maintain an effective metapopulation structure (Murphy and Noon 1992; Noon
and McKelvey 1996). This kind of management controversy will only become more
common with further fragmentation of existing wild lands.Special kinds of metapopulation dynamics occur when some patches are large
enough or productive enough to sustain permanent subpopulations, whereas other
patches are small enough or unproductive enough that local extinction is common.
If both the permanent patches and the transient patches can support positive
population growth, such an arrangement is termed a mainland–island system.
Examples include the checkerspot butterfly (Euphydryas editha bayensis), which
inhabits scattered patches of serpentine soil in coastal California (Harrison et al. 1988)
and spiders in the Bahama Islands (Schoener and Spiller 1987).
In other cases, only a fraction of patches can sustain positive subpopulation
growth, whereas individuals in other patches always experience higher rates of mor-
tality than birth. Such an arrangement is referred to as a source–sink system, with
source sites supplying a steady stream of dispersers that fan out to surrounding sinks
(Pulliam 1988; Pulliam and Danielson 1991). Despite the fact that sinks are incapable
of supporting viable local populations, through immigration from source patches they
can have substantial numbers of individuals. Beavers (Castor canadensis) inhabiting
shallow lakes in the mixed deciduous and boreal forest of southern Ontario provide
a good example of a mammalian species with source–sink dynamics (Fryxell 2001).
Beavers at a small fraction of colonies have sufficient food supplies to support sub-
stantial production of offspring year after year. These populate the surrounding area
when they disperse. Most of the other colonies rarely produce viable young.
Clearly, the conservation needs of mainland–island and source–sink systems differ
from those of classic metapopulations. Mainland or source sites take on dispropor-
tionate importance in sustaining viable populations over the larger landscape. Loss of
even small amounts of these critical source or mainland habitats could be unsustainable.The distribution is the area occupied by a population or species, the dispersion is
the pattern of spacing of the animals within it, and dispersal, migration, and local
movement are the actions that modify dispersion and distribution. Dispersion and
distribution are states; dispersal, migration, and local movement are processes. The
edge of the distribution is that point at which, on average, an individual just fails to
replace itself in the next generation. Its position may be set by climate, substrate,
food supply, habitat, predators, or pathogens. The limiting factor can often be
identified by the trend in density from the range boundary inward.
Dispersal plays a key role in dictating the rate of spread of a species reintroduced
into a new area or one recovering from catastrophic decline. Diffusion models are
often an effective means of modeling the spread of reintroduced species, particularly
if they incorporate both demographic and random walk processes. We demonstrate
the logical basis for the simplest random walk and diffusion models. Dispersal is also
integral to the dynamics of organisms occupying spatially subdivided habitats form-
ing metapopulations. Simple models demonstrate that the long-term persistence of
metapopulations depends on the relative probabilities of extinction versus dispersive
colonization. There is some empirical evidence for regular turnover of colonies and
high rates of extinction and colonization. Variations on the metapopulation theme
include source–sink systems, island–mainland systems, and metapopulations with inter-
nal territory structure.108 Chapter 7
7.7.4Source–sink
and island–mainland
systems