1893.] Absolute Thermal Conductivities of Copper and Iron. 151 



temperature. These observations confirmed the rectilinear relation 

 with the first three ; but in the case of the two alcohols, evidence was 

 obtained of molecular association, as also with acetic acid. It is pos- 

 sible to calculate the amount of association at any temperature in such 

 cases. For, assuming the constancy of k for the molecular surface of 

 the "normal " liquids, the equation 



where d is the differential coefficient of an associating liquid, and x 

 is the molecular aggregation, gives the number of simple molecules 

 which have united to form a compound at the temperature chosen. 

 Eor the alcohols at —90°, and for acetic acid at 20°, the association of 

 molecules approximates to (C 2 H 4 2 )4, (CH 4 0)4, and (C 2 H 6 0)4, 



We have thus a method by which it is possible to ascertain the 

 molecular complexity of undiluted liquids. The results with the 

 alcohols are shown to agree within reasonable limits with those ob- 

 tained from strong solutions by Raoult's method. 



It is incidentally shown in the course of the paper that there is no 

 angle of contact between liquid and glass, when the liquid surface is 

 in contact only with its own vapour. Ordinary measurements of 

 capillarity give inconstant, and probably inaccurate, results, for it is 

 not the surface tension of the liquid which is measured, but that of a 

 solution of air in the surface film of the liquid. 



The paper contains tables and curves exemplifying and illustrating 

 the statements given. 



IV. " The Absolute Thermal Conductivities of Copper and 

 Iron." By R. Wallace Stewart, B.Sc. (Lond.), Assistant 

 Lecturer and Demonstrator in Physics, University College, 

 Bangor. Communicated by Lord Kelvin, P.R.S. Re- 

 ceived March 2, 1893. 



(Abstract.) 



The experiments described in the paper were undertaken with the 

 object of determining the theoretical conductivity at different tem- 

 peratures of iron, and, in particular, of pure electrolytically deposited 

 copper. 



The method adopted was that due to Forbes, with two modifications. 



(a.) The thermo-electric method of determining temperature was 

 employed. The thermo-electric couple used was one of German 

 silver and iron, giving, between 0° C. and 200° C, a uniform deflec- 

 tion on the galvanometer scale of about four divisions for a difference 

 of one degree centigrade between the temperatures of its junctions. 



