254 Mr. Gr. A. Boulenger on (he Faces and 



first grade, we may place the var. chinensis in the second, 

 the var. ridibunda in the third, and the vais. saharica 

 and lessonce in the fourth. Looking at things from a 

 practical standpoint, we must regard the var. sahanca as 

 hut a slight modification, a geographical race distinguishable 

 from its nearest neighbour but impossible to define it' 

 specimens from the whole range of distribution of the 

 var. ridibunda are taken into consideration. The typical form 

 is completely connected with the var. ridibunda, and where 

 the two co-exist in a locality, annectant individuals may be 

 regarded as the result of crossing, such as undoubtedly 

 must take place ; but this explanation fails when we have 

 to deal with specimens from France, S.E. Europe, and 

 Asia, where individuals of uncertain identification like- 

 wise occur, although the discrimination of the two forms is 

 in most cases quite easy. It is no longer so when we come 

 to the typical form compared with the var. lessonce, and in 

 this case the naming of certain specimens is perfectly 

 arbitrary, as those who have had to deal with a considerable 

 material from places where the two forms co-exist fully 

 admit*; vet, the extreme, what some would call the " pure 

 lessonce" such as it occurred in the Cambridge fens and is 

 still found in a very few localities in Norfolk, is well entitled 

 to varietal rank, its structural characters being fixed and so 

 considerable in degree when compared with the typical form 

 that it would undoubtedly be looked upon by many as a 

 species were we not acquainted with the annectant examples 

 from the Continent. These extreme specimens of the var. 

 lessonce represent the terminus of an uninterrupted series 

 starting from the var. ridibunda and passing through what 

 is called the typical form. 



Another terminus form, in which the principal characters 

 of the var. lessonce are repeated, is the var. chinensis, which 

 in all probability is also derived from the var. ridibunda, 

 but the connecting-links of which have disappeared or are 

 still unknown. If we appeal to the existence of a hiatus 

 between forms as a sole criterion for deciding on what 

 is a species, then R. chinensis is entitled to stand as such ; 

 however, considering the many points of agreement, and 

 preferring to keep an eye on resemblances rather than 

 on differences, the rank of variety or subspecies appears to 

 me the more appropriate for this form, as it did to Lataste 

 many years ago j\ Owing to the state of things in the 



* Cf. "Wolterstorff, Sclir. Nat. Ges. Danzig, (2) xi. 1904, p. 46. 

 t Bull. Soc. Zool. France, 1880, p. 61. " Cette forme, que quelquea 

 auteurs regardeut comme une espece distincte, d'autres cuninie une 



