52 Trans. Acad. Sci. of St. Louis. 



from a woody base, and the floral leaves glandular punctate 

 and greatly produced at tips. This last character separates 

 it from all other species of the tribe. The sericeus of Graham 

 is described from "West Indies," though not noticed since 

 its description. 



SPECIES OF CENTRAL AND SOUTH AMERICA. 



In the study of these species attention must first be centered 

 on the exact determination of mollis. Dr. Gray followed 

 Grisebach, who knew it well, in considering it a distinct and 

 well-marked species, and our herbarium material amply sup- 

 ports this view. It seems to be entirely distinct from either 

 scaber or tomentosus^ which have the stem scapiform and the 

 pappus paleae long and acuminate, gradually narrowed into 

 the setae, while in mollis the stem is leafy without a rosette 

 of leaves at base, and the pappus paleae are of the nudatus 

 type — short and broad and suddenly narrowed into the 

 setae. One of the most characteristic things about mollis is 

 its vestiture. In general the plant is rather thickly clothed 

 with rough hairs, but the lower surface of the leaves is soft 

 tomentose or almost villose, its feeling to the touch being very 

 characteristic. Bang's No. 497 of the PlantaeBolivianae seems 

 to be a typical example, though it was distributed as scaher, 

 which it is totally unlike. Rusby's No. 1105 from Bolivia is 

 near mollis but apparently not that — certainly not scaber for 

 which it was distributed. Morong's No. 258 from Paraguay 

 is likewise neither mollis nor scaber^ but apparently an inter- 

 esting new thing. It seems likely that Martii represents a 

 variety of mollis^ as it is closer to it in many respects. 



One of the .most distinct of all the South American forms 

 is hirtifiorus^ which has a scapiform stem and most remark- 

 able pappus, the limbs being flattened, not only at base but 

 throughout their length, making it distinct in this regard 

 from all other Elephantopeae . The whole plant is villose 

 throughout — even to the corollas. The heads are of un- 

 usual length and the floral leaves are unmodified, which is 

 unusual for species of this habitus. Likewise the coalesced 

 and consequently coroniform paleae of vaginatus separates 

 that species distinctly from all others. 



