The field of apoptosis is unusual in several respects. Firstly, its general importance has been widely recognised only in the past few years and its surprising significance is still being evaluated in a number of areas of biology. Secondly, although apoptosis is now accepted as a critical element in the repertoire of potential cellular responses, the picture of the intra-cellular processes involved is probably still incomplete, not just in its details, but also in the basic outline of the process as a whole. It is therefore a very interesting and active area at present and is likely to progress rapidly in the next two or three years. This review emphasises recent work on the molecular mechanisms of apoptosis and, in particular, on the intracellular interactions which control this process. This latter area is of crucial importance since dysfunction of the normal control machinery is likely to have serious pathological consequences, probably including oncogenesis, autoimmunity and degenerative disease. The genetic analysis of programmed cell death during the development of the nematode Caenorhabditis elegans has proved very useful in identifying important events in the cell death programme. Recently defined genetic connections between C. elegans cell death and mammalian apoptosis have emphasised the value of this system as a model for cell death in mammalian cells, which, inevitably, is more complex. The signals inducing apoptosis are very varied and the same signals can induce differentiation and proliferation in other situations. However, some pathways appear to be of particular significance in the control of cell death; recent analysis of the apoptosis induced through the cell-surface Fas receptor has been especially important for immunology. Two gene families are dealt with in particular detail because of their likely importance in apoptosis control. These are, first, the genes encoding the interleukin-1 beta-converting enzyme family of cysteine proteases and, second, those related to the proto-oncogene bcl-2. Both of these families are homologous to cell death genes in C. elegans. In mammalian cells the number of members of both families which have been identified is growing rapidly and considerable effort is being directed towards establishing the roles played by each member and the ways in which they interact to regulate apoptosis. Other genes with established roles in the regulation of proliferation and differentiation are also important in controlling apoptosis. Several of these are known proto-oncogenes, e.g. c-myc, or tumour suppressors, e.g. p53, an observation which is consistent with the importance of defective apoptosis in the development of cancer. Viral manipulation of the apoptosis of host cells frequently involves interactions with these cellular proteins. Finally, the biochemistry of the closely controlled cellular self-destruction which ensues when the apoptosis programme has been engaged is also very important. The biochemical changes involved in inducing phagocytosis of the apoptotic cell, for example, allow the process to be neatly integrated within the tissues, under physiological conditions. Molecular defects in this area too may have important pathological consequences.