I 
! 
14 EDMUND W. SINNOTT AND IRVING W. BAILEY 
petiole. In some cases, as in a few leaves today (fig. 81), they stayed 
apart, though crowded close together, and diverged again at the base of 
the lamina (fig. 84). In others they partially fused by their margins 
(fig. 80). In most instances, however, they became joined together 
into a continuous cylinder (fig. 82) or were broken up into a ring of 
numerous strands. This petiolar system sometimes separated into 
approximately equal portions again at the base of the lamina but seems 
to have shown an increasing tendency to persist as a single strand, the 
midrib, sending off numerous branches on both sides. In other words, 
the midrib is the continuation of the petiole, to which it owes its 
origin. The approximation of the three traces in the petiole appar- 
ently extended down to the node and was responsible for the origin 
of the unilacunar nodal type which we have seen to be so intimately 
correlated with the simple pinnate leaf. Among woody plants the 
unilacunar type is largely confined to the tropics. This seems to be 
due to the fact that the tropical rain forest leaf is usually large 
and heavy and provided with a stout petiole, which for mechanical 
reasons is cylindrical and which thus has a circular leaf scar or point 
of attachment as opposed to the elongated scar of most temperate 
woody plants. This narrower point of insertion of course tends to 
pull the leaf traces more closely together and thus to develope the 
unilacunar node, which is also produced occasionally by the loss of 
the two lateral traces. 
We are in almost complete ignorance of the other factors operative 
in the evolution of the leaf. The development of the palmate simple 
leaf from the palmate lobed one may well have been due, at least in 
the first instance, to reduction consequent upon loss of vigor or to 
decrease in size for other reasons. The acquirement of a leathery, 
xerophytic texture by the leaf is operative in some way to change the 
lobed leaf into a simple one, as is shown by the almost complete 
absence of palmate-lobed leaves among woody plants in the tropics 
(see Table I and tropical species of Rubus, figs. 12 and 13). Tropical 
leaves are often large and are frequently compound but the ultimate 
foliar unit, whatever it is, is generally narrow in shape and has a stout 
midrib. The reason for this is not clear, but it may be connected in 
some way with tropical xerophily. 
As to what causes the development of compound leaves we are 
also uncertain. Such types, however, perhaps from their ability to 
orient themselves to sunlight, are well represented in warm, dry 
