32 THE SUBSECTION EU-CANINE OF THE GENUS ROSA 
0. y gnoses I L 
to be more frequent than R. spherica, but do not include it until 
it is confirmed by good authority. 
SYNTRICHOSTYLA 
Ripart MSS. ex Déséglise in Catalogue Raisonné du Genre Rosier 
p. 143 (1877). 
“A low shrub, with flexuose branches, bark green, stem- 
rickles strong, dilated at the base, hooked, those of the flowering 
above, paler beneath, oval-acuté or oval-elliptic, some obtuse 
simply dentate, the lower leaflets with some rare accessory teeth 
acute, divergent. Peduncles smooth, solitary, or in a corymb, 
with glabrous, lanceolate, or oval-cuspidate bracts at their base, as 
long as, or longer than, the peduncles. Calyx-tube ovoid, glabrous. 
Sepals appendiculate at the tips, glabrous, two entire with to- 
men es, three pinnatifid with narrow appendages, salient 
in bud, equalling the corolla, reflexed after flowering, not per- 
sistent on the fruit. Styles free, very villous, simulating a more 
or less salient column. Dise conical, Flowers white with yellow 
claw. Fruit small, ovoid, red when ripe.”—He adds that Ripart 
wrote him saying that it belongs to the group R. systyla, but 
Déséglise points out that the styles are free in Ripart’s own 
specimen; moreover, Ripart, though he places this in his Section 
Systyl@ of his key, remarks therein that it and most of its allies 
may equally well be placed in the next Section (i.e., RB. canina 
rei lato}, as their styles are merely agglutinated, and not truly 
united. 
I cannot trace that Ripart ever published a description of this 
or of any other species himself. His key, referred to on p. 8, only 
