Limnol. Oceanogr., 44(3, part 2), 1999, 932–940

Coral reefs are subject to a high frequency of recurrent biological and physical disturbances. The temporal and spatial scales of these are often large and difficult to study, so that most of our knowledge of disturbances on coral reefs comes from investigations conducted at one or a few sites, over short periods of time. We argue that studying single events in isolation can be misleading and that a longer term approach is necessary for understanding the responses of coral reef assemblages to multiple stressors. We present first a brief review of the impacts of physical disturbance (e.g., cyclones, hurricanes) on the community dynamics of coral reefs, with special attention to the effects of recurrent events. We then examine two unusually detailed, long-term data sets from Heron Island, Australia, and Jamaica which demonstrate some of the complexities of multiple stressors (broadly defined as natural or man-made disturbances). Both case studies illustrate that the effect of a particular disturbance often depends critically on the impact of previous perturbations. Consequently, even the same type of recurrent stressor can have different effects at different times, depending on history. Accordingly, when the added dimension of time is considered, the distinction between single and multiple stressors becomes blurred. Even a single event such as a hurricane can be viewed mechanistically as a multiple stressor, with shortand long-term impacts. We emphasize that multiple stressors often have significant effects on recruitment and regenerative processes of assemblages. These impacts are much less obvious than catastrophic or chronic mortality, but they play a crucial role in community dynamics over longer time scales. Importantly, chronic anthropogenic impacts can impede the ability of coral assemblages to recover from natural disasters, even where there is little detectable effect on rates of adult mortality. Once a reef has been degraded, it is usually impossible to ascertain retrospectively the precise mechanisms that were involved or the relative importance of different events. A single survey will provide a snapshot of the status of coral reefs, but a longer term approach is required to understand the processes underlying changes in assemblages. Stress is often considered in the coral reef literature as a sublethal effect on the physiology of an animal or plant, e.g., a decline in feeding, growth, or fecundity, or a biochemical change. The term has also been used at the population or demographic level, usually in terms of an acute or chronic disturbance that causes a decline in the number of organisms. Increasingly, scientists, reef managers, and the general public also refer to stressed or degraded assemblages of species or ecosystems affected by natural and anthropogenic impacts, and there is growing concern about the health of coral reefs (e.g., see Brown and Howard 1985; Ginsberg 1993; Richmond 1993). The concerns for the declining status of many coral reefs are legitimate, although we caution that some of the terminology which is now in general use can be misleading. For example, the notion of a stress being imposed on a coral reef may conjure up a disturbance to a system which is otherwise