154 ANTHOTAXY, OB, INFLORESCENCE. 



smaller H} T pericums, and notably in H. Sarothra, in which the 

 leaves are all reduced to bracts. It is not always easy to show 

 why this is not a true raceme. But the other bract of the pair, upon 

 that supposition, is unaccountably empty : the successive angular 

 divergence of each joint of the axis of inflorescence in the } T ounger 

 part, which commonly runs into a coil, finds explanation in the view 

 that each portion is the lateral branch from the axil of the subtend- 

 ing leaf: and occasionally the other axil produces a similar one, 

 thus revealing the cymose character. When the bract from the 

 axil of which the missing branch should come disappears also, as 

 sometimes it does, and uniformly on the same side, a state of 

 things like that of the upper part of Fig. 297 occurs. The same 

 figure may serve for the arrangement corresponding to that of 

 Fig. 296, only with alternate leaves. But then, close as the imi- 

 tation of a raceme here is, the position of each flower in respect 

 to the bract supplies a criterion. While in a true raceme the 

 flower stands in the axil of its bract, here it stands on the oppo- 

 site side of the axis, or at least is quite away from the axil. 



282. Sympodial forms. The explanation is that the axis of 

 inflorescence in such cases, continuous as it appears to be, is 

 not a simple one, is not a monopode, but a sympode (110, 116, 

 notes), i.e. consists of a series of seemingly superposed inter- 

 nodes which belong to successive generations of axes : each axis 

 bears a pair of leaves (Fig. 296) or a single leaf (Fig. 297), 

 is continued beyond into a peduncle (or pedicel in these 

 instances), and is terminated by a flower. From the axil 

 promptly springs a new axis or branch, vigorous enough soon 

 to throw the adjacent pedicel and flower to one side : this 

 bears its leaf or pair of leaves, and is terminated like its prede- 

 cessor with a flower ; and so on indefinitely. The fact that the 

 alternate leaves or bracts arc thrown more or less strictly to one 

 side and the flowers to the other, in Fig. 297, shows that these 

 leaves do not belong to one and the same axis ; for alternate 

 leaves are never one-ranked or disposed preponderatingly along 

 one side of an axis, as in this diagram, and as is seen in the 

 inflorescence of a Ilouseleek, &c. 



283. A further difficulty in the morphology of clusters of this 

 class comes from the early abortion or complete suppression of 

 bracts. This is not unknown in botryose inflorescence, occurring 

 in the racemes of almost all Crueiferae : it is very common in the 

 cymose of all varieties, and especially in the uniparous ones in 

 question, which characterize or abound in Borraginacese, Hydro- 

 phyllaceae, and other natural orders. In some genera or species, 

 the bracts are present, or at least the lower ones ; in others, 



