Community Ecology Processes, Models, and Applications

5.3 Metacommunity theory: resolving MacArthur’s paradox

Each of the four metacommunity perspectives pro-
vides important insights, and they should not be
viewed as strict alternatives but rather as different
ways to view problems of coexistence and diversi-
ty. ‘MacArthur’s paradox’ (Schoener 1989) refers to
the fact that Robert MacArthur played a significant
role in the simultaneous development of metacom-
munity models based on niche differences (Mac-
Arthur and Levins 1964) and those based on more
neutral perspectives (MacArthur and Wilson 1967).
Here, we emphasize that, rather than viewing these
perspectives as dichotomous alternatives, they il-
lustrate different ends of a continuum, and that
each approach is valid for problems at different
scales and for different questions. For example,
when the focus is at a relatively coarse scale, it
may be an acceptable first approximation to assume
that all species are equal. This could explain the
reasonable success that simple ‘neutral’ models
like the MacArthur and Wilson theory and their
derivatives have had for addressing questions at
coarser scales (Simberloff 1974; Whittaker and Fer-
na ́ndez-Palacios 2007), despite the abstraction of
much ecological detail that obviously is important
for local coexistence. When the focus is local, or
when the questions being addressed are fine-
scaled, these more abstract models often need to
be expanded to focus specifically on niche differ-
ences among species and how those differences
influence patterns of community structure (e.g.
Chase and Leibold 2003).
Any theoretical model of metacommunities – be
it based on neutral, patch dynamic, species sorting
or mass effect/dispersal processes – can only be a
cartoon of the processes that are actually occurring
in that metacommunity. We argue that, although it
is instructive to examine patterns observed in natu-
ral systems or from experiments in the context of
these model predictions (e.g. Table 5.1), this should
not be done to test the validity of one model over
the other. Instead, we echo several recent syntheses
that have explicitly recognized that, in any meta-
community, a variety of processes, including those
related to neutral models (e.g. stochasticity) and
those related to niche models (e.g. determinism),

are operating simultaneously (Tilman 2004; Chase

2005, 2007; Gravelet al. 2006; Leibold and McPeek

2006; Adleret al. 2007; Clarket al. 2007). The chal-

lenge for metacommunity ecologists is to disentan-

gle the relative importance of the different

processes, and, most importantly, to identify fea-

tures of a given habitat type that would make one

or the other set of processes exert a stronger influ-

ence on community patterns.

5.4 As easy asa,b,g: the importance of scale

Why is it necessary to explicitly consider the role of

space in a metacommunity? If spatial processes are

not important, then the factors that influence the

number and relative abundance of coexisting species

at any given locality should simply ‘add up’ to de-

scribe those features at the regional scale. If, on the

other hand, there is something inherent in the way

that space differentially influences patterns of local

and regional diversity, then scale (and thus space)

matters (e.g. Whittakeret al. 2001; Chase 2003).

This problem is illustrated by explicitly consider-

ing how scale influences patterns of diversity. At

local (small) spatial scales, species diversity (¼spe-

cies richness) is the number of species counted in

the area defined; known asa-diversity. Because

there is generally variation in the types of species

(species composition) observed from one locality to

the next, the turnover of species composition

among sites is known asb-diversity. Finally, re-

gional diversity, known as g-diversity, can be

derived by either an additive (g¼aþb(e.g.

Lande 1996)) or multiplicative (g¼ab(Whittaker

1972)) partitioning. This emphasizes how under-

standing patterns ofb-diversity is critical to under-

standing the scalar relationship between local and

regional diversity, which is the domain of meta-

community ecology.b-Diversity can emerge from

both deterministic and stochastic processes. First, if

localities differ environmentally, highb-diversity

can emerge from deterministic processes that fa-

vour different species in different environments,

i.e. species-sorting processes (e.g. Whittaker 1972;

Chase and Leibold 2003). Second, even when local-

ities have identical environmental conditions, high

INCREASING SPATIO-TEMPORAL SCALES 61