AND THE CUBIC COMPRESSIBILITY OF MERCURY. 51 



latter values are dependent upon the somewhat arbitrary and uncertain 

 theory of Poisson, because Amagat rejected entirely his work on the linear 

 compressibility of the metals, and was not even willing to record its 

 results. The present direct determinations have the advantage of being 

 free from mathematical assumptions ; and if they are accepted, the mathe- 

 matical method used by Amagat must obviously be modified in such a 

 way as to reduce the result by about 42 per cent in the cases of iron and 

 copper. This conclusion casts doubt upon the true magnitude of the coef- 

 ficient of Poisson, except in the case of amorphous substances like glass, 

 which gives fairly consistent results by the two methods. 



It is now possible to compute the compressibility of mercury between 

 100 and 500 atmospheres, by adding the differences between the com- 

 pressibilities of the solid metals and mercury to the values just given. 

 The results are recorded in the following table : 



Compressibility of mercury between 100 and 500 

 atmospheres. 



I. 0.000000385 + 0.000003325 = 0.00000371 

 II. 0.00000051 + 0.00000319 = 0.00000370 



Average 0.000003705 



The agreement of these results must be partly accidental, as it is greater 

 than would be expected from the large probable error of the result with 

 copper. The figure 3.71 obtained from iron is much the surer, and will be 

 taken as representing the compressibility of mercury. It is interesting and 

 satisfactory to note also that the result agrees well with the figures of Ama- 

 gat, when these are corrected for change of compressibility at high press- 

 ures. Amagat found 0.00000392 as the compressibility of mercury under 

 small pressures, these being measured in ordinary atmospheres. Corrected 

 for the decrease of compressibility with pressure, and transposed to the 

 metric standard of pressure, this value becomes 0.00000370,* essentially 

 identical with the new value given above. The work of very early experi- 

 menters is sufficiently commented upon by Aime, and of those preceding 

 Amagat by the latter physicist. The earlier values are not to be com- 

 pared with Amagat's in accuracy. The more recent value of de Metz,f 

 0.00000390, at 20 under small pressures, depends like that of Amagat 

 upon the assumptions of the theory of elasticity, and is therefore no more 

 certain than his. 



On this account, the result of the present investigation (0.00000371 

 expressed in kilograms per square centimeter, 0.00000379 expressed in 



* Amagat, Ann. de Chim. et de Phys., (3), 8, 257 (1843) ; Richards and Stull, 

 Publication Carnegie Institution of Washington No. 7, p. 20 (1903). 

 fWiedemann's Annalen, 47, 706 (1892). 



