158 THK LEAVES. 



the angles or lobes at the base diverge. The margins of the sinus 

 are sometimes brought into contact and united, when the leaf be- 

 comes inltate or shield-shaped (Fig. 248) ; the blade being attached 

 to the petiole, not by its apparent base, but by some part of the 

 lower surface. Two or three common species of Hydrocotyle 

 plainly exhibit the transition from common radiated leaves into the 

 peltate form. Thus, the leaf of H. Americana (Fig. 247) is round- 

 ish-reniform, with an open sinus at the base, while in II. inter- 

 rupta and H. umbellata (Fig. 248), the margins have grown to- 

 gether so as to oblitei-ate the sinus, and an orbicular peltate leaf is 

 produced. In nerved leaves, when the nerves run jjarallel from 

 the base to the apex, as in Grasses (Fig. 237), the leaf is necessa- 

 rily lineal", or nearly so ; but when they are more divergent in the 

 middle, or towards the base, the leaf becomes oblong, oval, or ovate, 

 &c. (Fig. 243). In one class of nerved or parallel-veined leaves, the 

 simple veins or nerves arise from a prolongation of the petiole in the 

 form of a thickened midrib, instead of the base of the blade, constitut- 

 ing the curvinerved leaves of De Candolle. This structure is almost 

 universal in the Ginger tribe, the Arrowroot tribe, in the Banana, and 

 other tropical plants ; and our common Pontederia, or Pickerel-weed 

 (Fig. 236), affords an illustration of it, in which the nerves are 

 curved backwards at the base, so as to produce a cordate outline. 



281. As to the margin and particular outline of leaves, they ex- 

 hibit every gradation between the case where the blade is entire, 

 that is, with the margin perfectly continuous and even (as in Fig. 

 243), and those where it is cleft or divided into separate portions. 

 The convenient hypothesis of De Candolle connects these forms 

 with the abundance or scantiness of the parenchyma, compared 

 with the divergence and the extent of the ribs or veins ; on the 

 supposition that, where the former is insufficient completely to fill 

 up the framework, lobes, incisions, or toothings are necessarily 

 produced, extending from the margin towards the centre. Thus, 

 in the white and the yellow species of Water Ranunculus, there 

 appears to be barely sufficient parenchyma to form a thin covering 

 for each vein and its branches (Fig. 251, the loAvest leaf) ; such 

 leaves are said to be jiliformhj dissected, that is, cut into threads ; 

 the nomenclature in all these cases being founded on the conven- 

 ient (but incorrect) supposition, that a leaf originally entire is cut 

 into teeth, lobes, divisions, &c. If, while the framework remains 

 the same as in the last instance, the parenchyma be more abun- 



