124 BOTANICAL GAZETTE [ FEBRUARY 
fact that the magnification of fig. 92 is but two thirds of that of 
JE. OI. 
At this time the cotyledons are nearly half as broad as long, 
and the tissues of the ovule are reduced to a thin sac, loosely 
covering the embryo. Ample protection for the growing embryo 
is provided by the hard walls of the ovary. Older stages of the 
embryo were not studied, as the increasing hardness of the ovary 
wall made the rapid fixing of the ovule and embryo more and 
more difficult. 
SUMMARY AND CONCLUSIONS. 
In the general features of its development, the foregoing 
account of the life history of Silphium agrees with those of the 
other Compositae. But so many facts are peculiar to the genus, 
that a résumé of the course of development is necessary, and a 
comparison with other Compositae. 
The flowers of Silphium in their adult form are of two kinds, 
the disk flowers being staminate and sterile, the rays being fertile 
and without stamens. 
The order of development of floral organs in the disk flower 
is corolla, stamens, carpels, calyx. The nectary appears just 
before the pappus. There is never any trace of ovary or of ovule. 
The pappus is almost entirely wanting. 
The floral organs of the ray flower, in the great majority of 
cases, appear in the same order as in the disk. The stamens 
almost always remain as mere papillae, but a few flowers were 
found in which they had formed pollen mother-cells. The ovary 
is formed as a cavity between the carpels. The ovule is terminal 
and cauline, as is shown by its position and bundle relations. 
This claim is not in accord with the other special accounts for 
Compositae, but agrees with the claim for a primitive cauline 
ovule in this family, made recently by Campbell (30). Proofs 
for this claim, however, are first presented in the present 
paper. 
In that the hypodermal archesporial cell develops a row of 
four potential megaspores, of which the inner one germinates 
to form the embryo sac, Silphium resembles all the other 
