E. L. EKMAN, WEST INDIAN VERNONIiE. 51 



unique e Martinique)). This specimen is in the Lamarck 

 Herbarium at Paris; no collector is given. Martinique, 

 Plee n. 957 (P). The specimen may be from Porto Rico, 

 where Plee also collected. Martinique, Duss n. 302 

 (KU); according to the label the plant was collected in 1890, 

 and since Duss that year collected in Guadeloupe, it is very 

 probable that his plant came from that island, inasmuch 

 as it agrees perfectly with other specimens from Guadeloupe, 

 for instance, Duss n. 2489. Martinique, Duss n. 4070 

 (KU). The label of this specimen reads: N:o 4070. Pere 

 Duss, Herbier de la Martinique. Vernonia longifolia Pers. 

 ex Gleason in Uteris. Casser coutelas. Haut de F/ 2 3*/ 2 

 m. fl. violacees ou violet pale ou blanches. Abondant, fl. de 

 Juin en 10-bre. Case Pilote. Alt. 2450 m. LeAoutl899. 

 According to this the plant should be common at Case Pilote 

 in Martinique. But then it is puzzling that it has not been 

 collected there more than once. Most probably the label in 

 question originally belonged to a specimen of V. arborescens 

 (L. ) Sw., and has been interchanged afterwards. Urban 

 states in his Notse biographical that there are many errors 

 of locality on the Dussian labels. Santa Lucia: leg. 

 Crudy (M); probably from St. Thomas, where Crudy also 

 collected. St. Vincent: leg. Anderson (K). The speci- 

 men resembles those from Guadeloupe, and may have been 

 collected there, or in another island near it. 



There is a specimen of V. albicaulis in the Paris Herba- 

 rium, the locality of which e Caroline, Th. Noisette , if 

 meant for the North American state by this name, clearly 

 must be an error. 



As indicated by the numerous synonyms, V. albicaulis 

 is highly variable. Outlying forms of it, such as V. punctata 

 Sw., V. longifolia Pers. and V. Sintenisii Gleason have been 

 described as distinct species. However, it is the same case 

 with V. albicaulis as with other Vernonia?, for instance, 

 V. arborescens (L. ) Sw. and V. divaricata Sw. : the whole set 

 of forms constitutes a variable, though always easily re- 

 cognizable, species with a particular geographical range, 

 each form considered as a species becomes impossible to 

 delimitate. The only consistent way is therefore to take 

 V. albicaulis as the wide species here conceived. 



It must be acknowledged, however, that the different 



