71. PASSEECULTJS. 677 



dusky, ashy along the inner edge : " hill pale brown heneath, dusky 

 above; feet light flesh-colour; iris brown" (Audubon). Total 

 length 5 inches, culmen 0-4, wing 2-7, tail 1*9, tarsus 0-8. 



Adult female. Not different from the male in plumage. Total 

 length 4-5 inches, culmen 04, wing 2-65, tail 1*75, tarsus 0*8. 



During the breeding-season the plumage becomes much abraded, 

 the light edges being worn off. This is especially the case with the 

 rufous margins to the wing-feathers. The black streaks on the 

 underparts are strongly developed, and form a more or less well- 

 defined spot in the centre of the breast. 



The ivinter plumage is altogether browner than the summer 

 plumage, and the uuder surface is more fulvescent, all the markings 

 of the body being obscured by pale margins. 



Young in first autumn plumage. Similar to the adults, but strongly 

 washed with yellow on the eyebrow and even over the entire crown ; 

 cheek-stripe and sides of face ochreous buff; throat also spotted with 

 black as well as the fore neck, which is pale ochreous buff. 



In a female procured on the 16th May there is still some of the 

 yellow tinge on the crown remaining, as well as some of the tiny 

 black spots on the throat. 



Like Messrs. Salvin and Godmau, I have been unable to find 

 definite characters for the different races into which P. sandwich- 

 ensis has been divided, as they all seem to merge into one another. 

 Eastern specimens (P. savanna) are darker than western ones, and 

 have more rufous on the wing ; but this depends a great deal on the 

 season of the year. Western specimens are greyer and less dis- 

 tinctly streaked below, the streaks not extending so far on the 

 breast and being blacker and more defined. Individuals from the 

 high north are often larger, and some from California are extremely 

 dark. More than one of these races were collected by Mr. Gaumer 

 in the islands of the Bay of Honduras, and altogether there does 

 not seem to be more variation in the species than is seen in the 

 Meadow-Pipit of Europe (Anthus pratensis). As many of the differ- 

 ences in a series of the latter species are known to be caused by 

 variation in the time of breeding, migration, and nesting-locality, 

 the same may be found to be the case with the Passercidi of North 

 America. I have failed to find in the series at my disposal the 

 definite characters laid down by American authors ; and I have come 

 to the conclusion that under the heading of P. sandivicJiensis there 

 is only one variable species, the divisions between the races not 

 being in any way definable. That size is not a character to be relied 

 upon is shown by the measurements of the series in the Museum, of 

 which I subjoin the length of wing : — 



24-5 inches. — California ; Cozumel Island. 

 2 - 50 inches. — Cozumel. 

 2-55 inches. — British Columbia ; N. Mexico. 

 2 - 60 inches. — District of Columbia ; British Columbia ; Cozumel. 

 2*65 inches. — District of Columbia ; Indiana ; Saskatchewan district ; 

 Nebraska; Utah; Nevada; California; Mexico; Cozumel. 



