312 
Proceedings of the Royal Society 
1*5 ohms at 10° C.). When the wires were reversed the deflections 
differed by about 1 per cent. The upper orifice of the gun was 
closed with a steel plug, of the form shown in the diagram, and 
secured by a slot passing through plug and gun. At the lower 
orifice was a tap to allow of a sudden release of pressure. The 
gauge for measuring the pressure has been already described, and a 
figure of it will be found in the “ Challenger ” Commission Report. 
In experimenting on solids our method was simply to fix tightly 
a piece of the substance 
two centimetres long by 
one square cm. in section 
on the inner junction. 
In the case of viscous 
solutions and liquids 
heavier than water, we 
arranged a small test tube 
within the coil, so that 
the inner junction dip- 
ped down into the tube, 
and was surrounded on 
all sides by the liquid. 
For liquids lighter than 
water we inverted the 
tube as shown in the 
figure, and having filled 
the gun with water, 
sucked out the air with 
a narrow U tube, and 
then admitted the liquid 
to be tested by a pipette. 
To Pump A pressure of three 
tons on the square inch 
could be applied almost 
instantaneously by two strokes of the pump handle; the deflec- 
tion of the galvanometer was simultaneously taken, and when the 
spot of light had steadied, the pressure was let off instantaneously, 
the deflection of the galvanometer being observed. The data for 
converting these deflections into degrees centigrade were obtained 
