Concerning the Thermodynamic Correction for an Air Thermometer

Abstract

IT seems noteworthy not only that the theoretical investigation on this subject, given in Maxwell's “Theory of Heat” and Tait's “Heat,” is fallacious, as has been pointed out by Mr. Rose-Innes in his letter which appeared in your issue of May 26, but that there is no indication in either of these text-books that the result given is not in accordance with the experimental facts. The investigation in question, which is practically the same as that first given by Lord Kelvin before the experiments were performed, leads to the result that the cooling effect should be proportional to the difference of the logarithms of the pressures before and after passing the plug, instead of proportional to the difference of the pressures themselves, as was actually found to be the case; the discrepancy was, of course, noted'by the experimenters. The methods used in all the theoretical discussions in Lord Kelvin's papers on this subject prior to the final one in Part iv. of the papers “On the Thermal Effects of Fluids in Motion,” seem equally open to objection. The numerical results derived by the-last method, which seems perfectly sound, were given apparently for the first time in the article “Heat” in the “Encyclopaedia Britannica”; the investigation and results are reproduced in Preston's “Theory of Heat,” and substantially in Parker's “Thermodynamics.” The correction obtained to the number representing the temperature of melting ice on the absolute scale by this method was + ˙70, the previous methods giving in succession + ˙83, –˙3 and + .87; the uncorrected estimate finally used was also different from that taken at first.