5.2.1 Neutral
Although Hubbell’s (2001) construct is the most
widely recognized neutral model, there are actually
many related ‘neutral’ theoretical constructs that
make no specific assumptions about species traits
or their responses to the environment (reviewed in
Chave 2004). The neutral model assumes that spe-
cies are neutral with respect to their interspecific
interactions as well as the underlying environment.
This means that the numbers of individuals and
species that occur in any given locality result from
purely stochastic processes (e.g. colonization and
extinction). MacArthur and Wilson’s (1967) Equi-
librium Theory of Island Biogeography is probably
the most widely recognized neutral model, and
similar concepts form the basis of Hubbell’s (2001)
neutral theory. In the MacArthur and Wilson theo-
ry, the number of species in a habitat is solely
determined by the balance between the coloniza-
tion rate of species from a species pool (usually
mainland) and the local extinction rates of species.
Hubbell’s (2001) model expands on the MacArthur
and Wilson theory in two important ways. First, the
stochastic processes of colonization and extinction
at the patch level from the MacArthur and Wilson
theory are transferred to the individual level so that
they denote birth–death processes, allowing it to
make predictions of species’ relative abundances.
Second, rather than defining a species pool (e.g.
from a mainland), Hubbell allows the species pool
to result from speciation, which is also derived
from stochastic processes. While specific details of
the neutral model and its assumptions have been
heavily debated (e.g. Ricklefs 2003; Etienneet al.
2007), neither Hubbell (2001) nor others who have
developed neutral theories (e.g. Chave 2004) truly
believe that the assumptions of the neutral model
can fully capture the true nature of these commu-
nities. Instead, the neutral models emphasize how
far one can go by making the simplest assumption
that communities are structured by stochastic pro-
cesses only.
5.2.2 Patch dynamics
This perspective is also borne out of stochastic pro-
cesses of colonization and extinction, but in this
case at the patch level. In Levins’ (1969) original
metapopulation model, patches are assumed to be
similar, and species can colonize and go extinct
from a patch at defined rates. However, patch dy-
namics models of metacommunities go beyond the
neutral model by assuming that species may dis-
play trade-offs in their relative abilities to colonize
and compete, in turn allowing coexistence under
some circumstances. There are many ways to depict
these sorts of colonization–competition and other
trade-offs, each of which gives slightly different
sets of predictions (Levins and Culver 1971; Hast-
ings 1980; Tilman 1994; Hanski and Gyllenberg
1997; Yu and Wilson 2001; Calcagnoet al. 2006).5.2.3 Species sorting
This framework leaves behind the stochastic pro-
cesses of colonization and extinction inherent in the
two frameworks described above, and instead fo-
cuses on deterministic processes that result from
differential responses of species to heterogeneous
environments; i.e. niche differences. Here, a species
will occur in a locality if the abiotic and biotic
environments are favourable. Trade-offs in toler-
ance to different environmental conditions are typ-
ically invoked here to suggest that species that are
favoured under some environmental conditions
will be disfavoured under others (Tilman 1982;
Chase and Leibold 2003).5.2.4 Mass effects
This framework combines aspects of patch
dynamics and species sorting by assuming that
species have differential responses to environ-
mental conditions and differential dispersal,
colonization and extinction rates (e.g. Mouquet
and Loreau 2003; Amarasekareet al.2004).Dis-
persal is assumed to be high enough to influence
local dynamics. The coexistence and relative abun-
dances of the species will depend on rates of dis-
persal and extinction, as well as on the source–sink
relationship between the different habitats for
each species. The existence of source patches is
crucial for a species to persist in a dispersal-driven
metacommunity.60 SPACE AND TIME