152 THE PLAJ\T WOULD. 



different parts of a flower head, or in different heads of the same 

 plant ; and in rare eases this resnlts in dioecism. Someho\\' or 

 other this componnding mnst he advantageous in the struggle for 

 existence; else the Compositse should not he the most ahundant 

 family of the flowering plants. Their only competitors for pri- 

 macy are the Graminese and the Cyperacese, which possess an 

 analogous system of reducing and crowding their flowers. 



Bentham's idea was that in the diversity of the involucre 

 and other parts we may have some key to the comparative an- 

 tiquity of the different orders of Composite'. Assuming as 

 probable that the decaying groups are older than the progressing 

 groups of the family; also assuming that decaying groups are 

 recognized by having the subordinate races very distinct in their 

 structure, by their restricted areas, and fewness of numbers, he 

 would place the sunflowers or Ilelianthoidese (at least some sub- 

 tribes of them) in the decaying stage, and probably the most 

 antique. Correlated with this he cites the fact that the great 

 consolidation and uniform structure of the floral organs of Com- 

 posite are least seen in that group ; and conversely are most pro- 

 nounced in the Cichoriese, which are the most modern, according 

 to his view. The outer bracts of the involucre are leaflike in 

 Helianthoidea^, the receptacular pales more like bracts, and the 

 pappus more calyx-like, than in other tribes ; all indicating that 

 they are old-fashioned and less specialized than Asters or Sene- 

 cios ; and perhaps suggesting why they are now retrograding. 

 The Cichoriese, on the other hand, are most specialized, and for 

 this reason, most puzzling. The uniformity of their organs of 

 reproduction is such that neither pistil nor andrecium, nor cor- 

 olla affords the slightest structural or sexual character among the 

 numerous species ; the pappus is not at all like a calyx-limb ; and 

 the pales of the receptacle are not like bracts, and are very fre- 

 quently mere rudiments or entirely suppressed. 



Some groups of Compositce take a higher order of special- 

 ization, having glomerules or heads of which the components are 

 not flowers, but are themselves flower-heads, further reduced in 

 size and provided with an outer wrapper, like an involucre of 



