ATMOSPHERIC PKESSUKE BETWEEN 104 C. AND 115 C. 415 



thermometer were taken at every even minute, the contact on the bridge-wire being 

 set at the nearest millimetre, and the deflection of the galvanometer noted on 

 reversal of the current. 



During every experiment several readings were taken with the contact made 

 at the next millimetre reading on the bridge which gave a galvanometer deflection 

 in the opposite direction. The height and temperature of the barometer were read 

 at intervals of about seven minutes, unless they were varying very rapidly, when 

 readings were taken more frequently. The approximate value of the current was 

 noted on the ammeter ; a knowledge of this was useful when adjusting the current 

 in the determination of the heating effect of the leads. The potential balances across 

 the standard cells, across the secondary standard resistance, and across the heating 

 coil were taken at intervals during the experiment. In addition, the temperature of 

 the standard cells and that of the oil bath in which the secondary standard was 

 placed, were noted. Since seven Leclanche cells were required in the maximum flow, 

 and only five in the medium and minimum flows, a small steady change took place in 

 the potentiometer reading for the standard cells during the short interval following 

 the change in the number of Leclauches used. The variation was only 3 or 4 parts 

 in 7000 in about half-an-hour, and after that period of time had elapsed the reading 

 remained quite constant. The potential balances across the heating coil and across 

 the secondary resistance were wonderfully steady, especially during experiments 

 with the two smaller flows. In the case of the maximum flow the irregular 

 fluctuations only amounted to about 30 divisions on each side of the balance-point on 

 the galvanometer scale, and a change in the potentiometer reading of 4 (in about 

 8000) was sufficient to throw the balance out by the same amount. At the end of 

 the second experiment with any flow, the electric current was switched off and the 

 weights of condensed steam determined. As soon as the temperature was again 

 nearly steady, another flask was put under the end of the condenser and a " check " 

 flow obtained. Meanwhile temperature readings were taken until the recorded 

 values were constant. This temperature reading is tabulated as " cold temperature 

 before." A current equal to that used in the main experiments was then passed 

 through each pair of current and potential leads, and the heating effect of the current 

 in each pair of leads measured. Then the cold temperature was again determined : 

 this value is tabulated as " cold temperature after." I have included specimen 

 tables showing all the observations made on one day, for thereby the remarkable 

 steadiness of the conditions is illustrated very clearly. 



Reduction of the Observations. 



In each experiment there are as many observations of the " hot temperature " as 

 there are minutes of time in the duration of the flow, and, in addition, there are the 

 two " cold temperatures," the barometric readings, the various temperature observa- 

 tions and the potential balances. The mean value of each set of readings is then 



VOL. CCXV. A. 3 K 



