2.')() On the Races and Variadon of (he Edible Frofj. 



varieties striata and nii/roniacuhita in I', urvalis *, and 

 let'iised to admit it as a subspecies. 



These observations on the fossorial habits of the vavs. 

 chijiensis and lessonce should be borne in mind by those 

 who appeal to the beliaviour of the Indian R. crassa, 

 compared to that of the typical R. tir/n'na, as an argument in 

 favour of its specific distinction f. These supposed species 

 offer a perfect parallel to the lines of evolution which can be 

 traced in R. esculenta, as I have recently pointed out %• 



Although we must not expect to fiml aniong tiie S[)ecies of 

 the j)resent day the actual types out of which their allies 

 have been evolved, yet I think it legitimate (ipecubitioii to 

 look uj)on certain species, or certain small groups of species, 

 as a sutiiciently near approximation to help us towards an 

 elucidation of the phylogenetic relationships, the expression 

 of which should be the aim of taxonomy. In this sense, and 

 with this reservation, I consider R. catesbiana, Shaw, and 

 JR. gri/Ho, Stejn., as representing the most primitive forms of 

 An)erica and Eurasia; the species that cluster round them, 

 J\. septentrional iS, Baird, A', damitans, Daud., R. onca, Oope, 

 R. vir()atipes, Cope, 11. montezunnv, Baird, would be derived 

 from the same stock ; they constitute a distinct section, wiiich 

 is perfectly natural, though not susceptible of a very strict 

 definition. From this section we may imagine the one of which 

 li. esculenta is the type to have been derived, and there is 

 little doubt in my mind that the Chinese K. planet/ i, Lataste, 

 is a connecting form, nearly allied to, but in most respects 

 less moditied than, R. esculenta, both having been evolved 

 out of the same ancestor, possibly related to the Oligocene- 

 Miocene Ji. meriani, IT. von Mey. The chief distinctive 

 features of jR. esculenta compared to li. planci/i reside in a 

 reduction of tiie nasal bones, the more obtuse fingers, and the 

 very peculiar external vocal sacs. By what steps this last 



* Such modifications represent varieties only in the sense taken by 

 liorticulturiets, and sboiild not be given names in scientific nomenclature. 

 Eliminating these cases, I apph' the term varietas to every division of 

 the system subordinate to the species, without any further consideration 

 of hierarchy, in order to avoid complicating nomenclature by the use of 

 tri-, quadri-, or even quinqueuomial?. In so doing, I simply adhere to 

 the Linnean method which has so long been followed, and is still used 

 by most of the botanists for whose work I have the greatest respect. 

 " Les varietfia des systematistes stSrieux sont les especes de M. Jordan, 



au moins du Jordan des Observationes et du Pur/illus Le mot 



vari6t^ employd par les botanistes pour designer une race iauvage laisse 

 peut-etre a d^sirer, viais il jonit ile la priorite." J. liriijuet, Questions 

 de Nomenclature, Bull. Herb. Boissier, ii. 1894, p. 84. 



t Annandale, liec. Ind. Mus. xv. 1918, p. 03, 



t Ilec. Ind. Mus. xv. 1918, p. ol. 



