584 BRIDGMAN. 



maximum falls at the values indicated by the smooth curves. This 

 adjustment of the maximum was always slight and could be made with 

 little uncertainty. Finally, each one of these smoothed adjusted 

 deviation curves was imiformly changed in scale by the factor neces- 

 sary to make its maximum coincide with the maximum deviation 

 found from the smoothed curves of maximum deviation against tem- 

 perature, and these curves are given in the following as the " deviation 

 curves." 



It will conduce to clearness to give an example or two showing the 

 combined use of the tables and the deviation curves. I^et us find, for 

 example, the resistance of tin at 5000 kg. and .50° in terms of its resis- 

 tance at 0° and kg.* Turning to Table II, we find the average 

 pressure coefficient at 50° is O.OsOSG and the resistance at kg. at 50*^ 

 is 1.2179. If the relation between pressure and resistance were linear, 

 the resistance at 50° would be 1.2179 (1 - 5000 X 0.05936) = 1.1609. 

 But from the deviation curve. Figure 4, we find the deviation at 50°, 

 and 5000 kg. to be 0.0046. The actual resistance at 50° and 5000 kg. 

 is therefore 1.1609 - 0.0046 = 1.1563. Or let us find the initial 

 pressure coefficient of lead at 75°. From Table V, the average coeffi- 

 cient at 75° is O.O4I243. By drawing a tangent to the deviation curve. 

 Figure 7, at 75° at the origin we find that the deviation for 1 kg. at 

 75° is O.O5302. But the initial resistance at 75° is 1.3127, so that the 

 initial deviation at 75° for 1 kg. in terms of unit resistance at 75° 

 is O.O5302/I.3127 = O.O523O. Adding this to the average coeflicient 

 gives O.O4I243 + O.O523O = O.O4I473 for the initial pressure coefficient 

 of lead at 75°. 



The column in the tables headed "coefficient at 12000 kg." requires 

 a word of explanation, the meaning of "coefficient" not alwaj^s being 

 unambiguous. This means the instantaneous rate of change of 

 resistance with pressure at the temperature in question divided by the 

 resistance at kg. at the temperature in question. In other words, it 

 is the slope of the line plotting resistance against pressure, drawn to 

 such a scale that the resistance at the temperature in question and kg. 

 is taken as unity. Later in this paper I shall discuss another " coeffi- 

 cient " as 12000 kg., this time " the instantaneous coefficient." By this 

 will be meant the rate of change of resistance with pressure divided by 

 the actual resistance at 12000 kg. and the temperature in question. 



* The pressures in the tables and diagrams are gauge pressures. To get 

 absolute pressures, add approximately one kg. The difference between abso- 

 lute and gauge pressure is in almost all cases far within the Limits of error. 



