130 Mr Miller's Description of a new Pyrometer. 



the increasing specific heat of platinum, for this might be 

 supposed nearly balanced by the increasing rate of expansion 

 of platinum in the other determination of 1141° with the re- 

 gister pyrometer. The increasing capacity, and the expan- 

 sion of metals in general, seem to bear a relation to each 

 other, and in platinum both are small. If the temperature 

 assigned to an ordinary fire by this pyrometer be really too 

 great, the error, I apprehend, can be ascribed only to pro- 

 gressive specific heat ; yet, no such increase was shewn in 

 the previous experiment with boiling oil, that gave a result 

 rather imder 600°, the temperature indicated by an open ther- 

 mometer. 



The fusing point of copper, and the white or welding heat 

 of iron, were also shewn at higher ranges, but much under 

 those assigned by Morveau in his corrections of Wedg e wood's 

 scale. 



It is worthy of remark, that the determinations by Daniell's 

 first pyrometer (described in Brande's Quarterly Journal) 

 were higher when pulleys were used to magnify expansion, 

 than those afterwards given by the register pyrometer, in 

 which a lever is used instead of pulleys. The melting point 

 of cast-iron by the former was 3479°, by the latter 2786°. 

 Dr Brewster, in his edition of Ferguson's Lectures, has 

 pointed out a singular oversight in the construction of Fei'- 

 guson's pyrometer, with respect to the lever employed to 

 increase expansion to the eye. He says, in a note (vol. i. p. 

 20), " It is wonderful how the author and other writers on 

 natural philosophy should have overlooked the striking de- 

 fect in the principle upon which this new pyrometer is con- 

 structed." The error is then pointed out, with references to 

 the drawing of the instrument in the book of plates. Dr 

 Brewster adds, "As the arms of the two levers, therefore^ 

 are continually changing their proportion, every pyrometer 

 constructed upon the principle of the lever must give a very 

 inaccurate result."* 



* The error in Ferguson's pyrometer might be rectified by a scale of equal 

 parts, laid down, not on the arc, but on a tangent to it, parallel to the expand- 

 ing bar. 



