1960] 



SPATHIPHYLLUM 



21 



Type collection: C. C. Parker s.n., "Demerara," British Guiana (lectotype 

 K). Apparently a unicate, collected in 1824. A Splitgerber collection from 

 Suriname was also cited by Schott in the original publication of this species, but 

 it has not been seen, or even located. 



Distribution : Swampy forested areas of British Guiana. 



BRITISH GUIANA: Mazaruni R., Appun 230 (K) ; Appun 848 (K) ; Smote Creek, Waini 

 R., Jul 1906, Beckett $ Kort right 8487 (K) ; Kaow Creek, Mazaruni Station, Jul 1939, 

 B. G. For. Dept. 2976 (K) ; Potare, Mahdia R., Jan 1943, B. G. For. Dept. 3904 (K), 3909 (K) ; 

 in swamp, Huradaiah, Moruka R., Pomeroon Dist., Aug 1921, de la Cruz 1026 (GH, NY, US) ; 

 Pomeroon R,, Pomeroon Dist., Jan 1923, de la Cruz 2940 (F, GH, MO, NY, US) ; Pomeroon 

 R., Dec 1922, de la Cruz 3128 (F, GH, US) ; Waini R., Northwest Dist., 8° 20' N, 59°40' W, 

 Apr 1923, de la Cruz 3610 (F, GH, MO, NY, US) ; along creek, Kartabo Station, junction 

 of Mazaruni R, & Cuyuni R., Jul 1924, Graham 310 (NY) ; swamp in forest by Kawow Creek, 

 Penal Settlement, Dec 1919, Hitchcock 17235 (GH, NY, S, US) ; Hooker s.n. (NY) ; Pomeroon 

 R., Mar 1884, Jenman 1985 (K) ; Demerara R., Jan 1896, Jenman 6868 (K) ; Barima R., Mar 

 1896, Jenman 6989 (K) ; Kamakusa, Nov 1922, Leng 49 (NY) ; Kamakusa, Dec 1922, Leng 

 353 (NY); wet places, Comaka, Demerara R., Mai 1923, Persaud 269 (F) ; in mora swamp, 

 Moraballi Creek, near Bartica, Essequibo R,, Aug 1929, Sandicith 153 (K). 



Engler and Krause (Pflanzenreich 4 23B : 128. 1908) cited Jenman 6989 as 

 S. cuspidal urn, but incorrectly stated its locality as "Roraima." It is a mixed 

 collection from along the Barima River; the New York sheet represents S. 

 cannaefolium, the two Kew sheets 8. cuspidatum. Engler described the locules 

 of the ovary of this species as biovulate ; my observations are that there are 4—2 

 ovules per locule. 



S. cuspidatum appears to be limited to British Guiana. However, Jonker & 

 Jonker (Fl. Suriname 1: 28, 29. 1953) included this species in their treatment 

 on the basis of a Splitgerber collection from Sarumassa, cited by Schott (Prodr. 

 Syst. Aroid. 428. 1860). Although no subsequent investigator appears to have 

 seen that specimen, it is possible that 8. cuspidatum occurs in Suriname. 



8. cuspidatum may represent a terminal element in phylogenetic develop- 

 ment in this genus. It varies very little, and is common within narrow geo- 

 graphical limits. It is most similar to 8. kalbreyeri, from which it differs in its 

 lanceolate or elliptic spathe which is somewhat broader than in the latter species. 

 There is also a tendency toward a slightly greater number of ovules per ovary 

 than is found in 8. kalbreyeri. 



Of considerable interest is the possible relationship between 8. cuspidatum 

 and S. humboldtii. which latter species exhibits such diverse ovary characteris- 

 tics as to be placed in a separate section. It is a reasonable conjecture that these 

 two species, as well as 8. gardneri, may have arisen from a common primal stock. 



8. S. kalbreyeri Bunting, sp. nov. Figure 5. 



Foliorum lamina e oblongo- ad laneeolato-ellipticam, 33-40 cm longa et 9-10.5 

 cm lata, ad apicem acuminato-cuspidata, ad basim acuta ; venis primariis multis 

 angulo 35-45° divergentibus ; petiolus 32-43 cm longa, ad % vel plus vaginatus; 

 geniculum ca. 2 cm longum. Pedunculus 40-58 cm longus; spatha oblonga, 13.5- 

 15 cm longa et 2.5-3 cm lata, apice attenuato-acuminato, basi acuto et leviter 

 complexo ; spadix 7-8.5 cm longus, stipite ca. 1 cm longo ; perianthii segmenta 

 separata; pistillum obpyramidale, apice truncato, solum stigmate papilliformi 

 perianthium superanti; ovarium 3-loculare, loculis 4—2 (-1 ) -ovulatis, ovulis ad 

 pulvinum subbasilarem in loculo quoque affixis, 9-5 in toto ovario; fructus 

 ignotus. 



