BY R. J. TILLYARD. 275 



Now there are, in any given Order, many diverse groups. 

 Some of tliCvSe will exhibit very high specialisations in one or 

 more directions: others wdll show unexpected archaisms, which 

 may make it exceedingly awkward to frame a comprehensive 

 definition of the Order. As an example of this, we may take 

 the evolution of the mouth-parts within the Order Lepidoptera. 

 In the great majority of the families within this Order, the 

 mandibles are absent, the maxillary palps absent, the galeae 

 produced into a long sucking-tube or haustellum, and the labial 

 palps present and more or less highly specialised. But, in the 

 Alirropterygidce, the mandibles are present, as are also the typical 

 maxillae of older Holometabolous Orders, with five-jointed palpi, 

 and unspecialised galeae and lacinise There is, therefore, no 

 reason, from the Phylogenetic view-point, why we should not 

 consider the Lepidoptera as, ot bottom, essentially a mandibulate 

 Order, when we come to discuss its relationship with the other 

 Orders of the Panorpoid Complex. 



It must, therefore, be evident that, for the purposes of this 

 paper, the usual definitions of Orders to be found in text-books 

 not only will not serve our purpose, but may actually be mislead- 

 ing. I propose to overcome this difficulty by the use of Arche- 

 types (German, Ur-typns). For each Order that comes under 

 review, we must define an archetype, which shall include in itself 

 all the most archaic characters found within the Order. Having 

 done this, we may reasonably discuss the relationships of the 

 archetypes of the various Orders, wuth some prospect of a suc- 

 cessful issue. 



In following out this line of argument, we can lav down two 

 guiding principles : — 



(1) The Phylogeny must not be determined from one set of 

 characters only (e.g., wing-venation), however important that set 

 of characters may be. But it must be determined by a review 

 of as many characters as possible. 



(2) The greatest care must be exercised in the determination 

 of the characters of the Archetype. For instance, in the case 

 of the jugum and frenulum in Lepidoptera, it is not suflBcient to 

 adopt the jugum as the more archaic character, merely because 



