140 BOTANICAL GAZETTE [FEBRUARY 
events has become fixed in the life-history, but arrest of develop- 
ment in the center of the receptacle due to contact with the apices 
of the involucral bracts might explain its origin both in individual 
and racial development. In the light of the foregoing it is evident 
that the terminal heads develop as in the normal Compositae, 
because the involucral bracts are not hemmed in by subtending 
structures and hence their tips probably do not come in contact 
with the receptacle. 
Growth in the marginal region is lateral as well as longitudinal, 
making the depression larger at the bottom and constricted above. 
Furthermore, the floor of the pit is never a uniformly concave sur- 
face, but is slightly elevated at its exact center (fig. 7). This 
elevation soon resumes growth (fig. 8) and gives rise to a septum 
dividing the depression into two chambers (fig. 9). Later this 
septum develops in its upper surface a cleft which may extend 
downward some distance (fig. 13). 
Many papillae early cover the marginal surface of the receptacle 
(figs. 8 and 9), asin the staminate head. Those of the latter, it will 
be recalled, give rise to the flowers and floral bracts, while these 
papillae of the pistillate head form hard hooked spines (fig. 13). 
That these spines are homologous with the floral bracts of the 
staminate head seems reasonably certain. It is hardly likely that 
flowers would become transformed into such structures by any 
process of evolution. These spines have the form and spiral 
arrangement of floral bracts. Furthermore, Rostowzew (13) has 
shown that the vascular anatomy is like that of bracts. To con- 
sider them new structures would be to assume the disappearance 
of floral bracts and the subsequent appearance of structures similar 
to them in exactly the same location. Such a substitution is not 
supported by the observed facts. It seems, therefore, that there is 
abundant evidence for considering these structures floral bracts. 
The margin of the head grows more rapidly on one side, 
causing one of the depressions to become deeper. This unequal 
growth also accounts for the peculiarity noted by ARTHUR (1) that 
the bur is flat on one side and curved on the other. The two 
flowers appear simultaneously, one at the bottom of each pit 
(fig. 8); but the lower, being in the deeper depression, is always 
larger, grows more rapidly, and at all stages of development is 
