A NEW CLASSIFICATION OF LEPIDOPTERA. 



819 



impovtnnt cliRvactPr mav bft wvonsflv snrmisprl, V)nt, so far as 

 we have tp<=!tpfl, no obscnritv is caused bv these devices, whilst 

 hMTidiness is immenselv increased. To tbose of us whose fir<t 

 definite ideas of Lepidoptera were obtained from Stainton's 

 'Manual.' and who find it difficult altogether to outgrow our 

 earlv lessons, the considerable changes in nomenclature are not 

 at all attractive ; but we must give as graceful an acquiescence as 

 we can where they are shown to be right, and therefore necessary. 

 To the young student the system of nomenclature will present no 

 greater difficulties than any other would have done. The amount 

 of good, sound, accurate woik on the part of the author must 

 have been very great, as he tells us that all structural characters 

 have been verified bv his own observations. 



We have little doubt that the book will be the Handbook of 

 the working lepidopterist in the same manner as the 'Manual' 

 has been for a third of a centurv. It is for this reason that we 

 think it desirable to consider whether the classification adopted, 

 and now " fully published for the first time," is sufficiently sound 

 and correct to be adopted, for perhaps another third of a century, 

 without any reservation. 



We find that the classification is based almost entirely on the 

 neuration ; otiier characters are mentioned in the descriptions, 

 but are usually severely neglected as grounds of classification. 

 It is undoubtedly true, as advanced by Corastock, that any one 

 character may be made use of for classification, and must give 

 correct results if correctly observed and fully understood ; but it is 

 precisely here that the study of a solitary character fails ; similar 

 modifications occur along parallel lines that are nevertheless 

 widely divergent phylogenetically ; very slight modifications will 

 occur over a large area in some families : great modifications 

 within the limits of a genus within others, and so on. These 

 difficulties can only be overcome by frequently considering many 

 other characters, and using them to enable us to interpret the 

 meaning of the one we have selected. The result is that a 

 classification based on one structural point gives us, to take a 

 botanical parallel, a Linnean and not a natural system, and has 

 a mechanical rather than a scientific character. In the work 

 before us there is evidence that this is the case, although the cases 

 in which a true result is reached are of course very numerous. 



It is a pitfall for the neurationist that the characters he adopts, 

 especially when he limits them as our author does, lend them- 

 selves to easy observation and simple numerical tabulation. 

 The limitation our author makes is a very curious one, and 

 throws over so weighty (phylogenetically) a feature of neuration 

 as to surprise one very much. 



This very important feature in the neuration of the Lepi- 

 doptera, which is several times referred to, but, so far as we can 

 see only to be dropped as of no importance, is the gradual 



2d2 



