12 ROBIN? ON. 



of them nodding on curved or flexuous pedicels; the leaves are ovate 

 and apparently membranaceous. 



Entirely glabrous and at the same time thin-leaved members of 

 § Exivibricata are not very numerous. Finding a plant of this general 

 nature among the specimens collected by Mr. H. H. Smith near Santa 

 Marta, Colombia, the writer (Proc. Am. Acad. liv. 315) ventured to 

 place it doubtfully in E. Domhcyamim with which it appeared to have 

 many points in common. However, further study of this and the 

 related plants from Peru renders it decidedly unlikely that the Colom- 

 bian plant can have anything to do with the original E. Bomhcyamnn. 

 It is accordingly characterized below as a new species under the name 

 E. psilodorum. 



Similar efforts to identify with E. Dombeyanum DC. certain Peru- 

 vian plants, notably Weberbauer's nos. 860, 2766, and 3253, have 

 likewise failed. No. 860 (described below as E. stidophyUum), 

 while possessing rather closely the leaf-contour of E. Dombeyanum 

 has considerably denser and corymbiform inflorescences, the young 

 stems, branches, and pedicels are pulverulent-puberulent to a degree 

 that DeCandolle would scarcely have described as " Glaberrimum ;" 

 the achenes are covered with sessile glands and the petioles are 

 muriculate, which is clearly not the case in the type of E. Dombeyanum 

 of which there is a photograph in the Gray Herbarium; finally, the 

 leaves are rather conspicuously dark-punctate beneath — a feature, 

 which had it been equally manifest in the type of E. Dombeyanum, 

 would almost certainly have been mentioned by so careful a writer 

 as DeCandolle. 



Weberbauer's nearly related nos. 2766 and 3253 (below described 

 as E. simulans) differ from E. Dombeyanum in having lanceolate 

 (rather than ovate) leaves, which are pinnately veined rather than 

 3-nerved from above the base; the stems are much more leafy, with 

 mostly short internodes; the branches, petioles, pedicels, and achenes 

 are all perceptibly granular-puberulent, and the involucral scales are 

 more pubescent than is indicated by DeCandolle in his character of 

 E. Dombeyanum. 



After prolonged effort to take into account all characters and make 

 reasonable allowance for indiA'idual variation, it has seemed impossible 

 to refer any of these specimens to E. Dombeyanum. On the other 

 hand, they are so close as to give added strength to the view that the 

 species will ultimately be found in Peru, where the greater part of 

 Dombey's South American collecting was accomplished. 



E. (§ Subimbricata) drepanoides, spec, nov., fruticosum usque 



