168 ANNALS NEW YORK ACADEMY OF SCIENCES 



animal on its side., the writing point shoiild not move up and down in a perpendicular 

 plane, but should move horizontally, back and forward. This, then, should give in 

 the kjTnograph reading a series of dots, repre.5enting the apes of each cur\-e. Such 

 a tracing was recorded on the drum of a kj-mograph. The ^■ibrations were 24 per 

 second. From this experiment it is seen that the muscle itself and the abdominal 

 tissue \-ibrate at exactly the same rate. 



Experiment XV. — Another experiment was made to show the effect of sub- 

 stituting an inflated rubber balloon for the air-bladder. The nimiber of vibrations 

 was 24 per second. It is thus evident that the \-ibrations produced in the presence 

 of the rubber balloon are the same as in the normal condition of the animal. 



The five preceding experiments agree in the rate of ^■ib^ations of the 

 abdominal part which is in immediate relation to the drumming muscles, 

 and which is directly connected ^vith sound production according to our 

 present views. It vras next necessary to record the contractions of the 

 muscle itself, and for this purpose the following two experiments were 

 performed. 



Experiment XVI. — An incision two inches long was made in the mid-ventral 

 line just posterior to the pectoral fins. Through this op>ening was inserted a slender 

 v,-ire ■n-ith a sharp hook on one end and an eye on the other. The hook v.-as fastened 

 directly into the fibers of the m. sonificus. The eye was attached to an ordinary 

 muscle-lever which was supplied with a writing point. The kymograph was then 

 placed so that it would receive the tracings made by the -^-riting point. 



In this experiment none of the \"iscera were disturbed and the noise produced 

 differed in no way from that of the normal animal. To measure the time, a tuning 

 fork was used, whose double vibrations of 100 per second were registered on the 

 revoh-ing drum. The rate of the contractions is 24 per second, which is identical 

 with the experiments made on the abdominal walls. As is shown in PI. 'STII, fig. 

 3, the ampUtude of the contractions is much more than in the experiments made 

 on the abdominal walls. This is undoubtedly due to the release in tension caused by 

 the separation of the right and left abdominal portions to which these muscles are 

 attached, together with some resistance caused by the rubber bands. To determine 

 this point another experiment was made in which the \"ibrations of the abdominal 

 wall were registerd by a wire hook attached at one end to the rectus abdominis and 

 the other to the muscle-lever. With this method the ampUtude of \-ibration is 

 nearly the same as that of the muscles. It was noticed, too, that when the ampli- 

 tude was the greatest the loudest sound was produced in both the experiments on 

 the abdominal waUs and on the drumming muscle. 



Experiment XVII. — Experiment X\T was repeated, except that the air-bladder 

 was punctured. The drumming noise stopped. The contractions of the drumming 

 muscles, registered as in the preceding experiment, are given on PI. ^'III, fig. 4. 

 The number of contractions computed from those of a tuning fork is 24 per second. 



It is veiy- e\ndent that there is no difference between the contractions 

 when the swim-bladder is full of air and when it is collapsed, and that this 

 organ has no effect upon the contractions of the drumming muscle. This 



