118 NEW SPECIES OF POLYGALA FROM SOUTHERN CHINA. 
tristichis carnosulis brevissime petiolatis ovalibus obtusis cum acumine 
opacis subtus paululum pallidioribus 3—7 lineas longis flavido-viridibus, 
racemis plerumque supra-axillaribus densissimis 12-20-floris folia 
zequantibus, pedunculo communi angulato, bracteolis minutis deciduis, 
floribus vix bilinealibus, alis glabris lanceolato-oblongis faleatis setaceo- 
acuminatis capsula bis longioribus flavo-viridibus lineis purpureo- 
brunneis percursis, carina cristata basi viridulo-alba apice cum crista 
clare et intense ezerulea, capsula lineam tantum longa ovali-compressa 
æqualiter angustissime marginata leviter emarginata margine vix 
ciliata, seminibus oblongis nigrescentibus dense albo-pilosis, arilli tri- 
lobi semine triplo brevioris lobis :equalibus. 
'This pretty little plaut was found by me in the beginning of October, 
1865, growing, not very abundantly, on a grassy hill on Danes' 
Island, Whampoa, and was very conspicuous by its yellowish tint, its 
dense clusters of flowers with variegated wings, and particularly by the 
beautiful deep clear blue of the crest, resembling in colour Anagallis 
cerulea, or some Gentian ; the foliage is not very dissimilar to that of 
the Algerian P. owycoccoides, Desf. Amongst the Chinese species 
own to me, it comes nearest to P.- glomerata, Lour., with living 
specimens of which I have compared it, but is readily distinguished by 
the stem branching all the way up, by the yellowish not bluish-green 
colour of its much smaller leaves, by the flowers being not half as 
large, arranged in longer and far denser clusters, with striped wings 
and a blue crest, whilst in the many specimens of P. glomerata I have 
seen, they have invariably been green-winged, and with a greenish- 
white keel ; and by the elliptic, less emarginate, narrow-margined and 
obscurely-ciliated capsule, which is less in size than one of the cells of 
that of Loureiro's species, the fruit of which is besides broadly orbicular, 
its greatest diameter being transverse, not vertical The nearest 
natural ally of my new plant is, however, evidently P. ¢elephioides, 
Willd. (nec Boiss. et Balansa), but that has the stem branched only 
from or at most just above the collum, the branches flexuose and 
shsrpiy-engid upwards ; its leaves are oblong and narrowed at the 
se; its racemes few-flowered ; the flowers still smaller and different 
T in colour; iis capsule broader, and seeds less densely silky. Both Mr. 
 Bentham and Prof. Miquel suggest that P. glomerata may be a form 
; arvensis, Willd., which, however, appears distinct enough, and 
Sn capsule, that of true P. glomerata 
