590 



REPORT — 1897. 



cast-iron lies on the same side of the copper line as that of wrought-iron, but a 

 little lower down. 



Curve III. shows the differences as hefore for the platinum-platinum-rhodium 

 couple. The scale here for the ahscissas is just one-fifth that in the other two cases, 

 and that of the ordiuates one-tenth. The greatest difierence will be seen to be at 

 150°, the differences being much greater below the 500° point than above. 



© Observations of H. and W. + Observatious by Plat. Pyr. 



The curve representing the formula of Holborn and Wien ' has been plotted, 

 and the differences from the same parabola are shown by the dotted line. Their 

 observations are necessarily much less accurate than those taken by the resistance 

 pyrometer, as is shown by the points marked ©, along the difference curve. This 

 is owing, in the first place, to the difficulties in air thermometry work, and also to 

 the fact that the temperature was changing when the observations were taken. 

 Their curve, though similar, shows the E.M.F. for the same difierence of tempera- 

 ture to be considerably higher than in the present case, due probably to a differenc* 

 m the composition of the alloy. Various other empirical formulas have been sug- 

 gested, but as they have no theoretical basis, they have not been considered. It 

 is obvious that a straight line does not fulfil the conditions unless it be between 

 points not very far separated. 



Tait's formula, as far as it holds, perhaps simply amounts to drawing the nearest 

 parabola, as the differences found in the present observations are quite beyond the 

 limits of possible experimental errors. There seems to be no reason, however, to 

 conclude that the formula may not represent the physical explanation of the efl'ect, 

 at least to a first approximation. 



Witd. Ann., 1892, p. 107. 



