102 BOTANICAL GAZETTE [FEBRUARY 
brushed out of the stamen tube. My preparations show no such 
device in Silphium. A comparison of figs. 9, 70, zz shows that 
while at one time the style is actually longer than the stamens 
(fig. z0), the stamens soon overtake the style again (fig. rz). 
As this is in a flower where the pollen-mother cells have reached 
a late synapsis stage, and the maturity of the pollen follows rap- 
idly, it would seem hardly probable that the style could become 
coiled before the dehiscense of the anthers. The stylar thrust 
seems to be accomplished rather by an actual increase in the 
length of the style, partially, at least, by mere cell elongation. 
It will be seen that there is no trace of an ovule, or even of 
an ovary, inthe disk flower. In other respects the resemblance 
to Helianthus (Sachs) is quite close. 
Ray flowers— The ray flowers occupy about two turns of 
a spiral upon the receptacle. As the individuals of the two 
rows alternate with each other, only one can ever appear in a 
single radial longitudinal section of a head. Moreover, in serial 
sections, never more than two rays, one each from the inner 
and outer rows, are cut near enough to the radial plane to be 
valuable for study. It will be seen that in the form of the 
young ray flower Silphium shows a very close resemblance to 
Aster and Solidago, especially when, as happens in the great 
majority of cases, rudimentary stamens are present. 
The ray flowers of the outer row (7, figs. 1, 2) are subtended 
by bracts of the involucre, which are so closely folded overt in 
the bud as to distort the young flower for a considerable period. 
This compression on the anterior side is plainly visible in fig. 
12, where the corolla is starting, and is still noticeable in fig. 23: 
where the stamens (sf) are just appearing. Fig. rg shows in 
addition the young carpels (4), and figs. 15, 16 show the begin- 
ning of the ovule (ov). At about the stage shown in jig. /4 the 
corolla tube begins to grow more rapidly on the anterior side, 
and is thus pushed away from the subtending bract. The 
opening of the tube is gradually made smaller by the growth of 
its lateral edges, which finally lap past each other and give rise 
to such appearances of the corolla as are shown in figs. 18-24) 
cre 
