48 



PROCEEDIXGS OF THE 



be specially variable. But, as 1 have said, it was not my object 

 to attempt, nor did 1 attempt, to suggest the evolutionary stages 

 in tlieir order, nor the pattern ot the hypothetical variety which 

 started the ininietic pattern. 



Now that the question has been raised I am quite ready to 

 express my opinion for what it is worth on so speculative a 

 subject, i believe that the evolution of the mimetic ])attern 

 could not have begun upon the upper surface until the appear- 

 ance of a variety iu whi(;h a reddish or orange tint liad largely 

 replaced the dark ground-colour, and could nut have begnu npon 

 the lower until a similar or paler tint had replaced the dark 

 choci)late-brovvn. The wide range of Viiriation in the reddisli 

 elements of tlie pattern referred to by Scudder and proved to 

 exist by Prufessor Abbott's measuremeuts requires to be greatly 

 exteuded beFore the mimetic pattern could begin to arise; but, 

 after that, the pattern was gradually elab(n*ated by the selection 

 of small variations. I believe that the hypothetical ancestor was 

 a butterflv which, as it sailed after ihe manner of a LimenUis, 

 appeared to be of an orange or reddish colour with a broad white 

 band crossing both wings. From this foundation I believe selec- 

 tion has gradually built up the detailed likeness which we see in 

 ol).soleta and stdl more perfected in archip/ms. 



Professor Abbott-, has shown bv measureuients of many indi- 

 viduals that in arthemis the white band is a very stalile structure 

 (p. 218), as indeed I should have expected from its long persis- 

 tence in Limenitls ohsohta with its beautifully mimetic pattern. 

 The comparison of the Danaine-mimicking North American forms 

 of Limenitis with L. astyanux, which, although a very close ally 

 of arthemis, is without the white band, strongly suggests that it 

 has been a much easier task to get rid of this feature altogether 

 than to mould it in the fore wing into the likeness of the model. 

 Evidence that it is most persistent and has nevertheless been so 

 moulded is afforded by the vestiges which still cling in out of the 

 way places to the mimetic pattern but do not contribute to it, 

 which indeed even slightly dimiriish the likeness to the model. A 

 feature of especial interest in this respect is the distinct trace of 

 the lower (inner marginal) end of the white band on the fore 

 wing of L. oh.wlefa. 



H. Mimicry betweex the DiFFEREJfi Insect Orders. 



The inadequacy of the hypothesis that mimetic resemblance 

 arose at a single bound can be shown by innumerable instances 

 of such likenesses between insects of diftVi'ent orders. I will 

 speak of only a single example, and choose it from among " the 

 species of Scaphura (a genus of Crickets) in South America 

 [which] resemble in a wonderful manner different Sand Wasps 

 of large size"*, because they are mentioned in Bates' historic 



* n. W. BatP9 in TrauK Linn. Soc, xxiii. 1802, p. .500. Scaphvra is now 

 placed among thv Locudidce {I'fitifffoiieiiridte) and not in the Aciictidce. 



