TRANSACTIONS OF SECTION A. 589 



The copper, cast-iron, and wrought-iron -were tinned together at one end and 

 heated in an oil bath, in which also was inserted the tube containing the pyro- 

 meter. The observations were taken only at perfectly steady temperatures. The 

 direct reading galvanometer, devised by Professor Oallendar for the purpose, was 

 used in taking temperatures with the pyrometer. A carefully calibrated resistance 

 box of the standard type was also used. 



The platinum temperatures were calculated from the formula 



Pt = (1^^") ^00 (1) 



and the air thermometer temperatures by means of the difference formula 



^-^'^=<(i4y-4) • • • '^^ 



The object being to show the relation between the experimental curve and Tait's 

 parabola, it is quite obvious that the ordinary method of plotting the temperature 

 and the E.M.F, is not sufficient. A difference method similar to that above — equa- 

 tion (2) — was therefore adopted. The values of t were taken as abscissas, and the 

 corresponding values i - ?c as ordinates, t^ denoting the couple temperature, where 



-y|^ is the temperature coefficient. 



The relation to the parabolic formula may be shown thus : — 



El = at + I3t- (Tait's formula) 

 then 



VVioo/ 100/, 



where S is a constant depending on the nature of the metals. The value 5 was 

 calculated from the observed difference, t — t,at 200°, With this value for 8 the 

 parabola was plotted, passing therefore through three points on the experimental 

 curve, those corresponding to 0°, 100°, and 200°. The accompanying curves show 

 the differences from this parabola plotted along the axes. 



In order to compare the Le Chatelier couple directly with the resistance pyro- 

 meter, the couple and the pyrometer were placed in the same porcelain tube. The 

 couple was the usual form, a pure platinum wire coupled with another of platinum 

 containing 10 per cent, rhodium. The wire was obtained from Messrs. Johnson, 

 Matthey & Co. The junction was placed so as to be directly under the pyrometer 

 coil, from which it was separated by strips of mica. 



The temperatures, as before, were taken only at steady points. These points 

 were obtained by varying the gas supply under a vessel containing molten tin, and 

 by taking the melting point of tin, the boiling point of sulphur, and the melting 

 point of silver. The curve in this case was also plotted against the parabola, the 

 value S being calculated from the observed difference of temperature , t - 1^, at 979°. 



The temperature coefficient of the junction was in this case taken as —522 go as to 



500, 

 give the nearest parabola throughout the whole range. 



Curve I. shows the differences for the copper wrought-iron couple. The ab- 

 scissas are temperatures by the air thermometer, the ordinates as before stated the 

 diHerences from the parabola. The difference is — 1°"1 at 50° and about + l°'l 

 at 150°. 



If the temperature of the neutral point be calculated from the difference equa- 

 tion, using for b the value found at 200, namely, 23-25, then Tq (the neutral 

 point) = 265°. The experimental curve shows it to be somewhat lower than this. 



Curve II. shows on the same scale the differences for the junction of cast-iron 

 and wrought-iron. The differences are smaller in this case, and the curve more 

 regular. The neutral point, calculated as before, is 917°, which lies entirely 

 beyond the limits of the experiments. The value of 8 is 5-765. 



By reference to the thermo-electric diagram it will be seen that the line for 



