Family and Generic Differ entiation. 337 



But why, I ask, fight shy of the very constant and 

 congenital colour-pattern, so conspicuously characteristic of 

 such genera as I have mentioned, and swallow the very 

 variable structural forms of the hills, which have been 

 impressed upon these various Eroliiue or Tringine species 

 through purely functional causes. Which is the oldest, the 

 most constant, and the most fundamental character ? 

 Which is the most reliable character for the purpose of 

 classification — one which had a direct and obvious relation 

 with function, or one which had not such an obvious 

 relation ? 



We get out of this dilemma by regarding the variations 

 of the bill in the genera just noted as specific. But are we 

 justified ? I do not say we are not — but are we? 



As a matter of fact, as I shall indicate later, my sympathies 

 are, in many instances, witli the lumper, but lumping genera 

 on too large a scale may, perhaps, be almost as inexpedient 

 a proceeding as splitting on too large a scale. To this 

 subject, however, I shall return later. 



(7) Colour-pattern as a phylogenetic clue. 



In this connection, as is only natural, we turn to a con- 

 sideration of the colour-pattern characterising the immature 

 or nestling examples of any particular group selected. 



I have here a very fine series of the nestling downy young 

 of the Dunlin association — the subfamily formerly known 

 as the Tringina^., but for which I presume to think the term 

 Eroliiuse might now be substituted, since the name Tringinse 

 is, by those who apply the principle of Linnean tautonymy, 

 ruled out of court. As far as it goes, and I think it goes a 

 very long way indeed, the colour-pattern common to all these 

 nestling species serves as an excellent control experiment, 

 whereby Ave are able to judge of the exact limits of the 

 Eroliine subfamily. This colour-pattern, which I have not 

 time to describe — you can see it for yourselves, — appears to 

 be absolutely diagnostic of the subfamily. In it, we have 

 a phylogenetic clue, pointing to the fact that in the Eroliinse 

 we have a very definitely diti'erentiated group. 



