434 BULLETIN 203, UNITED STATES NATIONAL MUSEUM 



of the underbrush and low growths, where it obtains most of its food, 

 it often selects a singing perch near the top of a fair-sized tree. 



Voice. — ^Aretas A. Saunders has sent me the following careful study 

 of the song of this warbler : "The song of the prairie warbler is the 

 most distinctive one I know in the genus Dendroica. It consists of a 

 series of notes rising gradually higher in pitch to the end of the song. 

 They are separated and distinct from each other and not run together 

 in a trill or linked in 2-note phrases. The quality is sibillant, but 

 pleasingly musical and of medium loudness. 



"Songs vary considerably in details. The change in pitch between 

 the lowest and the highest notes of the song varies all the way from one 

 to five and a half tones, averaging two and a half tones. While the 

 first note of the song is the lowest in pitch, it is not always true that 

 each succeeding note is a little higher, but often the first few notes are 

 all on the lowest pitch. There may be anywhere from 1 to 6 notes on 

 this lowest pitch before the rise in pitch begins. My 35 records show 

 that 2, 3, or 4 notes on the low pitch are more frequent than only 1. The 

 number of notes per song varies from 5 to 13, averaging 9. When 5 or 

 6 notes, at the beginning, are all on the same pitch, more than half of 

 the song is over before a rise in pitch begins. In all but 4 of my 

 records the last note stands alone as the very highest, but in 2 records 

 there are 2 highest notes at the end, and in 1 there are 4. Only a 

 single record has the next to the last note the highest, and in that 

 record the last note drops to a pitch lower than the note at the 

 beginning. 



"The pitch varies from A'" to G sharp"", half a tone less than an 

 octave. The upward grading of pitch is usually in half-tone steps, 

 but is sometimes less regular, and in 10 of my records are quarter tones, 

 that could not be played on a piano or any graded instrument. 



"In a majority of records the notes are all of equal length, but in 9 

 records the first 3 to 6 notes are longer and slow, and the last notes 

 shorter and about twice as fast. Songs vary greatly in the rapidity 

 of the notes. The length of songs varies from 1% to 2% seconds. 

 The longest song, in time, contains only 9 notes, whereas the largest 

 number of notes, 13, took 2% seconds. 



"The period of song lasts from the arrival of the bird in migration 

 to the middle of July. In only four seasons have I had opportunity 

 to hear this song in July. These show an average date of July 17 for 

 the last song, extreme dates being July 14 and July 24. Earely the 

 bird revives its song in early September. In most years it is appar- 

 ently entirely silent at this season." 



Francis H. Allen writes to me: "I found one singing in the same 

 pasture with a field sparrow and singing a song that puzzled me until 

 I got a view of the singer. Comparing the songs of the two birds, I 



