34 



MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



[VOL. 10 



BRAZIL : woods south of I.A.N., Belem, Para, Jan 1943, Archer 8204 (US); Glaziou 

 10126 (C, P) (type collection of S. glaziovii Engl.) ; Tepeirinha pr. Santarem, Iguape, Dec 

 1938, Marhgraf 3865 (RB). 



FRENCH GUIANA: Cayenne, Leprieur 1839 (G-DEL; ; sur la Montague en Mahuri 

 dans l'ile de Cayenne, Poiteau s.n. (L) ; Jul 1824, Poiteau s.n. (K); Maroni, ile Portal 1858, 

 Sagot 769 (BM); Maroni-en-fort, Aug 1910, Santini 7773 (L) ; "Cayenne, No. 12" (L) 

 (type collection of S. candolleanum Schott) ; "Guyane, herbier de M. Berbier" (P). 



PERU: Canchahuaya, Rio Ucayali, Loreto, Oct 1898, Ruber 1403 (MG) (type collection of 

 S. huberi Engl.). 



SURINAM: fluv. Saramacca sup., Feb 1922, Gonggrijp $ Stahel (Herb. Surin.) 5 (U) ; 

 Coppename R., Onobissi Kreek, Mar 1915, Sur. For. Bur. (B.W.) 1133a (U). 



VENEZUELA: iguarape forest of upper Rio Yaciba, Terr. Amazonas, Dec 1953, Maguire, 

 Wurdack, $ Bunting 36689 (NY, T T S). 



There is considerable variability in this species but the lack of discernible 

 pattern makes such variation of no taxonomic significance. Some specimens 

 referred to 8. humboldtii have bimerous flowers mixed with trimerous ones 

 on the same spadix. Engler apparently overlooked this feature in all specimens 

 except Huber 1403, which he described as type of 8. huberi. The flowers of 

 Hostmann 1154 are either bi- or trimerous in the same spadix, with four or five 

 perianth segments and a bi- or trilocular ovary. Notable variation also occurs in 

 the form of the base of the leaf -blade, which may be obtuse, rounded, or even 

 acute. 



Engler and Krause (Pflanzenreich 4 23B : 120. 1908) separated 8. candolleanum 

 from 8. humboldtii by its smaller size, ovary shape, and number of ovules per 

 locule. 8. glaziovii was set apart in that publication by the slightly different shape 

 of its leaf-blade and spathe, shorter petiole, and fewer ovules per locule. There 

 are few differences other than size and mery between Englers's descriptions of 

 St huberi and 8. humboldtii. I cannot find justification for maintaining these as 

 separate species, since their differences seem to fall within the expected natural 

 variation of a widespread species, in this case exemplifid by the differences be- 

 tween specimens of the type collection of 8. humboldtii. 



Jonker and Jonker (Fl. Suriname 1: 28-30. 1953) recognized both 8. hum- 

 boldtii and 8. huberi, separating them by the number of locules per ovary, and 

 ovules per locule, as Engler had done. I cannot maintain 8. huberi on these 

 grounds. Though Jonker and Jonker incorrectly considered the locality of 

 Huber 1403 to be in Para rather than Peru, their treatment 12 lends credence to 

 the possible conspecificity of the material collected in Peru and Suriname, and 

 suggests that 8. humboldtii might be widespread in the Amazon basin. 



The single Venezuelan collection identified as 8. humboldtii varies most from 

 the other material referred here, but it may simply represent a point in the 

 constant variance of this widespread species. Subsequent collections from the 

 Amazon drainage may prove it to represent an unknown species. 



8. humboldtii can be separated from all other species by its non-decurrent 

 spathe, long-stiped spadix, and elongate pistil with the style very prominently 

 exserted beyond the perianth. • 



Section 4. Spathiphyllum sect. Spathiphyllum. 



Massowia subg. Samowia K. Koch, Bonplandia 4:11. 1856. 



Spathiphyllum sect. Euspathiphyllum Engl, in DC. Monogr. Phan. 3: 220. 1879. 



1 -Jonker and Jonker considered two Suriname collections to be cohspeciflc Avith Huber 

 1403 (type of S. huberi). 



