1. A very complete analogy exists between a chemical reaction and the process of disinfection, one reagent being represented by the disinfectant, and the second by the protoplasm of the bacterium. 2. Three classes of disinfectants were studied, (a) metallic salts (HgCl 2 and AgNO 3 ), (b) phenol, and (c) emulsified disinfectants (disinfectant “A”). B. paratyphosus and spores of B. anthracis were chosen as types of vegetative and spore-bearing organisms respectively. 3. In the case of anthrax spores, the disinfection process proceeds in obedience to the well-known equation for a unimolecular reaction, if numbers expressing “concentration of reacting substance” are replaced by “numbers of surviving bacteria”. 4. Experiments with B. paratyphosus show a departure from the simple law owing to permanent differences in resistance to disinfectants among the individual organisms. The younger bacteria were proved to be the more resistant. 5. The process of disinfection is influenced by temperature in an orderly manner, and the well-known equation of Arrhenius can be applied. (a) Disinfection of B. paratyphosus by metallic salts is influenced by temperature to about the same degree as most chemical reactions, the reaction velocity being increased about three-fold for a rise in temperature of 10°C. (b) For disinfection of B. paratyphosus by phenol and the disinfectant “A” there was a much higher temperature coefficient, viz., seven to eight. In the case of phenol the effect of temperature was again found to be complicated by the want of uniformity among the individual bacteria. Disinfection of the younger, more resistant bacteria, was found to possess a higher temperature coefficient than that of the less resistant forms, the coefficient varying from ten to three, or two according to the age and number of the bacteria disinfected. 6. It follows from (5) that there is a very great advantage in the use of warm solutions for practical disinfection. 7. Experiments, made with varying concentrations of disinfectant, and using similar groups of bacteria from cultures of B. paratyphosus , showed a definite logarithmic relation, between the concentration of disinfectant and the mean reaction velocity of disinfection, to exist in the case of phenol and the disinfectant “A”. 8. In the case of silver nitrate, the same relation existed, but, in the case of mercuric chloride, numbers representing concentration of the salt had to be replaced by those representing concentration of the metallic ion. This confirms the theory that in disinfection with metallic salts the metallic ion is the real disinfecting agent. 9. This logarithmic relation is surprising in view of the simple proportionality existing in the case of chemical processes running the course of a unimolecular reaction, with which disinfection shows a close analogy. 10. Some evidence was obtained that, in disinfection with mercuric chloride, a toxic compound is formed between the metal and the substance of the bacterial cell. This compound prevents all further growth, but vitality can be restored by the administration of a large excess of soluble sulphide as an antidote. I am glad to have this opportunity of expressing my great indebtedness to Dr C. J. Martin, at whose suggestion the work was undertaken, and who has helped me throughout, not only with most valuable advice, but also with practical assistance in many of the experiments.