524 HERSEY AND BACKUS [CHAP. 13 



The signal available for analysis is the output voltage of the transducer. This 

 electrical signal may be fed to filter circuits or it may be recorded, for example, 

 on magnetic tape. Let it be e{t). Then remember that I{t) is given by 



m = pHt)ipc. 



Now 



e{t) = Bf^pit), 



where B/^ is the factor of proportionality between the acoustic pressure of a 

 wave, arriving along the direction of greatest sensitivity of the transducer, and 

 the open circuit voltage generated by the transducer. Therefore, 



m = eHt)lpcBf^^. 



J'hus, remembering that z = ctl2, equation (38) becomes 



The quantity e{t) is measured by comparing recordings of the reverberation 

 with recordings of a "white-noise" calibration signal introduced across a 

 resistor in series with the transducer. Other quantities are given or measurable ; 

 thus m{z) is obtainable with (40). 



Machlup and Hersey (1955) used this result to cpmpute the reverberation 

 due to several different hypothetical distributions of scatterers. They also 

 reported the results of computations of a series of oscillographic recordings 

 made directly at sea from the transducer through band-pass filters. A number 

 of tape recordings made since 1953 have been analyzed recently by the method 

 of equation (40) ; these results, which are pertinent to the general problem of 

 frequency dependence, are now discussed. 



5. Sound- Scattering Observations 



The data on which this discussion is based were collected in the areas in- 

 dicated in the index chart of Fig. 14. The observations used for these analyses 

 were made with a small charge (| pound of T.N.T. or 2| pounds of tetratol) 

 as the sound source fired at shallow depth (three feet or less). The hydrophone 

 was in most instances the QBG transducer. Fig. 15 shows directional properties 

 at 5 and 27.5 kc/s and the broad-band receiving response of this transducer 

 along its principal axis. The preamplifier was the "Suitcase" amplifier (see 

 Hersey, 1957) and the recorder was a Magnecord or Ampex magnetic tape 

 recorder used for high-speed direct recording. 



For examination of frequency behavior, the tapes were played to a Kay 

 Electric Company Sonagraph or Vibralyzer. These instruments present the 

 spectrum of a transient sound as relative blackening of a spark-type recording 

 paper in which the plane co-ordinates are frequency and time (Koenig, Dunn 

 and Lacy, 1946) (Fig. 16). The samples of scattered sound are analyzed over 



