246 DR. LOUIS VESSOT KING ON THE PROPAGATION OF SOUND IN THE FREE 
be expected that further work with the Webster phonometer will contribute 
materially to our knowledge of atmospheric structure. 
By examining the distribution of sound along circular arcs at different distances 
from the source it is concluded that the greater part of the atmospheric losses occur 
within half-a-mile, and very probably within a few hundred feet of the siren trumpet. 
To measure the very intense sound-waves at this close range a less sensitive phonometer 
without a resonator is recommended for future work. The acoustic wave-form within 
the trumpet itself should be studied in connection with the theory of the propagation 
of waves of large amplitude. Theory shows that, in an ideal medium, plane sound¬ 
waves transmitting only a small fraction (from 5 to 8 per cent.) of the energy 
available as compressed air across the area of the narrowest part of the trumpet 
could only travel about 10 feet before discontinuity sets in. In reality, the tendency 
to form a discontinuity will be retarded owing to the effect of viscosity and thermal 
conductivity and owing to the divergence of the waves in three dimensions. It is 
suggested, however, that the serious energy losses known to exist in the conversion 
of energy of compressed air into sound may be closely associated with phenomena 
of finite wave propagation. The experimental solution of this problem will be 
necessary before we are able to ascertain whether there is any limit (as in the case of 
ordinary heat engines) to the proportion of power capable of being converted into sound. 
In order to obtain some preliminary information on this question, of considerable 
practical importance in the design of fog-signal generators, the acoustic output of the 
diaphone was estimated by measuring the difference of temperature of the air on the 
high- and low-pressure side. Specially wound resistance thermometers of minimum 
heat capacity enabled temperature differences to be measured to 1/100° C. This 
temperature difference is a measure of the external work done by the compressed air, 
propagated away as sound. By means of a few simple thermodynamical formulae the 
acoustic output may be calculated in terms of this temperature difference. It 
was found that under the best conditions nearly two and a-half horse-power may be 
delivered by the diaphone as sound. In rating the efficiency of the diaphone it was 
necessary to adopt a standard of comparison. That chosen was an ideal siren 
operating on an adiabatic cycle and capable of converting (in some way as yet 
undiscovered) all the energy of compressed air into sound. In these terms the 
acoustic efficiency of the diaphone under the best conditions came out to be a little 
over 8 per cent. It is probable that the larger and more modern types of diaphones 
have acoustic efficiencies considerably greater than the above figure. The estimates 
of acoustic output obtained in these experiments must, therefore, be regarded as 
provisional, pending further work on the subject. 
There remains the pleasant duty of acknowledging valuable assistance and 
co-operation from many quarters. The writer is much indebted to Prof. A. G. 
Webster, of Clark University, for his kindness in undertaking to supervise the 
construction of a “phonometer” specially designed for the Father Point tests, and 
