ney (1993) model the relationship between the number and
intensity of connections among species (which they call de-
pendencies) on the impact of species removal. As the level
of dependencies among species goes up, the number of spe-
cies expected to go extinct when a species is removed rises
in an S-shaped curve. Their model does not distinguish pos-
itive and negative coactions, nor does it consider indirect
effects. They nevertheless come to the cautious conclusion
that communities that rarely experience species removals
will evolve to increasing levels of dependencies. This in turn
will make them extremely vulnerable to mass extinctions
if a major disturbance does occur. By contrast, frequently
perturbed communities will evolve toward lower levels of
dependency, and hence acquire more extinction resistance.
For conservationists, this implies that for a species of con-
servation concern we need to know about its dependencies
with other species to judge both (1) its community domi-
nance, and therefore its importance to the integrity of the
community as a whole, and (2) its vulnerability to nonspe-
cific disturbances.
Because rodents are often abundant in communities, they
will be strongly connected to other species and hence re-
latively dominant. Rare species may be connected to com-
mon ones through food competition and predator sharing.
A few examples will illustrate the importance of rodents in
diverse communities. Ostfeld and colleagues (Jones et al.
1998) have assembled a remarkable account of how white-
footed mice (Peromyscus leucopus) of eastern North Amer-
ican deciduous forests are complexly connected to white-
tailed deer (Odocoileus virginiana), gypsy moth (Lymantria
dispar), Lyme disease (Borrelia burgdorferi), and oak trees
(Quercus).Keesing (1998b) has shown how the small mam-
mal community in African savannahs interacts strongly with
the large herbivores. In an experimental investigation of
the effects of sheep grazing in Norwegian mountain pas-
tures, Steen et al. (2005) demonstrate that sheep strongly
depressed populations of Microtus agrestis,but had no ap-
parenteffectonClethrionomys glareolus. Rodents also have
been implicated as major dispersers of mycorrhizal spores
in coniferous forests (Maser et al. 1978). Many other ex-
amples of rodents being complexly connected within their
communities will undoubtedly be reported as investigators
focus more on this level of ecological analysis.
A more recent, but extremely important, insight has
been the realization that not only is the community context
critical for conservation action, but landscape variables
may also be important determinants of a species’ well-being
(Forman and Godron 1986; Forman 1995; Lidicker 1995;
Barrett and Peles 1999). Recently (Lidicker 2000), I argued
that the interaction between the ratio of optimal to mar-
ginal habitat (ROMPA) in a landscape combined with the
presence or absence of generalist and /or specialist predators
can determine the kind of demographic pattern shown by
species ofMicrotus. In general, the presence of various
predators will depend on the sizes of the habitat patches
and their proximity to other community types in which the
predators may reside. Moreover, spillover predationmay
impact species living in habitats insufficiently productive
or too small to support resident predators (Oksanen and
Schneider 1995; Lidicker 1999). Different combinations
of contiguous community types can lead to quite different
population dynamics in species resident in only one of the
patches (Danielson 1991; Lidicker 1995, 1999). For ex-
ample, a habitat patch suitable for some rodent species that
is surrounded by a matrix that absorbs emigrants from the
patch but does not supply immigrants in return (dispersal
sink; Lidicker 1975), will result in a chronically depressed
rodent population in the isolated patch (demographic sink;
Pulliam 1988, 1996). Conservation planning must also ac-
knowledge that migratory species will require preservation
of both winter and summer habitats. Even nonmigratory
species may benefit from or even require the juxtaposition
of two or more habitat types (Lidicker et al. 1992; Shell-
hammer 1998).
Another landscape feature receiving increased attention
is the nature of the ecotone (edge) that develops between
two community types (fig. 38.5), and the responses of or-
ganisms to this zone (Wiens 1992; Lidicker and Koenig
1996; Lidicker 1999; Lidicker and Peterson 1999; Wolff
1999, 2003). Sometimes ecotones have positive impacts on
biodiversity. In other situations they may serve as conduits
for the invasion of predators, competitors, and parasites
that are deleterious to the original community members.
Ecotones also can strongly decrease the effective size of the
habitat patches if edge effects penetrate deeply into the ad-
jacent communities, as they often do (Laurance, Bierregard
et al. 1997, Laurance, Laurance et al. 1997). As pointed out
earlier, it is important to know how target species respond
to ecotones, as this may significantly affect their success in
the landscape. Some species are ecotonalphobic, and hence
are unlikely to cross gaps in their habitat to find new suit-
able patches. They also will not readily be able to find cor-
ridors that could allow them to emigrate from their home
habitat fragment (Crome et al. 1994; Lidicker and Koenig
1996). Consequently, both colonization of empty patches
and genetic exchange among patches will be rare.
Corridors are also landscape features that can facilitate
connectedness among habitat fragments (Forman and Go-
dron 1986; Lidicker and Koenig 1996; Lidicker 1999), and
therefore may be critical assets in metapopulation or meta-
community persistence (fig. 38.6). Negative effects can also
be associated with corridors (Simberloff and Cox 1987).
Corridors can simply increase the amount of edge habitat
available to exotic invaders and thereby negatively impact
the habitat patches (Crooks and Soulé 1999; Hawkins et al.
1999; Panetta and Hopkins 1991). Or, corridors can facil-460 Chapter Thirty-Eight