WHITE-THROATED SPARROW 1381 



other types of sounds were not essential for species recognition. 

 Similarly, variations in loudness between notes were shown to be 

 unimportant, but songs that fell partly or wholly above or below the 

 range of pitch Borror and Gunn (1965) found in their sample elicited 

 very weak responses. Falls summarized the results of his experi- 

 ments as follows: "To be effective, a white- throated sparrow's song 

 must consist of unvarying pure tones within a certain range of pitch. 

 Less important is the presence of notes of different pitch arranged in 

 a certain pattern. The notes should be of a certain minimum length, 

 and the intervals between notes should not exceed a certain maxi- 

 mum." This study dealt only with reactions of territorial males to 

 song, and properties of song that appeared to be unimportant in this 

 context may serve other functions. 



To determine which characteristics of song are important for 

 individual recognition, Brooks and Falls (unpublished data) played 

 altered neighbor's songs to territorial males. The birds responded 

 normally (see above) to songs altered in length, but strongly to songs 

 altered in pitch, which they apparently no longer recognized as 

 neighbors' songs. Thus pitch appears to be more important than 

 timing for individual recognition. 



Regarding inheritance of song, Borror and Gunn (1965) state, 

 "Certain general features are common to all the White-throat song 

 patterns we found (clear, whistled notes steady in pitch or nearly so, 

 uually one or two pitch changes through the song, and the song 

 ending in triplets), so it may be assumed that these features are 

 hereditary." They suggest further that the different patterns are 

 learned, and that the 11 uncommon patterns may be inaccurate copies 

 of the four more common patterns. 



Bruce Thorneycroft (unpublished data) raised a sample of nestling 

 white-throats in isolation where they could not hear songs of adults. 

 In the songs these birds developed none of the characteristics Borror 

 and Gunn list was invariably present, and none of the songs contained 

 triplets. This experimental evidence implies that the inherited basis 

 of song in this species may be less than Borror and Gunn suggest. 



In view of the unusual plumage variation in this species (see 

 Plumages), it is of interest to consider the singing behavior of the 

 various color types. No differences were noted between songs of 

 white-striped and tan-striped males, and a recording of each elicited 

 similar responses from both sexes. White-striped males, however, sing 

 more often than tan-striped males both spontaneously and in response 

 to recorded songs. In five-minute counts of spontaneous singing early 

 in the morning, Lowther (MS.) found an average of 5.27 songs for 

 55 white-striped males compared with 1.34 for 44 tan-striped males. 

 White-striped females occasionally sing early in the breeding season 



