INVESTIGATIONS ON THE PHYLOGENY OF THE ANGIOSPERMS II 
is as yet not definitely settled. All evidence now available, however, 
as to the structure of both vegetative and reproductive organs in the 
various families of the inclusive order Amentiferae seems to indicate 
clearly that this great group of simple-flowered plants is not ancient 
in type, as has frequently been supposed, but is rather to be looked 
upon as reduced and specialized. The prevalence of the simple 
pinnate leaf among its members is therefore no indication of the 
primitiveness of this foliar type. Those Dicotyledons which both 
from their wood structure and reproductive morphology are today 
winning recognition from phylogenists as being among the most 
ancient members of the whole class are included mainly among the 
simpler types of the Ranales, Rosales and Mai vales. The very great 
frequency of the palmate leaf, either simple, lobed or compound, among 
the Ranunculaceae, Saxifragaceae, Hamamelidaceae, Malvaceae and 
their immediate allies (over half of the palmate lobed leaves in the 
floras investigated belong here) thus furnishes the most important 
evidence that modern phylogeny can contribute to a solution of the 
present problem and confirms the evidence from other sources that 
the palmate type is indeed the' most ancient among Dicotyledons. 
As to which of the three palmate forms is the oldest we cannot be 
quite sure. It is possible that the simple pinnate leaf with closed 
venation and the veins all converging at the tip of the lamina, a type 
which is predominant among Monocotyledons and the seed-leaves of 
Dicotyledons, may be the most ancient condition, from which the 
lobed form has arisen by a separation and divergence of the veins. 
The high percentage of palmate lobed leaves in the more ancient fossil 
floras, however; the primitiveness of the trilacunar node; the numerous 
cases of reversion of palmate simple to palmate lobed leaves, and the 
predominance of the palmate lobed form in the more primitive orders 
whereas the palmate simple and palmate compound types are often 
characteristic of more highly specialized groups, all point to the 
conclusion that a three-lobed, palmately three-veined leaf is one of at 
least very high antiquity. In such a leaf the three independent leaf 
trace bundles, widely separated at their insertion, probably remained 
so throughout their course into the lamina. In modern leaves of this 
type we usually find either three separate bundles in the petiole 
(fig. 8i) or a single three-lobed one (fig. 80) . Other leaf types generally 
possess a much more complicated petiolar system. If the primitive 
leaf was sessile or nearly so, as most writers agree, it is probable that 
