216 
DR. H. T. BARNES ON THE CAPACITY FOR HEAT OF WATER 
requisite number of bridge resistance-coils required to compensate the rise in tem¬ 
perature in the water, we will term the “ hot ” reading. The conduction errors at the 
ends of the apparatus remained the same for both readings, except (as has already 
been pointed out) the conduction error due to the rise in temperature, d9, in the out¬ 
flow end. It was therefore possible for any temperature over the scale between 
0° and 100° to eliminate the conduction errors at the ends of the apparatus by the 
“ cold " readings in a, very simple manner, and reduce the only conduction error to be 
considered, i.e., that due to the rise (16, to the same amount all over the scale. Unless 
this procedure had been followed, large errors would have been incurred, especially at 
the inflow end where the conduction effect was the largest. 
The adjustment of the electric current to any given flow was made by varying the 
number of accumulators, or by inserting a number of platinoid strips, ’02 ohm, in 
series in the rheostat. While the conditions became steady in the calorimeter after 
turning on the electric current, requiring 10 to 15 minutes, the weights of the empty 
flasks were obtained as already described. One of these was then affixed to one of 
the nozzles on the calorimeter. The temperature of the cooler was then adjusted, and 
the preliminary readings of the potential difference across the standard resistance and 
calorimeter obtained. The chronograph being started, the accurate balance of the 
two Clark cells in series was obtained, together with the temperature of the Clark-cell 
bath. The other conditions remaining steady (including the thermo-regulator in the 
tank, the jacket circulation and the different water circulations to the resistance oil- 
bath, cooler, constant-level device and steam-jacket when used, as well as the thermo¬ 
regulation in the Clark-cell bath) the flow was switched over into the weighed flask 
at a given moment, which was recorded automatically on the chronograph. The 
complete set of observations then followed in order for 15 minutes. These, besides 
the temperature of the cooler at the beginning and end of the interval, were made 
every minute—first minute the deflection of the galvanometer at the nearest millimetre 
on the bridge-wire for the balance-point of the differential thermometers, then, in 
succession at every even minute, the potential balance of the standard resistance, the 
reading of the differential thermometers, potential across the calorimeter, reading of 
the thermometers, potential across standard resistance, and so on, including at the 
half-minutes the reading of temperature of standard-resistance oil-bath, jacket water, 
air temperature and temperature of thermometer resistance box, although this last 
was not really necessary. 
At the end of the 15 minutes, to the nearest second by the watch used in starting 
the interval, the flow was switched over to the other nozzle, and the time automatically 
recorded. The balance-point of the Clark cells was then obtained, together with their 
temperature. On changing the full flask for a second empty one, a second set of 
readings for 15 minutes was obtained, without otherwise altering the conditions. The 
extreme steadiness of the potential balance for the Clark cells made it quite unnecessary 
to have it recorded oftener than just before and just after the 15-minute interval. 
