36 THE VOYAGE OF H.M.S. CHALLENGER. 



be described, that, for the object I had in view, nothing was to be gained by prolonging the exposure 

 of the instruments to pressure. A very slight additional compression might probably have resulted ; 

 but it would in all cases have produced much less effect on the thermometers than that due to 

 changes of temperature in the room towards afternoon, especially with several persons working for a 

 considerable time round the apparatus. This was the case when we were working at an initial 

 temperature of 40° F., when water is not heated by pressure. When we worked at temperatures of 

 50° F. or upwards, it would have been vain to expect anything from protracted pressure ; for the 

 sudden rise of temperature in the water is soon greatly diminished by the good conducting power of 

 the steel gun, and the large capacity of three tons of steel and iron, as compared with that of 25 lbs. 

 of water. Thus the mercury in the thermometers falls away from the index, — so that, even if a 

 farther compression took place under continued pressure, the index would not be affected by it. 

 This cooling explains, to a great extent, the apparent leakage described in last paragraph. 



After the pressure had been let off, our most formidable difficulty presented itself, viz.. the 

 extraction of the plug with as little as possible of a jerk, and (especially when there were air-gauges 

 in the instrument) with as little exhaustion of pressure as possible. To do this with perfect steadi- 

 ness, and with the requisite slowness imposed by the great length and very small bore of the lower 

 aperture, a powerful screw-jack would have been required. But, though I have an instrument of the 

 kind, I determined to do as well as I could without it, as the necessary fittings would have been not 

 only expensive, but exceedingly cumbersome, and would have greatly extended the time required for 

 each experiment. The method adopted was to haul the tackle tight, but not so tight as to start the 

 plug ; and then, by pinching two laps of the chain together, to produce the desired result. There was 

 always about 1th or ^th of the air sucked out in this way from our air-gauges, except when we took 

 the precaution of putting into the chamber before commencing operations a large inverted vessel full 

 of air. This works well enough in some respects, but it is objectionable for several reasons, especially 

 the heat developed in compressing air. Another mode was to force out the plug by reapplying pres- 

 sure after the key had been extracted. This was, of course, a very tedious operation as, even when 

 no air-gauges were in the apparatus, at least 900 strokes of the pump were required : — for the section 

 of the plug is 16 square inches, and it had to be raised 5'6 inches, the pump inserting .j th of a cubic 

 inch of water at each stroke. Any other mode of meeting this difficulty would have involved a 

 weakening of the apparatus, which could not be permitted. 



The smaller pressure apparatus, already alluded to, is figured in section in the woodcut. Its bore 

 is one inch in diameter ; and its content, when the plug is in, is about nine cubic inches. A single stroke 

 or two of the pump only is required to produce in it a pressure of three to four tons. The important 

 feature in its construction is the large flange by which the lower end, with its fittings, is attached. 

 Between the flanges two large leather washers (carefully soaked in wax) are compressed by means of 

 six powerful screws. Their object is to enable us to insert a thermo-electric junction in the pressure- 

 chamber, the other junction of the circuit being outside. In the sketch the covered wires (copper and 

 iron of 23 gauge, two of each metal) are seen twisted together and extending up the chamber in a 

 cork-screw form with the junction at the top. This arrangement enables the experimenter to raise the 

 junction above the top of the cylinder when he wishes to fit it into a mass of any substance which is 

 to be tested for the heat developed by compression. The wires pass out, each by itself, and are laid 

 in a serpentine form between the leather washers. A day or two after it was first set up, this apparatus 

 leaked considerably at the flange, but by tightening the screws a second time it was made, and still 

 remains, almost perfectly water-tight even up to five and six tons pressure. 



