132 DRS. GUY BARLOW AND H. B. KEENE ON THE ANALYSIS OF SOUND. 
Page 
Analysis of Sound in Water.. . . . 146 
(a) Sources of Sound ..146 
(1) Cylindrical Sounder .146 
(2) Double Diaphragm Sounder .147 
(3) Single Diaphragm Sounder .148 
(4) Evinrude Row-boat Motor . 149 
(b) Receivers .149 
(1) Metal Diaphragm Receiver . 149 
(2) Rubber Diaphragm Receiver with Adjustable Natural Frequency. Listening Arrangements 150 
(o) Reservoir Experiments with Sounders .151 
Natural Frequency of Sounders and Receivers .151 
Metal Diaphragm Receiver —variation of amplitude with distance and depth .... 152 
Rubber Diaphragm Receiver —disturbances—resonance—-variation of amplitude with distance 153 
(d) Sound Spectrum of Evinrude Motor (Reservoir Experiments).154 
Variation of Spectrum with Motor Speed .155 
Variation of Spectrum with Direction . .156 
Variation of Spectrum with Depth . 156 
Variation of Amplitude with Distance .156 
Analysis by Telephone .158 
Introduction. 
A method of analysing an alternating current termed “ Analysis by Periodic Inter¬ 
ruption ” was worked out by G. Barlow in May, 1916, and is described by him in some 
detail in Part II. The present paper, Part I., gives an account of certain experiments 
in which this method has been applied to the analysis of sound vibrations in air and 
water with the object of obtaining “ sound spectra.” The principle of the method 
may be stated as follows. The alternating current circuit contains a direct current 
galvanometer and also an interrupter of which the speed can be varied over the whole 
range of frequency to be investigated. Generally the type of interrupter used is such 
that the intervals dining which the circuit is open 
(a) A ±\— 7 /*\ yT \ —7 and closed are equal. When the interruptions 
synchronize with any component A sin 2 rent of the 
current, fig. 1 (a), the galvanometer responds by 
giving a steady deflection of magnitude depending 
on the phase difference. Fig. 1 ( b ) shows interrup- 
n m 
V////////A 
(h) 
n. -r 
v///////MMw/A ' 
(C) 
V) 
tions and current in the same phase. Practically 
it is better toallow a slight difference in frequency ; 
the galvanometer then oscillates slowly to and fro 
as the phase alters. The maximum amplitude of 
Fig. 1. Interruption of a simple harmonic 
current at frequencies n, »/ 2, »/3. the galvanometer swings is then proportional to the 
amplitude A of the component current—actually 
it measures A Jti. In making the analysis the frequency of interruption is slowly increased 
over the whole range. The approach to the condition of synchronism is indicated by 
