REPORT ON THE PRESSURE ERRORS OF THE THERMOMETERS. 35 



admirably. Copper washers are employed at smaller junctions. The tubes connecting the apparatus 

 with the pump and with the external gauge are of copper, half-an-inch in diameter, and -j^-th inch 

 in internal bore. The pump has a bore of - 25 inch ; and the piston, which is a solid steel rod with 

 a sharp cup-shaped end (like the large plug) has a stroke of 2T6 inches. At the usual rate of working 

 of the gas-engine, there are 44 strokes per minute. All these fittings were executed in Edinburgh 

 after the arrival of the main tube from Woolwich. 



The plug was supported by block and tackle from a strong beam fixed in the walls of the apart- 

 ment, 5 feet above the upper end of the pressure apparatus. 



The key was originally planed true to the slot, but it was deemed prudent to reduce (very 

 slightly) its depth ; lest, under great pressures, it might flex at the plug and become permanently 

 jammed. 



During the winter session, when the temperature conditions were most favourable, I was in 

 general unable to find time for more than one experiment each day. But the great capacity of the 

 pressure-chamber enabled me to operate on five thermometers at once, three gauges and sometimes 

 other apparatus being also introduced. At least one thermometer was common to every two batches 

 of five thus operated on. 



The mode of conducting an experiment was as follows : — 



The thermometers, gauges, &c, to be operated on had been left all night, in light tinned-iron 

 cans, in the pressure-chamber, which was full of water. The cans were lifted out, full of water, 

 and the thermometers and gauges were then taken out one by one and read, after the indices had 

 been adjusted by the external magnet. The instruments having been restored to the cans, these were 

 at once lowered into the pressure-chamber, which occasionally required to have a small additional 

 quantity of water put in. This was taken from a vessel which had been kept standing beside the 

 pressure vessel all night. The plug, carefully coated with a mixture of tallow and oil, was then let 

 down into the cylinder, and pressed down by hand as far as possible. Then one of the working party 

 mounted (by means of the tackle) on the top of the plug, which was thus gradually forced in by his 

 weight, till it just passed the slot. At that instant another of the party opened the screw-tap below, 

 to allow the escape of water. A third mounted on the platform, and with a marlin-spike occasionally 

 gave a slight rotation to the plug in its descent ; so that when it was home there should be a clear 

 passage for the large steel key through the slots in the walls of the tube and in the plug. The moment 

 that the key was shot in, the screw-tap below was closed, the external gauge read, and the gas- 

 engine turned on. After a little practice the observer at the external gauge could give a signal to 

 throw off the engine, so that the pumping should be stopped exactly when the desired pressure was 

 reached. 



The thermometers, &c. were then left under full pressure for about three minutes only ; during 

 which interval the pressure, when originally three tons, lost at the utmost about T5 per cent. — 

 usually, however, not more than about - 8 per cent. A pressure of three tons, when there were no 

 air-gauges in the pressure-chamber, was generally reached in six or seven minutes. After the three 

 minutes' interval the screw -tap was very slowly opened, so that the relaxation of pressure usually 

 occupied from one and a half to two minutes. [When the tap was less slowly opened, the issuing 

 stream of water was at a temperature many degrees higher than that of the iron vessel — an excellent 

 instance of what was said in the text above about the heat developed by friction in narrow 

 channels.] 



I have satisfied myself, by trials both with this large apparatus and with the smaller one soon to 



