204 DIAGNOSES SPECIERUM NOVARUM. 
Thee Dianthera sinensis, W. W.Sm. Sp. nov. 
WP 
Species affinis D. collinae, Clarke a qua sepalis glanduloso- 
pilosis inter alia signa differt. 
Planta herbacea circ. I m. alta basi procumbens nodis 
radicans, deinde erecta vel suberecta. Caules sulcati pilis cris- 
patis plus minusve dense induti. Folia petiolo ad 2.5 cm. longo 
crispato-piloso suffulta; lamina foliorum majorum 6-8 cm. 
longa, 4-4.5 cm. lata, ovata, apice acuminata vel nunc obtuse 
breviterque angustata, basi breviter et late cuneata, membra- 
nacea, laete viridis, utrinque pilis paleaceis conspersa, nervis 
tenuibus circ. 8-jugis. Flores in paniculam terminalem com- 
positi, albi, pro genere magni; axis dense albido-pilosa ; brac- 
teae ultimae et bracteolae lineares 2-3 mm. longae; pedicelli 
pilosi longiores ad 5 mm. longi. Calyx viridis pilis longiusculis 
glanduloso-capitatis patentibus se ee fere ad imum in 
lobos lineares divisus, circ. 6 mm. longus. Corolla circ. 2.5 
cm. longa, glabra, bilabiata, labio ‘inferiore maculis brunneo- 
flavidis ornato. Stamina 2 filamentis glabris antheris muticis 
altero altius affixo, polline globoso minutissime echinato. 
Ovarium glabrum stylo glabro, stigmate subcapitato obscurissime 
bilobulo. Fructus simillimus ei D. leftostachyae, Benth. circ. 
12 mm. longus, seminibus quatuor. 
“West China :—Shweli-Salween divide, Yunnan, in open 
situations amongst rocks. Lat. 25°20’ N. Alt. goooft. Plant 
of 3-4 ft. Flowers creamy-white. July 1918.” G. Forrest. 
No. 17,574. 
This species is closely allied to Justicia and the species of 
Dianthera which appear in Hooker’s Flora Brit. Ind., iv (1885), 
542. Inits large pedicelled flowers it comes nearest to D. collina, 
Clarke. The anther-cells are also superposed by almost the 
entire length of the anther. Later, in King and Gamble’s - 
Materials for a Flora Malay. Penins., Clarke instituted a new 
genus Leda in which he placed the Indian and Malayan species 
of Dianthera, keeping them apart from the American species. I 
do not find that D. sinensis quite conforms to the characters 
laid down by Clarke for Leda. The pollen of this species is 
globose, not ellipsoid, and minutely echinate. It has the pollen 
neither of Justicia nor of Leda. The material available of the 
allied species is not sufficient to warrant a definite decision and 
I am meanwhile relegating the plant to the genus Dianthera 
as characterised in the Genera Plantarum and the Flora of 
British India. 
Elsholtzia pygmaea, W. W. Sm. Sp. nov. 
Species affinis E. Bodiniert, Vaniot a qua habitu annuo, 
foliis sublinearibus, bracteis diversis inter alia recedit. 
