with the conception we have here developed of the course of the process of excitation and its physical nature. But another important point is evolved from the observations we have made of the nature of the process of self-regulation. The process of self-regulation is founded on the same principle as that which governs the taking place of all chemical equilibrium, for metabolic equilibrium is merely a special kind of a chemical equilibrium. The development of a chemical equilibrium between reacting substances and reaction products has, as known, a characteristic course in regard to its duration. If the rapidity with which the equilibrium is reached is expressed by a curve in which the abscissa represents the time, while the ordinates signify the number of contacts of the interacting molecules, the rapidity of reaction is altered with the approach to the equilibrium in the form of a logarithmic curve; that is, the approach to the state of equilibrium, which is represented by ordinate value zero, takes place at first very rapidly, then with more and more decreasing speed, for with the decrease of the number of reacting molecules and the increase of the amount of products of reaction, the contact of the interacting molecules and with this the opportunity for the reaction occurs less and less frequently. Although the self-regulation of metabolic equilibrium is by no means such a simple process as, for instance, that of the well-known example of the forming of ethylester from acetic acid and æthyl alcohol, we have still in every case to deal with the taking place of a chemical mass equilibrium. Hence the progress to the metabolic equilibrium must likewise correspond with a logarithmic curve, i.e., restitution after a disturbance of the equilibrium must take place at first rapidly, then at a constantly decreasing rate. For reasons readily to be understood the special form of this restitution curve has so far not been accurately ascertained for any kind of living substance. Even in those cases where the restitution occurs very slowly we meet with the difficulty that, when the tests are applied which are necessary to determine the restitution at different intervals, with each testing stimulus irritability is each time reduced. Hence the construction of the restitution curve can only be achieved by indirect means, and we must content ourselves with the ascertainment of a smaller number of its points from which by interpolation its form can be constructed. Indeed in this connection a certain number of results have already been gained quite sufficient to experimentally confirm the correctness of these types of curves, primarily obtained by purely theoretical deductions. That irritability very gradually reaches its maximal height has been already shown, as previously mentioned by _Bowditch_[133] in his investigations on the influence of rhythmical induction shocks on the apex of the heart of the frog.