The above data relating to the antistaphylococcus phage and single strain of S. aureus with which previous papers have been concerned (9, 10, 11, 12), bring out the following points. (a) For live, resting, susceptible B suspended in broth as well as for B in a P-B mixture during the logarithmic phases of B growth and P formation, P is distributed in a manner typical of numerous materials soluble in both phases of a two phase system, i.e., distribution proceeds in accordance with the equation C(b)/C(a) = K where C(b) = extracellular P/ml. of broth and C(a) = intracellular P/ml. of B. The combination is quantitatively reversible. (b) With heat-killed, susceptible B, P distribution is of the adsorptive type, expressible in the form of the adsorption isotherm equation a = kC(1/n). The average value of 1/n is 0.80 in agreement with the results of von Angerer (2). Under ordinary conditions dead B take up much more P than do live B, the reaction proceeding to > 99 per cent completion. The combination of P with dead B is not demonstrably reversible and with high initial P/B ratios saturation of B with P is effected. Bacteria killed at 65 degrees C., 80 degrees C. and 100 degrees C. show no differences in adsorptive ability. (c) The rates at which live, resting, susceptible B and heat-killed, susceptible B remove P from solution do not differ significantly. Velocity constants of the process calculated from See PDF for Equation agree satisfactorily. It is shown that the time interval consumed is concerned with an actual reaction between P and B and not with diffusion of P through the broth to B. (d) P determinations have been found to serve as satisfactory indicators for B growth in P-B mixtures where [B] is to be maintained at a constant level. Very small increments in [B] give rise to measurable increases in P by virtue of the fact that dP/dt is proportional to a power of the rate dB/dt (9). (e) Similarly [P] estimations will detect death of B cells in P-live B suspensions. Dead B take up large amounts of P irreversibly; such P cannot function in the titration and the result is a sharp drop in [P] of controls.