266 Mr Johnston's Remarks on Mr Putter's Paper 



the same capacity for heat ;" and, consequently, that the pro- 

 duct of the atomic weight into the specific heat is in all simple 

 substances a constant quantity ; or S W = C, where S denotes 

 the specific heat, W the atomic weight, and C the constant. 



In the preceding number, (No. ix.) of this Journal, page 

 75, is inserted a paper by Mr Potter, in which he professes to 

 give " an exposition of some erroneous determinations of MM. 

 Dulong and Petit, and an examination as to the metals," of the 

 above important law. On this paper I shall take the liberty of 

 making a few observations. 



Mr Potter endeavoured to determine the specific heats of 

 eight different metals, by the old method of mixing known 

 weights of the metals and of water of different temperatures, 

 and marking the resulting temperature of the mixture. His 

 results agree essentially with those of Dulong and Petit, except 

 in regard to the three metals, silver, gold, and bismuth, in 

 which the differences are as follow : 



Potter. Dulong & Petit. Difference. 



Specific heat of Silver, = .063 .0557 + .0073 



Gold, = .046 .0298 + .0162 



Bismuth, = .039 .0288 + .0102 



These differences form the sum and substance of Mr Potter's 

 paper. They are all in excess, and are the more deserving of 

 examination, as they represent the specific heats of gold and 

 bismuth as nearly one-half greater than they have hitherto been 

 esteemed. 



The old method adopted by Mr Potter is the easiest, but it 

 is liable to many sources of error, some of which are ably point- 

 ed out by Biot in the fourth volume of his Traite de Physique. 

 To escape from these sources of error, Messrs Lavoisier and 

 Laplace invented their calorimeter, and Messrs Dulong and 

 Petit adopted the expedient of employing a small cylinder of 

 silver filled with the metal to be experimented upon in the state 

 of powder. All have agreed in considering the method of mix- 

 tures as incapable of yielding any precise results, and therefore 

 we are led, a priori, to regard these high determinations of Mr 

 Potter with some degree of distrust. 



This distrust is increased when we inspect the results obtain- 

 ed by Mr Potter by inverse methods. These differences for 



