270 Mr Johnston's remarks on Mr Potter's paper 



heats obtained respectively by MM. Dulong and Petit, and by 

 Mr Potter, let us see how the constant C comes out in each. 



If the law in question be true, the value of C obtained from 

 each metal must necessarily approach very nearly to the mean 

 of all the values. Now the mean of column fourth is 3.057, 

 and all the values of C in that column are so near this mean as 

 to be completely within the limits of error, the greatest difference 

 being that of zinc, which is + .101. The mean value of C in 

 column fifth, deduced from Mr Potter's specific heats, is 3.366, 

 but the differences are more irregular than in column fourth ; 

 and the greatest is that of gold, which is 1.233 above a third of 

 the whole. In the one of these columns then, the value of C or 

 SW is nearly constant, in the other it is very irregular; we are 

 forced, therefore, to the conclusion, that either the law of equal 

 capacities does not hold, or Mr Potter's results are inaccurate. 



But in column fifth there are three numbers which are con- 

 spicuously beyond the mean of the other five, and which, there- 

 fore, in other cases would be rejected in endeavouring to deduce 

 an accurate mean. These numbers stand opposite to the three 

 metals silver, gold, and bismuth. The mean of the other five 

 values is 3.1 96, which is so near the mean deduced from column 

 fifth that we may safely rely upon the truth of the law to which 

 we have so often adverted — though, so far from showing it " to 

 hold with singular fidelity," we have seen that were Mr Potter's 

 results to be depended upon, the law, in the present state of 

 our knowledge, would be wholly untenable. 



Let us turn now to the specific heat of bismuth, from which 



