50 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
so that it stood out horizontally about. 1 meter (TO in.) in the plane of that # end 
(fig. 2). From the free end of the beam a bass-viol string was stretched to the 
opposite side of the sounding-board. This string could be tightened by a bolt and 
nut at the free end of the beam, and it was made to pass over a bridge placed near 
the middle of the sounding-board. The length of the string from the attached end 
on the sounding-board to the bridge was 25 cm. (10 in.), and from the bridge to 
the attachment near the free end of the beam 1.15 meters (45 in.). Thus the end 
of the aquarium might be regarded as something like a large one-stringed bass viol 
resting sidewise, with the sounding-board for a body and the beam for a neck. 
When the string was tightened and plucked or bowed a good tone was obtained, 
which was transmitted directly through the sounding-board to the water within the 
aquarium. On keying the string up to a good clear tone, I found by writing off its 
vibrations on a revolving cylinder that it produced on an average 40 per second, 
and I retained this pitch by frequent adjustment for the experiments that I subse- 
quently performed. I was led to adopt this low tone since most of the noises that. I 
have heard fishes make were in the nature of low-pitched grunts. 
Each time the string was plucked the note began with maximum intensity and 
then gradually died away. It was, consequently, impossible to get any very signifi- 
cant record of the intensity, but I endeavored to use the apparatus in a uniform way 
by drawing the string out a constant distance from its position of rest each time 
I plucked it. The distance usually employed was about 1.5 cm. (0.6 in.). The 
amount of weight required, when hung at the middle of the longer segment of the 
string, to depress it 1.5 cm. was found to be about 2.15 kilograms (4.75 pounds), 
so that each time the string was liberated on being plucked in the usual way, it 
moved forward with an initial force equal to the pull of 2.15 kilograms, a rough 
measure of the maximum intensity of the sound produced. 
The fishes to be experimented upon were not allowed to swim unrestricted' in 
the aquarium, but they were placed in a small cage (fig. 1) suspended from a cord 
attached at its ends to the Avails of the room. Thus the support for the fish cage 
Avas entirely independent of the walls of the aquarium and any vibrat ion that reached 
the fishes must have done so almost entirely through the water. The cage could be 
moved in a horizontal direction back and forth on the cord, and thus the fish could 
be placed at any desired distance from the sounding-board up to 75 cm. (30 in.). 
The inside measurements of the cage Avere as follows: Height, 10 cm. (4 in. ) ; length, 
20 cm. (8 in.); and breadth, about 10 cm. (4 in.). The bottom of the cage Avas wood, 
padded on the inside Avith cotton wool covered with cloth to provide a deadened 
surface on which the fishes might rest. The top and three sides were glass; the 
fourth side was made of coarse netting to retain the fish but to interfere as little as 
possible Avith the entrance of sound, and this side Avas always directed toward the 
sounding-board. As the fishes averaged about 7 cm. (2f in.) in length, the cage 
gave them ample room for moving about. 
My plan was to introduce fishes in various conditions into the cage, and, after 
they had become accustomed to their surroundings, to subject them to stimulation 
by sound and observe their reactions. I found it desirable to experiment Avith three 
classes of fishes; first, normal ones for a basis of comparison; secondly, fishes from 
which the ears had been removed; finally, fishes in which the general integument 
had been rendered insensitive, but in which the ears were intact. The methods of 
