208 
JOSEPH F. ROCK 
According to the right interpretation of Cyrtandra triflora Gaud, and 
C. polyantha C. B. Clarke, Hillebrand's variety grandifolia of his C. triflora 
cannot be retained, but must be referred as a variety to Cyrtandra laxiflora. 
Even should Hillebrand's determination or interpretation of Gaudichaud's 
C. triflora have been correct, it could not possibly have been retained as a 
variety of C. triflora Gaud., but would have had to be referred to C. laxiflora 
to which it comes so close as to make even its varietal rank questionable. 
It differs from C. laxiflora mainly in the smaller flowers and perhaps larger 
leaves. 
Cyrtandra polyantha C. B. Clarke in DC. Monogr. Phan. 5: 220. 
1883-1887. 
Cyrtandra triflora Hillebr. Fl. Haw. Isl. 332. 1888, in part, not Gaud. 
Cyrtandra gracilis Drake Cast. Illustr. Fl. Ins. Mar. Pacif. 7: 253. 1892, 
not Hillebr. 
A small shrub, i m. high; branches subterete, glabrous, young portions 
yellowish-villous to woolly; leaves opposite, of equal size, 12 cm. long, 
2.5-3.5 cm. wide according to type (5-7 cm. broad teste C. B. Clarke), 
closely denticulate, elliptical, acuminate at both ends, coriaceous, slightly 
scabrous above, shortly yellowish-silky beneath; cyme dichotomously 
branching, densely many-flowered; peduncle 5-8 mm., almost glabrous, 
bracts 6 mm., lanceolate; pedicels up to 8 mm., glabra te; calyx divided to 
the middle, the lobes 2 to 3 mm.; corolla 12 mm., cylindrical, hirsute or 
villous; ovary glabrous, the style partly glandular-pilose. 
Oahu: Mountains of Ewa, Hillebrand in herb. Kew, Berlin, and Gray 
Herbarium, clastotype in herb. College of Hawaii; main ridge of Niu Valley, 
elevation 1700 feet, flowering August 22, 1909, Rock no. 4815 in herb. 
College of Hawaii. 
The specimen in the Hillebrand Herbarium in the Berlin Botanical 
Museum bears in C. B. Clarke's handwriting the name C. polyantha C. B. 
Clarke, and in Hillebrand's handwriting the name C. triflora, Oahu, Mts. 
of Ewa. 
Drake Del Castillo is absolutely wrong in citing C. polyantha as a syn- 
onym of C. gracilis; the plant has nothing in common with it. In Cyrtandra 
gracilis the calycine lobes are very long and subulate, and the whole aspect 
of the plant is different ; besides the leaves in C. polyantha are densely silky- 
villous below. 
The writer's specimens came from an exposed ridge or spur in Niu 
Valley and agree quite well with the specimen marked as C. polyantha by 
C. B. Clarke in the Berlin Herbarium, with the exception of the peduncles 
which are up to 2 cm. in length. There is, however, no doubt that it is 
C. polyantha. 
