JOHNSTON. — NEW TLANTS FROM VENEZUELAN ISLANDS. 601 



Stamens unknown : carpels shortly 3-aristate, minutely puberulent, back 

 convex, areolate ; sides fiat; base narrowed; seeds triangular-reniform, 

 apex rounded, base acute, puberulent. — San Pedro, growing in the deep 

 sand near the graveyard, Coche, Johnston, no. 18, Aug. 5, 1903. Allied 

 to P. /iiimifusa, A. St. Ilil. 



Casearia spiralis, n. sp. Tree 5 m. high : leaves alternate, ellipti- 

 cal, 4 to 10 cm. long and 3 to 5 cm. wide, membranous, punctate with 

 pellucid dots and lines, crenate-serrate, indentations glanduliferous ; 

 apex shortly acute ; base obtuse or truncate ; petiole glabrous, 1 cm. 

 long ; stipule lanceolate-setaceous, 4 mm. long ; inflorescence fasciculate, 

 fascicles sessile or shortly pedunculate, compound, the branches bracteate 

 at the base and at the nodes ; bracts and bracteoles glabrous, as much as 

 1.5 cm. long; calyx corolline, 5- to 7-lobed ; lobes unequal, imbricated, 

 the smaller outside, pellucid-punctate, glabrous; corolla none; stamens 

 about 22, slightly perigynous ; connective of anthers produced into a 

 short acumen ; single, very short staraiuodia clothed with a whitish pubes- 

 cence alternating with the stamens: ovary ovoid, pubescent; the three 

 parietal placentas having many ovules; style short, thick; stigma 3-lobed; 

 fruit unknown. — El Valle : river trail, Johnston, no. 283, Aug. 30, 

 1903. 



The spiral arrangement of the sepals distinguishes this plant at once 

 from all other Cuseariae so far as can be determined from their descrip- 

 tions. According to Warburg in Engler and Prantl's Pflanzeufamilien, 

 III. 6a, 13 this spiral condition would place this plant in the Erytho- 

 spermeae, a group consisting of four African genera and a monotypic 

 Chilian genus. Though the plants of these genera have no staminodia, 

 some of them have scale-like appendages at the base of the inside sepals. 

 The plant under consideration, Casearia spiralis, has none of these ap- 

 pendages to the floral envelope, but has the staminodia characteristic of 

 the genus Casearia. In such groups as these in which the calyx is im- 

 bricated and varying in the number of its parts, from 4 to 6, the spiral 

 or cyclic character would seem to be a less distinctive characteristic than 

 the presence or absence of calyx-appendages or of staminodia. Thus the 

 plant seems to be more nearly affiliated to the Caseareae than to any of 

 the Erythospermeae , though showing affinities to botii. As to whether 

 it is actually a Casearia may be a question. 



None of the species of Casearia, so far as described, show any marked 

 irregularity of sepals or any definite spiral arrangement. Nor have any 

 of them as many as seven lobes and very few as many as 22 stamens, 

 which is the case in C. spiralis. Nevertheless, the form of anthers, the 



