768 TRUEBLOOD. 



passing in unit time), 8Q is the heat leakage in unit time (counted 

 positive if heat is received by the fluid), Cp is the specific heat at 

 constant pressure and AT and Aj) respectively the temperature and 

 pressure drops, assumed to be so small that /jl and Cp are sensibly 

 constant over the temperature-pressure interval covered in the 

 experiment. 



To eliminate the effect of heat leakage, experiments at approxi- 

 mately the same pressure and temperature, but with different rates 

 of flow, must be conducted. The flow may be varied without varying 

 the temperature drop by using different plugs, and if it may be assumed 

 that the various circumstances which affect the total heat-leakage 

 remain the same for all members of a set of experiments made in this 

 manner, it will be possible to eliminate the effect of heat-leak from 

 the results. A second method of varying the relative magnitude of 

 that portion of the observed effect which is due to heat-leak is to vary 

 the pressure drop and the flow together, using the same plug through- 

 out the set of experiments. 



The first of these methods involves more labor and greater experi- 

 mental difficulty than the other, and although, in the work under 

 discussion, a number of plugs giving widely different flows at the same 

 pressure drop have been used, the second method is the one which has 

 been employed in arriving at a result that is believed to be free from 

 leakage errors. 



The application of this method to the elimination of the heat leak 

 effect depends, 1°, upon an experimental fact, and 2°, upon an assump- 

 tion. 



1°. The experimental fact is that when a set of fj.' 's obtained with 

 a single set-up of the apparatus are plotted against the reciprocals of 

 the corresponding flows, the curve so determined is a straight line, 

 within the limits of experimental error. (One or two possible excep- 

 tions to this statement will be noted below.) It follows from this 

 that, within the range of experiment, the leakage term in equation (2) 

 must be of the form 



(2a) -J^=A + ^ 



in which A and B are constants. B is evidently the negative of the 

 slope of the experimentally determined straight line just mentioned. 



Plots of n' vs. - are shown in Figs. 12 and 13. Table III contains 



