62 Journal of Comparative Neurology and Psychology. 



changes in the syrinx. Observation of the birds leads me to believe 

 that they have no control whatever over the breaking of the voice ; 

 it is purely mechanical. 



First appearance of the coo. — The first attempts at cooing usually 

 appear much later than the alarm and the kah. Only in one case, 

 in a bird which showed other signs of precocity, did the song originate 

 on' the same day as the kah, the twenty-seventh day. In the nest- 

 mate of this bird the coo was not heard until the fortieth day ; in 

 another bird not until the forty-seventh day, and not decisively until 

 the fiftieth day. The first coo is thus very variable in the time of 

 its appearance. And it is equally variable in its character. The 

 varial)le character of the early cooing is shoAvn in this quotation from 

 my notes. The young bird '"takes few hasty steps toward mother on 

 perch, head directed toward her, giving kah in squeaky voice. He 

 repeats this about three times, then stands up straight and stiff (in 

 attitude of male in the up phase of the bowing-coo), then he bows, 

 down and up several times, making not a sound. He goes through 

 much the same performance two or three times. Little later he 

 gives coo in a purely squeaky voice (pitch a") without bowing." 

 In this case, then, the perch-coo and the bowing-coo apparently de- 

 veloped at the same time ; Imt, while the perch-coo was audible, the 

 bowing-coo, if sounding at all, had not passed the threshold of audi- 

 bility. In many cases, however, I think that the bowing-coo precedes 

 the perch-coo by a day at least. This was tnie of the nest-mate of 

 the bird just referred to: "After giving the kah in its squeaky 

 voice, bird went through bowing motion, roughly, making just a single 

 short note now and then, pitch g', sometimes two notes, with time- 

 interval between, and second higher than first by one tone or less." 

 In another case the first coo was a "perfectly nondescript sound. 

 Much resembled its attempts at kah, but notes more irregular, some 

 of them more prolonged." All these accounts go to show that the 

 coo at its first appearance is not only variable but extremely imper- 

 fect. While the alarm-note springs into being perfectly formed, 

 as it were, and tlie kah, at its inception, is almost as perfect, the coo, 

 at first, is an insignificant fragment which does not in the least 

 suggest the sound it is ultimately to assume. It seems that in those 



