22 
The effect of temperature variation in the environment on the 
temperature sensing system is not known. The temperature of the 
usual environment is quite constant, and thus this difficulty is not 
important. In any event, the use of the automatic calibration re- 
moves these sources of inaccuracy. 
The pressure transducer which indicates depth has errors 
due to nonlinearity and hysteresis specified to be 3/4 percent of 
full scale reading (1000 ft). In practice, only a small interval is 
used during a test run, and hence nonlinearity and hysteresis are 
greatly reduced. The depth gage in the submarine control room is 
used to calibrate depth above and below the thermocline (say an in- 
terval of 50 ft) and at one or more points in the thermocline. Not 
enough work has been done to give the precision or accuracy of 
these measurements; the error in the relative depth is probably a 
fraction of a foot. 
DATA REDUCTION 
It is clear from the above descriptions that the record ob- 
tained gives the indicated temperature at several points on the 
strut, as well as the strut depth. The object of the data reduction 
described here is to yield the depths of one or more true isotherms 
as a function of position along the track of the recording submarine. 
The analysis of such a signal (see fig. 13 for an example) is not 
considered here. 
LINEARIZATION OF THE OUTPUT OF THE 
THERMISTOR BRIDGES 
When the techniques of Appendix B are properly applied, the 
output voltage is so very nearly a linear function of the temperature 
(over a restricted range) that no correction need be applied. If the 
bridge parameters are not quite optimum or if too large a temper- 
ature range must be spanned, there is a departure from linearity. 
