276 THE PANORPOID COMPLEX (Introduction), 



it is found in those Lepidoptera which have the most archaic 

 wing-venation. If a mistake is made here, the whole Phylogeny 

 is bound to go wrong. As a corollary, it should follow that 

 characters in which the evidence of archaism or specialisation 

 may be uncertain, with a division of opinion upon the point 

 amongst those entomologists who have studied it, must be either 

 definitely cleared up by means of new evidence, or they must be 

 entirely omitted from the argument. 



The Detej'mination of Archetypes. 



We ma)^ exhibit this principle as follows : — 



Let a, b, c, d, etc., represent a series of characters which 



are constant throughout the Order in question. 



Let p, q, r, s, . . . . etc., represent a further series of characters 

 which are not constant throughout the Order, but show definite 

 lines of evolutionary advance in one or more directions. 



Taking any one of these characters ]>, after examination of its 

 structure in various families, let F represent the most archaic 

 condition, while p', ]"■> ^t^., represent various specialised condi- 

 tions of the same. 



Any family within the Order can now be defined by the 

 characters 



a, b, c, d, etc., p, q, r, s, etc., 



where p can have any of the values F, p', p'\ etc., q any of the 

 values (j>>, 5', q\ etc., and so on. 



For instance, one family may be defined by 

 «, ^, c, f/, ;y, q\ R, s"\ 



The Archetype of the Order, which may or may not still be 

 existing (the greater the number of characters taken, the less 

 likely is it to be still existing) will be defined by 

 «, 6, c, c/, F, Q, F, 'S, 



77te Determination of the Fhylogeuy of separate Orders. 



The same principle as above may be followed in this case, pro- 

 vided we take the characters of the ArcJietype as the characters of 

 the Order it represents, for the purposes of the Phylogeny. 



In this case, the characters a, b, c, d, etc., which were 



constants throughout any single Order, will now be variables for 



