COMPARISON OF PLATINUM AND GAS THERMO>fETER8. 97 



from the law previously found by one of us that the departure of the coefficient from 

 that of a perfect gas varies proportionally to the initial pressure, we have 



a = 0-003 666 3. 



This coefficient would give for temperatures near the sulphur point values about 0'2 

 higher than those deduced by employing the one directly observed. 



We have nevertheless adhered to the latter for the calculation of the temperatures 

 of this series of comparisons, in order to avoid the introduction of any hypothesis. 



XLI. EXPLANATION OF THE TABLES OF RESULTS. 



The results of the whole of the comparisons made are given in the tables for each 

 thermometer at the end. In these the experiments are arranged in order of ascending 

 temperature. The first three columns give for each experiment the progressive 

 number, the number in our note-books and the date. Columns IV. and V. give pt 

 and d, the value for d being that deduced from the Callendar formula given on p. 39, 

 assuming the value for 8 as determined for each thermometer at the sulphur point, 

 and taking our new value for the boiling-point of sulphur at 760 millims. pressure, 

 namely 445'27, given later on p. 101. Column VI. gives the equivalent on the 

 nitrogen scale of the observed pt, as thus calculated, and Column VII. the tempera- 

 ture on the nitrogen scale as given by the gas thermometer. Column VIII. shows 

 the difference between the calculated and observed values, and Column IX. the 

 constancy of the temperature in each experiment as given by the indication of the 

 platinum thermometer. 



XLH. DETERMINATION OF THE BOILING-POINT OF SULPHUR. 



After ascertaining that it was possible, by means of the bath of fused nitrates, to 

 make accurate comparisons between the platinum and gas thermometers at tempera- 

 tures up to about 600, we saw that by making alternately a determination of the 

 resistance of a platinum thermometer at the boiling-point of sulphur, and a 

 comparison with the gas thermometer near the same temperature, we had a means 

 of obtaining a new determination of the boiling-point on the nitrogen scale. We 

 accordingly made, in an apparatus of the form described by CALLENDAR and 

 GRIFFITHS as the " Meyer tube," a number of determinations of the platinum 

 tnuperature of sulphur- vapour boiling freely under atmospheric pressure. Readings 

 of the barometer were taken simultaneously with those of the platinum thermometer. 

 The reservoir of the platinum thermometer was protected from contact with any 

 < iidensed sulphur which might flow down to it from the cooler part of the thermo- 

 meter above, by surrounding it with an asbestos cone perforated with several holes 



VOL. OXCIV. A. O 



