276 ADAMS AND JOHNSTON: STANDARD TEMPERATURE SCALE 



in the temperature scale resulting from the new and more accurate gas 

 thermometer determinations at temperatures up to 500.° 



Too much reliance can not be placed on the readings of copper-con- 

 stantan thermoelements at the zinc-point (419?4), for some diminution 

 of electromotive force sets in, not serious, but sufficient to preclude the 

 most accurate measurement; for this reason we have ceased to make 

 use of the zinc point as a calibration temperature. 



The calibration temperatures, expressed in the corrected scale, 

 together with the corresponding values of electromotive force of the 

 standard element, are given in Table I. 



TABLE I 

 Calibration Temperatures 



* The benzophenone used was from Merck, which melts at (46?9) 0?3 lower, and 

 boils 0?2 higher, than that obtained from Kahlbaum (Waidner and Burgess, Bull. 

 Bur. Standards, 7: 3-11; 1910). The f. p. of our Merck benzophenone was also 

 46?9; consequently we have added 0?12 to the accepted b. p. (305?9) of Kahlbaum 

 benzophenone. 



To reproduce the above data, a quadratic equation is insufficient, 

 except over a very short range, and so is the cubic equation 3 of the form 

 t = A e + B e 2 -\- C e z ; but they can, as we found, be fitted very closely 

 by the inverse form of function, e = A t + B t 2 + C t 3 . Accordingly, 

 on this basis a least square solution for all the points in Table I was 

 made; this resulted in the equation 



e = 38.105 t + 0.04442 t 2 - 0.00002856 P, 

 from which the figures in the third column of Table I have been com- 

 puted. The agreement is excellent; it can not, however, be used as a 

 valid argument in favor of the accuracy of either the temperature scale 

 or of the measurements, as anyone can readily convince himself by work- 

 ing with a number of similar cubics. By computation from this equa- 

 tion, a table giving t in terms of e was constructed. 4 



* This was tried because its use would have saved so much trouble in calculating 

 the most convenient form of table — that giving t for round values of e. 



4 This table is presented in full in the more complete paper in Am. Jl. Sci. 



