Most communities go through some kind of succession (change with time) following a major disturbance, with certain species characteristic of the early stages and certain other species characteristic of the later stages of succession. Herbivores can arrest succession in the early stages, and prevent the later stages from developing. One way for this to happen is if the herbivores move about from patch to patch, where they feed destructively, eventually returning to the same patch after it has partially recovered from the effects of their feeding. If the "compression hypothesis" is the correct explanation of "the elephant problem" mentioned earlier, then this is exactly what elephant foraging would have done prior to the advent of technological man. By destroying trees in one habitat patch, and then moving on to another, elephants could allow the early to middle stages of succession to take place, but their eventual return to the same patch would prevent tha later successional stages from developing.

Another way in which an early successional community can be maintained, despite intensive herbivory, is if herbivore feeding is reduced on a seasonal basis. Perhaps the best example of this phenomenon comes from the marine environment. I said earlier that much of the Northwest Atlantic subtidal zone is heavily grazed by sea urchins, and so the amount of seaweed present is relatively low. But in late winter and spring, when sea states are quite rough, the urchins cannot feed as effectively as they can during the calmer sea states of summer and autumn. So in the spring there develops a community of early successional annual seaweeds, which may then persist well into the summer. This phenomenon sometimes also happens on South African shores, when limpets are prevented from feeding effectively during extended stormy periods, and patches of sea lettuce and other seaweeds develop.