254 BOTANICAL GAZETTE [Marcu 
traces pass out surrounded by parenchyma. Fig. 6 is a more 
highly magnified view of one of the lateral leaf traces, showing that 
even in the upper aerial region of this persistently woody species 
confronting as well as flanking parenchyma is present. Fig. 6 also 
elucidates the relatively greater size of the leaf trace in proportion 
to the segment, as compared with fig. 2. Allowing for the difference 
in the size of the leaf trace and for the inevitable thinning of the 
cylinder in the more slender portions of this particular stem, as 
well as in herbaceous forms generally, one can readily see how, if 
the cylinder were sufficiently reduced in size, the subtending storage 
tissue would be confined to the sides of the foliar strand even in the 
region of the node. Lower in the internode there would normally 
be only flanking parenchyma, since the central region of the con- 
fronting parenchyma is transformed below into a characteristically 
woody segment composed of vessels, fibers, etc. Eames (Joc. cit.) 
has made it clear that in the Rosaceae in which both woody and 
herbaceous types occur the latter has undoubtedly come from the 
former by the parenchymatous transformation of more and more 
secondary wood in proximity to the leaf traces. Hence the bundles 
of the herbaceous stem represent ordinary woody segments intet- 
spersed with other segments which have undergone, the paren- 
chymatous metamorphosis described. 
It would seem fairly clear from the preceding descriptions that 
the highest dicotyledons present the same general conditions of 
modification of the originally woody cylinder in relation to storage 
and the leaf trace as is found in the lower groups, such as the 
Betulaceae, Fagaceae, etc. It is quite clear that the herbaceous 
type has originated in a similar manner and largely as the result 
of the thinning of the woody cylinder. Solidago obviously illus- 
trates this transition stage, since it shows in its herbaceous regions 
the same topographical relation to the cylinder as is found in the 
typically woody part of the axis. To elucidate this point, and 
chiefly for the sake of making the situation indubitably clear, the 
accompanying diagrams have been introduced. 
In text fig. 1, A represents the woody portion of Solidago, and 
would also hold equally well for the woody dicotyledons with large 
foliar rays. On the other hand, B elucidates the characteristic 
