266 DR. LOUIS YESSOT KING ON THE PROPAGATION OF SOUND IN THE FREE 
In fig. (i.) are drawn on a large scale the gradients of phonometer readings obtained on August 30, 
1913. According to test 1 of Appendix III., the acoustic output corresponding to an operating pressure 
of 19-9 lbs./sq. in. is l - 6 H.P. Expressing the distance R in feet, we obtain for the phonometer reading 
d the expression 
d (mm.) x [R(feet)/1000] = 19'4 .(iii.) 
from which the theoretical gradient shown in fig. (i.) has been drawn. It will be noticed from the 
Fig. (i.)- Acoustic gradient in the neighbourhood of Father Point diaphone, August 30, 1913. 
comparison of the two curves that a very large proportion of the atmospheric losses occur within a radius 
of l mile. When once a sound-wave has traversed this distance, it is evident from fig. (i.) that on a calm 
day the subsequent losses are small. This conclusion is supported from an inspection of the acoustic 
gradients observed on other days. 
The curious undulatory character of the gradient is well shown in fig. (i.), and lends support to 
Taylor’s theory respecting the effect of atmospheric eddies on the attenuation of sound. The eddies may 
be supposed to be travelling with the average speed of the wind (3 to 4 statute miles per hour): the 
undulations are produced when the observing ship crosses successive eddies. The velocity of the ship was 
about 7 statute miles per hour. By a simple calculation the diameter of the eddies corresponding to a 
light breeze of 3'5 statute miles per hour, comes out to be about 700 feet or 230 metres, a result which 
may not be impossibly large.* Future observations in this connection should be carried out with a sound- 
generator (if possible of adjustable pitch) giving a continuous note, preferably from a ship anchored at a 
considerable distance from the land. Fluctuations of sound amplitude should then be measured by means 
of a phonometer situated at fixed distances directly to leeward and windward. In addition to obtaining 
wind-velocity, direction, and other meteorological data by self-registering instruments, means should be 
taken to observe the velocity gradient of the wind and its gustiness, from which the characteristics of the 
* According to Taylob (“Eddy Motion in tlie Atmosphere,” ‘ Phil. Trans.,’ A, 215, 1915, p. 22), the average diameter 
of an eddy in a wind of velocity 7 metres/sec. (21 statute miles per hour) is given as 40 metres. In light winds the diameter 
increases. As yet data on the subject are very incomplete. 
