412 



Let S be the specific resistance of the material, or the resistance 

 of a bar one foot long, and one square foot in section ; let G be the 



resistance of the cylindrical cover of a length of cable L ; let -= be 



the ratio of the external to the internal diameter of the covering ; 

 then 



The resistance G was calculated from cables of various lengths, 

 lying in iron wells at the works of R. S. Newall and Co., Birkenhead. 

 The cables were not wet ; but direct experiment proved that cover- 

 ing a sound iron-covered cable with water has no effect on the loss. 

 The details of this experiment are given in the paper. 



The resistance G was obtained in the following manner. The 

 copper conductor of the cable to be tested was arranged so as to form 

 a complete metallic arc with a battery of 72 cells and a tangent gal- 

 vanometer : the deflection on this galvanometer was read and entered 

 as the continuity test. Deflections were then read on the same gal- 

 vanometer with the battery and several known resistances in circuit, 

 for the purpose of measuring the resistance and electromotive force 

 of the battery, in the manner described by Pouillet. The deflection 

 caused by the loss was next read on a second tangent galvanometer : 

 the same battery was used. This deflection was entered as the insu- 

 lation test. The temperature of the tank containing the cable was 

 observed by means of a thermometer inserted in a metal tube, ex- 

 tending from the circumference into the mass of the coil. 



The relative delicacy of the galvanometers was ascertained by 

 experiment, or, in other words, the coefficient was found by which 

 the tangents of the deflections of the first were multiplied to render 

 them directly comparable with the tangents of the deflections of the 

 second galvanometer. 



The resistances of the galvanometer coils, of the artificial resistance 

 coils, and of the copper conductor of the cable were measured by 

 Wheatstone's differential arrangement. Special experiments were 

 made by means of this differential arrangement to determine the 

 change of resistance of the copper conductor in the cable, produced 

 by a change of temperature. 



