278 Profs. Dewar arid Fleming on the Electrical 



The process of measuring the mean diameter of any wire 

 was then as follows : — The appropriate length of the wire 

 being cut off and measured against a standard-metre scale, it 

 was drawn through the field of the microscope, the smallest 

 gauge, F or G, being placed in the same focal plane and 

 parallel with the wire. The gauge was placed with the flat 

 side of the handle parallel with the screen a ad the centre of 

 the steel gauge in the centre of the objective field. We had 

 then on the screen the images or shadows of two objects, one 

 a wire of steel of known diameter, and the other a wire of 

 unknown diameter ; the diametral planes of each wire being 

 in the same focal plane of the lens. The diameters of the 

 images of the steel gauge and the wire were then carefully 

 measured with dividers and expressed in terms of some arbi- 

 trary unit. The ratio of these apparent image-diameters was 

 taken as the ratio of the real diameters of the wire and the 

 gauge. 



The measurement of the diameter of the wire was taken 

 about every five centimetres in the case of very uniformly 

 drawn wires, and more frequently in other cases. Generally 

 about twenty, and sometimes sixty diametral measurements 

 were taken in the metre length of the wire to be measured, 

 and as the wire was turned round its long axis in passing 

 through the field of the microscope the diametral measure- 

 ments were taken in numerous directions. The mean of all 

 these measurements gave a close approximation to the mean 

 diameter of the wire. As an instance of the close agreement 

 of the results by this method, and those obtained by the 

 density method, we may give the following comparison of the 

 determination made of the mean diameter of a certain high- 

 conductivity copper wire about '01 inch in diameter. Mr. J. 

 W. Swan sent us this wire, and determined its mean diameter 

 by a very careful measurement of its length and mean density. 

 Our measurement of the mean diameter of this wire was 

 0*02609 centim., and his was 0*02608 centim., agreeing thus 

 within 1 part in 2600. By the use of a suitable objective 

 this method was applicable to the smallest wires used as well 

 as to the largest. Each wire having had its mean diameter 

 and length thus ascertained, the mean cross-sectional area was 

 calculated, and a reduction-factor obtained by which the 

 reduced true resistance of the wire in ohms was multiplied to 

 deduce therefrom the volume-specific resistance as above 

 defined. 



§ 5. Furnished with these wires, the next step was to pre- 

 pare suitable resistance-coils by which the resistance of the 

 wires could be taken in a bath of liquid gas. In our former 



