THE ANGIOSPERMAE : LEAVES 1037 



spiny. The axillary buds, on the other hand, develop their basal leaves pre- 

 cociously, so that a group of foliage leaves appears to be axillary to a group 

 of spines (Fig. 1029). The groups of three spines at each node on the stem 

 of the Gooseberry, Ribesura-crispa, are superficially similar to those of 

 Berberis, but they arise below the foliage leaves and are probably outgrowths 

 of the leaf base (Fig. 1030). 



The opposite condition is illustrated by Citrus, where the leaves of the 

 long shoots are normal, but one or two of the basal leaves of the axillar}- buds 

 become spines, which appear, therefore, to be axillary to the leaves and were 

 formerly classed as branch thorns (Fig. 103 1). 



Purely stipular spines are not uncommon. They occur in the succulent 

 species of Euphorbia and also in the arboreal Leguminosae, e.g., Rohinia 

 (Fig. 1032), and especially Acacia, in some species of which the spines become 

 very large and hollow, and afford shelter to ants (see Volume IV). 



2. Leaf Tendrils. — Either the whole leaf, or any part of it, including the 

 stipules, may become modified into a tendril. Sometimes a part may act as 

 a tendril with little or no modification from the normal form, as for example 

 the petioles of Tropaeohim and of Clematis, or the elongated leaf tips of 

 Gloriosa, but often there is some degree of reduction. 



It is not unusual to find that the terminal leaflet, or a number of leaflets 

 of a pinnate leaf, become reduced and more or less elongated into tendril 

 form. In Cobaea scandens several leaflets are thus altered, their stalks forming 

 the actual tendrils, while the reduced blade of each leaflet forms a small hook 

 at the end of each tendril (Fig. 1033). The Garden Pea affords another 

 familiar example of such a modification of part of a compound leaf (Fig. 1034), 

 while in Viciafaba the terminal tendril has itself become vestigial. In these and 

 in most other such cases the blade of the tendril leaflet has disappeared, but 

 in Miitisia the laminar development is unaffected and the tendril appears to 

 be a prolongation of the midrib of an otherwise normal leaf, an example of 

 progressive, as contrasted with regressive modification (Fig. 1035). 



Less common is the modification of the entire leaf into a tendril. In 

 Corydalis clavicidata the change from leaves to tendrils takes place by stages, 

 from the base of the shoots upwards, so that a proportion of the leaves remain 

 normally functional, but in Lathyrus aphaca all the leaves have become 

 tendrils, and the assimilatory function has been taken over by the enlarged 

 stipules (Fig. 1036). 



The morphology of the tendrils in the Cucurbitaceae has been much 

 discussed, but there seems reason to regard them as, in some cases at least, 

 leaf tendrils, which represent modifications of the basal cataphyll or two 

 basal cataphylls of the axillary bud. Another suggested explanation is that 

 they are themselves axillary shoots which end in a leaf tendril. Abnormally 

 they may form a terminal leaf blade, which is not discordant with either view. 



One of the most puzzling cases of tendril formation is that found in the 

 genus Smilax, in which two tendrils are attached to each leaf, one on each side 

 of the petiole (Fig. 1037). Each tendril contains a ring of bundles essentially 

 like that of the petiole itself, and according to Arber's interpretation they are 



