12 Remarks on Dr. Mills's Researches on Thermometry. 



servations of the position of the zero recorded in the other 

 columns were made. The numbers in the other columns are, 

 in all cases, the depressions of the zero measured in milli- 

 metres below the original position observed before the bulb 

 had been heated for the first time, i. e. in the first series of 

 experiments. 



These observations, then, show that the depression of the 

 zero of an open thermometer diminishes (1) with the lapse of 

 time after it has been disturbed by heating, (2) with nu- 

 merous and rapidly recurring heatings and coolings. The 

 latter of these two phenomena is identical with what we 

 understand Dr. Mills to mean by an ascent of the zero; and 

 we have therefore proved (as has indeed been noted by other 

 observers) that this can be produced by frequent heating ; but 

 the fact that it has thus been observed in an open tube is 

 not in accordance with Dr. Mills's explanation of these 

 movements. 



His experiments differ from ours in the fact that the bulb 

 was raised, not to the same, but to continually increasing 

 temperatures; but before any certain conclusions can be 

 drawn, it would be necessary to know something about the 

 past history of the thermometers employed. The effect of 

 heating is "almost invariably" to depress the zero; the effect 

 of frequent heating is to diminish this depression; and thus 

 the effect of any particular heating is dependent in part on 

 those which preceded it. It has generally been supposed that 

 by repeated heating and cooling the fluctuation of the zero could 

 be reduced to a very small amount; and Dr. Mills's experi- 

 ments agree with this, in that thermometer 455 (which, as far 

 as Ms paper shows, was heated most frequently) is that which, 

 according to his formula, was most nearly approaching a 

 stationary point. However this may be. it is evident that it 

 cannot be the atmospheric pressure which causes the zero of 

 an open tube to rise when it is frequently heated, and there- 

 fore that it need not be atmospheric pressure which produces 

 the same effect in a closed thermometer. 



Summing up, then, — our criticisms on Dr. Mills's paper are 

 as follows: — 



(1) The method of comparing one factor only of the ex- 

 posure correction for different thermometers, viz. oc, instead 

 of the whole correction y, tends to exaggerate the importance 

 of the variation of the corrections. In the case of three out 

 of the four thermometers observed, it is, within the limits of 

 the error of experiment, the same for the largest possible ex- 

 posure and for a temperature-difference which is probably 

 greater than any of those for which observations were made, 



