RUDDER TRIALS, U. S. S. STERETT. 323 



table. The load was then reduced by steps to 300 lbs., when the force- 

 diagram pen again coincided with the fixed base-line. Throughout the 

 calibration, both the base-plate and table were rapped and jarred continu- 

 ously to remove any binding effect of friction on base-plate or in pins. The 

 straining head was stopped at each successive load, both increasing and 

 decreasing, long enough to get a good force-diagram at constant load, from 

 which to measure the ordinates for the calibration curves shown on Fig. 8, 

 Plate 118. 



These curves all show a deflection of zero at 300 lbs. load, the abscissae 

 being pounds load, while the ordinates are linear units (centimeters for the 

 particular cross-section paper used), and were taken directly from the con- 

 tinuous force-diagrams obtained on calibration. The calibration curves 

 were subsequently used to pick out the maximum and steady values from 

 the force-diagrams obtained both on the turning trials and on the friction 

 tests. Care was taken in all cases to pick out increasing and decreasing 

 stresses on the proper curves, the latter lying uniformly above the former. 



The average deviation of the points from the calibration curves is about 

 .05 cm., which amounts to an error of less than one per cent of the maximum 

 ordinate. The precision of reading the scale of the testing machine is less 

 than one-quarter per^cent. 



