INVESTIGATIONS ON THE PHYLOGENY OF THE ANGIOSPERMS 1 3 
pinnate simple by a more or less complete lateral dissection of the 
lamina, the lobes really representing very large leaf teeth. Transi- 
tions may be found in many herbaceous families (figs. i6 and 17). 
The pinnate compound leaf evidently originated from the pal- 
mate compound, and transitions may frequently be observed in 
Acer (figs. 8 to 11), Rubus (figs. 14 and 15) and many other genera. 
In each case the two lowest leaflets are homologous with the lateral 
leaflets of the trifoliate type and the terminal leaflet has cut off basally 
one or more additional pairs. In practically all woody (and therefore 
more ancient) Dicotyledons the pinnately compound leaf seems to 
have arisen through this almost meristematic activity of the terminal 
leaflet. It is essentially a modified palmate leaf rather than a pinnate 
one, a fact which may explain what we have previously noted, its 
almost invariable correlation with a multilacunar node. In certain 
of the Proteaceae, however, and a great many herbaceous and semi- 
herbaceous species leaves which are essentially pinnate compound 
have had their origin in an actual dissection of the lamina of a sim- 
ple pinnate leaf. 
In the even-pinnate Leguminosae (and possibly in other families 
as well) the pinnately compound condition seems to have arisen from 
a palmately two-lobed rather than from a palmately three-lobed type. 
Transitions may be found from the simple palmate leaf of Cercis 
(fig. 22) through various species of Bauhinia (fig. 23) and related 
genera to forms with two distinct leaflets (fig. 24) which in still others 
have cut off pairs of leaflets basally to produce the typical even- 
pinnate type. 
Factors in Foliar Evolution 
The most important factor operative in altering the ancient three- 
lobed, three-veined palmate leaf seems to have been the origin of the 
petiole. It is generally agreed that this organ, the last part of the 
leaf to appear in ontogeny, is also more recent in evolutionary origin 
than the rest. It may well owe its development to the fact that the 
early Angiosperm leaf, ever increasing in breadth of lamina, was in 
need of a greater flexibility than was afforded by a broad sessile base 
in order to guard against havoc by winds and also, perhaps, to increase 
transpiration. The primitive leaf with its three traces widely separ- 
ated in origin passing directly from node to lamina, was thus con- 
stricted at its base and its three bundles forced close together in the 
