HOW THE TESTS WERE CONDUCTED. 29 



The calibration of the indicator springs used on the tests 

 reported in this volume has in many cases been carried on by 

 testing them under the action of dead weights, and correcting 

 the result thus found by a percentage of allowance for the 

 reduced tension caused by the heat of the steam in which they 

 ordinarily work. The author's testing-apparatus consists of a 

 scale-beam mounted on knife edges, on one end of which the 

 weights are suspended. The movement of the beam at the 

 other end is transmitted upward by means of a vertical adjust- 

 able rod extending to the under side of the indicator piston. 

 The tests are made with the highest pressure to which the 

 springs are subjected, and from this point down to the atmos- 

 phere at uniform reductions. The apparatus is operated so as 

 to get an average reading, whether the pressure is going up or 

 going down. This is done each time by pushing the scale-beam 

 down as far as it will go, and drawing a line on the indicator- 

 card, then, without changing the weight, pushing the same 

 upward as far as it will go, and marking another line. When 

 the lines are measured, the mean of the two is selected as the 

 true indication. The springs are in some cases compared under 

 different pressures with a correct steam gauge, admitting the 

 steam directly into the indicator, and subjecting it as near as 

 possible to its working conditions of temperature. In making 

 calibrations under steam, difficulties are often experienced in 

 obtaining satisfactory indications, owing to the friction of the 

 piston of the indicator under the action of the continuously 

 applied pressure. This is overcome, provided the pressure is 

 maintained at a constant point, by drawing two lines with the 

 instrument, one when the pencil-arm is pushed down with the 

 finger as far as it will go, and the second when the arm is 

 pushed up as far as it will go, the true indication then being 

 taken as the mean of the two. When a set of indicator springs 

 has once been calibrated, and their exact scales obtained, the 

 dead-weight apparatus above referred to furnishes a much more 

 satisfactory means for future determinations, and for showing 

 the changes in the scale which may take place under continued 

 use, than the steam-testing apparatus, for the reason of its 



