( 455 ) 



Illustration. — The islands of Sambawa, Alor, Wetter, Timor, Letti, tiic Teniuiljer 

 Islands, North Australia, and the New Ilebridos, are each iuliabited by a subspecies 

 of Papilio canopus; the subspecies differ from one another in the shape of the 

 hindwing, and in the extent, presence, and partial absence of the wing-markings. 

 Latel}' Mr. Rothschild received a specimen of Papilio from the island of Sumba 

 difTeriug from the Sambawa and the Alor forms oi canopus in a similar way as these 

 do from one another and from the Timor form, namely in the jjartial absence 

 and in the extent of the markings. The only logically possible way, accepted by 

 Mr. Rothschild, was to treat the Sumba specimen also as a form of canopus. 



When the evidence leads to the conclusion that the diiferences exhibited by a 

 number of sjiecimens, or forms represented by specimens, are not specific, it is self- 

 evident that the various forms belong to one species. This species, then, consists of 

 a number of different varieties, every single individual of which, however aberrant it 

 may be, represents the species, and every single peculiarity of any individual is a 

 pectiliarity of the species : all the specimens of all the various forms taken together 

 are " the species " as ojiposed to every other sjiecies. The diagnosis of this species, 

 which must not be confounded with tlie description of the species, is therefore a 

 diagnosis of a sum of varieties ; and as a diagnosis is analogous to the definition of 

 a term, it must contain all the distinguishing characters comm(jn to all the 

 specimens, and hence must ap])ly to each single specimen. Besides the specific 

 distinguishing characters each variety has one or more characters of its own which 

 form the diagnosis of each respective variety. The discovery of a new variet}-, 

 which was hitherto unknown on account of the incompleteness of our knowledge, 

 or which has sprung uj) in consequence of the area of the' species having recently 

 become extended (Pieris rapae ab. loc. nocangliup, for example), may necessitate an 

 alteration of the diagnosis of the species to which the new variety belongs. As 

 the forms diagnosticated for the sake of convenience are fixed by a name given 

 to each of them, it is obvious that the only way logically possible to name a 

 species and its subordinate com])ouents is to give a name of its own to the species, 

 one to each subspecies as subordinate to the species, and one to each indi\'idnal 

 aberration as subordinate to the subspecies. Diagrammatieally it can be illustrated 

 thus : — 



Species A <- subsp. •« \jjij ^. 



ab. £■ 



subsp. B^Zlab. E» 



\ab. C 



= species A. 



Every individual forms together with other individuals a grouji cliaractcriscd by 

 a ])eculiarity not met with in the rest of the individuals, wliich, tlierefbre, form 

 another group (or more), termed here inditidual aherration = ab. A number of 

 individual aberrations are the components of a subspecies, and a number of sub- 

 species the components of a species. Each specimen will require in this case a 

 specific, a subsjiecific, and an aberrational name. In many cases, however, the 

 nomenclature will become much simplified by aberrational names being unnecessary, 

 because the individuals do not vary to sucli an extent witliiu a subspecies that 

 aberrational names are required ; the simplification will be still greater when, 

 besides the aberrational names, the subspecific names are not required. In order 



