18 THE JOURNAL OF BOTANY 



immediate neiglibourhood, although the character of the leaves not 

 extending beyond the Rowers — an important character which should 

 certainly he included in the definition of B. arvensis, exceptions being 

 very rare — might suggest such a parentage. Crosses between 

 a. arvensis and H. stylosa have been described, but they are quite 

 unlike B. j^tychojjln/Ua. 



As stated by Burnat and Gremh (Roses Alp. Marit. p. 34), 

 ** L'hypothese de Torigine hybride d'une Rose ne doit etre admise 

 que rarement et apres mur examen." The present form deserves 

 attention and the only course open to me is to treat it, provisionally, 

 as a distinct species, much as I deplore the excessive splitting-up by 

 which some systematists have increased the difficulties of a most 

 perplexing subject of study. 



Rosa aevensis, var. major Coste. 



The name var. umhellata Godet should be applied to the remark- 

 ably robust form often called B. hihracfeata Bastard in this country, 

 but for the fact that there is an earlier B. iinibellata (a variety of 

 B. ruhiginosa^ ; Wolley-Dod has suggested the appropriate name 

 major Coste, as Rouy and Foucaud regard B. hihracteata as a 

 B. seinpervirens X stylosa, an opinion which is probably correct. 

 Some of the specimens I have seen in British herbaria are probably 

 hybrids B. arvensis x stylosa. 



Several bushes grow on the side of the road from Studland to 

 Corfe Castle ; I have found the same form in Surrey, at Clandon. 



The bush has much the same trailing habit as the typical 

 B. arvensis, but the stems are much stouter, also purple, and the 

 prickles on the flowering branches are large and strongly hooked, 

 instead of small and often straight or nearly so. The folioles are 

 large, dark green and rather shiny above, glabrous, 7, rarely 5 in 

 number, 1| to 2 times as long as broad, the terminal 21 to 39 mm. by 

 14 to 24, mostly acutely pointed, with 8 to 15 shnple teeth on each side, 

 some of which may bear one or two glands ; petioles glandular and 

 prickly, the prickles sometimes extending under the mid-rib of the 

 folioles. Crepin has attached an undue importance to the nvimber 

 of folioles on the middle leaves of the flowering branches in his keys to 

 the sections and species : thus B. micrantlia has more often 5 than 7 

 in this country and in Belgium, though not so in Switzerland, and 

 yet is placed in the Canince, defined as having the middle leaves 

 7-foliolate, whilst 7 as against 5 is given as a useful chamcter for 

 distinguishing B. arvensis from B. semjjervirens. It is well to state 

 that 5 is the rule in B. arvensis, in the South of England at least ; 

 in Smith's English Flora, ii. p. 397, it is described as with "leaflets 5, 

 rarely 7." 



Flowers large, white, up to 50 mm. in diameter, in clusters 

 of 3 to 9, extending beyond the leaves ; pedicels smooth to densely 

 glandular (in the same cluster), 3 to 6 times as long as the calyx- 

 tube, which is naked; sepals short, pubescent on the inner side, 

 with stipitate glands on the borders, with one or two short, simple 

 pinnae. Although I have carefully examined quite a hundred bushes 



