DRS. GUY BARLOW AND H. B. KEENE ON THE ANALYSIS OF SOUND. 155 
The Rubber Diaphragm Receiver was found to be much more satisfactory than any 
other type tried, and was used in all the experiments described below. 
Variation of Spectrum with Motor Speed .—It was soon found that even under closely 
identical conditions the character of the spectrum often differed greatly on repeating 
the experiments. It has not been found possible to explain this variation completely, 
but it is partly due to change in motor frequency. Experiment showed that quite small 
variations in the motor frequency—variations originally regarded as of little conse¬ 
quence—may greatly affect the relative intensities of the “ lines ” in the spectrum. 
There are several ways in which this may take place :— 
(i.) In the experiments the motor speed was always changed by altering the phase 
of the ignition ; this may result in a change in the character of the vibration 
of the motor itself. 
(ii.) Mere change of speed may bring certain harmonics into, or out of, resonance 
with natural frequencies of the boat. 
(iii.) Change of speed may bring certain harmonics into resonance with the receiver. 
Experiments show that (iii.) at least is an important factor. Unfortunately, the 
frequency of the Evinrude motor (with boat fixed) can only be varied by about 10 
per cent., i.e. from 10-11/sec. 
Three examples of the variations in the spectrum are given in fig. 12, where the 
Fig. 12. Variation of spectrum of Evinrude with motor speed and natural frequency of receiver. 
Depth, 15 feet. Receiver on bottom. Distance 13 feet; broadside on. 
VOL. CCXX1I.-A. 
Z 
