208 



MEMOIRS OF THE NEW FORK BOTANICAL GARDEN 



[vol. 10 



divisis; calycis tubo 3-4 mm longo 5 mm lato; corollis 40-55 mm longis late 

 infundibuliformibus basi tubo 3.5-8 mm longo 4—5 mm lato extus dense sericeo- 

 veliitinis intus infra zonam barbatam 4—8 mm glabro vel papillato, supra tubum 

 angustum subito ampliatis, limbo 21-30 mm longo superne 18-30 mm lato extus 

 mollissime tomentuloso, lobis 5 late ovatis apice rotundatis 12-18 mm longis 



10- 19 mm latis extus tomentulosis intus glabris uno lobo basi puberulo excepto ; 

 staminibus inaequalibus duobus brevioribus, antheris 5-7 mm longis, filamentis 



11- 16 mm longis. 



As the original description of H. oblonga was inadequately drawn from fruit- 

 ing material, a more complete emended description, supplemented by flowering 

 material, has been "riven above. 



Distribution. Southwestern Venezuela, Territorio Federal Amazonas, along 

 the Rio Atabapo and Carlo Catua and tributaries. 



VENEZUELA. Amazonas: frequent on the Rio Atabapo. an affluent of the 

 Orinoco near the mouth of the Casiquiare ; small tree of about 18 ft in height, 

 with spreading branches; Jun 1854, Spruce 3702 (photo NY) ; along Cano Catua 

 at northwestern base of Cerro Yapacana ; occasional ; small or medium riverine 

 tree ; leaves 4 per node ; buds pink ; sepals pale tinged pink ; corolla flushed pink 

 with orange-red spots and median line within ; alt. 125 m ; 19 Nov 1953, Maguire, 

 Wurdack, t(- Bunting 36557 (NY) ; Atabapo, cerca de la desembocadura del Rio 

 Atacavi ; 4 Sep 1960, Foldats 3706 (VEN). 



The Maguire, Wurdack, &' Bunting 36557 collection differs in several details 

 from that of Foldats 3706 in having 4-verticillate leaves with much longer, more 

 puberulent petioles 15-30 mm long, no apparent petiolar glands, larger corollas 

 48-55 mm long instead of 40-42 mm, and anthers 5 instead of 7 mm long. The 

 larger corolla appears to be due to harder pressing. The longer, more puberulent 

 petioles may prove at a future time to have taxonomic significance, but, at present 

 writing, are being interpreted as within the limits of variability manifested by 

 this taxon. 



All the specimens examined of H. nitida show predominantly large oblong 

 leaf-blades broadly rounded or subtruncate to broadly obtuse at apex, often 

 retuse or mucronulate, and with the lateral nerves on the lower leaf-surface 

 prominent or manifest. When Spruce described H. oblonga (loc. cit. 297), based 

 on his number 3702 from the Rio Atabapo of southwestern Venezuela, he stated 

 that it differed from H. nitida in having smaller leaf-blades with relatively longer 

 petioles, and less shining leaf-blades, paler below and with the lateral nerves 

 scarcel/ evident. He admitted, however, that his H. oblonga, collected only in the 

 fruiting stage, might prove to be possibly a variety only of H. nitida, which he 

 collected in the flowering state. Two recent collections, both in anthesis, which 

 I have judged to be conspecific with H. oblonga, are Foldats 3706 from near the 

 type locality of the Rio Atabapo station and Maguire, Wurdack, <f- Bunting 36557. 

 Both collections show the smaller leaf-blades, paler beneath with the lateral nerves 

 of the lower leaf surface scarcely developed and not obvious, and match the 

 characters given in the original description as well as the photograph of the 

 holotype from Kew Herbarium deposited in The New York Botanical Garden 

 Herbarium. The leaves of the Foldats 3706 collection are 3-verticillate, those of 

 the Maguire, Wurdack, & Bunting 36557 collection 4-verticillate. Apparently, 

 3-verticillate leaves are represented by the holotype specimen photographed at K. 



Although it might appear that these smaller-leaved trees with the lateral 

 nerves scarcely evident on the lower surface of the leaves represent a distinct 

 species, intergradation occurs in leaf size between H. nitida and H. oblonga. No 



