Problem (a) can be solved by proper layout of the sound path, by 

 suppressing spurious oscillations (as in Problem (d) ) and by gating the 

 receiver "on" during a predetermined "window" only. 



Noise, Problem (b) , can be overcome by providing sufficient 

 signal. 



The influence of attenuation, Problem (c) , can be eliminated 

 by an automatic gain control circuit to adjust either the transmitted 

 output power or the receiver amplification so as to deliver a constant peak 

 amplitude signal to the threshold circuit. The T.R.G. under development 

 by NUS Corporation will include A.G.C. 



In view of these possible problems and the relatively large 

 investment of time in each run, it seems therefore advisable to use two 

 velocimeters together whenever possible for oceanographic survey purposes. 

 3. Expendable Velocimeters 



For the specific purpose of sonar range prediction, a velocimeter 

 should be designed to be able to distinguish the fine structure, so that the 

 first derivative of the velocity profile can be reliably computed. The 

 absolute values of depths and velocity, i.e. scale of the velocity profile, 

 are of lesser importance. 



It is in this respect that the B.T. methods can prove unsatis- 

 factory, because they cannot take into account any salinity anomalies and 

 therefore may not always reproduce the velocity profile in the upper layers 

 above the main thermocline correctly in shallow water. 



Any instrument useful for sonar-range prediction must therefore 

 have high sensitivity to small changes in velocity, small hysteresis, i.e. 

 good short term reproducibility, and its speed of response must be appro- 



45 



Arthur 2D JUttlr Jnr. 



