several smaller predators (e.g. striped skunk, gray fox, feral cats) that in turn have
reduced the diversity of scrub-breeding birds (Crooks and Soule 1999).
A more complex situation is found on Marion Island off South Africa (Huyser
et al. 2000). There feral cats fed on exotic house mice (Mus musculus) and burrow-
ing petrels, reducing petrel numbers considerably. Mice consumed terrestrial macro-
invertebrates. The endemic lesser sheathbill (Chionis minor), an aberrant shorebird
of the Chionididae related to plovers, also depends on these macro-invertebrates
in winter. Macro-invertebrates are maintained by the feces from burrowing petrels,
and the loss of petrels resulted in the loss of invertebrates. Removal of cats caused
an increase in house mice (but not petrels because their habitat for burrows had
changed), and a further decrease in macro-invertebrates. The sheathbill population
collapsed on that island in comparison to that on neighboring Prince Edward Island
where mice were at much lower density. Thus, non-linear indirect effects involved
an increase in competition between mice and sheathbills for invertebrates through
meso-predator release.
In all of these examples ecosystem disturbances through introduction of a new species
resulted in unpredicted outcomes due to indirect food web interactions, with unex-
pected consequences for conservation.Within groups of co-occurring species some of the species play a larger role in defining
the group. In terrestrial environments, it is the plants that dictate the structure and
function of that community, and these plants are determined by the abiotic environ-
ment. Thus, in arctic and alpine areas we get tundra with herbs, grasses, and shrubs
determining a low-lying structure. In contrast, in very wet temperate and tropical
areas we find forests with tall, woody plants creating a complex three-dimensional
structure. Thus, we say that the most common plants (most numerous, highest biomass),
such as white spruce in boreal forest, are dominantspecies that determine the struc-
ture and function of the ecosystem.
Some species can have a considerable influence on the community even though
they are relatively rare. Often these are predators and they can determine not just
the species composition of prey but, indirectly, many other components of the
ecosystem as well. Such species have been called keystone(Paine 1969; Power et al.
1996), and are defined as those that have a greater role in maintaining ecosystem
structure and function than one would predict based on their abundance or biomass.
Top predators are often presented as keystone species: for example, the presence or
absence of sea otters (Enhydra lutris) as top predators of inshore marine communit-
ies determines the abundance and species composition of other members (Estes and
Duggins 1995). Herbivores, however, can also act as keystones: for example, rodents
can structure desert plant communities (Brown and Heske 1990), snowshoe hares
(Lepus americanus) structure boreal forest vertebrate communities (Krebs et al.
2001b), and wildebeest (Connochaetes taurinus) structure the Serengeti ecosystem
(Sinclair 2003).
There are two major problems with the keystone concept. First, there are
operational problems with identifying keystone species. We need to define which
parameter we measure – abundance, biomass, species composition, or something else.
We need to specify what is the degree of change in the community expected from
losing a keystone species. Communities are open ended and we must state how far
into the food web we should trace the impacts. Thus, the impacts of top predators368 Chapter 21
21.6 Community features and management consequences
21.6.1Dominant and
keystone species