in the Evolution of Herbaceous Angiosperms. 529 
Thus,manyherbaceousCentrospermae,Contortae,Tubiflorae,Campanulaceae, 
&c., with slender aerial stems, are devoid of ‘ foliar storage rays \ The leaf- 
traces are not confronted, even at the nodes, by wide sheets of parenchyma 
(Figs. 14, 1 6, and 17), nor are they flanked by high multiseriate rays. The 
stele is continuous and unbroken, except for the presence of foliar gaps at 
the nodes. In certain of these forms (Fig. 14), particularly plants with 
a typical decussate phyllotaxy, the secondary xylem confronting the leaf- 
traces may be devoid of vessels for a considerable distance below the node. 
This phenomenon, however, is not fundamentally a concomitant of the 
transformation of woody tissue into parenchyma, but is correlated with the 
conduction of water to the upper levels of the stem. 
It is to be emphasized, accordingly, that not only do trees have foliar 
rays, but that many herbs are devoid of masses of 4 foliar storage paren- 
chyma \ Such herbaceous Dicotyledons obviously have not 4 developed 
from arboreal dicotyledonous types by the formation of storage rays about 
the leaf-traces ’. 
A more detailed study of the comparative anatomy of the Dicotyledons 
fully justifies our contentions that the aerial stems of herbs are similar to 
the young stems of their arborescent and fruticose relatives, and that the 
reduction of secondary growth is the most significant anatomical concomi- 
tant of the herbaceous habit. Certain phenomena, such as the widening 
and lengthening of medullary rays, the disappearance of secondary vessels 
in certain segments of the woody cylinder, an increase of parenchyma in the 
nodal regions, &c., may be accentuated in certain groups of herbs, vines, and 
lianas, but they are not essential concomitants of herbaceousness, and are 
often present in arborescent forms. Furthermore, Jeffrey and Torrey admit 
the essential correctness of our conclusion that the stele of many-bundled 
herbs is dissected by interfascicular masses of parenchyma which are not 
subtended by primary zuood on their centrad sides. They state : 
4 The foliar ray surrounding and subtending the leaf-trace is charac- 
teristic of the less advanced dicotyledonous herbs and of the lower, more 
woody regions of those higher in the herbaceous sequence. This type of 
ray gives place by the later thinning of the woody cylinder to one in which 
the storage tissue is confined to the flanks of the traces. The long vertical 
extension of these flanking rays results in the division of the originally con- 
tinuous woody cylinder of the ancestral Dicotyledons into a circle of separate 
strands, the fibro-vascular bundles/ 
The last sentence of this quotation is in fact an admission of the 
correctness of the view which we expressed in our previous paper, that the 
flanking rather than the confronting parenchyma is responsible for the 
breaking up of the woody cylinder into separate bundles. It should be 
noted in this connexion that, according to the present position of Jeffrey 
and Torrey, the confronting portion of the foliar ray, emphasized by them 
