certain Orders of the Ranales. 655 
impossible to imagine the stem remaining parvifoliate in its vascular struc- 
ture while that of the leaf becomes modified in a grandifoliate direction ; in 
other words, that a condition like that found in many Umbelliferae and 
other plants could by this method be attained ; for no conceivable reason 
can be adduced for such a change. On the other hand, if the plant is 
developing out of a grandifoliate into a parvifoliate habit, it is quite 
conceivable that the stem, for the reasons above given, should become 
modified in the parvifoliate direction, while the leaf retains its primitive 
grandifoliate structure. It is also conceivable that the grandifoliate stem, 
owing to certain structural peculiarities, such as the possession of a thick 
sclerotic zone, which is adequate to resist bending-strains should, on eventual 
elongation, retain the grandifoliate structure while the leaf has begun to 
assume that of the parvifoliate plant. The above inferences strongly 
favour the truth of my present theory. 
In such cases as that of the Umbelliferae the stem-structure is not 
quite typically parvifoliate, for the primitive ranked or scattered arrange- 
ment of the bundles can still be traced there, for although all the bundles 
may be practically at one level, i. e. in a single ring or series, they are 
of three or four different sizes, the largest in some, but not all, cases slightly 
projecting into the pith, while the smallest are outermost and often quite 
rudimentary. This latter character is due to the fact that in the grandifoliate 
ancestor all the ranks of bundles would be well-developed and serviceable 
to the large leaf-base, and therefore to the stem ; but on the acquirement 
by the latter of the cylindric arrangement, owing to the elongation of the 
internodes and the reduction in comparative size of the leaves, the bundle- 
ranks would become condensed and pressed outwards to form a single rank, 
and the smallest ones, being the most external, would be the first to be 
obliterated. This can be the only interpretation of the phenomenon ; for 
it is impossible to conceive of the possibility that such plants are modifying 
their structure in the opposite and converse direction, viz. towards the 
grandifoliate condition. This is supported by the following facts: In the 
Ranunculaceae, plants with a primitive floral structure, the grandifoliate 
habit and vascular structure is pronounced, at any rate in many cases, 
while the traces of it are universal. There is no evidence, nor is there 
anywhere any sign, that the Ranunculaceae have been reduced and 
modified from woody parvifoliate ancestors. The Nymphaeaceae, also with 
a primitive floral structure, have the same grandifoliate habit and vascular 
structure which have been retained probably owing to their having acquired 
an aquatic habit. I cannot agree with Henslow’s view that the grandifoliate 
vascular structure is the residt of an aquatic habit either in the plant itself 
or in its ancestry, for the general effect of this habit upon the vascular 
structure of stems, as seen in the cases of Hippuris , Hottonia , Ceratophyllum , 
Potamogeton , &c., is extreme reduction and concentration of the bundles to 
