The Zoologist— February, 1870. 2009 



compressed and acute ; the inner edge of the middle is more or less 

 dilated; the middle is always the largest, except in two genera, 

 which present the peculiarity of having a very large serai-circular 

 inner lateral claw, which, moreover, lies horizontally instead of 

 vertically. 



That rigid adherence to the type of structure just described which 

 all the species maintain, while facilitating the recognition of the family 

 as a family, is a serious obstacle in the way of defining its subdivisions 

 with precision. With no very abrupt transition from one form to 

 another, and without any very marked modification of general features, 

 the minor groups seem to be formed mainly by the varying combina- 

 tion of the few differences in structure which obtain in the family. 

 The assemblage of characters, latlier than the presence or absence of 

 particular features, in most cases determines the genera ; and no two 

 species are absolutely alike in all points of form. 



" Facies non omnibus una, 



Nee diversa tainen, qualis decet esse sororuin." 



In one of the ablest papers that has appeared upon this subject, 

 Professor Brandt divides the Alcidae into two subfamilies — those with 

 feathered and those with naked nostrils. In this arrangement the 

 guillemots stand next to the typical auk (A. torda). Viewed from any 

 other standpoint the two forms appear to represent the extremes of 

 structure in the famil}' ; particularly in regard to the bill, cultriform in 

 one, subulate in the other. The two types are by most authors placed 

 at opposite ends of the generic chain, and separated by all the 

 Starikis. Attentive consideration of all the bearings of the case may 

 very likely result in the opinion, held by the present writer, that the 

 difference between the views of Professor Brandt and other writers is 

 rather apparent than real. It should be borne in mind that the 

 Alcidae are a family very rigidly circumscribed, and one showing no 

 tendency to aberration, or to connect itself ultimately with the 

 families standing next to it on either side. Whether as cause or 

 consequence of this, the fact is indisputable, that the genera of Alcidas 

 are not strung along in a chain whose ends seem, as it were, to be 

 linked with the genera of other families: they tend, on the contrary, 

 to aggregation in a circle about a common centre. We may take any 

 genus, — it matters not which, — we shall find its closest ally to the 

 right and to the left; and the circuit shall be complete when all the 

 genera have been considered. To illustrate this point. Prof. Brandt, 

 like all other writers, takes the typical Alca as his starting point. 



SECOND SERIES VOL. V. I 



