Conduction of Heat in Liquids. 



17 



lower temperature exceeded by fully 3 per cent, the mean 

 deduced from all the experiments of that series. 



I have applied Lorberg's method to the other specimen 

 tables given by Weber, which supply the observations taken 

 in a single experiment on each of the three fluids glycerine, 

 benzine, and mercury. Employing s and r as in Lorberg's 

 tables on water, my results are as follows: — 





Jc. 

 After Weber. 



After Lorberg. 



s. 



T. 





•0402 

 •0200 

 •9094 



•043 



•0235 



•925 



14-0 



23-6 



95 



c 



0-8 

 1-4 

 0-5 













I have also tried by an approximate method what variation 

 of the conductivity with the temperature would account for 

 the variations in these tables, and got a = '0045 for glycerine 

 and =*038 for benzine. By actual experiment at a higher 

 temperature, Weber got for glycerine a = '00423 ; he does 

 not give its value for benzine or mercury. Thus, in the case 

 of glycerine, Weber's results seem at least consistent; yet 

 Lorberg's values are quite distinct, and agree with the data 

 quite as well as in the case of water, though they require t to 

 have a value which could hardly escape detection. As to 

 benzine, the change in the conductivity could hardly be big 

 enough to explain the phenomena, though the decrease of 

 M/a 2 in the table seems very regular. As Lorberg's theory 

 would require t=1°*4, it can scarcely be considered more 

 satisfactory. 



There is still one point in Weber's experiments that calls 

 for remark. He observed that even when the liquid originally 

 occupied the space between the copper plates so fully as to 

 have its surface very decidedly convex outwards, before the 

 experiment ended the surface would become concave. This 

 he ascribed to the diminution in the volume of the liquid con- 

 sequent on the fall of temperature. It is to be noticed that 

 the area of the plates and the distance between them must, for 

 the same reason, be diminished ; so that what is observed is 

 a difference-effect. The rate of expansion of most liquids is, 

 however, much greater than that of copper or of the material 

 separating the plates ; thus the contraction of the liquid would 

 produce nearly its full effect. A consideration of the actual 

 rate of expansion of the liquids examined would scarcely lead 



Phil. Mag. S. 5. Vol. 24. No. 146. July 1887. C 



