OF PLANTS CALLED COMPOSITE. 273 



to make a few remarks on the more usual modes of inflo- 

 rescence. 



It is well known that in an absolutely simple spike the 

 expansion of the flowers is ascendent ; that is, begins at 

 the base of the spike and proceeds regularly upwards. To 

 this order very few real exceptions occur, several of the 

 apparent deviations being connected with some degree of 

 composition in the spike. 



It is also known that in a compound spike, while the [92 

 expansion of each partial spike is ascendent, that of the 

 spikes, with relation to each other, is descendent ; the ter- 

 minal spike expanding first, and the others in a regular 

 succession downwards. This order, indeed, admits of a 

 greater number of exceptions than that of the simple 

 spike ; several of them apparently depending on the den- 

 sity or imperfect composition of the spike ; and the more 

 usual deviation consisting in the expansion beginning 

 below the apex, and proceeding in opposite directions from 

 the point of commencement ; the upper portion following 

 the order of the simple, the lower that of the compound 

 spike. 1 



The simple racemus and corymbus are obviously very 

 slight modifications of the spike, and in their expansion 

 obey the same law. 



A syngenesious compound flower, or capitidum as it 

 may be termed, is merely a spike with a shortened and 

 generally depressed axis. In cases where this capitulum 

 is unquestionably simple, the expansion of its flowers is 

 uniformly from circumference to centre, or in the order of 

 the simple spike. Where the capitula are disposed in a 

 corymbus, which is their usual mode of combination, the 

 order of the compound spike is observed ; their expansion 

 with relation to each other being from centre to circum- 

 ference. In their denser aggregations, whether forming a 



1 The most remarkable exception to the order of the compound spike exists 

 in the compound umbel of Umbellifera?, of which the outer umbellultc expand 

 somewhat earlier than the central; and as this order of expansion seems to 

 extend through the whole natural family, Astranlia, in which the terminating 

 umbel expands much earlier than those of the lateral branches, cannot be con- 

 sidered as havinff a compound umbel. 



18 



