THE GREWIAS OF ROXBURGH 335 



p. 511); the writer suspects that this may have been in part the 

 ^'arborea" of Eoxburgh, but as Mr. Barber's plant was a low shrub 

 analogous in habit to G. sapida, it is out of the question to make 

 any definite assumption on the subject. There is a small tree, 

 said to be plentiful in the Sewalik region from the Jamna to 

 North-west Bengal, which has drupes closely resembling those in 

 the drawing of " O. tilicBfolia," but the writer can find no sufficient 

 proof so far that this form was ever seen by Eoxburgh. ''Arborea" 

 as a name for any of the asiatica-tiliafolia group could not stand 

 in any case, for it is appropriated, as already stated, to the true 

 "excelsa" of the Symbolce. 



The practical deduction is that, while the tilicsfolia of the Flora 

 Indica was meant to include Vahl's species, it included at least 

 one other form. We cannot be sure that it covered only one 

 other, and, if only one, we are at present in the dark as to its 

 identity. 



To return now briefly to G. asiatica. Eoxburgh had seen the 

 true Linnean plant, but under this again the Flora Indica merges 

 at least two distinct forms, as indeed the description might have 

 led us to expect ; in any case, there are Eoxburghian examples of 

 " asiatica " from the Calcutta Garden in the Kew Herbarium that 

 are plainly referable to a distinct form which the writer supposes 

 to be the subincBquaUs of De Candolle. 



As regards tilicefoUa, the substitution by Eoxburgh of Vahl's 

 name for his own " arborea " will not bear the stress which Wight 

 and Arnott's remarks would lay upon it. If we had the letters 

 which Eoxburgh received from his mentor in England we should 

 probably attach much less importance to these amendments than 

 has sometimes been assigned to them. 



Assuming then that the Calcutta Garden asiatica is to be dis- 

 tinguished from the true Linnean form of Southern India, and 

 that the former corresponds to De Candolle's subincEqualis, we 

 find under Eoxburgh's asiatica and tilicsfolia, taken together, the 

 following forms, not attempting for the moment to decide whether 

 these are in each case "good species," viz. : — 



(1) G. asiatica Linn, (typica !), cultivated only. 



(2) G. subincequalis DC, cultivated (and spontaneous?). 



(3) G. tilicefolia Vahl (vera !), spontaneous in South India. 



(4) A species not yet satisfactorily identified, possibly the 

 small tree of the Sewalik belt already mentioned. 



We may next consider whether any other members of the 

 tiliafolia group were dealt with by Eoxburgh. It seems far 

 from likely that he should never have come across the tree 

 which Wallich subsequently named G. vestita, and in fact the 

 writer has no doubt at all that this was the species from Eastern 

 Bengal which he was unfortunately led to refer to Vahl's cxcelsa. 

 He had evidently seen only coppiced shoots, and hence described 

 what is normally a tree as "shrubby." The type of cxccha Eoxl). 

 non Vahl (nee Masters) is to be found, in the writer's judgment, 

 in a specimen collected by Buchanan in 1800 or earlier, and pre- 



