1909] on Experiments at High Temperatures and Pressures. 545 



of the physical state— and has thus been able to plot out many inter- 

 esting diagrams of condition. The apparatus consists of a screw 

 press by which a piston of ebonite is driven down a steel cylinder of 

 small known cross-section. The cylinder is filled with oil, and the 

 ebonite piston fits practically oil- tight. The oil communicates with 

 the oil contained in a strong steel vessel, which also incloses a glass 

 tube open at the lower end, containing the substance and dipping 

 below the level of mercury contained in a dish. The oil occupies 

 the rest of the space. The steel vessel is placed in a thermostat so 

 that its temperature can lie ascertained. The oil pressure is measured 

 by a Bourdon gauge which it was possible to standardise, thanks to 

 the previous work of Amagat and Tait. In order to construct a 

 diagram of condition it is necessary and sufficient to find a number 

 of points separating the liquid from the solid area— or separating the 

 areas corresponding to different crystalline forms in the case where 

 the transformation of one sort of crystal into another is under 

 investigation. To understand how this is done, it is best to take a 

 special case. If we have a quantity of a substance under a known 

 pressure and temperature in the piezometer and suddenly increase 

 the pressure, so that there is not time for heat to pass in or out to 

 any appreciable extent before the pressure gauge can be read, we 

 have practically adiabatic compression. If the apparatus be then 

 left to itself, the heat which we may suppose to be liberated by the 

 pressure will slowly diffuse outwards, and the pressure will fall as 

 time goes on. If we happen to start from a point on the m.p. curve 

 before the pressure is raised — then the final result will be that we 

 shall thaw or freeze more or less of the material, and the original 

 pressure will be exactly regained, the change of state compensating 

 the impressed change of volume. If, however, the increase of pressure 

 has been so great that a change of state of the whole mass has been 

 brought about, then the after variation of pressure will be so much 

 greater that it is easy to distinguish this case from the previous one. 



The accompanying diagram (p. 546), taken from Prof. Tammann's 

 book, shows how the equililirium curve can be located in the case of 

 carbon dioxide and naphthalene. In the former case the temperature 

 was 0*31°C. The pressure was 3800 kilograms per sq, cm., or 

 24" 13 tons per sq. inch. (157 '49 kilograms per sq. cm. = 1 ton 

 per sq. inch = 152 '38 atmospheres.) 



The pressure w^as raised adiabatically to 4400 kg./cm.'-^ (27*93 

 tons/sq. inch) and the subsequent fall of pressure plotted against a 

 time scale for ten minutes. The pressure was then adiabatically reduced 

 to 3550 kg. /cm.- and the recovery curve again plotted. The equi- 

 librium pressure must lie between the pressures approached asymp- 

 totically on the diagram, i.e. between 3825 and 3792 kg./cm.^. A 

 repetition between narrower pressm-e limits enables the pressure to be 

 fixed at between 3808 and 3797 kg. /cm.'-. A similar procedure fixed 

 the pressure of the m.p. of naphthalene between 3090 and 3080 



