woodland area and isolation. Overall, they found a
mix of area-sensitivity, isolation-sensitivity, and
compensatory area-isolation effects. But they also
reported significant variation in the form of the
incidence functions, which they interpreted as most
likely being the consequence of differences between
the three landscapes in terms of the habitat matrix
in which the woodlands were embedded (cf.
Ricketts 2001). Watson et al.’s study illustrates that
area–incidence functions cannot be assumed to be
consistent features within the range of a species.
In conclusion, incidence functions can be useful
tools, particularly if repeat surveys are available to
establish temporal variability in their form, but
they do not of themselves provide sufficient data
on which to base conservation policies. Notions of
threshold population sizes and area requirements,
as ever in island ecology, must be offset against dis-
persal efficacy, habitat requirements, and specific
factors influencing mortality of particular species.
Watson et al.’s (2005) findings also suggest that the
‘island–sea’ analogy is at fault, and that we should
be paying greater attention to the ecology of the
matrix (below).
10.4 Metapopulation dynamics
Most species are patchily distributed. Many
occupy geographically separated patches that are
interconnected by occasional movements of indi-
viduals and gametes, constituting a metapopula-
tion. The first metapopulation models were
constructed by Richard Levins in papers published
in 1969 and 1970 (Gotelli 1991). The basic idea can
be understood as follows. Imagine that you have a
collection of populations, each existing on patches
of suitable habitat. Each patch is separated from
other nearby habitat patches by unsuitable terrain.
Although these separate populations each have
their own fairly independent dynamics, as soon as
one crashes to a low level, or indeed disappears,
that patch will provide relatively uncontested
space for individuals from one of the nearby
patches, which will soon colonize. Thus, according
to Harrison et al. (1988), within a metapopulation,
member populations may change in size independ-
ently but their probabilities of existing at a given
time are not independent of one another, being
linked by mutual recolonization following periodic
extinctions, on time scales of the order of 10–100
generations.
Studies of the checkerspot butterfly (Euphydryas
editha bayensis) in the Jasper Ridge Preserve (USA)
provide one of the better empirical illustrations of
metapopulation dynamics (Harrison et al. 1988;
Ehrlich and Hanski 2004). The checkerspot butter-
fly is dependent on food plants found in serpenti-
nite grasslands. The study area is of 15 30 km and
includes one large patch (2000 ha) and 60 small
patches of suitable habitat. The large ‘mainland’
patch supports hundreds of thousands of adults
and is effectively a permanent population. The
smaller populations are subject to extinctions, prin-
cipally due to fluctuations in weather, but patch
occupancy may also be influenced by habitat qual-
ity. A severe drought in 1975–1977 is known to have
caused extinctions from three of the patches,
including the second largest patch, and so was
assumed to have eliminated all but the largest pop-
ulation. Unfortunately, only partial survey data
were available from the period and this assumption
cannot be tested. By 1987, eight patches had been
recolonized. Small patches over 4.5 km from the
‘mainland’ patch were found to be unoccupied.
Harrisonet al. (1988) showed that the distribution
of populations described an apparent ‘threshold’
relationship both to habitat quality and distance.
Patches had to be both good enough habitat and
near enough to the ‘mainland’ in order to be
inhabited at that time. A similar relationship, but
from a separate study, is shown in Fig. 10.3 (and see
Ehrlich and Hanski 2004).
From models of the dispersal behaviour of the
populations, assuming a post-1977 recolonization of
the patches, it was predicted that more patches
would become occupied in time. How many
depends on the assumptions of the model. They
offered two differing scenarios. The first assumed
that colonization continues until reversed by the
next severe drought, an event with an approximately
50-year periodicity in the study area. The second
assumed that some extinction events occur between
such severe events, thereby requiring a continuous
extinction model and producing an equilibriumMETAPOPULATION DYNAMICS 259