerous 
additional examples are on record. GOEBEL (20), in discussing the 
many gametophytes of Casuarina, suggests that ‘‘biologically 
this repeats the case of the embryos of the Abietineae, where, of 
the many embryos which arise from one egg, only one develops.”’ 
His interpretation of the significance of polyembryony in conifers 
is treated as a process of correlation, where he compares it to the 
