LKSSON ll.j RACEMK, COUYMB, UMBKL, ETC. 



79 



by intorm«'<li:ite pradations of every port. F<ir instance, if we 

 lengthen the lower pedicels of a raceme, and keep the main axis 

 rather short, it is converted into 



203. A Corymb (Fig. ir>8). This is the same as a raceme, except 

 that it is fiat and i)r()a<l, either convex, or level-topped, as in the 

 Ilawfhorn, owing to the lengthening of the lower pedicels while the 

 uppermost remain shorter. 



204. The main axis of a corymb is short, at least in comparison 

 witii the lower pedicels. Only suppose it to be so much contracted 

 that th(! bracts are all brought into a cluster or circle, and the 

 corymb becomes 



205. All I'inbcl (Fig. 150), — as in the Milkweed and Primrose, 

 — a sort of Hower-eluster Avhere the jK'dicels all spring apparently 

 from the sanie point, from the top of the peduncle, so as to resemble, 

 when spreading, the rays of an imibrella, whence the name. Here 

 the pedicels are sometimes called the rays of the unil)el. ^nd the 

 bracts, when brought in this way into a cluster or circle, form what 

 is called an involKcre. 



I 



20G. For the snme reason that the order of blossoming in a ra- 

 ceme is ascending (201), in the corymb and umbel it is centripetal, 

 that is, it proceeds from the margin or circumference regularly to- 

 wards the centre; the lower flowers of the former answering to the 

 outer ones of the latter. Indeterminate inflorescence, therefore, is 

 said to be centri|)efal in evolution. And by having this order of 

 blossoming, all the sorts may be distinguished from those of the 

 other, or the determinate class. In all the foregoing cases the 

 flowere are raised on pciliccls. Tliese, however, are very short in 

 many instances, or arc wnntiiig altogether; when the flowers are 

 tessilc (2(M>). They are so in 



FKi. IjT. a raronio. l.'>8. A forynil). i:>0. An iiinhcl 



