STRUCTURE.] DE CANDOLLE's THEORY. 271 



jointed with their support, nor deciduous. A leaf having 

 such segments is called dissected, 



" If the lobes are united near the base around the origin of 

 these veins, we name them partitions, and the leaf is said to 

 be parted. 



" Supposing the lobes to be united as far as the middle, 

 they become divisions, their recesses are fissures, and the ad- 

 jectives formed from these are made to end in fid, as multifid, 

 quinquefid, &c.; this should not be applied to any cases in which 

 the divisions extend below the middle of the veins ; it is, 

 however, frequently applied to cases of a division as deep as 

 the midrib. 



" Finally, if the adhesion of the lobes is complete, and if 

 the parenchym which separates the extremity only of the 

 veins is not extended to the extremity of the principal veins, 

 or beyond them, the leaf is merely toothed (dentate] ; the 

 salient parts are toothings. When the toothings or teeth, are 

 rounded, they become crenels, and the leaf is crenelled (or 

 crenate). This form of leaf is not very important because it 

 is not connected with the arrangement of the primary veins, 

 while that of the lobes, already mentioned, always is. 



' ' The terms that express precisely the important subdivi- 

 sions of the leaf are combined with those which indicate 

 venation. Thus a feather-veined leaf (pennivenium) may be 

 either pennatisected, or pennatiparted, or pennatifid, according 

 as it has segments, partitions, or fissures. In like manner a 

 palm-veined leaf (this is what I call radiating, p. 267.) may be 

 palmatisected, palmatiparted, or palmatifid ; and so on. 



" In like manner we say that a leaf is trisected, trifid, or 

 triparted, when we would draw attention to the number and 

 depth of the lobes of a leaf, rather than to the relation they 

 bear to the veins. And, on the other hand, we may, by 

 neglecting the number of the lobes, simply indicate their 

 presence by saying that a leaf is pennatilobed, palmatilobed, 

 and so on. 



" The lobes themselves are sometimes subdivided upon the 

 same principle as the leaf itself. We then say that a leaf is 

 bipennatisected, bipennatiparted, &c.; if the subdivisions of 

 the lobes are themselves lobed, we may say tripennatisected, 



