28 ARKIV FOR B0TANIK. BAND 13. N:0 15. 



has its origin there, though in the West Indies showing a 

 particular and independent evolution. 



Within the vast assemblage of species constituting the 

 Arborescentes smaller groups can be distinguished. They 

 do not appear as sharply definable taxonomic units, the exist- 

 ence of which could not be denied; on the contrary, they are 

 but vaguely indicated, and very difficult to characterize. Sub- 

 jective susceptibility may also play some part in the percep- 

 tion of these delicate taxonomic sensations. However, con- 

 sidering every method of arranging species according to their 

 natural relationships preferable to an artificial arrangement, 

 I have tried to give an exposition of the Arborescentes based 

 on their supposed relationships. The species have been 

 arranged according to characters about the value of which 

 for recognizing groups in Vernonia very little is known, such 

 as consistence and relative length of the pappus, pubescence 

 of the corolla, length of the anthers etc. Or, mutatis mu- 

 tandis, the groups recognized have been found to differ in 

 those points. Now such characters are not convenient for 

 identifying specimens the names of which are required. I 

 have, therefore, given an artificial key in order to facilitate 

 such identifications. 



The smaller groups concerned have not been named. 

 They are based upon too small a material and are, as already 

 stated, too vaguely indicated to deserve that certificate of taxo- 

 nomic importance which a name betokens. However, a few 

 words upon their characteristic features and their geogra- 

 phical ranges may be advisable. 



1. The allies of V. argyropappa Buek. 



One species of Trinidad. As already mentioned it ap- 

 proaches very nearly to certain Mexican and South American 

 species, the best known of which is V. argyropappa Buek. 

 Another member of this group is the widely distributed, annual 

 V. remoti flora L. C. Rich. By their floral characters these spe- 

 cies cannot be distinguished from the allies of V. arborescens 

 (L. ) Sw. The differences are chiefly habitual ones, such as 

 larger involucres, subulate, spreading scales, the outer of which 

 are slightly recurved, etc. 



The striking resemblance between V. Trinitatis and 

 V. arborescens has made me believe that the latter has taken 

 its origin from V. Trinitatis or from some species nearly allied 



