170 Mr. Woods on the British Sjiecies of Rosa. 



species these continue unchanged, or nearly so, in those leaves 

 which accompany the inflorescence, and no stipuliE are found 

 unaccompanied by leaves; in others the leaflets gradually dimi- 

 nish in number, till at last they are entirely deficient, and the 

 two stipulae unite and form a bractea ; in others, again, before 

 this process is complete, the stipulae increase very remarkably in 

 breadth, and the first bractea formed is perhaps subrotund, though 

 arising from an alteration of strictly linear stipulae ; but in the 

 cymes of flowers the bracteae are repeated, growing gradually 

 smaller and somewhat narrower ; still, however, retaining traces 

 of their original increase in width. The description of the bractece 

 is therefore taken from the usual form of the first, which are 

 found entirely devoid of leaflets; and the circumstance affords a 

 very good distinction between two tribes of Roses, the family of 

 li. cinnamomea possessing them in a remarkable degree, which [ 

 therefore describe as bracteata ; and those of the family of R. spi- 

 nosissima preserving the stipulae nearly unaltered, which I have 

 therefore called ebracteatce. 



This appropriation of terms is not the only liberty for which I 

 have to apologize in this essay. I must acknowledge that I have 

 described plants as species, of which I can hardly say that I really 

 believe them to be distinct; but when this is the case, it is be- 

 cause I did not know with what species to join them. In enu- 

 merating them as species, I hope likewise to provoke the at- 

 tention necessary to rectify the error; while, if placed as varie- 

 ties, they would have less chance of being attended to. Another 

 circumstance in which I have deviated from the usual practice 

 of British botanists, though in this I am supported by the au- 

 thority of Willdenow, is, that I have given names to many of 

 the most remarkable varieties ; and this practice has been 



adopted 



