BISHOP. — THERMAL CONDUCTIVITY OF LEAD. 685 



Sources of Error in Measuring Conductivity. 



1. Purity of lead. — The lead was obtained as pure as possible in 

 the granulated form. Every precaution was taken in casting to keep it 

 free from impurities. Clean chips from both parts of the shell gave 

 99.16 per cent pure. Impurities, iron and aluminium. Density: 11.313 

 gms. ± 0.018 per c.c. at 22° C. 



2. Bloiv-holes. — Probably the most difficult correction to determine 

 was that due to blow-holes. Every precaution was taken in the casting 

 to make it free from blow-holes. In the turning, constant and careful 

 watch failed to show the least trace of blow-holes, although a thickness 

 of about 6 mm. was taken from both inside and outside. 



The presence of large blow-holes in either half would show in the 

 weight. A special balance was constructed for this purpose, which was 

 sensitive to 50 mmg. under full load. The weight was found in this 

 way to be : — 



Top 14,640.98 gms. 



Bottom .... 14,642.66 gms. 



Difference ... 1.68 gms. 



This would indicate that the variation in the amount of lead in the two 

 parts is less than .013 per cent. This, together with the close agree- 

 ment of the volumes as found by measurement, would indicate the 

 probable absence of large blow-holes. 



3. Variation of conductivity due to the two halves not being close to- 

 gether. — That this is negligible is shown as follows : The surface of the 

 cavity from which heat is conducted equals 172 cm 2 . If the two parts 

 are separated by a non-conducting space of .1 mm., the amount of sur- 

 face in the cavity would be decreased by only .2 cm 2 . The surfaces fitted 

 so well that it was practically impossible to slide one over the other. In 

 order to be certain that no oil flowed in or out, the outside junction of 

 the two hemispheres was sealed with shellac. 



4. Conduction through hole carrying shaft and wires. — This hole at 

 the surface of the cavity was 4.8 mm. in diameter. If this hole was 

 filled with a non-conducting material the error would amount to less 

 than .2 per cent. But the hole was partly filled with copper, nickel, and 

 steel, all of which have a greater conductivity than lead, and partly by 

 fire-clay, which has a smaller conductivity than lead. These would 

 tend to neutralize each other and make this error negligible. 



