Chap. 12] MISCELLANEOUS GEOPHYSICAL METHODS 941 



the sound velocity must be known to about 1 per cent or better. It is 

 in regard to the latter that the chief difficulties arise, since the effects of 

 wind, temperature, and humidity, their rate 6i change in vertical direc- 

 tion, and the corresponding curvature of the sound rays are not known. 

 To eliminate this uncertainty as much as possible, another sound-ranging 

 station may be set up between the front and the main base line, with 

 microphones arranged in a circle about a point at which small charges are 

 exploded from time to time and meteorological elements are recorded 

 simultaneously. In this manner, the relation between sound velocity and 

 meteorological data is established empirically. 



The location of the sound source is established by measuring the time 

 differences in the arrival of the sound at the various stations 0, 1, 2, 3, 

 and so on. If the sound arrives at station in <o (unknown) seconds 

 after its initiation at the source, in ti seconds at station 1, in ^2 seconds at 

 station 2, and so on, circles may be drawn about station 1 with the radius 

 v(^i — to), about station 2 with the radius v(<2 — ^o), and so on, with v 

 the velocity of sound corrected for wind velocity, temperature, and hu- 

 midity, or determined experimentally by the auxiliary ranging setup 

 described above. The source is the center of a circle with radius v<o 

 passing through and touching the circles about stations 1, 2, 3, and so on. 

 In practice, the source is located by the intersection of hyperbolas drawn 

 about the receiving station as foci. 



To eliminate calculations, a series of hyperbolas are plotted previously 

 on a large map with the time differences to be expected, and the source is 

 located by interpolation. The recordmg apparatus and amplifiers are 

 started by a sentry located between the front lines and the base, when he 

 hears the sound of the gun, or they are set in motion automatically by a 

 microphone in forward location. With some practice it is possible to 

 identify the type of gun from the character of the record; if shell bursts 

 from the same gun have been recorded, its range and thus its caliber can 

 be deduced. 



6. Direction-finding, noise-detection. Atmospheric-acoustic direction- 

 finding is concerned largel}^ with the detection of enemy scouting or 

 bombing planes when unfavorable weather or light conditions preclude 

 other ways of detection. The sound emitted by a flying plane is com- 

 posed of the exhaust noise, the ship's vibration, and the propeller noise 

 (the latter being the most predominant), and comprises a wide range of 

 frequencies. Most suitable for detection is probably the band from 300 to 

 600 cycles, corresponding to a wave length of 1 to ^ meters. Since the 



"^ The appearance of the record is much similar to that of a seismic refraction 

 record (see Fig. 9-91, first part of record) except that the impulses are of shorter 

 duration. See also plate on p. 16 (article: "Sound") Ency. Brit., 14th ed. 



