637 



"ad sepes et in dumetis passim." (See Raii Syn, edition 2, p. 214, 

 anno 1696). Afterwards, in the third edition of the same work, 1724, 

 Dillenius mentioned and figured another plant, as " a variety of Viola 

 canina, if not a different species, observed by Du Bois, and much 

 smaller than the common plant in all its parts." The small variety, 

 or different species, was a Surrey violet, stated to have been found in 

 pastures near Mitcham. This, the only locality recorded by Dillenius, 

 is not without some value as evidence in the question of specific 

 names. 



We may pass over the intermediate authors on British Botany, by 

 a long leap to the ' English Flora' of Smith, published just one hun- 

 dred years after Dillenius had published the third edition of Ray. 

 Smith's work became, as is well known, a standard authority for sub- 

 sequent writers to follow and copy from, for the last quarter of a century. 

 Here we find Gerarde's identical Viola canina sylvestris still desig- 

 nated Viola canina by Smith, under the belief that Linnaeus had like- 

 wise so designated it ; while the Dillenian small variety or different 

 species is described under the name of Viola flavicornis. Besides 

 these two, Smith still kept up his own V. lactea, as a third species, 

 distinct from both the others. 



Since the death of Smith, very i'ew of the writers on British botany 

 seem to have clearly understood and distinguished his three species, 

 individual forms or varieties of one species having often been referred 

 to another of them. Dwarf forms of V. canina (Smith) have thus been 

 referred to V. flavicornis ; and examples of the latter, in its turn, have 

 been supposed to represent V. lactea. And, again, in changing the 

 name of Gerarde's species, from V. canina into V. sylvatica, the latter 

 name has unfortunately not been applied to the common species in 

 its totality, but only to its more luxuriant forms. This unfortunately 

 partial change and application is proved, among other evidences, by 

 the author of the Manual (the work in which the change is made) in- 

 dicating his V. canina to be " common," and his V. sylvatica only 

 questionably " common ?" But the V. canina of Gerarde and Smith 

 (the V. sylvatica of various continental authors) is a hundred times 

 more common in England, than is the other species (the V. canina of 

 continental authors) figured by Dillenius, and described by Smith, 

 from small examples, under name of V. flavicornis. 



Through the mis-references of particular forms to the wrong spe- 

 cies, and the partial change of name above mentioned, the present 

 application of " Viola canina" is becoming very vague ; and the use 

 of the name is too likely to confuse the ideas of readers and writers, 



