ROSA ANDEGAVENSIS 47 
some eet its forms might be — but, as a rule, it presents 
distinct characteristics. Its ma n prickles (not only those on 
branches) are opuelly straightish ee often rather slender. The 
petioles, though always glabrous, are a good deal glandular. The 
narrowed to each end; they are almost always large. The tooth- 
ing is coarse and open, not connivent, as in R. lutetiana. The 
peduncles are usually short and considerably glandular hispid. 
The fruit varies a good deal in shape. Its most common form is 
ovoid; it is also sometimes eae but perhaps more frequently 
tends tow as subglobose, though it usually is somewhat narr 
below so as to be broadly obovoid. It is yery seldom urceolate. 
oe is areas hispid- — at the base, and sometimes so all 
i sepals are usually more or less glandular on the back ; 
iss are rarely plabeount in Continental examples, as Mr. Baker 
ays. The pinne are usually well developed, z.e. broad as well 
as often again incised. The styles usually project and vary con- 
siderably in hairiness; they are sometimes quite glabrous, in 
which case I do not nee why Déséglise has not placed 
the specimen to Rf. agra 
here is a specimen collected by Bastard in herb. Gay at Kew. 
It has prickles. iow elle curved. Leaflets spaced on the 
id. 
hispid all over. Sepals reflexed, Snes set mow on backs, 
pinne well developed. ake hispid. 
Déséglise, 1. c., quotes his own No. 17 as one of his types, 
but, as is frequently ‘be case in other species, the specimen he 
cites appears to be considerably off type, in fruit and in leaflets 
at any rate. This No. 17 is only a small specimen. The leaflets 
are rather small and irregularly serrate, elliptical oval, not rather 
elongate-diamond-shaped as is usual. The petioles are glandular. 
The sepals are not, or are only very obscurely Oya on the 
backs. The peduncles are short, mostly in threes; fruit narrow, 
longish, rather hispid. Styles salient, thinly aii: 
ere are no British examples in herb. Déséglise. 
Déséglise, in Cat. Raison. p. 181, places the undermentioned five 
Species after R. andegavensis with the remark: “RB. -é ae 
presents the following forms. We will not assign separate 
bers to them, but ae wishing by that to diminish their 
importance in the lea I am unable to see a reason for this 
treatment. Tb is true ‘that except for R. agraria Rip. there is only 
a line or two of published description for each, but yuieme fe from 
Specimens, they are just as remote from (or similar to ande- 
gavensis as most species to ack Déséglise does not give eo 
treatment, while two of them have biserrate leaflets :— 
R. agraria Rip. Leaflets less acuminate, uniserrate. Sti- 
Pules large. Sepals angen on Sbesk: Styles glabrous or 
