64 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 
Buddleia variabilis (F 424). — This is usually very poor in these 
parts, where it occupies the river-shingles in special profusion. 
However, about Kwanting, in the Nan Ho Valley, it seems to 
have longer spikes and smaller leaves and better colours than 
elsewhere. Accordingly, on the faint chance of such remaining 
constant, I have collected seed from the biggest -tailed bushes. 
Above Naindzai I saw one specimen which I believe to be a 
hybrid of B. variabilis X B. F 100. But B. variabilis does 
indeed earn its name, growing especially neat and small and 
dainty-leaved and brilliant in pure river silt (it would be well 
worth while to try to preserve such forms by similar cultiva- 
tion at home), and then developing in hedgerows &c. volumes 
of limp and flopping leafage, terminating in wretched drooping 
rat-tails of absurdly disproportionate, small, dull blossoms. In 
all this region it blooms in July-August. 
Callianthemum Farreri sp. nov. (F 73). — This, as collected in May from 
damp, cool ledges in the Satanee Alps, had low outspreading foliage, 
very glaucous and beautiful, with outlying stems of 2-3 inches, 
and very large flowers of a most melting China-blue, suggesting 
a discarnate Anemone blanda. The seed sent out under this 
number was collected from a plant of precisely similar habit, 
abounding in the earthier parts of the big limestone screes on 
Thundercrown ; I have none but a philosophic doubt that the)/ 
should prove identical. Seed is very hard to get ; the carpels 
fall while yet green, and you have to quest around each clump 
to detect the green nutlings lurking here and there in the chinks 
of the shingle ; and then the catching them becomes an agitating 
business, for if not caught at the first pounce they dive deeper 
and deeper among the pebbles every moment, and are soon 
completely buried from view. So hard are these wee nuts, and 
so evident their purpose in dropping prematurely, in order that 
the husk may wilt and rot below ground, and give the nucleus 
full time to sprout, that in the artificial conditions of the garden 
it would be well, I think, cautiously to split open the nut and 
extract the kernel to be sown. 
Caryopteris sp. (F 350). — This is a speciality of the J6-ni district, 
extending down the Nan Ho Valley, but on all the hot bare 
open loess hills about J6-ni making clouds of azure haze in August, 
and even flaunting its blue films from the crumbling walls of 
the dilapidated little city itself. It is a most lovely thing, and, 
as J6-ni stands at 8,000 feet, should be very happy in dry aspects 
in England ; it is always a low flopping mass of some 15 inches 
high, and I can by no means believe it not to be quite distinct 
from C. Mastacanthus. 
Celtis sp. (F 335).— This, I think, is the most beautiful deciduous 
tree I know, when well developed, having the look of a great 
many-trunked, smooth, grey beech, with the bark minutely 
parqueted with cracks and tiny peelings that bring a note of 
