112 A DECADE OF NEW CHINESE PLANTS. 
After full consideration I do not think that this plant, which is re- 
markable by its elliptic, usually 1-seeded silique, can be regarded 
otherwise than as a somewhat aberrant Cardamine ; for, in the first 
place, the foliage is precisely similar to that of C. Africana, L., and 
secondly, the American C. rotundifolia, Mx., not unfrequently has a 
much abbreviated fruit; whilst all botanists are familiar with the great 
differences in this respect between very close relatives in the allied 
genus Nasturtium, especially in the section Brachylobos,—as, for 
instance, N. terrestre, R. Br., and N. Niloticum, Boiss. 
v^ 3. Pterospermum (Velaga) heterophyllum, m. sp.; arbor 40-50- 
pedalis, ramulis petiolisque rufo-tomentosis, foliis arboris juvenilis 
atque turionum peltatis basi conspicue truncatis circumscriptione sub- 
semiorbiculatis profunde 5-palmatis 8-11-nerviis 9-12 pollices diametro 
petiolo limbo subzequilongo suffultis reliquis basi obliquis rotundatis 
truncatis vel obscure cordatis 4—6-nerviis nervis infimis marginalibus 
ac preterea costulatis nervis omnibus subtus prominulis oblongis v. 
ovato-oblongis integerrimis caudato-acuminatis supra glaberrimis sub- 
tus dense fulvescenti-tomentosis 3—7-polliearibus petiolo semipollicari 
haud peltatim affixo suffultis, stipulis oblongis acutis petiolo parum 
brevioribus (floribus non visis), capsula crasse stipitata cylindracea 
acuta 2$ poll. longa furfure rufo dense tomentosa, seminibus superne 
in alam iis triplo longiorem productis in quovis loculo quaternis.—In 
silvis densis cirea coenobium buddhisticum Fi-loi-tsz, prov. Cantoni- 
ensis, d. 18 Sept. 1866, detegerunt Sampson et Hance. (Exsice. n. 
am very imperfectly acquainted with this genus, but the Indian 
peninsular species have been so well and carefully described by Wight 
and Arnott, that there can be little doubt of the present one being dis- 
tinct; and, though Miquel has unfortunately given very incomplete 
diagnoses of the Archipelagie species, it seems equally to differ from 
any of them. 
yide 4. Abrus Cantoniensis, n. sp.; diffusa, 13-9-pedalis, ramulis petio- 
lisque strigoso-pilosis, foliis 8-11 jugis foliolis oblongis v. obovoideis 
obtusis setaceo-apiculatis 23-43 lin. longis 11-21 lin. latis utrinque 
elevato-reticulatis supra parce pilosulis subtus appresse strigosis, race- 
mis axillaribus brevibus, floribus 3-linealibus rubellis vexilli ungue 
limbo 4-plo breviore vaginam stamineam semiamplexante, leguminibus 
oblongis compressissimis pilosulis 4—5-spermis.—Ad radices montium 
