120 Seward and Gowan .— The Maidenhair 
almost horizontal or more or less steeply inclined towards the 
petiole; the venation is of the Cyclopteridis type, the lowest 
vein in the right and left half of the lamina follows a course 
parallel to the edge, and gives off branches which fork 
repeatedly as they spread in a palmate manner towards the 
upper margin of the lamina 1 . The short secretory canals, 
which have their origin in the forks of the veins, are often 
clearly marked as short dark lines traversing the mesophyll. 
These canals are occasionally visible in fossil Ginkgo leaves, 
and by one author have been described as patches of a 
Fungus 2 . A striking feature of Ginkgo leaves is the variation 
in their size and shape, a fact insufficiently recognized by 
palaeobotanical writers 3 . The deeply cut lamina with more 
than two long and narrow segments (Figs. 63 and 66) is met 
with more particularly in the large leaves of vigorous long 
shoots ; between this type and the small entire leaves char¬ 
acteristic of flowering shoots the difference is considerable. 
An unusually large bilobed leaf obtained from a tree growing 
in the temperate house of the Glasgow Botanic Garden 
measured 20 cm. across, and a still larger leaf, 21 cm. broad, 
is figured under the name Salisburia adiantifolia var. macro* 
phylla laciniata in the tenth volume of the ‘Flore des SerresV 
The few examples of leaf variation represented in PL X, 
Figs. 63-67 and 71, are chosen principally in illustration of 
the resemblance between Ginkgo biloba and various Mesozoic 
species of the genus. More than one writer has called atten¬ 
tion to the resemblance between the more deeply lobed leaves 
on the vigorous shoots of Ginkgo and certain Mesozoic forms ; 
Bailey, for example, speaks of such leaves as the ‘fitful 
recollections of an ancient state V 
A study of the anatomy of Ginkgo leaves reveals some 
features of interest from the point of view of comparison 
with Cycads and Conifers. The epidermis consists of cells 
1 Drude in Schenk (’ 79 ), vol. i, p. 654; Bertrand (’ 81 ), p. 178, Fig. 129. 
3 Massalongo (’ 59 ), PI. I, Fig. 1, p. 87. 
3 Gardner (’ 83 ), p. 45. 
4 Flore des Serres (’ 54 ), p. 119. 
5 Bailey (’ 96 ), p. 97 ; vide also Ettingshausen (’ 90 ), PL VII. 
