E. L. EKMAN, WEST INDIAN VERNONLE. 5 



Most probably Vernonia lepidota Grisebach is to be made 

 the representative of a new genus. 



Vernonia Milleri Johnston from the island of Margarita 

 was described in the year 1905. It is evidently an Oliganthes, 

 having few-flowered heads, narrow involucres, and broad, 

 somewhat curled scales of the inner series of pappus. 



As for the delimitation of the sections of the genus, I have 

 adopted the treatment given by Hoffman in Engler and 

 Prantl: Die Naturlichen Pflanzenfamilien, with the except- 

 ions already mentioned. Save Piptocarpha tetrantha Ur- 

 ban, and some introduced species, all West Indian Verno- 

 nia) belong to the section Lepidaploa (Cass.) DC. 



Within this section Lepidaploa the species have been 

 before exclusively artificially grouped. The first attempt 

 to arrange the species according to their relationships was 

 made by Gleason in his Revision of the North American 

 Vernonieae. He correctly recognized several groups of 

 species, for instance, the Fruticosce and the Argyropappce, 

 but most of his other groups include species of different ori- 

 gin. He also has proceeded too far in splitting up the divi- 

 sions of Lepidaploa, placing, for instance, V. albicaulis Pers. 

 ( V. longi folia Pers.) in a particular group Longifolice, 

 separated from its nearest relative, V . icosantha DC. (= V. 

 arbor escens Sw\ of this paper). Especially unfortunate is 

 he in arranging the forms of V. divaricata Sw. , some of which 

 he takes to be distinct species, putting them into three 

 different species -groups (V. arborescens Gleason, not Sw. 

 among his Arbor escentes, V. albicoma among the Divaricate, 

 V. permollis and V. intonsa constituting a particular spe- 

 cies-group). He is, further, absolutely wrong in assigning V. 

 lepidota Griseb. to the Scorpioidece reductce, V. pallescens 

 Gleason to the Havanenses, V. canescens H. B. K. to the Dep- 

 peanaz, many other mistakes not to be mentioned. 



I have tried, in this paper, to arrange the species accord- 

 ing to their relationships in subsections based in the first 

 place on the characteristics of the inflorescense, in the second 

 place on the pubescence of achenes and corollas and on the 

 structure of the pappus. The arrangement proposed is by 

 no means definitive. All delimitation of taxonomic units 

 in Vernonia is extremely difficult, from that of the species 

 up to the genus. 



