( xcviii ) 



The various kinds of similarities shortly rel'orred to are met with ajcain in 

 other groups of insects. Tlioy demonstrate, we think, not only that there are 

 varions ways to arrive at and different causes for resemblance, bnt also that 

 there is abundant material of similarity in details ujion which selection may 

 seize, and modify and an.^nnent the incipient resemblance, and produce those 

 wonderful likenesses which illnstrate that fascinating phenomenon called 



]\Iimicry. 



A study of resemblance from this ]ioint of view is incidentally also a study 

 of relationship, and therefore an essential ]nivt of classification. It is well known 

 that the older students of Entomology were frequently taken in by similari- 

 ties which do not indicate relationshij), associating often widely different species 

 on account of superficial resemblance. The older systematic works, and old 

 collectious which have not been disturbed, testify to this abundantly. But even 

 modern work, or rather work of recent date, is not free from striking blunders 

 of this kind, although the flourishing study of Mimicry has, or should have, 

 made everybody suspicious of mere resemblances in aspect. We well remember 

 trying in vain to convince a famous Lepidojiterist, now dead, that the streaked 

 Indian Papilio macareus, xenocles, etc., are not nearly related to the streaked 

 form of Pnpilio cli/tia ; and we notice in a recent number of Lepidoptera Tndica 

 that the mimetic rapilio rhcteiwr and its models are housed in the same 

 Moorean genus. The association of the Sphingidae with the Aegeriidae and 

 Ziigavnidac from Linue down to recent times was due to a certain similarity in 

 the shape of the antennae being erroneously interpreted as meaning relationship. 

 The older writers were the more convinced of the correctness of the association 

 of the Aegeriidae with the Hawk Moths, since the clear-winged Aegeriids 

 appeared to them to be coimected with the ordinary Sphiiapdae by the equally 

 clear-winged species of the Sphingid genera Ceplionodes and Haemorrhagia. 

 The fallacy of the conclusion has now become evident to every student of 

 Lepidoptera, though perhaps not to every collector of " flies." The similarity 

 of Pseitdosphinx tetrio to species of Protopmrce, the " Bombycine " appearance of 

 Arctonotus lacidtis and Lapara, the resemblance of Deldamia inscriptum and 

 other PliilampeUnae to certain Ambulicinae (= Smerinthinae auct.), the agree- 

 ment in ai)pearance between Proserpinus flavoj'asciata and Hae,7norrhagia, of 

 Nycergx hjposticta and Amph/pterus, of Akbesia davidi and Proserpinus, of 

 Sesia and Lcucostrophux, etc., etc., have frequently misled even modern classifiers 

 of Spkingidae. 



The student who tries to build u]) a classification based on relationship has 

 to beware especially of two kinds of similarities referred to above: (1) similarity 

 as tlie outcome of reduction, and (2) simihirity as the result of adaptation to 

 similar extraneous objects or to eacii other. 



The first point is the more frequently overlooked in classification, and re(|uires 

 some further explanation. 



If a, h, e, d are four organs of a sjiecies or a genus, and «', b', c, d' the 

 same in their most reduced state ; if, further, the species or genus develops 



