COMPOSITAE OF PERAUSTRAL AMERICA. 153 



higher order; they are Composites in duplicate ratio. Such 

 occur among the Mutisiea?. 



The outcome of this condition of matters as to the task of 

 classiiication is embarrassing ; somewhat like what a zoologist en- 

 counters in classifying the birds. It is very difficult to find dis- 

 tinctive characters; and even when found they are subject to 

 variations of degree, and to numerous exceptions. Shutting off 

 the Cichoriese, as a separate tribe, or as even a family, limited by 

 their ligiilate flowers and milky sap ; a dozen tribes remain, two 

 of which (Vernoniese and Eupatoriese) have uniformly herma- 

 phrodite, tubular flowers ; and the remaining ten have one or 

 more rows of female flowers in the perij)hery, with exceptions. 

 The styles of the female flowers are very uniform, 2-branched 

 (save in sterile flowers, which have undivided styles) ; yet the 

 tribes are limited chiefly by differences of the styles, also by 

 dift'erences of the bases and tips of the anthers, and by the differ- 

 ences of the pappus, and the form of the involucre. 



Thus we obtain a trial-analysis of the tribes (subject, how- 

 ever, to many exceptions), viz. : 



A. Tuhifloral, and homogamous. All the flowers tubular. 



(1) Vernojiieae, Anthers basi-sagittate, styles slender. 



(2) Eupatorieae. Anthers basi-entire. Styles long, 



obtuse. 



(3) Cynaroideae. Thistles. Anthers tailed. Styles 



short, obtuse. 



B. Heads heterogamous, radiate, or abortively discoid. 



a. Anthers basi-obtuse, apically appendaged. 



b. lieceptacle mostly naked. 



c. Involucres imbricated, 



(■i) Astereae. Style branch flat. Pappus setaceous. 



(5) Helenioideae. Style branches truncate or hair- 



tipped. Pappus palaeceous. 

 c2. Involucre scales 1-2-seriate, subequal, scarcely im- 

 bricate. 



(6) Senecioneae. Pappus setose. 



b2. Peceptacle with chaff about the achenes. 



