These theories of Berthelot are probably partly true and must be taken into consideration. I prefer, however, a thermodynamical solution of this problem because it is more general and more profoundly scientific. Unfortu- nately, this solution is based on the subdivision of the total energy of a system into available energy (susceptible of being transformed into work) and unavailable energy (capable of manifesting itself only in the form of heat), and on a series of theorems of which it is impossible to give any idea in these ' Outlines of Physical Chemistry.' I can, therefore, only give the conclusions of this special study and say that, according to thermodynamics, the majority of reactions must take place with evolution of heat — in conformity with Berthelot 1 s principle ; but in certain special cases, notably changes at high temperatures or between diluted substances (gaseous or dissolved), the normal thermal effect may be negative.
The law of maximum work has not, therefore, the value of a principle. It is only a more or less general rule to which we know and foresee many exceptions. In order the better to sustain this conclusion it will be well to recall Maupertuis's principle (Le Chatelier's theorem) which states that : any change in the factors of equilibrium from outside is followed by a reverse change within the system. In thermo -chemistry we translate this to : if, by a change we give rise to a transposition whose thermal effect is opposed to the change of temperature. In cooling the system we favour exothermic exchanges ; on the other hand, on heating we favour endothermic transpositions. These latter appear, therefore, in the natural order of things, and seem necessarily to be realised especially at a
1 The question is, for example, the equilibrium of a reversible reaction : aa, + bb 1 J^ ab + AjBj.
2 The formation of acetylene in the path of the electric arc in an atmosphere of hydrogen may serve here as an example. Amongst
�� �
