What is the Keystone Herbivore Hypothesis?

Owen-Smith offers a different approach to multicausality in his "keystone herbivore" hypothesis, one which relies on the known ecological effects of truly large herbivores (species larger than 1000 kg) in places where they still exist today, like Africa. Very large mammals, like elephants, white rhinos, and hippos, are prodigious environmental engineers. When they exist in large numbers ("level of saturation"), they may radically affect local environments in their quest for food: "Elephants fell trees, and uproot or break shrubs and saplings. Thereby they create open glades in forests, shrublands where there were woodlands and, with the [external] agency of fire, grassy savanna where there was wooded savanna." Environmental modification is beneficial rather than detrimental for a host of smaller ungulates that literally depend on megaherbivores to "open up" closed forest and other formations so that they can enter and feed. In that sense, megaherbivores perform a crucial, "keystone" role for a variety of other species. It follows from this that when megaherbivore populations are reduced, previous levels of environmental engineering performed by them no longer obtain, and the vegetation will begin to close up. In turn, this means that open-country grazers and browsers will become increasingly confined to areas in which open, productive habitats remain (e.g., floodplains), where succession to closed forest is prevented. This is not necessarily problematic for their survival if the amount of food available is adequate to maintain them in reasonable numbers, and if predation pressure is not excessive. However, if the food supply is decreasing, and if predation pressure is in fact high, a gradual slide to oblivion is possible for these species.

How did the loss of "keystone herbivores" cause extinctions?

Owen-Smith believes that these concepts have applicability to the problem of understanding the Late Quaternary extinctions. For North America, the most prominent example, he notes the following points:

1. There were a number of megaherbivores, including mammoths, mastodonts, gomphotheres, ground sloths, horses, and very large species within groups that are not normally thought of as megafaunal (e.g., rodents). Given their number and kinds, it is highly probable that they affected landscapes and vegetation in the same manner that the megaherbivores of Africa do today.

2. Like surviving megaherbivores, the megaherbivores of the American Quaternary would have had few, if any, serious natural predators. Thus they probably reached, and maintained, saturation levels limited only by the supply of food.

3. Open-country browsers, such as equids, camelids, and smaller ungulates, would have benefitted from the level of disturbance maintained by the megaherbivores.

4. This dynamic system was capable of continuance so long as a crash did not occur among the megaherbivore component of the community. Following the loss of megaherbivores, the open-country forms would have been increasingly restricted to habitats maintained by edaphic factors.

The crash did, of course, occur. Owen-Smith states that "there seems to be no valid reason to doubt that human overkill was directly responsible for the extermination" of the American megaherbivores. However, the loss of other, smaller species was primarily due to the deterioration and elimination of the environments they required. Later human hunting may have also played a subsidiary role. In this very important sense, Owen-Smith's argument is multicausal.