130 
under the name of Cacara litorea (t. 141), C. gladiata as Lobus 
machaeroides (t. 135). C. turgida which occurs in Amboina but 
was not sanggeer figured by Rumphius was identified by 
Merrill with t. 141. Im 1784 Thunberg publishe ed his Flora 
Japonica i in Ponta C. lineata, the only wild species in the eae 
is recognised under the name of Dolichos lineatus (p. 280) a 
growing on the sea-shore round Nagasaki, while the ulti 
Natta mame, appears as Dolichos ensiformis (p. 279). 
It was now 50 years before any further additions were made 
to the genus by workers in the field. But meanwhile the group 
had been recognised as distinct from Dolichos by Adanson 
in 1763 (Adans. Fam. ii. 325) and established as Canavali, 
enumerating most of the characters now relied on to distinguish 
it from neighbouring genera. Under a Latinised form, Canavalia, 
Aug. De Candolle collected all the species known in 1825 (Prodr. 
ii. 404) and defined Adanson’s genus with greater accuracy 
(Mem. Leg. 375). It was about this time that various botanical 
writers, endeavouring to reconcile the figures, descriptions and 
C. obtusifolia. And it is largely this mistake and its prolific 
progeny which has impelled us to undertake a critical review 
of the whole subject and to offer in the following pages our final 
conclusions on the question of the nomenclature of the species. 
It is not necessary to lay stress on the progressive stages of the 
confusion. It may be clearly traced in our synonymy of each 
of the older species. 
In 1832 Roxburgh (Fl. Ind. iii. 300) enumerated all the 
species known to under Dolichos, though two of the maritime 
forms are probably included under the name of D. rotundifolius, 
Vahl under a misapprehension as to the true meaning o 8 
name. ears later Wight and Arnott (Prodr. ii. 253) 
correctly deecribed the 8. Indian species under Canavalia, but 
in endeavouring to adjust the synonymy introduced the prolific 
source of confusion centring on the inclusion of Lamarck’s name 
of Dolichos obtusifolius as a synonym of the sea-shore species. 
Distribution.—The most widely distributed member of the 
genus is C. rosea, which encircles the globe, abounding on the 
hot, sandy shores of the tropics and seldom found beyond them. 
Away from the sea it is represented in the interior of Africa 
y C. regalis towards sw north and by C. ferruginea in the south. 
Another allied species C. plagiosperma has been described as @ 
cultivated plant probably from Mauritius. Just north of the 
range of C. rosea on the China coast we find its near ally 
C’. obcordata, in the 8. Pacific on the very south of its range it is 
represented by the closely allied C. Baueriana in Norfolk Island, 
while in India and Ceylon it is replaced by C. podocarpa. 
