BALFOUR—NEW SPECIES OF RHODODENDRON. 81 
Filaments of stamens puberulous. 
ae ae mee ae ae floccose 
Core lla . tublar-campanulate dark black- 
rimso haemaleum. 
Pedicels ae ovaries glandular. 
Filaments of stamens puberulous, 
Calyx-lobes setulose - glandular on back, 
gined ro . : i : é . roseotinctum. 
Scale-leaves of the foliage-buds persistent after 
Pedicels and ovaries Scart a 
Filaments of stamens glabro 
Calyx-lobes yellow membraiious deciduous, 
Corolla tubular-campanulate rose. ‘ cloiophorum. 
Pedicels and ovaries glandular. 
Filaments of stamens puberulous 
Calyx-lobes greenish membranous setulose- 
glandular and floccose on back, setulose- 
gland-ciliate. Corolla campanulate bright ci 
lemon-yellow (rose-coloured sometimes) .  citriniflorum. 
These plants are all dwarf alpines growing on ledges of clifis 
or amongst boulders in open situations at elevations of 13,000- 
14,000 ft. north of 28° lat. The type occurs also further south, 
sometimes in plants of larger growth. To it belongs Rh. nerii- 
florum, Franch.—an aggregate seemingly which awaits analysis— 
with a centre of distribution on the eastern flank of the Tali 
Range, lat. 25° 40’ N., at elevations from 9000-11,000 ft., and 
another centre on the Shweli-Salween divide in lat. 25° 30’ N. 
at I10,000-11,000 ft. elevation. Rh. dichroanthum, Diels, repre- 
sents it also on the eastern flank of the Tali Range at elevations 
of gooo—11,000 ft. in lat. 25° 40’ N. On the western flank of 
the Tali Range at 10,000 ft. in lat. 25° 40’ N. there is Rh. 
dimitrum, Balf. f. et Forrest, and further south comes RA, 
apodectum, Balf. f. et W. W. Sm., at 10,000-11,000 ft. elevation 
in lat. 25° 30’ N. on the Shweli- Salween divide. On the Yung- 
peh Mountains, at an elevation of gooo ft. in lat. 26° 45’ N., 
occurs Rh. pilovittatum, Balf. f. et W. W. Sm., an exceptionally 
tall relation often ro ft. high. Then in Eastern Upper Burma, 
on the Nwai divide, at 12,000-13,000 ft., there is the dwarf- 
Spreading Rh. herpesticum, Balf. f. et Ward. We have here a 
number of species which seem to constitute a’ natural group of 
Rhododendrons, and may expect that future exploration will 
add to the number. 
At the same time we must not overlook many near relation- 
ships to other known species so close as to suggest that a correct 
