1906-7.] The Electric Conductivity of Sea-Water. 235 
The Thermostat . — The tank used for keeping the sea- water at 25° during 
the measurement was a thermostat of several gallons’ capacity of the 
Ostwald-Luther pattern. As a matter of fact, the temperature of the bath 
was found to vary by about one-fifth of a degree, rising slightly and slowly 
during the day, because the temperature of the room was comparatively 
high ; consequently, the temperature of the bath was always read for each 
observation and the necessary correction applied. It was found that over 
small ranges of temperature (2°-4°), one division of the bridge wire cor- 
responded to 0*2° rise or fall in temperature. The correction was the same 
for seven samples which included one each of high and low salinity. The 
correction applied was thus taken as 0‘5 mm. of bridge wire for each tenth 
of a degree above or below 25 c C. The cell was held in position by a clamp. 
The sea- water was not placed directly into the cell, but was first pipetted 
into a Jena glass tube fitted with a rubber cork, which was suspended in 
the thermostat. This assumed the temperature required and was then 
placed in the cell, which was first thoroughly rinsed with it. The cell was 
then allowed to stand in the thermostat for a few minutes, and meanwhile 
the Jena tube was filled with the next sample of water. No trouble was 
experienced from the formation of air-bubbles in the cell after this treatment. 
The cell assumed the temperature of the thermostat extremely quickly 
owing to the thinness of the glass and the smallness of its contents. 
Method of Determining the Conductivities . — The cell was rinsed three 
or four times with the sea- water to be examined and the removable electrode 
fixed in position. The cell was then clamped in the thermostat, and in a 
few minutes the battery connection made and the induction coil started. 
The slider was then adjusted without any alteration of the standard 
resistance, and a reading of the bridge taken. The thermometer in the 
thermostat was read, and a second reading of the bridge made after re- 
adjusting the slider. A third balance was similarly made and the mean of 
these taken as the true value. (The bridge readings were taken with the 
coils and cell in both the left and right gaps of the bridge respectively.) 
The temperature of the coils was then read. The observations were 
recorded as follows: — 
Number. 
Bridge Readings. 
Temperature 
of Cell. 
Temperature 
of Coils. 
Mean Readings 
of Bridge 
corrected 
to 25°. 
Cell in 
Left Gap. 
Cell in 
Right Gap. 
1 
521-5 
521- 5 
522- 0 
495-0 
495-5 
495-0 
25-35° 
55 
55 
20-5° 
55 
55 
\ 521-5 495-4 
j diff. = 26-1 
