150 



THE SCIENTIFIC PAPERS OF 



fixed ratio indications of the higher temperatures. It has enabled 

 me to test the general accuracy of the ratio of increase of electri- 

 cal resistance beyond the limits of the more correct tests obtained 

 at the lower temperatures. The accuracy of these corroborative 

 results depends upon the supposition that the specific heat of the 

 metal ball is the same at high and low temperatures ; but, although 

 this may not be, strictly speaking, the case, there is evidence to 

 show that the variations are not of serious import, except probably 

 in nearing the melting points. 



The following are some comparative results which have been 

 obtained by placing in the same heated chamber a copper ball of 

 known capacity of heat, and a coil of platinum wire wound in the 

 spiral grooves of a porcelain cylinder and protected from injury 

 by a cylindrical casing of platinum ; both the copper ball and the 

 protected spiral wire were placed inside the heated chamber in a 

 piece of wrought-iron tubing, to ensure more complete identity of 

 temperature, when the resistance of the spiral was taken, and the 

 copper ball dropped into the apparatus just described. 



It remains to be proved whether the law of increase of electrical 

 resistance, which I have here ventured to put forward, holds good 

 for all conductors ; and whether it may be trusted at tempera- 

 tures approaching either the point of absolute zero or the melting 

 point of the metal under consideration. The whole subject, 

 indeed, requires further and fuller investigation than I could devote 

 to it with the principal object of my investigation in view, which, 

 having been the construction of a reliable instrument for measur- 

 ing low and high temperatures by electrical resistance, I have 

 followed up this branch of the enquiry only to such a point as 

 to supply a tolerably reliable basis for such practicable purposes. 



