184 
DR. H. T. BARNES ON THE CAPACITY FOR HEAT OF WATER 
convenient length of bulb, a smaller wire was adopted than is usually employed for 
thermometric work. A thermometer was finally obtained which gave the required 
sensitiveness, and had a length and diameter of bulb quite suitable for the 
calorimeter. In point of steadiness and accuracy, the two thermometers forming the 
differential pair made in this way could hardly be surpassed. 
I propose to describe the tests made on the various thermometers used during the 
course of this series of experiments. Before doing so, I must briefly describe the 
resistance boxes and method used for compensating the change in resistance in the 
wire due to a change of temperature. 
The general plan of Wheatstone’s bridge connections for the thermometer-circuit 
is already familiar. The wires leading to the bulb of the thermometer are 
compensated for a change in resistance due to a change in temperature by similar 
wires placed side by side with them, but connected to the opposite arms of the 
bridge circuit. The change in resistance in the thermometer is compensated by 
resistance coils on an opposite arm of the bridge, and a final adjustment made on a 
short bridge-wire, of which the coils are suitable multiples. A change in resistance 
is referred to a change in units of the box, rather than measured in ohms. It is 
evident that a change in the temperature of the resistance coils, while compensating 
a change in resistance in the thermometer, will produce an apparent change in the 
thermometer reading. This can be corrected for either by taking the temperature of 
the coils in air, or by immersing them in oil at a constant temperature. For very 
accurate work, however, it is better to introduce a different arrangement. If each 
resistance coil on the bridge is wound with another coil, which has the same 
temperature coefficient, but a different specific resistance, then if these second coils 
are connected with an opposite arm of the bridge system, any change in temperature 
of the bridge coils cannot affect the balance point on the bridge wire. This method, 
which was devised by Professor Callendar, works exceedingly well. 
Through the kindness of Professor Callendar I have had the use of such a 
compensated resistance box throughout the greater part of my measurements. This 
box was exhibited to the Poyal Society in June, 1893, by Professor Callendar. 
Besides the compensated resistance coils, the special features of this box are the 
bridge-wire scale, which has a compensating device for a change in length due to a 
change in temperature, so that the galvanometer contact point always reads at- the 
same point on the scale, and mercury cup contacts for each set of coils. The 
resistance coils were multiples of the bridge-wire, commencing from the smallest coil, 
which was equivalent to 10 centims. of bridge-wire, and doubling always as the coils 
became larger, i.e., 10, 20, 40, 80, 160, Ac., up to 2580. The resistance of the bridge- 
wire was '0088 ohm per centim., so that the ten coil was rather less than T ohm. 
The bridge-wire scale was of brass, very carefully divided to half-millims.., and 
supplied with a vernier with lens reading to '01 millim. The total length of bridge- 
