Destructive grazing by large terrestrial herbivores also occurs, although in most cases it is an artefact of human interference. Perhaps one of the best examples of this phenomenon is the "elephant problem" faced by all wildlife conservation areas in Africa whenever poaching is not a serious problem.

Elephants often push over, uproot or ring-bark trees, so they can be a serious nuisance, even at lower densities than would be expected to cause serious overgrazing of all vegetation and consequent soil erosion. They prefer to feed on young shoots regenerating from the rootstock of felled trees, as well as young saplings, and so they can effectively prevent the regeneration of damaged woodland.

Once the woodland has been damaged by elephants, many of the other species which depend on trees for habitat and food, also decrease in abundance or disappear. Of course, grassland and bush species might increase, but this represents a complete change of the ecosystem from one state (woodland) to another state (bush and scrub). Because smaller reserves offer fewer refuges for woodland species during times of severe elephant damage, smaller reserves are more likely to suffer the complete extinction of woodland plants and animals within the reserve.

Some people believe that "the elephant problem" is actually a natural phenomenon, and that elephants and woodland cycle with one another due either to climatic cycles, or to the population dynamics of the elephants themselves. However, it is more likely that "the elephant problem" is of human creation. In 1978 Malpas pointed out the necessity of examining "the elephant problem" on a large enough scale. The basis of Malpas' "compression hypothesis" is the idea that elephants range over a large area, the size of a small state. Within this range there would be areas of good elephant habitat and areas of poor elephant habitat. Given the destructive habits of elephants, if they stay in a good area for too long then destructive grazing will result. If, however, their migratory routes are not restricted, Malpas postulates that elephants will move to another area of favourable habitat, leaving the damaged area to recover secondary growth. S.K. Eltringham has suggested that such secondary forest is superior habitat for elephants. If this is so, it is easy to see that by partially wrecking a forest, and then moving on, only to return when secondary regeneration has occurred, elephants might have been able to maintain favourable areas of habitat within their range.

However, with the colonisation of Africa, the superior habitat patches were either converted to farmland, used for forestry, or surrounded by game fences. The migratory routes of elephants were cut off, and what probably evolved as a highly beneficial behaviour became a destructive behaviour, harming elephants, vegetation and other animals alike. We may never know if this is the correct explanation for "the elephant problem," for it is unlikely that elephant migration routes will ever be restored, and that would be the only way to test this hypothesis.

Nobody likes to shoot elephants, but it may be necessary to prevent habitat destruction when they are confined to reserves. The photo shows an area of mopane woodland that has been turned into an elephant lawn by the destructive feeding techniques of elephants.

Elephants are not the only animals to cause habitat destruction when confined and protected in wildlife parks. The confined activities of all herbivores can lead to overgrazing. In the Urungwe Safari Area in the Zambezi Valley in Zimbabwe, I have seen repeated signs that a combination of fire and overgrazing have led to the destruction of vegetation cover, and consequently, to soil erosion. No one species of herbivore is responsible; rather, the total numbers of all herbivores is probably higher than the optimum for the area. The grazing problem is compounded by the spread of man-made fires into the area.

Herbivores such as hippopotamus can also be implicated in serious destructive grazing in reserves, especially in combination with other herbivores. Whether this would also occur under completely natural circumstances is not known.

We have seen how an overabundance of one or more species of herbivore can cause destructive grazing, or the switch of a community from one alternate state to another. We shall now explore some of the more subtle effects of herbivores on natural communities. It should be emphasized that the ecological effects of herbivores are mostly beneficial to the community, although perhaps the concept of "benefit" is anthropocentric and entirely inappropriate in this context.