in the Root and Stem of Dicotyledons. 265 
into a detailed criticism of his statements we may admit at 
once that his view is so far justified, that in many cases the 
medullary phloem groups show no clear relation, as seen in 
transverse section, to individual vascular bundles. But their 
longitudinal course shows sufficiently to which bundle or 
system of bundles they belong. The term vascular bundle 
itself does not always or even generally represent a well- 
defined unit. The idea is taken from certain plants, especially 
Monocotyledons, in which the strands of associated tracheae 
and sieve-tubes are really distinct. Botanists have found it 
convenient to extend the idea to vascular tissue generally, 
although in a large number of the higher dicotyledons, so far 
at least as the stem is concerned, the limits between the 
bundles may be impossible to trace. And so it is with the 
special case of bicollaterality. So long as an internal phloem- 
strand has the same longitudinal course as the neighbouring 
bundles of the leaf- trace there is no serious objection to 
regarding them as parts of the same formation. We do so, 
however, rather as a matter of convenience than of principle 
and without expressing any general opinion as to the order of 
development, which certainly varies in different cases. 
I. Relation between Stem and Root-structure 
in Plants with bicollateral Bundles. 
1. Brozvallia viscosa, H. B. and Kth. (Salpiglossideae). 
The bicollateral structure of the vascular bundles in many 
Salpiglossideae has been demonstrated by Vesque, Petersen, 
and Solereder. The two last-named authors have pointed out 
that this anatomical character, which appears to be constant 
for the tribe, seems to confirm the opinion of Bentham and 
Hooker, that these plants belong to the Solanaceae rather 
than to the Scrophulariaceae. Browallia itself was among 
the genera investigated by Petersen 1 . 
In the above species, the transverse section of a young 
1 Loc. cit, p. 382. 
U 2 
