358 



STRUCTURE AND CLASSIFICATION OF BIRDS 



can see, a further modification of the gull type, but further 

 from them than they are from the Limicolae. 



We may, therefore, probably regard the gulls as a slightly 

 modified offshoot of the typical Limicolae (Chionis, perhaps, 

 being slightly intermediate), leading towards the Alcae. The 

 arrangement inter se of the remaining families brings us to 

 the broader question of the affinities of the Limicolae as a 

 group. It appears to me that the only other groups which 

 need be specially considered in this connection are the 

 Grues, Otides, and Ralli. It may be convenient to preface the 

 discussion with a tabular statement of the actual points of 

 likeness between these several groups. 



The very difficulty of finding any characters, greatly 

 noteworthy, in which the groups in .question vary is an 

 index of how closely allied all four are. There can, to my mind, 

 be no doubt of their common origin. The Limicolae on the 

 whole come nearest to the Grues, and especially to the true 

 cranes, whose distinctive characters are a little swamped in 

 the above table, owing to the aberrant cranes (e.g. Ehinochetus, 

 Psopliia), whose peculiarities have naturally modified that 

 table. Taking this point of view, it seems to follow that of the 

 Limicolae the most primitive section is that of the Chara- 

 driida3 proper ; for it is among them that the forms with 

 the greatest number of points of resemblance to the cranes 

 occur. I should consider these birds to be slightly lower in 

 the scale than the cranes. 



