ON THB SPECIFIC HEAT OF WATER. 



213 



Tliis method eliminated errors in the ratio coils, and temperature errors in the 

 I .ridge, leaving only the possibility of casual errors in the range of comparison coils of 

 the bridge. The result of the series of values obtained in the calibration of the 

 mercury thermometer-resistance, where the maximum error found was less than 

 1 part in 20,000 (which error included errors in reading the thermometer graduations 

 of the resistances), was taken as showing that the bridge itself was in sufficiently 

 good working order, and needed only the application of the proper corrective factors 

 to reduce bridge ohms to international ohms. 



The Mercury Thermometer-resistance M,. When in use, this resistance had a 

 stream of cold water flowing over the spiral, so that its temperature range was small. 

 Within this range it was found that the temperature-resistance curve could be taken 

 as a straight line. One calibration is given below : 



TABLE I. Calibration of Thermometer-resistance M,. Experiment 35s. 



The calculated values are from the expression 



R= 1'8 1567 + 0-001 7m,. 



The bridge factor was about unity, so that the result is in international ohms. 

 The differences between observed and calculated values show that the resistance can 

 be calculated from the graduations of the mercury thermometer tube with an 

 accuracy of within about 1 in 40,000. But for practical purposes the constant was 

 calculated only to the fourth decimal place, giving a possible error of the order of 



1 in 18,000. 



Ttie Mercury Thermometer-resistance M a . This served as the electrical heater in 

 the calorimeter, and was calibrated under the actual conditions of subsequent use, 

 that is to say, the whole apparatus was installed as for a specific heat observation, 

 the calorimeter contained the requisite amount of water, the heater was itself used 

 to raise the temperature step by step from C. to 80 C., and the obturator 

 (through which pass the large mercury leads of the heater) was kept at 20 C. 

 whilst the temperature of the calorimeter rose from C. to 10 C., and was thence- 

 forward kept at 10 C. above the temperature of the calorimeter. In Table II. the 



