174 Proceedings of the Biological Society of Washington. 



Therefore to refuse any recognition to a subspecies merely be- 

 cause its characters are not entirely constant, is to obscure a 

 fact that may have important bearing on other biological 

 problems. 



As a matter of fact, Larus hyperboreus barrovianus is readily 

 distinguishable by either the color of the upper parts or by the 

 size of its bill. Our previous comparisons 1 were, of course, made 

 with only adult birds and with those in comparable state of 

 plumage. We should, however, have emphasized the fact that 

 the darkest birds of Larus hyperboreus hyperboreus are much 

 paler than the darkest birds of Larus hyperboreus barrovianus, 

 and that the average color of the upper surface is so much 

 deeper in the latter form that with a series it is appreciable at 

 a glance. That occasional examples of Larus hyperboreus bar- 

 rovianus become pale in worn plumage certainly does not invali- 

 date this race, particularly in view of the fact that such speci- 

 mens are nearly always darker than the corresponding plumage 

 of Larus hyperboreus hyperboreus. 



There is so much difference in size of bill in Larus hyperboreus 

 barrovianus as compared with Larus hyperboreus hyperboreus 

 that there usually is no difficulty at all in identifying sexed 

 specimens. It is true, of course, that some individuals are 

 intermediate in this character, but as mentioned above, this 

 should not cause the rejection of Larus hyperboreus barrovianus 

 as a subspecies. The length of the bill in this race is consider- 

 ably (4%) less than in Larus hyperboreus hyperboreus. It is, 

 nevertheless, not the best size character, for in the height of the 

 bill there is much greater relative as well as actual divergence; 

 this being 10% less in Larus hyperboreus barrovianus. This 

 point has evidently been overlooked by Dr. Dwight, as his figure 

 shows. 2 We, of course, should have placed more emphasis on 

 this difference, as it forms one of the best characters for dis- 

 tinguishing Larus hyperboreus barrovianus. 



The measurements of the bill in our former paper were taken 

 from ten males and ten females of each race, all adult and 

 properly sexed. Dr. Dwight's diagram itself 3 fully explains 

 the reason for the discrepancy between his measurement of 



1 The Auk, XXXV, No. 4, October, 1918, pp. 467-474. 



2 The Auk, XXXVI, No. 2, April, 1919, p. 247. 

 S The Auk, XXXVI, No. 2, April, 1919, p. 247. 



