20 THE JOriiXAL OF BOTA>-Y 



EosA CA>iXA, var. OBLOKGA. 



Sosa ohlonga Desegl. & Eip. is mentioned by "NVolley-Dod among 

 the foreign *' species " which should be looked for in this country. A 

 bush at Studland answers the dehnition, and I have been able to 

 compare a tlowering branch with the specimens in the Peseglise 

 herbarium — a rather mixed lot, I think ; there is, however, practical 

 identitv with the small-leaved specimen (no. 892) from Marmao:ne 

 (Cher). 



A lax bush, 4 ft. high, with small leaves, some tinged with 

 pm-plish red, with purplish-red stipules, small, pale rose solitar}^ 

 tlowers. and strong, straight or feebly curved prickles. Growing 

 isolated. 



Folioles 5 or 7, 1| to 2 times as long as brpad, the largest 20 mm. 

 in length, acutely j^ointed, rounded at the base, petiolulate, glabrous, 

 ■with stipitate glands on the mid-rib, biserrate (13 to 16 double long 

 and sharp teeth on each side), eglandular on the teeth or only here 

 and there with a stipitate gland, except at the base ; petioles with 

 stipitate glands and uncinate prickles. Corolla 35 mm. in diameter ; 

 sepals nearly as long as the petals, eglandular, with two long, simple 

 pinna? ; pedicel smooth, not longer than the ovate calyx-tube. 



But for the more obscm-ely gland-tipped secondary serrations of 

 the folioles, this Rose agrees with Baker's definition of the var. hi- 

 serrafa, in the spion^nny of Avhich his B. n'nacea — of which I have 

 examined types from near Thirsk, in Herb. Kew — is placed: 

 " Scarcely different from the last [var. diimalis^ but the serrations 

 open and compound, the petioles more glandular-setose, and the 

 glands extending a little to the midrib beneath. My vinacea has 

 oblong fruit, narrow sharp-pointed leaves and bracts, branches and 

 stipules suffused with vinous red." It seems to me midway between 

 jff. vinacea, with which it is probably connected by specimens such 

 as Mr. Eogers's from Luccombe Chine, alluded to by Major Wollcy- 

 Dod, and the small-flowered and small-leaved, but mostly though not 

 invariably nniserrate var. acipln/Ua Lindley {nee Eau) which should 

 probably bear the name <?.r//^s Crepin (ISOS) into which it grades; 

 I have observed Crepin's var. aciphyJJa growing in Belgium (Han- 

 sur-Lesse and "Wavreille). 



The three allied varieties, as here understood, may be thus con- 

 trasted : — 



Serratnre compound, with abundant glands; petioles and sepals 

 glandular var. vinacea. 



Serratui-e double, without or with very few glands ; petioles glan- 

 dular ; sepals eglandular var. ohlonga. 



SeiTature simple or irregularly double, without or with very few 

 glands ; petioles eglandulai" or with very few glands ; sepals 

 eglandular var. exiJis. 



The third variety leads to the typical R. canina {hifetiana) in 

 which the leaves are larger and quite eglandular. 



