TOWER, PRODUCTION OF SOUND IN CERTAIN FISHES 171 



Experiment XXI. — The swim-bladder was removed from a fresh specimen and 

 laid upon the operating table. The nerves and the muscles of the bladder were 

 then stimulated successively as in experiment XX. In each case there was a grunt 

 of the same pitch and intensity as is produced by the normal animal. 



This shoAvs very clearly that the sound-producing mechanism of the 

 sea-robin is entirely within the bladder and its intrinsic muscles. This 

 mechanism, then, stands in direct contrast to that of the drumfishes, just 

 discussed. 



Experiment XXII. — The swim-bladder was removed from a sea-robin. The 

 muscle was stimulated and an audible grunt was produced. The bladder was then 

 placed on an improvised registering apparatus, so arranged that the bladder was 

 connected with a muscle lever and writing point. The muscle was then stimulated. 

 An audible grvmt resulted. The vibration of the bladder was registered on the drum 

 of the kymograph. The grunt is produced by one single sharp contraction of the 

 intrinsic muscle (PI. VIII, fig. 6). This was repeated each time that the muscle was 

 stimulated. 



a. One of the lobes of the bladder was now punctured. Both lobes collapsed. 

 Through the opening was inserted the rubber balloon (collapsed). This was in- 

 flated; the muscle was thus superimposed over the inflated rubber balloon. The 

 muscle was then stimulated as before. It contracted and produced a grunt the same 

 as in the isolated bladder full of air. Moreau (1876) concluded that it was the 

 vibration of the perforated internal septum which was the direct cause of phonation. 

 That this septum vibrates is true, but from the foregoing experiment it would seem 

 that the walls of the air-bladder are the chief Aabrating organ. In the sea-robin 

 the left lobe only possessed the internal septum, but it made no difference with the 

 sound produced whether the right lobe or the left lobe was used for the experiment. 



b. The uninjured lobe was filled vnth salt water and closed by a ligature. The 

 muscle was then stimulated by a current from an induction coil. A grunt occurred 

 as when the swim-bladder was filled \vith air, although not so loud. These con- 

 tractions were recorded by means of a kymograph and are given on PI. VIII, fig. 7. 

 On comparing the record with those given on PI. VIII, fig. 6, it is evident that the 

 curves have about the same amplitude, but are not so well sustained. 



Experiment XXIV. — The swim-bladder was removed from a sea-robin as 

 quickly as possible. The muscle was stimulated by a current from an induction 

 coil. An audible grunt resulted. This sound was more intense when the bladder 

 rested on the table. It is interesting to note that this particular animal did not 

 produce any noise while alive. The isolated bladder was then placed on the regis- 

 tering apparatus, and records were obtained under single stimulations and also by 

 stimulations continued for several seconds. The records are given on PI. VIII, fig. 8. 

 The character of the curve is changed by the continued stimulation, the muscles 

 going into incomplete and then complete tetanus. Tetanic contraction does not 

 appear to be the normal procedure, but is produced by artificial stimulation. And 

 as far as could be determined, the sound was produced at the beginning of the tetanus, 

 i. e. at the first up-stroke of the lever, and died out during the remainder of the 

 contraction. The loudest grunts were produced at single full contractions of the 

 intrinsic muscles. The sovmd produced starts with a grunt, which gradually dies 

 out. It does not resemble drumming. 



