288 
E. J. Salisbury. 
with the increase in size of the leaf there is a gradually increased 
ratio. This may well be due to the fact referred to above that even 
in moving air the leaf is to some extent covered by a vapour cap so 
that the evaporation is proportional to some figure whose value lies 
between that of the periphery and the area. 
In the great majority of Dicotyledonous plants the increase of 
xylem within the petiole accompanying increase in size of the leaf 
is brought about by secondary thickening. But in such plants as 
the Polygonaceae the activity of the cambium is either very slight 
or absent altogether so that with the increase of leaf-area we find 
an increase in the number of bundles which the petiole contains. 
In the following table the number of vascular strands in leaves of 
Rheum and Rumex of different sizes is given. As will be seen there 
is a general correspondence which if the size of the individual 
bundles had been considered would have been found to be very 
much closer. 
Size of Leaf. 
No. of Bundles. 
« tfl 
h u 
Rheum. « > 
s-s 
(immature) 
il 
d 
e 
ft 
8 
6247 sq. cms. 
4513 „ „ 
3523 ,, ,, 
2155 „ , r 
690 ,, ,, 
510 ,, ,, 
284 „ „ 
300 
272 
246 
191 
90 
83 
93 
Rumex 
a 
b 
c 
d 
320 „ „ 
175 „ „ 
100 „ „ 
100 ,, „ 
30 
17 
15 
16 
Darwin and Phillips shewed that the bending of a stem had a 
marked effect upon the transpiration current, a fact which we find 
expressed in the large size of the conducting channels of climbing 
plants. It is therefore of importance to the plant that the petioles 
should withstand strains, which will be more or less proportional to 
the areas of the leaves they bear. 
For purposes of comparison the strength of the petiole has 
been assumed to be proportional to its breaking strain. In order 
to calculate the latter the petiole was supported by a slotted plate, 
and a band the width of the slot passed over it and attached to a 
vessel of known weight. Water was gradually added in such a way 
that immediately the petiole broke the stream passed outside the 
vessel owing to its descent. The water was then measured and its 
