PEIXCrPLES OF THE MECHANICAL THEORY OF HEAT. 27o 



in .1 ■\vidcr glass tube ; this outer tube contained enough quicksilver for the 



reservoir of the ether thermometer to be wholly surrounded Ijy it. 



This thermometer was n(»\v plunged, together with a cylindrical tube filled 



with air, into Oersted's compressing apparatus, which was tilled partly with 



water, partly with pieces of pure ice. By means of a ring of lead, care was 



taken to keep the water of the compressing vessel free from ice at that part 



of the thermometer on which the readings were to be made. A pressure of 



8.1, and again of 16.8 atmospheres, produced a sinking of the thermometer by 



7 5 

 7^ and 16j of the divisions of the scale, aiid thus by — '- =0.106° F., and 



71 



16.5 



= 0.232° F. ; which very nearly coincides with the theoretically calculated 



Fig. 13. 



F!g. 14. 



Fig. 15. 



71 



depression of 0.109° and 0.227° F. 



From the above developments and observations it might be expected that water 

 under very high pressure nuist remain fluid at relatively low temperatures. That 

 this, indeed, is the case, is confirmed by the experiments long since conducted by 

 Williams ill' Quebec, in order to measure 

 the force with which freezing water ex- 

 pands. He exposed to intense cold 

 thick iron bomb-shells, filled with water 

 and closed by means of an iron plug firmly 

 driven in. At a very low temperature 

 the stopper was either driven out and 

 then an icicle was projected from the 

 opening, (Fig. 13,) or the bomb was burst, and in that case a sheet of ice pro- 

 truded from the fissure, (Fig. 14.) The form of these extruded j)ieces of ice 

 indicated conclusively that the water at a veiy low temperature still remained 

 iluid, and was first converted into ice at the moment when it gainied additional 

 space. 



In fine, Mousson has shown (Pogg. Annal., cv) that at a very low temperature 

 ice may be rendered fluid by great pressure. The appa- 

 ratus of which he availed himself for this is represented in 

 section in Fig. 15, and on a smaller scale in side elevation 

 in Fig. 16. Through the axis of a massive prism A of 

 the best steel, four-cornered below and furnished above 

 with the worm of a screw, a cylindrical cavity, 7.12 milli- 

 metres Avide, is drilled, which, in its upper part, widens from 

 h to a, in a slightly conical form, so that the month at a 

 has a diameter of 8.61 millimetres. From above is driven 

 into the cylindrical cavity a {)iece of pure coppery, some- 

 wliat conical at first, and fitting into the cavity a b so as 

 to form abov'e a perfect closure of the same. To the cop[)('r 

 cylinder f/ is ailixed a steel prolongation D, of like diameter 

 with the cavity i c, and which, by application of the female 

 screw E, can be pressed downwards so as to drive the cop[)('r 

 cylinder /7 further into the cavity b c. Underneath the cylin- 

 drical cavity 6cis alsoaconical but rapidly widening cavity, 

 into which fits the copper cone/, which, by means of the 

 steel screw C, can l)e firmly pressed into its cavity. 



In order to perform the experiment, the screw (J and the 

 copper cone/ were first removed, the whole apparatus was 

 inverted, so that E was below, A above, and the free part 

 of the cavity Z^c, above .7, was filled with water that had 

 been boiled ; the copper index d was now lowered into this 

 water. With the position unchanged, while d thus stood 

 upon g, the whole apparatus was exposed to a low tempera- 

 18 s 



