1380 U.S. NATIONAL MUSEUM BULLETIN 23 7 part 3 



lies in the relative incidence of the various song patterns just de- 

 scribed. It is usually easy to assign songs of individual birds to one 

 of these patterns, but differences are recognizable among songs of the 

 same pattern. Borror and Gunn (1965) state that, "songs of each 

 pattern varied in the types of notes in the songs, the pitch of the 

 first note, the amount of pitch change between notes, the length of 

 the various notes, the length of the song and the number of notes it 

 contained, and the relative loudness of different notes in the song." 

 As already mentioned, songs of a given individual may vary con- 

 siderably in length, but in other respects they vary much less than the 

 songs of different birds. Thus songs of individual birds can often be 

 recognized even from others of the same pattern, and the character- 

 istic features appear to be retained from year to year. D. J. Borror 

 was able to recognize seven of nine birds occupying a 30-acre point at 

 Hog Island, Maine, by their songs, and the songs of the other two 

 proved to be different when studied with the sound spectrograph. 

 Several birds occupying the same general area from year to year had 

 songs of the same pattern and pitch characteristics each year. Al- 

 though not individually marked, they were probably the same birds. 

 The experience of Falls and his students with songs of individual 

 birds has been similar. 



To determine whether white-throats were capable of recognizing 

 different individuals by song, Ronald Brooks (unpublished data) 

 played recorded songs of neighboring and distant individuals to 

 males in their territories. Reactions were stronger to the strange 

 songs, even if they were of the same pattern as the neighbor's song. 



About three percent of the birds Borror and Gunn studied had two 

 songs usually of different patterns. Harold Axtell noted the same 

 phenomenon and writes (in litt.), "Usually the two songs are very 

 different and one is given much less frequently than the other." 

 Though an observer is apt to assume that the occasional quite differ- 

 ent song comes from another bird, close observation might show this 

 to be a more common situation than the above percentage indicates. 



In order to determine which characteristics of the white-throat's 

 song were important in species recognition, Falls (1963) played 15 

 artificial songs, generated by means of an audio-oscillator, to a large 

 sample of males on their breeding territories. The test songs varied 

 in the character of the notes, pitch, pattern of pitch change, and 

 the length of the notes and intervals between them. Birds responded 

 to a normal song by giving songs and calls and approaching the 

 speaker. The importance of the different characteristics was judged 

 by comparing the responses to test songs with responses to normal 

 songs and to those of other species. As a song consisting only of 

 continuous notes elicited normal responses, it was concluded that 



