THE GEEWIAS OF ROXBURGH 359 



= G. flavescens Juss. in Ann. Mus. iv. 91 {—G. commutata DC. 

 Prodr. i. 511). This is also = G. carinnifoUa Mast, in Hook. fil. 

 Fl. Br. Incl. i. 387, non Juss., and in part = G. pilosa Mast, in 

 Hook. fil. FI. Br. Ind. i. 388, non Lamarck. Lamarck's plant was 

 simply G. orientalis L. Sp. PI. ii. 964 (excluding synonyms from 

 Eheede, Hort. Malab.), which is not G. orientalis Eoxb. Fl. Ind. 

 ii. 586, nor of Masters in Hook. fil. Fl. Br. Ind. i. 384, but is 

 G. columnaris Smith in Bees Cycl. xvii. G. caijnnifolia Juss. 

 is a purely African type, characteristic of the Guinea coast, 

 extending to other parts of West Africa, but unknown, so far as 

 present information goes, in East x'Vfrica and Asia, G. flavescens 

 Juss. is a widely distributed species in the dry hot zone of the 

 Old World from Senegambia to Coromandel. 



(2) No. 2, Fl. Ind. ii. 584, " scabrojjJujUa E." The proper 

 citation is G. scleropliylla Eoxb. ex G. Don, Gen. Syst. i. 550 

 = G. sclerophylla Eoxb. of Prain's Bengal Plants, i. 283 := Eoxb. 

 tab. ined. No. 1859. 



(3) No. 10, Fl. Ind. ii. 587, " tilicefolia Vahl, Symb. i. 35," is 

 in part only =: the true tilicefolia of Vahl, which is a large tree 

 abundant in S. India, but not found so far elsewhere ; with this 

 Eoxburgh included certain other races, of which one may be the 

 spontaneous form of G. subincequalis DC. Prodr. i. 511. G. tilice- 

 folia Masters in Hook. fil. Fl. Br. Ind. i. 386 also includes forms 

 which are not referable to Vahl's species. 



The Behar habitat in Bengal Plants, i. 283, almost certainly 

 belongs to a distinct though allied species. 



(4) No. 5, Fl. Ind. ii. 586, " G. excelsa Vahl, Symb. i. 35." 

 This was doubtless = G. vestita Wall. Cat. 1105, which again is 

 probably a form of G. celticlifoUa Juss. Ann. Mus. iv. 93. 



G. excelsa Vahl belongs to a different section of the genus, and 

 is a native of the Arabian peninsula, where it was discovered by 

 Forskal, who described it (Fl. iEgypt. 105) as Gliadara arborea = 

 Grewia arborea Lamarck, Encycl. Meth. iii. 45 (1789) ; having lately 

 been rediscovered by M. Deflers it was redescribed and named by 

 him as G. clubia (in Bull. Bot. Soc. de France, xlii. 30). 



G. excelsa of Prain, Bengal Plants, i. 283, as regards the Chota 

 Nagpur and Singhbum localities is probably = G. Rothii DC. 

 (Prodr. i. 509) : the Chittagong habitat belongs to " G. vestita 

 Wall." G. excelsa Mast, in Hook. fil. Fl. Br. Ind. i. 385 consisted 

 partly of G. Bothii DC. and partly of G. vestita Wall., so far as 

 the Asiatic material is concerned : the African citation relates to 

 a quite different species. The true excelsa seems to l)e restricted 

 to Yemen, and G. Bothii to southern and central India. The 

 G. excelsa of Masters (non Vahl) was identified by him as the 

 G. salvifolia of Eoxb. Fl. Ind. ii. 587, which is G. Bothii DC, and 

 as regards the bulk of his material this is correct, but the G. sal- 

 vifolia of Eoxburgh (in the Flora Inclica) is not the G. salvifolia 

 of the younger Linne (which was an Alangiuvi), nor of Heyne 

 (ex Eoth, Nov. Spec. 239) which is G. Damine Gaertn. (Do Fruct. 

 ii. 113, t. 106). G. salvifolia of Masters (Fl. Br. Ind. i. 386) is in 

 part ^^ G. Damine Gaertn. " (r. salvicefolia Heyne" of Prain, 



