432 On the Comparison of Platinum Thermometers. 



and equally, before the constants are determined. If the 

 pyrometer is used frequently for high temperatures, it is 

 occasionally necessary to redetermine the zero in the most 

 accurate work to allow for minute progressive strains. 



10. Effect of Wire-drawing. — In order to test the effect 

 of the mechanical working of the wire on its thermometric 

 scale, I had a piece of wire of diameter 6 mils drawn down 

 to 4 mils, and a double pyrometer made from these two. 

 The wire was softened by passing quickly through a bunsen- 

 flame before winding, but not annealed after winding. On 

 determining the constants, it was found that the platinum 

 temperatures for the boiling-point of sulphur differed by 

 l o- 30. The pyrometer was then heated to 1000° for an 

 hour, and slowly cooled, and the constants again determined. 

 The wires then agreed at the boiling-point of sulphur to '01 

 of a degree, and on taking the melting-point of silver a 

 proportionately good agreement was obtained. The difference 

 in the first case was certainly due to insufficient annealing. 

 The pyrometer was tested again and again, and the results 

 were always in complete agreement. These observations on 

 samples of the same wire led to the conclusion that the 

 constants determined after the high temperatures and not 

 before, as suggested by Heycock and Neville, should be 

 used in the calculations (provided that the wire is not strained 

 in cooling), and this has been uniformly done, but the 

 difference in any case is very slight. 



The wires used differed greatly in specific resistance. In 

 all cases, the wires with the smallest temperature-coefficients 

 had the largest specific resistance, and, with the exception of 

 No. 4, gave the lowest platinum temperatures. The wire 

 d= 1*670 had a specific resistance about 30 per cent, greater 

 than that of the standard, and the others nearly as great, but 

 the measurements of specific resistance could not be made 

 very accurately on wires of so small a diameter. 



In conclusion, I must thank Prof. Callendar for kind 

 assistance, Miss Harriet Brooks, B.A., of the Royal Victoria 

 College, and Mr. N. M. Yuile, M.Sc, of the Mining Depart- 

 ment, McGill College, for assistance in taking observations. 



McGill College, April 19th, 1900. 



