[ 46 ] 



II. On the Thermal Conductivity of Water. By S. R. MlLNER, 

 B.Sc, 1851 Research Scholar, and A. P. Chattock, Pro- 

 fessor of Physics, University College, Bristol*. 



A KNOWLEDGE of the thermal conductivities of liquids 

 is of value in relation to their chemical constitution ; 

 and as the comparison of conductivities may be more easily 

 made than the determination of their absolute values, it is of 

 importance that the absolute conductivity of some standard 

 substance, such as water, should be known as accurately as 

 possible. The present paper is the account of an attempt 

 made to measure this quantity by the direct determination of 

 the quantities entering into its definition ; the heat flowing- 

 through the water being generated electrically f, and so 

 accurately measurable. The principle of the method will be 

 best understood by reference to the diagram, fig. 1. Here the 



Fig-. 1. 



■Halancing curre/x 

 —^Beatiny current 



water (shaded) forms a thin layer between the plane parallel 

 bottoms of the vessels A and B. Over the bottom of A 

 (which must be constructed of thin glass or some bad con- 

 ductor to prevent horizontal flow of heat) is a flat coil or 

 zigzag D of well-insulated wire which forms the " heating- 

 coil/'' and the wires of which are shown in section by the dotted 

 line. The heat generated in D when traversed by a current 

 flows downwards through the water layer into the sink C (a 



* Communicated by the Authors, having been read at the Bristol 

 Meeting of the British Association, 1898. 



t At the time of designing our apparatus we were unacquainted with 

 the work of Dr. C. H. Lees, who also makes use of an electrical method 

 of generating heat for conductivity measurements (see Brit. Assoc. Beport 

 Edinburgh, p. 647, also a later paper, Phil. Trans. 1898, p. 399). Except 

 in this particular, however, our method is quite distinct from his. 



