Dr. Arnott on the Genus Rhizophora. 371 



genus by the one-celled ovarium, agreeing in that respect 

 with Kandelia only, but in other points differing widely. To 

 this genus (Carallia) I have in the Nova Act. Ac. Nat. Cur., 

 xviii. p. 334. added two species, one which appears to be the 

 Diatoma bracchiata of Loureiro, the other from Ceylon. This 

 last I have also received lately from Dr. Wight, collected at 

 Quilon on the Malabar coast, accompanied with specimens, 

 procured at the same place, of another form which Dr. Wight 

 considers perfectly distinct and has named C. corymbosa : this 

 is almost intermediate between the Chinese and Ceylon plants, 

 if indeed the three be not varieties of one species. They all 

 have leaves with recurved margins, which although some- 

 times obscurely crenulate or serrulate towards the point, may, 

 in comparison with Roxburgh's species, be said to be quite en- 

 tire ; about seven orbicular slightly retuse but apiculate pe- 

 tals, broader than long, and cut and curled on the margin ; a 

 stigma of usually four, but occasionally five or seven emargi- 

 nate lobes, and the teeth of the calyx are about equal in length 

 to the petals ; they may be distinguished from each other as 

 follows : 



1. C. Ceylanica, (Am.) ; foliis cuneato-obovatis obtusissimis subretusis la- 

 titudine subdimidio-longioribus. — Am. I. c. ; Wight. Cat. n. 2446 — 

 C. obcordata, Wight, in litt. 



2. C. corymbosa (Wight) ; foliis oblongo-obovalibus obtusis vel obsoleto 

 ac obtuse acuminatis latitudine 2 — 2^-plo longioribus. — Wight. Cat. 

 n. 2447. — Rheed. H. Mai. v. t. 13, (inflorescentia ac floribus pessime 

 delineatis.) 



3. C. Sinensis ( Arn.) ; foliis cuneato-obovatis breviter ac obtuse acuminatis 

 latitudine subduplo longioribus. — Arn. I. c. (cum syn.) 



These obviously differ very slightly : in C. Ceylanica and C. 

 Sinensis the leaves are decidedly cuneate at the base, the mar- 

 gin presenting a rather concave curve ; in C, corymbosa the 

 cuneiform appearance is less evident from the curvilineal mar- 

 gin being usually slightly convex. What C. integerrima, DC. 

 is, or from what part of India it was obtained, I have not ascer- 

 tained ; if distinct from all the above, it may be thus defined : 



4. C. integerrima (DeC.) ; foliis ovalibus subacuminatis latitudine duplo 

 longioribus. 



C. Baraldeia,W .&A. {Baraldeia MadagascariensiSyVet.Th.) 

 has slightly serrated leaves, and according to the author's de- 



2 b 2 



