220 MESSRS. W. . BOUSFIELP AND W. ERIC BOUSFlELD ; 



differences from which are set out in the last column. Most of these differences must 

 down to experimental inaccuracies, resulting from the difficul ty _de to the 

 rise of tempefature in the short interval which is necessary to get the readings 

 a balance is obtained. There appears, however, to be a shght d.fference m h 

 ratios as between no current and the full current of 5 amperes of the order of about 

 - parts in 10,000. Even this error is possibly not genuine, but due to a 

 thermoid effect in the resistances of 100,000 and 19,000 winch were used m 

 the other arms of the bridge. But in the case of the manganm resistance, the 

 curve I) (fir. 5, p. 217), shows an error of 10 parts in 10,000 for a smnlar range of 

 runvnt There is, therefore, no doubt that the mercury resistance is much bet 

 adapted as a standard to carry heavy currents than the manganm resistance. 



7. Heating due to Obturator, Stirring, cfeo.-The obturator, situated about 3 cm. 

 above the level of the liquid in the calorimeter, and kept at a temperature of 10 

 above that of the liquid, constitutes a source of heat which it was necessary to 

 determine accurately. This determination also includes various sources of heat gain 

 or loss which must be enumerated : 



(a) Radiation from obturator, which is the main effect. 



(b) Conduction from the obturator through the mercury leads of the heater. The 

 portion of the leads passing through the obturator, being 10 C. higher than the 

 contents of the calorimeter, thus constituted a source of heat by conduction. 



(c) Stirring. The stirrer was driven from an alternating-current motor running at 

 a constant speed. The rate of revolution was frequently counted and found to be 

 practically constant at about 340 revolutions per minute. The heat due to stirring 

 is included in the determinations. 



(d) Evaporation. Any escape of vapour would of course constitute a heat loss 

 which might be considerable. It was the function of the obturator, kept at 10 C. 

 above the temperature of the water, to minimise this. In a preliminary series of 

 experiments, in which only 2 litres of liquid were placed in the calorimeter, thus 

 leaving an air space of over a litre above the surface of the water, it was found that 

 the heat loss due to this cause became sensible at 60 C. and considerable at 80 C. 

 Hence the volume of water used in the experiments described in this paper was 

 increased to nearly 3 litres, and the obturator was lowered so as to reduce the air 

 space to the smallest dimensions consistent with allowing for the expansion of the 

 contents on heating. With this alteration the observations show that this element 

 of heat loss introduces no uncertainty until a temperature of about 75 C. is reached. 



Detomination of Heat Gain and Loss. The joint effect of all these elements was 

 determined in a series of about 30 experiments in which all the conditions were the 

 same as during a run, except that there was no electrical heating of the contents of 



* A short note of further experiments on the thermoid effect is added at the end of this paper (p. 242), Imt 

 the matter is being made the subject of further investigation, the results of which we hope to communicate. 



