This article evaluates features of leaf and ¯ower senescence that are shared with, or are different from, those of other terminal events in plant development. Alterations of plastid structure and function in senescence are often reversible and it is argued that such changes represent a process of transdifferentiation or metaplasia rather than deterioration. It may be that the irreversible senescence of many ¯owers and some leaves represents the loss of ancestral plasticity during evolution. Reversibility serves to distinguish senescence fundamentally from programmed cell death (PCD), as does the fact that viability is essential for the initiation and progress of cell senescence. Senescence (particularly its timing and location) requires new gene transcription, but the syndrome is also subject to signi®cant posttranscriptional and post-translational regulation. The reversibility of senescence must relate to the plastic, facultative nature of underlying molecular controls. Senescence appears to be cell-autonomous, though de®nitive evidence is required to substantiate this. The vacuole plays at least three key roles in the development of senescing cells: it defends the cell against biotic and abiotic damage, thus preserving viability, it accumulates metabolites with other functions, such as animal attractants, and it terminates senescence by becoming autolytic and facilitating true cell death. The mechanisms of PCD in plants bear a certain relation to those of apoptosis, and some processes, such as nucleic acid degradation, are super®cially similar to aspects of the senescence syndrome. It is concluded that, in terms of physiological components and their controls, senescence and PCD are at best only distantly related.