104 BOTANICAL GAZETTE [FEBRUARY 
? 
in Aster and Solidago, to the posterior or ‘‘upper’”’ carpel. The 
Anlage of the ovule is said to be excentric, and nearer to the 
posterior carpel, a position which Warming figures for Senecio, 
but which is hard to explain if, as all the earlier writers agreed, 
the ovule belongs to the anterior carpel. 
I find that the ovule of Silphium is generally posterior in 
origin (fig. 16), but it may sometimes arise anteriorly (jig. 17). 
This fact suggests that the excentric origin of the ovule is of but 
slight significance, since it may vary in the same species. On 
comparing the course of the bundles with that described for 
Aster and Solidago, I find that the axial bundle, instead of end- 
ing abruptly just below the edge of the funiculus, passes directly 
into the ovule (figs. 24, 26, 27). Whatever may be the condi- 
tion in other Compositae, the ovule of Silphium, as shown by 
its origin and its bundle relations, is the termination of the floral 
axis, that is, the ovule is cauline. The oblique position of its 
Anlage is correlated with the position which must sooner or later 
be taken by the funiculus of an anatropous ovule in an ovary of 
limited size. 
It should be noted, in passing, that the bundle does not tef- 
minate in the upper end of the ovule, but continues in the integ- 
ument to a point on a level with the embryo sac (fig. 27)- This 
is not in agreement with the recent work of Mlle. Goldflus (27) 
upon Composite ovules, according to which the bundle ends 
directly above the chalazal region, and is connected to the antip- 
odals by a strand of elongated conducting cells. 
Nectaries.— It has been suggested that the nectaries of Com- 
positae are modified stamen rudiments. The disk flower of 
Silphium offers no conclusive evidence for or against this view, 
as the nectar disk occupies a position within the stamen whorl, 
and might be interpreted as consisting of the confluent rudiments 
of an inner whorl of stamens. The nectary in the ray flower is 
also a ring, less prominent than in the disk flower, but eve? 
more closely related to the base of the style. The general 
absence of maturing stamens here makes a comparison of the 
two rather difficult. But in the two cases of older stamen® 
