A full understanding of the mechanism of action of an antibiotic involves not only the identification of the drug-sensitive biochemical targets, but also the pathway by which the inhibition of the target reaction(s) leads to the eventual physiological consequences, i.e. inhibition of the reproductive capacity of the whole cell. The complexity of such a pathway may be illustrated by the case of the penicillin response of pneumococcal mutants defective in the activity of murein hydrolases, i.e. enzymes that do not seem to react directly with the penicillin molecule. These hydrolase-defective mutants seem to contain all the `normal' penicillin-binding proteins; penicillin treatment causes typical morphological and biochemical effects (interference with cell wall metabolism) and the minimal inhibitory concentration as well as dose response to penicillin are identical in the hydrolase-defective mutant cells and in the wild-type bacteria. In spite of all these similarities, the eventual physiological response of mutants is strikingly different from that of the wild-type cells: in contrast to the wild-type cells, mutants do not lyse during penicillin treatment and their rate of loss of viability is greatly suppressed. These observations suggest that inhibition of the primary biochemical targets of penicillin (penicillin-binding proteins) is a necessary but not sufficient condition for the induction of the pharmacologically most useful irreversible effects of <latex>$\beta$</latex>-lactams. During the past 4-5 years, our laboratory has been engaged in an intensive effort to unravel the mechanism of penicillin-induced death and lysis in several species of bacteria and I shall summarize briefly some of our conclusions here, using mainly observations obtained in two experimental systems: the Gram-positive pneumococci and the Gram-negative Escherichia coli. Our approach to the mode of action of <latex>$\beta$</latex>-lactams involves an attempt to identify cellular factors that are directly responsible for the events of cell death and lysis. In my mind, such studies complement the other approaches (biochemical, enzymological and genetic analysis of penicillin sensitive enzymes and binding proteins), summarized in the talks of Professor Strominger, Professor Ghuysen and Dr Spratt at this meeting.