1G2 



characters which separate R. ArkansiDia^ Porter, from R. blanday 

 Ait, says of the sepals: "If the presence of appendices are 

 more frequent in R. Arkansana than in R. blanda, this char- 

 acter it appears to me has not the importance that Dr. Wat- 

 son ascribes to it, and I even think that it has no specific vaUie 

 at all"* The autumnal coloration is often observed in A^ Caro- 

 lina and (as I am informed) in R, nitida\ it Is occasionally seen 

 in R. htmiilis. 



So far as the mode of vegetation goes, there are two types 

 met with in what M. Crepin recognizes as R. Jiwnilis ; in one 

 the bush perishes after two or three years, its life depending in a 

 measure on the severity of the winters, to be replaced by a new 

 growth given off from its roots; the other under favorable con- 

 ditions lives for years, the bark frequently becoming gray, new 

 stems arising from the base of the old, forming clumps like the 

 stems in R. Carolina. M. Crepin seems not aware that the latter 

 type is the usual one in what I have called var. hicida. 



Abundant opportunities to investigate the in-ground growth 

 of these roses have shown me that all the forms are more or less 

 surculose. If the stems of R. Carolina be rased to the ground, 

 the following season will witness a crop, usually copious, of young 

 shoots from every part of their predecessor's roots ; the terminal 

 portion is often transformed into an ascending axis. Here is a 

 question of degree, not of quality. The bushes of i?. hinnihs zxt 

 not so hardy, so durable, more hkely to be winter-killed ; a"d» *' 

 so, stems will spring up from any part of the roots, often bearing 

 flow^ers the same season, a thing seldom, if ever, observed in ^■ 

 Carolijia. 



i 



In the herbarium of the Academy of Natural Sciences, I^h^^' 

 adelphia, I saw a specimen from Kew Gardens w^hich corre- 



a 



spondcd very closely with the description given of the *' true 

 R. bicida. The sepals entire, branches erect-ascending, stipules 

 broad and margins toothed, leaves mostly nine-foliolate, some how- 

 ever seven follolate ; when the former, a forcing was noticeable, in 

 that the lower pair was smaller and sometimes one leaflet of this 

 pair rudimentary. Specimens collected by Mr, J. H. Redfield at 

 Mt Desert, Maine, when compared with the Kew specimen, made 



*Ninivelles Rcinarques sur les Roses Americaiucb-la seance du 12 Mars, 1S87- 



