PITTIER PLANTS FROM COLOMBTA AND CENTRAL AMERICA. 113 



Besides this, their habit is not exactly the same, the first plant being 

 voluble and more or less spreading, whereas in 0. cordifolium the 

 stems are ligneous, almost straight and simply climbing, with erect 

 branchlets. The leaves are decidedly heart-shaped in 0. cordifolium, 

 this detail being so marked as to have suggested the specific name, 

 whereas they are lanceolate in 0. riparium. The inflorescences of 

 this latter species are prevailingly axillary, whereas they are more 

 slender and elongate, and often terminal in the first. The shape of 

 the crown leaflets is not the same in both species, and lastly, they pro- 

 ceed from distant localities, 0. cordifolium from Bogota and 0. 

 riparium from Western Colombia, the altitudes being 2,000 and 2,000 

 meters, respectively. 



If there is only one West Indian species, its identity with 0. 

 riparium is very doubtful. Six flowers of Wright's no. 406 from 

 Cuba, showed uniform characters and the leaflets of the crown 

 would be defined as "late cuneatis, apice flabellatis distincte trilo- 

 batis," whereas Kunth describes the same part of his species as 

 "cuneatis apice dilatatis obliquis emarginatis aut bilobis, crenu- 

 latis ..." a definition well exemplified in the drawings given in 

 plate 231 of the cited work. The Cuban plant also differs in a 

 few other details, and moreover it is unlikely that a purely Andean 

 species would appeal' again near the sea level in the Northern West 

 Indies. 



In conclusion as to these two points it seems reasonable to admit: 

 First, that Oxypetalum cordifolium, (Vent.), not Schlechter and 0. 

 riparium II. B. K. are distinct species and, second, that 0. riparium 

 Kunth bears no relation to the Antillean species. I do not intend, 

 however, to insist here on this last proposition. 



Of 0. lindenianum Turcz. we can only say that its description is too 

 incomplete to admit of its being correctly placed among the other 

 species. Its apparent connections are with 0. riparium Kunth, and 

 it differs by several good characters from the new 0. Ivuilense Pittier. 



After much hesitation I have come to the conclusion that the form 

 collected in Costa Rica under no. 10826 (Instituto fis.-geog. Costa 

 Rica) corresponds to Gothofreda cordifolia Vent., which thus becomes 

 Oxypetalum cordifolium (Vent.) Schlechter, but exclusive of 0. riparium 

 Kunth and of the Antillean species. This species is carefully rede- 

 scribed hereafter, and the Guatemalan specimens distributed by 

 Captain Donnell-Smith are referred to it. 



Aside from the type station, as given by Kunth, Oxypetalum 

 riparium is indicated in two localities of the State of Vera Cruz in 

 Mexico. Until the original specimens have been examined anew or 

 new materials collected, it is not possible to decide whether this Mexi- 

 can form belongs to the Cuban or to the Central American species. 



« See note b, p. 112. 



