OF ARTS AND SCIENCES. 139 



measure, and hence their ratios are accurately known. The only 

 ratio, then, that remains is that of heat to one of the others, and this 

 must be determined by experiment alone. 



But although we cannot measure f, g, h in general, yet we can 

 often measure off equal amounts of energy of these kinds. Thus, 

 although we cannot predict what quantities of heat are produced when 

 two atoms of different substances unite, yet, when the same quantities 

 of the same substances unite to produce the same compound, we are 

 safe in assuming that the same quantity of chemical energy comes 

 into play. 



According to these principles, I have divided the methods into direct 

 and indirect. 



Direct methods are those where b is converted directly or indirectly 

 into a, c, d, or e, or vice versa. 



Indirect methods are those where some kind of energy, as g, is con- 

 verted into b, and also into a, c, d, or e. 



In this classification I have made the arrangement with respect to 

 the kinds of energy which are measured, and not to the intermediate 

 steps. Thus Joule's method with the magneto-electric machine would 

 be classed as mechanical energy into heat, although it is first con- 

 verted into electrical energy. The table does not pretend to be com- 

 plete, but gives, as it were, a bird's-eye view of the subject. It could 

 be extended by including more complicated transformations; and, in- 

 deed, the symmetrical form in which it is placed suggests many other 

 transformations. As it stands, however, it includes all methods so far 

 used, besides many more. 



In the table of indirect methods, the kind of energy mentioned first 

 is to be eliminated from the result by measuring it both in terms of 

 heat and one of the other kinds of energy, whose value is known in 

 absolute or mechanical units. 



It is to be noted that, although it is theoretically possible to measure 

 magnetic energy in absolute units, yet it cannot be done practically 

 with any great accuracy, and is thus useless in the determination of 

 the equivalent. It could be thus left out from the direct methods 

 without harm, as also out of the next to last term in the indirect 

 methods. 



