A LIST OF BRITISH ROSES 6 



many years differ startlingly in appearance and in technical 

 characters. Though the unravelHng of the tangle will be a diffi- 

 cult business, it is one which should not be shirked by British 

 botanists. 



I have made no attempt to revise nomenclature. The name 

 of each group or subgroup is, as a rule, that of or derived from 

 that of its most important member, but it must not be supposed 

 that all the other members have equal rank. I have given the 

 names as I find them, without prejudice to revision, either of 

 nomenclature or grouping, at some future date, and have refrained 

 for the present from making more than an occasional expression 

 of opinion as to their relative rank. It is of course anomalous, 

 for example, that such a name as B. tomentosa var. uncinata Lees 

 should appear in the group Omissa, but if I alter it to B. omissa 

 var. uncinata W.-Dod, there is not only no probability of finality, 

 but an extreme improbability thereof, so that another combination 

 would have been unnecessarily added to the nomenclature. 



I have made frequent references to my two former papers in 

 this Journal, viz. " The Subsection Eit-canina " and " The British 

 Roses," published as Supplements in 1908 and 1910 respectively. 

 For the sake of brevity I have cited these as " E." and " B. E." 



I have now fulfilled my promise to deposit such duplicates as 

 I have, with notes on some of them, in tlie British Museum Her- 

 barium at South Kensington. It is much to be hoped that these will 

 be added to by other botanists, but I would suggest that discretion 

 be exercised, and only such examples be deposited as can be more 

 or less definitely named. A heterogeneous collection, either un- 

 named, or with very doubtful names, is more confusing than none 

 at all, unless fully annotated as to differences from type. I must, 

 however, plead guilty to some degree of uncertainty attaching to 

 almost all my names up to the present, but the specimens at least 

 represent the interpretation I place upon the species and varieties 

 up to the time of depositing them. 



SECTION SYNSTYL^. 

 GROUP ARVENSIS. 



Rosa arvensis Huds. Fl. Angl. p. 192. Unmistakable, at 

 least as an aggregate, but I have definite records only from v.-c. 13, 

 16, 17, 32, 55, and 58. 



[B. dibracteata Bast, in DC. & Lam. Fl. Fr. v. p. 537. I can 

 throw no further light on this species, and think it should be 

 excluded from our list. All the large strong-growing forms 

 originally refeiTcd to it may equally well l)e placed to var. major 

 Coste.] 



R. arvensis var. major Coste in Pons & Coste, Annot. Herb. Ros. 

 iii. p. 14. To this variety must, I suspect, be referred, as just stated, 

 all the material formerly labelled B. dibracteata Bast., which is 

 regarded by many continental botanists as a hybrid. Such forms 

 are quite frequent and general. V.-c. 2, 3, 13 or 14, 17, 62. 



62 



