56 EVOLUTION OF BIRD-SONG 



strength, the phrase is sometimes modified in 

 length and in force ; and is often modulated in 

 approximately the same successive intervals of 

 pitch, so that a kind of musical strain is produced. 

 The length of a phrase would obviously be de- 

 pendent, firstly, upon the number of breaths neces- 

 sary to its expression ; thus, if a bird employed but 

 one breath, as a cock does in crowing, the length 

 of the phrase could not be great ; but if respiration 

 could be continued during the song, as it is during 

 the extended phrases of the skylark, starling, grass- 

 hopper-warbler, and many others, the utterance 

 would be limited only by fatigue of voice. It 

 appears that many extended phrases are accom- 

 panied by a movement of the lungs similar to that 

 which occurs in us during laughter ; and this seems 

 to induce a proportionately rapid movement of the 

 wings, for when the notes are uttered slowly, as, e.g.^ 

 in the comparatively slow song of the willow -wren 

 and the long notes of the wood -wren, the wings 

 are very slightly expanded during or after each 

 note, just as they are during each caw of the rook ; 

 but when the wood -wren sings its sibilous song, 

 which is only a rapid repetition of one cry, the 

 wings are shaken with increasing rapidity as the 

 phrase proceeds to its ecstatic close, which is 



