472 
PROFESSORS A. W. REINOLD AND A. W. RUCKER ON 
with the front detached for the purpose of showing the paddles, B, B, B, B, for stirring 
the water and producing uniformity of temperature throughout. These are moved 
up and down by means of strings passing over a pulley. The rectangular chamber 
A, A, A, A, is 33 centims, long, 21 centims. high, and 20 centims. deep. It is sur¬ 
rounded on four sides by water, and is closed on the remaining sides, viz.: front and 
back, by glass windows. In these experiments the window at the back was not used 
as such, being covered by thick felt. The window in front is provided with felt 
packing, and is held firmly in its place by four nuts. The bath is supported on an 
iron stand about a foot from the table. 
Four tubes C, C, D, D, each about 2 - 5 centims. in diameter, pass through the base 
of the apparatus and are closed by caoutchouc stoppers. Through C, C pass the 
battery wires, and through D, D two other wires connecting with the electrometer. 
Similarly three tubes E, E, E pass through the upper part of the apparatus, the 
standard thermometer F being supported in the centre one. H, H are mercury cups 
for connecting the electrodes of the tube with the battery wires ; and K, K similiar 
cups for making the electrometer connexions. By placing Bunsen burners beneath 
the lateral bent tubes G, G', which are also made of tin plate, the water could be 
heated. As a rule the range of temperature employed was between 10° and 35°; 
within these limits it was easy to obtain any required temperature and to maintain 
it constant during an interval sufficiently long for two or three complete sets of 
observations to be taken. No observations were made until the thermometers ff 
gave the same readings and the thermometer F indicated a temperature not differing 
by more than two or three tenths of a degree from that of the others. The mean 
between the readings of f,f and of F was then assumed to be the temperature of the 
liquid in the tube. 
The mode of connecting with the electrometer, first the wires in the liquid and 
then the points in the circuit on either side of the known resistance, was identically 
the same as that already described as employed in the measurement of the resistance 
of a film. The required temperature having been reached, the resistance in the box 
of coils was adjusted to give an electrometer deflection not very different from that 
of the liquid. The potential difference of the points a, a' was then examined, the 
battery circuit being broken meanwhile. As in the case of the films tins difference 
was found to be very irregular both in magnitude and direction. As a rule the 
deflection produced by it did not exceed 2 divisions of the scale, but occasionally it 
rose to as much as 6 or 7 divisions. The circuit w^as then completed and after an 
interval of 3 or 4 minutes, when the rate of weakening of the current (due to 
polarisation, &c.) appeared to be uniform, the electrometer readings right and left 
were taken 
(1) When the liquid was connected with the electrometer. 
(2) ,, resistance 
(3) ,, liquid again 
5 ? 
