OF ARTS AND SCIENCES. 231 



the element of memory enters, since the pitch of each sound must 

 he carried in the mind. With a direct comparison of the sound 

 heard through the telephone line with that given by a tuning-fork 

 sounded by the observer at the receiving end, it is probable that a 

 still larger percentage of results would be correct. This remains, 

 however, to be actually proved by experiment. 



It will also be noticed from the tables that in some cases it seems 

 easier to judge correctly when the two tones make a small interval 

 with each other than when they make a greater one. For example, 

 see the comparisons of C 3 and D 3 in Table VIII. That this is true 

 when the greater interval is an octave might well be expected from 

 the well known liability to confuse notes of this interval ; but it is 

 true of notes of other intervals which are not likely to be mistaken 

 for each other. 



It is our intention to extend these observations to briefer and 

 longer durations of the sound, and to ascertain what number of 

 vibrations is necessary to recognize the different musical intervals 

 when these are entirely unknown beforehand. To obtain louder 

 souuds it may be possible to use a microphone, although we have 

 avoided this because of the great liability to disturbance of such a 

 transmitter when strong vibrations actuate it. It has also occurred 

 to us that we may use a sinusoidal electrical wave, generated by a 

 peculiar form of alternating dynamo machine, especially devised for 

 the production of such a current by Mr. F. A. Laws, of the Rogers 

 Laboratory. This will give a very loud and at the same time very 

 pure tone. We also intend to study the comparative accuracy of 

 estimation when each sound is heard only once, as in the observa- 

 tions detailed in this paper, with that which is obtained when the 

 sound is repeated two or more times. 



The question may be raised, regarding the method employed by 

 us, whether the sound given out by the diaphragm was really of as 

 brief duration as we have assumed it to be. It might be questioned 

 whether the vibration of the diaphragm was not in fact prolonged; 

 so that while the duration of the electric current was, for example, 

 considerably less than the time occupied by two complete vibrations, 

 yet more than two complete sound-waves were actually produced. 



Of course, the instant the diaphragm ceases to be actuated by the 

 electro-magnet of the receiving telephone, it will begin to assume 

 its natural rate of free vibration; but how minute a fraction of a 

 second will elapse before it ceases to move at substantially the rate 

 impressed upon it might seem doubtful. But under the circurn- 



