396 
AMERICAN JOURNAL OF BOTANY 
[Vol. 8 
vessels. None of the bundles, even those with lignified cells in their upper 
portions, can be followed to their origin in the floral axis. The posterior 
bundle (fig. 4, m) can be followed down the farthest, that is, into the cortex 
of the pedicel or receptacle. It does appear that the perianth bundles are 
abortive in the lowest part of their courses. 
Payer (14) and Zinger (24) describe and figure the cup-like perianth as 
having slightly developed anterior and posterior lobes. Anatomically, the 
author found no difference in the lobed regions as compared with the 
remainder of the perianth except the fact that the most prominent bundle 
to the perianth is the posterior bundle. 
A feature of the pedicellar stele not yet described is the presence of 
regions suggestive of bundles (PI. XX, figs. 3, 4, x). Such a condition 
described in the preceding species w^as looked upon as one demonstrating 
abortive bundles. Such faint bundles and others not recognizable may pass 
into the perianth and become lignified in their upper parts only, a condition 
such that they can be followed. Also in the upper lateral ovary wall there 
are faint bundles with delicate, lignified cells (PI. XX, fig. 5, n) which cannot 
be followed to their origin. These must be either branches of the dorsal 
carpellary bundles or strands continuing from the pedicel. If the latter, 
they arise similarly to the two dorsal carpellary bundles and therefore 
suggest abortive carpellary bundles to suppressed carpels. 
The stele in the pedicel of the staminate flower is very different. It 
has many small strands which organize into five strands in the base of the 
flower. These pass out of the axis and each separates immediately into 
strands to the perianth and to the stamens. There are no signs of abortive 
strands to the suppressed carpels in the writer's experience. Likewise in 
the pistillate flower, the stamens are suppressed and no vestiges of vascular 
supply are present. 
Pritchard (16) concludes from his experiments on the hemp plant that 
both the male and the female flow^ers are potentially hermaphroditic and 
that the unisexual condition is the result not of different zygotic constitu- 
tion, but of the lack of food supply. At what time in the life of the hemp 
plant the suggested feeding must be begun in order to establish organs that 
are suppressed, even in vascular supply, is an interesting problem to a plant 
anatomist. It may indicate that the unisexual nature of the hemp flower 
is not well established. Some of its congeners in the order still have bisexual 
flowers, 
Humuliis Lupulus L. The pedicel of the pistillate flower duplicates in 
structure that of Cannabis except that it has fewer vessels in each of its 
four bundles. Anteriorly, a large trace passes out and branches profusely 
in a large bract which envelops the flower (PI. XX, fig. 9, hr\ fig. 10, a). 
The three remaining strands in the pedicel, as in Cannabis, pass into the 
pistil. Strands (figs. 14-18, 1,1^) pass up the dorsal sides of the tw^o carpels 
into the styles respectively. These are the dorsal carpellary bundles. The 
