364 



K. S. LASHLEY 



upon the piano and violin. The bird sang readily enough in 

 response to these sounds but did not reproduce their pitch. 

 More frequently the response consisted of several notes, often 

 with the addition of words as, "Oh, little birdie, Oh," sung with 

 five changes in pitch. This is in accord with his manner of 

 singing alone and it was found so difficult to obtain single notes 

 that singing tones were soon abandoned. 



Whistling tones were next used as stimuli. The bird's whis- 

 tling register is somewhat more extensive than his singing one. 

 It extends upward from F m for about two octaves. The re- 

 sponse to a single whistling tone was usually a single note also, 

 but sometimes several notes were sounded. In the latter case 

 the first note only was compared with the stimulus. 



The results of this experiment have been combined in a cor- 

 relation table, printed below. In this table the intervals rep- 

 resented are each two half tones on the chromatic scale, begin- 

 ning with F 384 . 



Stimuli 

 1 2 3 4 5 6 7 8 9 10 11 12 13 . 



Table of Correlation in Pitch Between Stimulus and Response 



The coefficient of correlation obtained from this table is 

 0.658 ±.059. This expresses a degree of likeness between stim- 

 ulus tone and response that is far too great to have been due 

 to chance alone, and indicates that the bird actually tends to 

 reproduce the pitch of the stimulus. 



EXPERIMENT 4. MELODY 



A number of attempts were made to get the bird to reproduce 

 several successive notes. Stimulation by short melodies usually 



