516 Johnston, and Adams — Influence of Pressure on the 



and V 8 are the volumes of 1 gram of the metal at the melting- 

 point in the liquid and solid states respectively, and q is the 

 latent heat of melting in calories per gram. 



A number of measurements not agreeing among themselves 

 are recorded in the literature on the latent heat and volume 

 change on melting of Cd, Pb, Bi, and Sn. We have made use 

 of the data of Person* for the latent heat of melting, and of 

 Vicentini and Omodeif for the change of volume at the melt- 

 ing point, to calculate by means of the formula, the change 

 of melting point per 1000 atmospheres for each of the four 

 metals. These data were chosen for the sake of uniformity 

 and because we believed them to be the best available. The 

 results of the calculation follow : 



I 



II 



III 



IV 



V 



Latent 

 heat : cal. 

 per gram. 

 Metal q 



Vol. change 

 on melting 

 ec per gram. 

 Vi — V, 



(H per 1000 

 atm. calc. from 

 vol. change and 

 latent heat 



6t\ per 1000 



atm. cale. from 



obs. change of 



m. p. with 



pressure 



Probable 

 error§ of 

 coefficient 

 in col. IV 



Sn 14-25 



0-003894 



+ 3-34 



+ 3'28 



± 0-02 



Cd 13-7 



0-00564 



+ 5-91 



+ 6-29 



± 0-04 



Pb 5-37 



0-003076 



+ 8-32 



+ 8-03 



± 03 



Bi 12-6 



0-00342 



— 3-56 



— 3-55 



± 0-08 



The agreement between the values 



given in columns III 



and IV is close — in fact, closer than might be expected when 

 we consider the uncertainty in the latent heat and volume 

 change of the metals involved. 



The change of melting point with pressure of tin and bismuth 

 has been measured by Tammann.|| He found for dt / dp 'per 

 1000 kg /cm 2 , 2-2 for tin and 3*9 for bismuth. His calculated 

 values (by the Clausius-Clapeyron equation and using, as it 

 happens, the same data for latent heat and volume change as 

 those we have employed) are 3-3 and 3-3 for tin and bismuth 

 respectively.!" It is worth while noting that Tammann made 

 his temperature measurements with a thermoelement of Pt — 

 Pt Rh and a direct-reading galvanometer. This fact is sufficient 

 to explain the discrepancy between his results for tin and bis- 

 muth and ours for the same metals. 



*Ann. Physik. Ixxv, 462; lxxvi, 432, 596, 597. 



fBeibl. Ann. Physik, xii, 176. 



% These coefficients, are, of course, the dt ( x 1000) of the third column 

 of Table III. 



§ For calculation of probable error of the coefficient cf. Merriman, Method 

 of Least Squares, 6th ed., Chapter on the Precision of Observations. 



I Zs. anorg. Chem., xl, 54, 1904. 



*[f 3 - 3° per 1000 kg /cm 2 is equivalent to 3'4° per 1000 atm. Apparently, 

 Tammann's calculated value for bismuth is in error by about 2 units in the 

 second significant figure. 



