264 BOTANICAL GAZETTE [Marcu 
when in flower, 25 mm. with fruit; flowers very small, sepals 
broadly ovate, acute, about 1 mm., puberulous; petals ovate, acute, 
2.5 mm.; stamens about one-third the length of the petals, all of 
equal size; ovary deeply lobed, puberulous; capsule puberulous, 
thin papery, only one locule usually maturing, often with one or 
two abortive ones, locules divided almost to the base but still 
united; seeds not seen. 
Kava1.—Swampy forest on the high plateau of Waimea, elevation 4500 ft., 
collected in company with Mr. Francis Gay of Kauai, who knew the plant 
by its native name “ Kaleiohiiaka”; fruiting March 3, 1909, Rock, type no. 1072 
in College of Hawaii Herbarium; flowering September 1909, same locality, 
Rock no. 5285 in College of Feawati Herbarium 
HILLEBRAND’S specimen in the Berlin Makeun:, Knudsen no. 165, belongs 
here; a portion of this specimen is deposited also in the College of Hawaii 
Herbarium. 
PELEA CINEREA (Gray) Hbd. var. rubra, n. var.—P. oblongifolia 
Gray 8 var. (?) Hbd. Fl. Haw. Isl. 65. 1888; P. cinerea (Gray) 
Hbd. var. 6 Rock, not Hbd. in Indig. Trees Haw. Isl. 239. 1913.-— 
Shrub with rambling branches; leaves elliptical to elliptical-oblong, 
thick chartaceous, glabrous on both sides, dull green, acute at apex, 
mucronate, rounded at base or subcordate, 7 .5—9 cm. long, 3-4 -5 cm. 
wide, on petioles 1.5 cm. long; peduncle 3 mm., stout and quad- 
rangular; the single hirsute pedicel 2 mm.; flowers unknown; 
capsule hirsute with reddish hair, the cocci (separated). divided to 
the base, cohering only at the very base, 1.5 cm. long, nearly 1 cm. 
high, strongly nerved, endocarp hirsute, pale yellowish, and free. 
Hawaul.—Lava beds of Huehue, North Kona, fruiting June 6, 1909; 
Rock no. 3565 type in College of Hawaii Herbarium. 
This variety of P. cinerea was doubtfully referred to P. oblongi- 
folia by HItLEBRAND, who collected it in South Kona, Hawaii. Of 
his fruiting specimen in the Berlin Herbarium only one coccus of 
each of the two capsules is developed, the others are abortive. This 
might have misled him, as it cannot be determined clearly whether 
the cocci are cohering or not. In the writer’s specimen, which is 
identical with that of HILLEBRAND, the carpels are discrete and 
therefore must be referred to P. cinerea (Gray) Hbd. The writer 
had previously referred it to HILLEBRAND’s var. 6 of that species. 
It is apparently intermediate between P. cinerea and P. elliptica. 
