IOWA ACADEMY OF SCIENCE 
155 
oped on the inner than on the outer side. In another marginal flower 
(fig. 11) the parts on the outer side were developed like those of the 
normal pistillate flower. On the inner side, however, they took the form 
of the staminate flower, the eoralla and stamens being fully formed. 
This modification even extended to the style which bore brush hairs on 
the inner (staminate) side. On the lateral side of the flower there was 
a gradual transition between the two conditions (fig. 12), two stamens 
aborting at the mother-cell stage. On the other lateral side an abrupt 
change from the pistillate to the staminate form occurred. At this 
point there was present an opening in the ovarian wall between the 
eoralla and the base of the style. The abnormal flowers suggest that 
the normal pistillate flower possesses both abortive stamens and an 
abortive eoralla, and that the staminate flower possesses an abortive 
pistil, which indicates the derivation of both forms from the perfect 
flower. 
DISCUSSION AND SUMMARY. 
The study of floral development in Iva xanthiifolia, Nutt, reveals 
strong evidence that the capitulum is, as Warning (15) held, phylogeneti- 
cally a contracted spike. The meristematic region in the center of the 
head is suggestive of apical growth. The existence of floral bracts 
within the head points to the previous arrangement of flowers in the 
axils of subtending leaves. And the vascular system, in so far as it is 
dependent upon recapitulation for its form, is likewise indicative of 
axial organization. 
Considerable difference of opinion has arisen as to whether the ances- 
tral form of the Compositae possessed perfect or diclinous flowers. 
Lecoq, Deipino, Dammers and Muller contended that hermaphrodite 
flowers were derived from the unisexual forms. Spruce, Bentham, Dar- 
win, Hildebrand, Warming and Uexkull-Gyllenband held that perfect 
flowers represent the primitive condition and that monosporangiate 
flowers have arisen by the abortion of stamens or pistils. In Iva the 
abortive pistil which still functions in opening the anthers, indicates 
the derivation of the staminate flower from the hermaphrodite. In like 
manner the abortive stamens, which occasionally develop into pollen- 
bearing members, suggest a similar origin for the pistillate flower. So 
that the evidence presented in this study favors the view that the 
unisexual condition is derived. 
Assuming, then, that the pistillate and staminate flowers have arisen 
from the perfect, one seeks an explanation of this differentiation. The 
conditions surrounding the staminate flowers are in several ways unlike 
those of the pistillate. The opening of the involucral bracts exposes the 
