EXPERIMENTAL KNOWLEDGE OF THE PROPERTIES OF MATTER, 519 



In a paper by Tait, of Dec. 1887, the ultimate loss of volume of water 

 under infinite pressure is estimated at about 25 p.c. ; in a paper of 

 ]\Iarch 5, 1888, the same author finds the compressibility of lead-glass 

 per atmosphere as '0,527 ; and the increase of compressibility per degree 

 Centigrade as "072. 



Melting-points — GarneUey and Carleton- Williams. 



In the application of heat to solid bodies which are not decomposed 

 by heat at the temperatures employed, the effect after reaching a certain 

 temperature is that the body will change its state, either by liquefying or 

 volatilising ; or both effects may take place together, especially in a case 

 such as that of iodine, where there is not a great interval between the 

 point of liquefaction, or melting-point, and the boiling-point. 



Both of these points, however, are dependent on external pressure ; 

 the effect of pressure on the boiling-point is very considerable compared 

 with the effect on the melting-point, which is in general hardly 

 appreciably altered by a great alteration of pressure. But whether a 

 substance shall volatilise without melting depends on the external 

 pressure. 



The experiments of Cai-nelley are concerned with the melting-points 

 of bodies which melt at atmospheric pressure ; a multitude of cases of 

 inorganic bodies, chiefly compounds, have been determined with great 

 care and accuracy, and the experiments and results published in a 

 number of papers in the ' Trans, of the Chem. Soc' ' and in papers con- 

 tributed to the ' Phil. Mag.' and the ' Proc. Roy. Soc' 



In many cases a knowledge of Carnelley's melting-points makes it 

 possible to determine within moderate limits the boiling-point of a body 

 when this is high, without recourse to an air-thermometer ; this is done 

 by finding two bodies, near to one another in their melting-points, one of 

 which is and the other is not fused in the vapour of the boiling substance. 

 This is, in fact, the method which was used by Carnelley and Carleton- 

 Williams ^ for the appi'oximate determination of the boiling-points of some 

 salts and other bodies. By establishing relations connecting the melting- 

 points with the atomic weights in the case of elements ^ and of compounds, 

 Carnelley has predicted melting-points of bodies which had not been 

 determined, and found a very fair agreement between the calculated and 

 the afterwards-observed temperatures. 



A noteworthy example of this is in the cases of beryllium chloride and 

 bromide, the melting-temperatures of which they observed and found 

 to agree * well enough with what they should be if calculated for the 

 formulse BeCl2, BeBrj, but to be enormously too higli for the formulae 

 BeCls, BeBrg, or for these formulae doubled. It is evident from this 

 example that Carnelley's method is an additional and valuable aid towards 

 deciding between proposed formulae of metallic chlorides and bromides, 

 and so between two proposed numbers for the atomic weight of an 

 element and the molecular weight of a compound. 



p. 165 ; 31, p. 1000; 33, p. 644. Braun, Wied. Ann. 30, p. 250; 82, p. 504 ; 33, p. 239 ; 

 C. S. J. 1887, Aba. p. 436. Tait, Proc. Roy. Soc. 1883, Dec. 1887 and March 1888 ; 

 Nature, Jan. and March 1888. 



' C. S. J. 1876, p. 489 ; 1878, p. 273 ; 1879, p. 563 ; 1880, p. 125 ; 1884, p. 409, 



= Ibid. 1878, p. 281 ; and 1879, p. 563. 



3 Proc. Roy. ,*c. No. 197, 1879, p. 190. 



* Phil. Mag. Nov. 1879, p. 371 ; Proc. Roy. Soc. 1879, vol. xxix. 190. 



