A kinetic analysis of cell-mediated cytotoxicity (CMC) at the population level based on theoretical models has been performed. This analysis considers that the binding process and the kinetics of the lytic process occur through different types of conjugates: LTn, i.e., conjugates containing one effector cell and n target cells, and LmT conjugates which contain m effector cells bound to one target cell. This allowed us to provide a quantitative description of the conjugation process, and of the binding capacities of the effector and target cells. Thus, it has been shown that these processes are governed by an equilibrium in which are involved the unbound effector and target cells, and the conjugates formed. This implies that, when the equilibrium concentrations are reached, the total number of conjugates cannot be increased from the unbound effector and target cells. However, it does not mean that the free effector cells are nonbinding, and so, when the conditions of equilibrium are perturbated (as occurs for example in CMC), all effector cells, virgin and those who have already participated in the lytic process, are able to form new conjugates. The existence of this equilibrium has also important consequences when different subpopulations are separated from an effector-target system. Thus, it explains the observation reported in the literature that, although cytometric techniques can be used to detect and count different types of conjugates, the conjugates formed cannot be separated by cell sorting (unless special precautions are taken). Finally, we have found that the number of target cells killed by one effector cell, which has been previously considered as the recycling capacity of an effector population, is in reality the result of two different mechanisms. One of these mechanisms is due to the recycling process, whereas the other (which has the same effect) is due to the multiple killing capacity of the LTn conjugates which kill more than one target cell. The average number of target cells killed per conjugate has been determined, and this allowed us to obtain the relative contributions of these two mechanisms to the total killing capacity of one effector cell. It appears that the recycling capacities of effector cell populations previously determined had been 1- to 3-fold overestimated.