of the Theory of Heat. 27 1 



heats furnished by the observations of those of the particles them- 

 selves, it is sufficient to divide the former by the number of par- 

 ticles contained in the same weight of the substances which we 

 compare : but it is obvious that the number of particles for equal 

 weights of matter are reciprocally proportional to the density of 

 the atoms. We shall, therefore, obtain the desired result by mul- 

 tiplying each of the capacities deduced from experiment by the 

 weight of the corresponding atom. These different products are 

 presented in the last column of the table. 



The approximation apparent on a bare inspection is too remark- 

 able by its simplicity, not to indicate the existence of a physical 

 law capable of being generalized and extended to all elementary 

 substances. These products, which express the capacities of the 

 different atoms, approach so near equality, that the slight differ- 

 ences must be ov/ing to trifling errors either in the measurement 

 of the capacities or in the chemical analyses ; especially if we 

 consider that, in certain cases, these errors, derived from these 

 two sources, may be on the same side, and, consequently, be 

 found multiplied in the result. The number and variety of the 

 substances on which we operated not allowing us to consider the 

 relation thus indicated as merely accidental, we are authorized 

 to deduce from them the following law : — The atoms of all sim- 

 ple bodies fiave precisely the same capacity for heat- 

 By recollecting what has been stated respecting the kind of 

 uncertainty that exists in fixing the specific weight of the atoms, 

 it may be easily conceived that the law which we have just esta- 

 blished will change, if we adopt for the density of the particles a 

 supposition different from what we have chosen ; but, in every 

 case, the law will exhibit a simple ratio between the weights and 

 the specific heats of the elementary atoms ; and it is obvious that, 

 when we had to choose among hypotheses equally probable, we 

 should naturally be led to prefer that which established the most 

 simple relation between the elements which we compared. But 

 whatever opinion be adopted respecting this relation, it will en- 

 able us afterwards to control tiie results of chemical analysis ; 

 and, in certain cases, will give us the most exact method of ar- 

 riving at the knowledge of the proj)ortions of certain combina- 

 tions : but if, in our subsequent experiments, no fact occur to in- 

 validate the probability of the opinion we now hold, we shall find, 

 in this method, the advantage of fixing in a certain and uniform 

 manner the specific weight of the atoms of all simple bodies that 

 can be suljmitted to direct observations. 



The law we have announced, seems to l)e independent of the 

 form which bodies assume, provided that wo always consider them 

 under the same circumstances. This, at least, is a consequence 

 deducible from the experiments of MM. Laroche and Berard ou 



the 



