RICHARDS AND LAMB. — SPECIFIC HEAT OF LIQUIDS. 661 



results were obtained which were concordant to within one to two-tenths 

 of a per cent. 



Almost simultaneously with Thomsen, Pfeundler* published measure- 

 ments of the specific heat of sulphuric acid at various dilutions, and 

 Marignac [)ublished an extensive series of measurements of the specific 

 heats of salt solutions, duplicating most of Thomson's, and including many 

 others. Both of these investigators employed the method of Andrews, 

 but in an improved form. They interposed a blank experiment upon 

 water between every experiment upon an unknown solution. In this 

 way they were able to correct, both for the heat exchange with the sur- 

 roundings, and for the lag of the thermometer in the calorifer. The 

 efficacy of the former correction is, as was recognized, dependent on the 

 constancy of the external conditions. The published results of Marignac 

 are means of a number of determinations, and for these he claims an 

 accuracy of from one to two one-tenths of a per cent. They agree, on the 

 average, with the measurements of Thomsen to within about two or three 

 one-tenths of a per cent. 



Pfauiidler, f before this, had devised an electrical method for the 

 measurement of the specific heats of liquids. Here a known amount of 

 heat is furnished to the liquids by the conversion of electrical energy 

 into heat. Two wires of the same resistance were prepared, and one 

 coiled up in a calorimeter containing the liquid of unknown specific heat. 

 These coils were arranged in series, and a current passed through them. 

 The rise in temperature in the two li(piids was of course inversely pro- 

 portional to the heat capacity. If these capacities were nearly the same, 

 the correction for radiation practically canceled out. If they differed 

 widely, the resistances were so altered that the temperature rise was 

 nearly the same with the same bulk of liquid. 



For use with electrolytes, the heating coils must of course be insulated, 

 and Pfaundler was unable to find a satisfactory coating for them. He 

 finally t resorted to coiled glass capillaries containing mercury. The 

 resistance of mercury, however, has so large a temperature-coefficient 

 that he was obliged to measure tlie resistance of the coils at frequent 

 intervals during a determination. His results were not especially accu- 

 rate. In the few results which he published the error is at least as great 

 as one one-tenth of a per cent, and this in spite of the fact that, by means 

 of a delicate thermopile, which measured the difference between the 



* Pfaundler, Sitzun<:sber. d. Akad. d. Wissen., Wien, 62-2, 379 (1870). 



t Pfaundler, Ibid. 59-2, 145 (1869). t Pfaundler, Ibid., 100-2a, 352 (1891). 



VOL. XL. — 42 



