8o PETER GUTHRIE TAIT 



one direction and then in the other along a piece of thin platinum foil which 

 was cut away towards the centre until it became very narrow, he hoped to be 

 able to witness the shift of the glow at this narrowest part. When he got the 

 Gramme Dynamo about 1877, one of the first experiments he tried was to pass 

 the current from the Gramme machine along an iron bar when it had been 

 brought to a steady gradient of temperature along its length, after the 

 manner of Forbes' experiment in thermal conductivity. He hoped to detect 

 a change in the gradient of temperature ; but here again there was no success, 

 the current density not being great enough. 



Another line of experiments on related effects, at which A. Macfarlane, 

 C. M. Smith, and I worked, was the coordination of the striking phenomena 

 which occur in iron about the dull red heat, namely, the loss of magnetic 

 susceptibility, the reglow as the iron wire cooled, the change of sign of 

 the Thomson Effect, and the change in the law of alteration of electrical 

 resistance with temperature, all of which Tait proved to be in the neighbour- 

 hood of the same temperature. In one of these experiments iron and platinum 

 wires were led through a white-hot iron cylinder side by side, while to the 

 middle of the iron were attached the M and N platinum-iridium wires. As 

 the whole gradually cooled, observations were taken in rapid succession of the 

 resistances of the iron and platinum wires and the thermoelectric currents in 

 the N-iron and N-M circuits. The method was no doubt rough and ready 

 and not susceptible of great accuracy, but it was effective enough to establish 

 conclusions which more carefully designed experiments of later date have 

 fully corroborated. 



Among Maxwell's letters to Tait about this time the following quaint 

 remark was found written in three lines on a long strip of paper. 



" If your straight lines, parabolas, &c. have no resemblance at all to those things 

 which men call by those names, I would as soon be J. Stuart Mill as call them so. 

 But if they differ very slightly, then T' is enrolled among the Boyle and Charles 

 of 0H 1 who remain unhurt by Regnault, &c. But in Physics we must equally 

 avoid confounding the properties and dividing the substance. In the one case we 

 fall into the sin of rectification (Eccl. i. 15) and in the other we see in every zigzag 

 a proof of transubstantiation." 



Although himself greatly taken up with the thermoelectric experiments, Tait 

 never lost sight of the investigation into the thermal conductivity of metal bars, 

 which was the first serious piece of experimental work he tackled in Edinburgh. 

 This following up of Forbes' important researches was begun under the 



1 The Greek initials of Thermo-Electricity. 



