198 MB. G. BENTHAM ON EUPHORBIACEiE. 



is not even restoring an old name ; for the specific adjective is not 

 of itself the name of a plant. Ask a seedsman for some Canari- 

 ensis and he will probably give you Tropasolum peregrinum, not 

 Phalaris canariensis. A generic name is sufficiently indicated by 

 one substantive, for no two genera in the vegetable kingdom are 

 allowed to have the same name ; but for a species the combina- 

 tion of the substantive and adjective is absolutely necessary, the 

 two-worded specific name is one and indivisible ; and the com- 

 bining the substantive of one name with the adjective of another 

 is not preserving either of them, but creates an absolutely new 

 name, which ought not to stand unless the previous ones were 

 vicious in themselves, or preoccupied, or referred to a wrong 

 genus. It is probably from not perceiving the difference between 

 making and changing a name that the practice objected to has been 

 adopted by some of the first among recent botanists, such as Wed- 

 dell, though under protest (see the note in DO. Prod. xvii. 1. 73). 

 To give a couple of instances among hundreds that have lately pre- 

 sented themselves to me : "Wight published a Nilgherry plant which 

 he believed to be new, and was certainly a new genus, under the 

 name of Chamabainia cuspidata, in all respects a legitimate name; 

 and he could not be expected to identify it with TJrtica squamigera 



" *— *"r4i'a < Paffllnrrnp ' *q tlia nlnnf. is nnh an TJrtica. Wight's 



Wall 



Weddell 



but in the ' Prodromus' he thought himself obliged, in spite of 

 his better sense, to call it Chamabainia squamigera, which is neither 

 Wallich's faulty name nor Wight's correct one, but an entirely 

 new name, to be rejected by the law of priority, which requires 

 the adoption of the oldest correct name. So, again, an Indian 

 grass was first named and described by Willdenow as Coix arundi- 

 nacea, then named in the ' Hortus Benghalensis ' and distributed 

 by Roxburgh as Coix barbata, and entered in Sprengel's ' Systems' 



Willdenow's character as Coix Kcenig 



names 



were defective as referring to a wrong genus. Brown corrected 

 the error by creating the new genus Chionachnc, and selected 

 Roxburgh's specific name as the one most generally known and 

 the least liable to misinterpretation ; and Brown's Chionachne 



first 



Keen 



, — „„_ ~y, — * 



useless name, which falls by the law of priority. It should be well 



with 



out affording any aid to science, is only an additional impediment- 



