Jeffrey and Torrey . — Transitional Herbaceous Dicotyledons. 245 
Conclusions. 
The various photographic illustrations and stereodiagrams supplied in 
connexion with the earlier pages of the present article seem clearly to 
justify quite definite conclusions as to the mode of origin of the herbaceous 
type of axis in the Dicotyledons from that of woody or arboreal texture. 
It seems obvious, as a consequence of the comparison of nearly related 
arboreal, woody herbaceous, and slender herbaceous stems, that the 
structures which we have called foliar rays are in the first place developed 
in woody herbs as a result of the clustering of ordinary rays of the wood in 
relation to the incoming leaf-traces. The clusters or congeries of rays, to 
which one of us has applied the name aggregate rays, are often characterized 
by the fact that vessels are eliminated in the bands of longitudinal woody 
elements separating the constituent members of the foliar aggregation of 
rays from one another. We have illustrated this condition above, in the 
case of the genus Hibiscus. It is of wide occurrence in woody or transi- 
tional herbs. A next step in the development of the herbaceous type is 
the transformation of the strands of fibres separating the storage units of the 
aggregate ray from one another into rows of parenchymatous elements 
which become more and more assimilated to the ordinary radial parenchyma 
both in their dimensions and in the relations of their axes. This condition 
of the foliar ray we have called the compound ray, to distinguish it from the 
a gg re £ ate ray from which it takes its origin. Quite often when the com- 
pound ray has been thoroughly established it still betrays its derivation 
from the aggregate ray by the fact that its lower portion is still largely in 
a condition of aggregation. As the herbaceous condition becomes more 
and more established, the foliar rays become not only more homogeneous, 
by reason of the more complete loss of identity of their originally diverse 
woody elements, but also more elongated in the vertical direction. 
The last condition often expresses itself in transverse section, where 
the traces are sufficiently numerous and their accompanying foliar rays are 
sufficiently developed in the long axis, by a series of separate strands, more 
or less regularly alternating as to size, in which certain members are deep 
radially and woody in structure, while the alternating segments consist of 
relatively slender bundles subtended radially by massive storage tissues. 
The storage tissues also occur on the flanks of the slender bundles just 
referred to, which are the foliar traces in their course in the stem. This 
state of affairs is extremely common, for example, in the more woody lower 
region of the aerial stem in the herbaceous Compositae, particularly if the 
foliar traces are numerous and the foliar rays greatly extended longitudi- 
nally. Favourable objects on which to test the truth of this general statement 
are the genera Helianthus , Aster , Lactuca , &c., &c. Although the organiza- 
