ROSA GRACILIS 33 
. 
margins and also sometimes on the midrib, but never to my know- 
rf 
hemispherical. Fruit globose. I have not seen it ripe.— Besides 
the marks enumerated under R. Doniana by which this species 
may be distinguished, the peculiar length, slenderness, an 
oods’s No. 21 from between Keswick and Lorton has its 
main prickles only slightly curved, and though rather stout-based, 
they do not differ materially from those of other segregates. The 
leaflets are hairy both sides, fully glandular-biserrate, with glandu- 
lar midrib. Petioles hairy, ve landular, but scarcely prickly. 
dial 
striking one. It may of course be more evident on the growing 
plant, and have become straightened in drying, but at the best it 
1s not a character which appeals tome. Differences in the length 
_ A specimen at Kew labelled by Woods as coming from his 
original bush has long stout main ‘prickles, subulate rather than 
i g, but those on the 
flowering branches are quite curved with more elevated bases 
than is usual in the group. The acicles are very small. Leaflets 
rather large, broadly oval, rounded or obtuse aé apex, fully glan- 
dular-biserrate, quite hairy both sides, but not densely so beneath, 
slightly glandular on midrib. Petioles rather densely hairy and 
glandular, with few small straight not falcate prickles. Peduncles 
ag 
characteristic differences. Fruit globose, almost smooth. Sepals 
peered aciculate, three out of the seven visible bearing one pinna 
each, 
Specimens in Déséglise’s herbarium from Woods, gathered 
c i ickle one 
airy, compoundly serrate. Stipules narrower, and with longer 
more acute auricles than is usual in the group. Peduncles in 
Journat or Borany, Marcu, 1910. (Suppnement.] d 
