MUSEUM OF COMPARATIVE ZOOLOGY. 205 



Variation in the relative Length of the Primary Quills. — From the 

 great stress laid upon the relative length of the outer primaries by de- 

 scriptive ornithologists in determining genera and species, one would be 

 led to expect but a slight amount of variation in this respect in speci- 

 mens of the same species. On the contrary, however, it is soon found, 

 on giving special attention to this character, that a considerable amount 

 of individual variation in this regard really exists. That the wing 

 formula, so generally introduced of late years into specific diagnoses, is 

 in a great degree unreliable as a specific character, is sufficiently shown 

 by the subjoined table (Table F, p. 20G) of the relative proportions of 

 the primaries. The comparison, extended in the table to only a few 

 species has been carried to scores of others with similar results. 



In general, in species of the Oscines which have the second primary 

 usually the longest, it is sometimes the first and sometimes the third 

 that is the longest. In those which have the third ordinarily the longest, 

 the second and third, the third and the fourth, or the second, third, and 

 fourth are frequently equal. In those in which the first (or the second 

 when the first is very short) is intermediate to the second and fourth 

 or to the third and fifth, it may be equal to or longer than the second or 

 third, or only equal to the fourth or fifth. 



Variation in the Form of the Tail, and in the Number of the 

 Rectrices. — Individual variation in the form of the tail is often quite 

 marked. In species with the tail deeply forked, different specimens 

 vary considerably in respect to the depth of the fork. Those with the tail 

 rounded and much graduated differ greatly in respect to the amount the 

 middle feathers exceed the outer ones in length. In species with a nor- 

 mally nearly even tail, the tail is sometimes distinctly emarginate, and 

 sometimes as distinctly rounded in different specimens of the same species. 



In regard to the number of rectrices, in those groups in which the 

 number exceeds twelve, as in the Rasores, the Lamellirostres, etc., the 

 number is frequently variable. The rectrices of the common ruffed 

 grouse (Bonasa umhellus) are usually eighteen in number, but an ex- 

 amination of numerous specimens shows that the number varies from 

 sixteen to twenty. The usual number in Tetrao canadensis is sixteen, 

 but the number varies from fourteen to eighteen. In Cupidonia 

 cupido, and in other species of grouse, similar variations also occur. 

 They are also frequent in the Anserina*. In Bernicla canadensis, 

 for example, the usual number of rectrices is eighteen, but the number 



