384 BOTANICAL GAZETTE [DECEMBER 
bud, he maintained that its terminal portion is a highly specialized 
adaptation, acquired through the action of natural selection. 
On the contrary, RETTIG and Fresric assert that the prostoma is 
merely the youngest or less highly differentiated portion of the 
groove, and that it is produced by the pressure of the axillary bud 
and other growth phenomena in the elongating internodes. The 
former investigator is of the opinion that the ants are deterred from 
excavating in the lower portion of the groove, not by mechanical 
obstructions, but by the occurrence of “‘laticiferous vessels’? which 
are absent in the prostoma. It is to be emphasized in this connec- 
tion, however, that Retric’s and Friesric’s statements do not 
necessarily invalidate ScHimPER’s conclusion that the diaphragm 
is an adaptation which originated as a modification of a previously 
existing structure. 
Under most growth conditions Cecropia angulata does not form 
a shallow groove which terminates in a conspicuous circular pit 
(text fig. 8). The more or less fusiform depression or rill is some- 
what deeper in the upper than in the lower portions of the inter- 
node, but the differentiation of specialized or mechanical types of 
tissue is retarded throughout its extension (fig. 1). As the walls 
of the internode increase in thickness, after the initiation of second- 
ary growth, this fusiform diaphragm of delicate parenchyma is 
slowly réenforced by tougher and denser tissues (fig. 3). The 
metamorphosis begins at the base of the groove, and gradually 
extends upward, but the ants excavate their exits before these 
changes have progressed very far. It is evident, therefore, that in 
the case of the myrmecophytic C. angulata, the whole groove is a 
potential prostoma. The exact location of the aperture is deter- 
mined, not by the presence or absence of resistant tissues or of 
“Jatex”’ vessels, but by the relative thickness of different portions 
of the diaphragm. That the whole groove is a potential prostoma, 
and that the ants merely excavate their exits and entrances in its 
thinnest and most easily perforable portion, are indicated by the 
behavior of young queens in juvenile plants. In many instances, 
several queens attempt to occupy the same internode, and as many 
as five entrances were found cut at different levels of a single groove. 
Not only do the queens cut through the basal portions of the 
diaphragms under such conditions, but they may even excavate 
