DR. O. STAPF ON THE FLORA OF MOUNT KINABALU. 151 
nutans, sub anthesi strictus, 1 poll. longus, gracilis, fulvo-tomentellus, basi atque ad 
medium vel supra vel basi tantum bracteolatus. Sepala vix connata, triangularia, 
1-13 lin. longa, extus fulvo-tomentella. Petala exteriora ovata, 5-6 lin. longa, 3—4 
lin. lata, obtusiuscula crassa, marginibus intus depressis, fere 1 lin. latis, extus pilis 
adpressis densissime fulvo-subsericea, intus minute griseo-puberula; interiora 
ovata, 2 lin. longa, acuta, valde convexa, extus griseo-fulvo-tomentella, intus 
glabra. Stamina numerosa; filamenta brevissima; antherarum loculi contigui ; 
connectivum glabrum, ovatum, apice inflexum. Ovaria numerosa, oblique ovoidea, 
fulvo-hirsuta; stylus brevis; stigma pilis superatum; ovula 4 in serie unica. 
Penokok, alt. 3000 feet (Haviland, 1310). 
This species is closely allied to Melodorum fulgens, Hook. fil. The indumentum is 
exactly of the same character as in M. fulgens, but less shining on the outer petals. 
VIOLACEJE, 
VIOLA SERPENS, Wall. in Roxb. Fl. Ind. ed. Wall. ii. 449, and in DC. Prodr. i. 256; 
Hook. fil. & Thoms. in Hook. fil. Fl. Brit. Ind. i. 184. 
On the Dahombang River, alt. 3000 feet ( Haviland, 1275). 
Distribution : Throughout tropical Asia, except the Philippines and East Malaya; 
temperate Himalaya. 
According to Dr. Haviland the flowers are slightly scented and white, and the lip 
streaked with violet. The leaves are perfectly glabrous, even in a young state, and the 
stipules fimbriate. The valves of the mature capsule are about 4 lin. long; besides, 
there is a capsule on Dr. Haviland's specimens which is globose and 23 lin. long. It 
developed evidently from a cleistogamic flower. I think, if we take Y. serpens in the 
broad sense of the *Flora of British India, we should include also V. distans, Wall. 
 '[he characters taken from the fruit are, in my opinion, reducible to heterocarpy, the 
capsules derived from normal flowers being always more or less oblong, those from 
cleistogamic flowers globose. A similar heteromorphy seems to exist with regard to the 
shape of the stigma, as Hook. fil. & Thoms. have suggested. But, apart from this, it is not 
unlikely that several species of close affinity are included in V. serpens, taken in this 
broad sense. Yet it seems to me quite hopeless to decide that question from dry 
material only. In any case, it is evident that one species—if not several nearly 
related species—is spread from the temperate region of the Himalaya and from South 
China all over the mountain-region of the Western and Eastern Peninsula, of Ceylon 
and Malaya, and that this species, or this group of species, is most closely allied 
to forms which are typically Boreal. 
POLYGALACE.E. 
POLYGALA VENENOSA, Juss. in Poir. Dict. v. 493. 
On the Tampassuk River near Kiau, from 2000 to 3400 feet (Low, Haviland, 1349). 
Distribution : West Malaya to the Philippines. 
