234 J e (F r ey and Torrey . — Transitional Herbaceous Dicotyledons. 
Fig. 24, we find in Fig. 25, PL XII, a leaf-trace subtended by a remarkably 
well developed foliar ray. The species used here is the herbaceous Melilotus 
alba. The occurrence of foliar rays is practically universal in herbaceous 
and vine-like representatives of the Leguminosae, and they consequently 
supply good evidence in favour of the coincidence of foliar rays with the 
herbaceous type. Here, as elsewhere, when the aerial stem is extremely 
slender, a radially well developed foliar ray is found only in the thick 
perennial subterranean axis. 
In a group like the Umbelliferae, the representatives of which in 
temperate regions are almost exclusively tender herbs, it is usually possible 
to discover radially developed foliar rays only in the persistent hypogaeous 
axis. The same statement holds true for a large number of other highly 
herbaceous orders. Obviously, in seeking a derivation of the herbaceous 
type from the woody, it is necessary to choose types of herbs in the first 
instance which are more or less woody in order to successfully trace the 
stages from the arboreal to the herbaceous type. Our critics have made 
the mistake of comparing extreme herbs with trees, and have failed to 
investigate with even moderate care the more important transitional forms 
which unite trees and herbs. 
Fig. 26, PI. XIII, illustrates the organization of the nodal region of the 
Scrophulariaceous genus Gerardia . The leaves are opposite, and correspond- 
ing to these are two opposite foliar traces. The trace on the lower side of 
the figure is making its way in towards the centre and is consequently sub- 
tended inwardly by the foliar gap. The upper trace by contrast has reached 
its final position on the margin of the pith and subtends radially a well- 
marked parenchymatous band, the foliar ray. In Fig. 27, PL XIII, is shown 
a detail of the foliar ray, and its contrasted structure may be compared with 
the ordinary wood on either hand. Almost any of the Scrophulariaceae 
show similar conditions as to the development of foliar rays, notably the 
common Mullein ( Verbascum) and the Toad-flax ( Linaria ). 
In conclusion, we may deal with the Compositae. Fig. 28, PL XIII, 
illustrates a transverse section of the foliar ray in Solidago canadensis from 
a preparation made by Miss Edith Whitaker, who has recently published 
an interesting contribution on the anatomy of the Golden-rods ( 4 ). A well- | 
developed radial mass of storage parenchyma is revealed in the section. 
Fig. 29, PL XIII, also from a preparation by Miss Whitaker, reveals the 
tangential aspect of the foliar ray. The trace in this instance, as is indeed ; 
common in the Compositae, lies well to the top of the ray. In Fig. 30, 
PL XIII, is shown the radial view of the foliar ray. This illustration is par- 
ticularly instructive. Above is shown the ordinary structure of the wood 
which is normally found above the region of exit of the trace in stems 
which are not too extremely herbaceous. Below the foliar trace, which 
shows as a structure pursuing an oblique and rising curved course towards 
