of organs like tliosi' occur indeed in niaiiv ^'cnera which staml far apart. Loss, 

 rednetion, an<l fusion may occur indejjendently everywliore. An instructive 

 illustration of erroneous classitioatioii of this kind outside the Sphifu/idae are 

 the Itliomiinue (= Xt'olrojiiiinr), which are divided by Salvin and (todraan and 

 later on bj- Scha*/. into two groups, the one containing the forms with strongly 

 rodnced forelegs, tlie otlier those with less reduced forelegs. The groups as 

 they thus stand comiirise i(uite lieterogeneous material ; they are not a division 

 based on relationship, but on similarity in that one organ, each group containing 

 genera which have their nearest relatives in the other, the species or genera 

 with reduced forelegs being derivations from such with more fully developed 

 forelegs. The striking resemblance in pattern between certain members of the 

 two groups is exjilained, at least in some cases, by the similar species being 

 nearly related, and not by having developed convergently. 



A form which has lost an organ becomes not only similar in this respect 

 to another which lias equally lost it, but also to those which never did possess 

 the structure. The abdominal spination, which is apparently confined to the 

 Sphiiiqidne, is lost in several genera, which therefore agree in the non-spinose 

 abdomen with the Notodontidac, for instance. Tiie loss in many Sphingidae of 

 the midtarsal comb, of the foreco.xal scent-organ, and of the friction-organ of 

 the clasper constitutes again similarity with Notodontidae, Satirrniidae, and other 

 Moths, and not relationship. 



There is another point illustrated by the pedigrees of Sphingidae. which is 

 of equal classificatory importance. A genus and a species are a composition of 

 many organs ; and since each organ becomes in most cases modified independ- 

 ently from the other, it is evident that a species or genus may be specialised 

 in some and generalised in other organs. It is indeed a rule holding good in 

 nearly all Sphingidae — we say " nearly " in order to be on the safe side — that a 

 species or genus is farther advanced than its nearest allies in some respects 

 and at the same time left behind in other details. This being so, it is clear 

 that a genus or species is not (or rarely) a direct derivation from another. 

 Though we have directly connected the genera with one another in the 

 pedigrees, we do not mean to say that, for instance, I'roserpinus is a derivation 

 from Rhodosoma as this is now, but wish to indicate only that Tihodomma is 

 of all recent genera the one which comes nearest in the characters mentioned 

 in the pedigree to the ancestral genus of which Proserpinus, Amphion, and 

 Sphecodina are developments. Each genus is, so to speak, the end-development 

 of its own branch. In the pedigree facing p. 499 the genera Enpinanga, Rethera, 

 Cizara, and Rhodosoma are thus connected : — 



I 

 Kliudosoma 



I Rethera 



Cizara I 



Enjnnawja 



