in the Evolution of Herbaceous Angiosperms. 525 
extends for a considerable distance radially, but is short longitudinally. In 
the more advanced herbs, where the cylinder is thin, the portion of the ray 
confronting the leaf-trace is reduced and disappears, but the flanking portions 
are well developed and extend a considerable distance downwards. In 
certain cases, woody segments of the stem which are devoid of vessels, or 
which are provided with more numerous or larger rays than are the adjoin- 
ing segments, are referred to as foliar rays. In pther cases, masses of 
parenchyma which are formed after the leaves have fallen are similarly 
designated. Thus, the term foliar ray appears to be a conveniently 
elastic one. 
In the following pages, we shall show that (1) the foliar ray, as thus 
defined, includes several morphologically distinct structures ; (2) many 
arborescent and fruticose Dicotyledons do have ‘ foliar storage rays 5 , and 
have steles which are dissected into discrete woody strands, such as are 
considered by Jeffrey and Torrey to be characteristic of advanced types of 
dicotyledonous herbs ; (3) many slender herbaceous stems have continuous 
vascular cylinders and are devoid of c foliar storage rays 5 ; and (4) Jeffrey 
and Torrey’s work, in ultimate analysis, is actually a confirmation of our 
contention that the stele of many-bundled herbs is dissected by interfascicular 
masses of parenchyma which are not subtended by typical xylem elements 
on their centrad sides. 
Before we present evidence in support of these .contentions, it is essential 
to discuss one aspect of our critics 5 argument upon which they lay much 
emphasis. They state that 4 obviously a clear conception of the differences 
of anatomical organization between a woody and an herbaceous stem can best 
be obtained by comparing the nodal regions of nearly related trees and 
herbs 5 , and in conformity with this attitude they particularly emphasize 
nodal conditions, citing facts and drawing conclusions chiefly from this 
portion of the stem. Such a restriction seems to us to be extremely illogical, 
for any hypothesis which attempts to account for the breaking up of the 
woody cylinder into separate bundles in the aerial stem of herbs must 
explain the dissection of the cylinder in the long internodal regions as well 
as in the comparatively short portion in the vicinity of the node. Conse- 
quently, the evidence which we brought forward in our previous paper, and 
that which we shall present here, is drawn from the aerial stem as a whole 
rather than from any selected region thereof. 
That there are several distinct morphological phenomena which are 
significant in any general discussion of foliar rays, is shown by a comparative 
study of the aerial stems of numerous representatives of the various orders 
and families of Dicotyledons. Such an investigation reveals the following 
facts: (1) The radially disposed sheets of parenchyma, which are commonly 
known as medullary rays, vary greatly in number, width, height, and distri- 
bution in different Dicotyledons. In one group of trees, shrubs, vines, and 
