20 Mr. C. Chree on the 



measures of the conductivity. It is, however, based on a 

 rather imperfect mathematical treatment ; and no clear idea 

 is given of how the experimental work may be carried out. 

 Some more delicate method of determining the temperatures 

 would be desirable. Since no results seem to have been 

 actually obtained by this formula, further explanation is 

 unnecessary. 



The formula actually employed is the second one of the two 

 given above. Thus Christiansen's results theoretically apply 

 to a perfectly determinate temperature ; and if the relation 

 between the conductivity and the temperature be linear, they 

 are not open to the charge of implicitly neglecting the varia- 

 tion of the conductivity. 



The method has some obvious advantages. It seems very 

 simple in execution, and allows the conductivities of any two 

 liquids to be compared directly through a single experiment. 

 Further, as the temperature is kept stationary, the readings 

 can be taken with great accuracy, and no errors consequent on 

 slight mistakes in measuring time-intervals can come in. For 

 the same reason, much less objection could be taken to the 

 assumption that no discontinuity of temperature exists in 

 passing across the common surface of a liquid and metal. 

 There is at least one rather serious objection, however. In 

 the actual experiments there existed a difference of from 11° 

 to 19° between the temperatures of the two fluid layers — a 

 difference varying with the nature of the fluid. Thus the 

 conductivities do not refer to any standard temperature; and 

 if the variation of the conductivity with the temperature be 

 as large as found by Weber, the comparison instituted is 

 somewhat misleading. 



With one exception, the upper fluid was air ; while different 

 liquids were successively introduced between the two lower 

 plates, and in neither layer was the mean temperature the 

 same in any two experiments. In one experiment only was 

 the air replaced by a liquid, leading to a direct comparison 

 between the conductivities in glycerine and olive-oil. The 

 value obtained for the ratio was 1*845 : 1; while that deduced 

 from the separate experiments with air was 1*87 : 1. Con- 

 sidering the complete difference in the temperatures to which 

 the conductivities refer, this coincidence is not so convincing 

 as Christiansen seems to think. No doubt the difficulty as to 

 the variation in temperature could be surmounted by experi- 

 ments in which both layers are filled with the same fluid. 

 This would enable the dependence of the conductivity on the 

 temperature to be directly determined for each fluid separately. 

 This being found, a single comparison between any two fluids 



