constrained by dynamics continue to figure promi-
nently in our ways of attempting to understand
food web patterns. The various aspects of theory
point to the ways that topological patterns can
emerge from dynamics (see Chapter 1), and the
equally important ways that topology can influence
dynamics (see Chapters 1 and 3).
One of the continuing challenges in community
ecology is to link the dynamics predicted by models
of complex community features, such as food webs,
with the dynamics of actual systems. For a number
of pragmatic reasons, models are used to draw in-
ferences about the dynamics of complex systems
that differ in structure, often because the dynamics
of real systems are difficult to observe. This means
that the statics, or structure, of natural networks are
used to constrain the form of models, which in turn
are used to draw inferences about how the dynam-
ics of natural systems might vary as the details of
structure vary. To do this, it is usually assumed that
natural systems exhibit stable dynamics, and then,
given that constraint and others, it is possible to
estimate other network properties, such as the
strengths of interactions among species in the net-
work (de Ruiteret al. 1995; Neutelet al. 2002, 2007;
Mooreet al. 2003; Beckermanet al. 2006; Petchey
et al. 2008). In other cases, it is possible to estimate
how deletions of species from networks will result
in the extinction of other species that depend solely
on the deleted species for energy or nutrients (Sole ́
and Montoya 2001; Montoyaet al. 2006).
The interesting feature of such studies is that in
most cases the actual population dynamics of the
species in question remain unknown. Partly, this is
because dynamics are difficult to measure. Partly, it
can also be difficult to selectively remove one spe-
cies from a network and see how others respond.
Nonetheless, it is surprising that we still know the
detailed dynamics of only a very few food webs,
and these food webs are often relatively simple in
structure. This is a fundamental gap in our knowl-
edge of complex ecological networks.
14.1.5 Evolutionary community dynamics
Evolution obviously has the potential to modify in-
terspecific interactions, along the way influencing
food web topology, and driving large-scale diversity
patterns (Yoshidaet al. 2003; see Chapters 11 and 12).
Nonetheless, evolution remains poorly integrated
into community ecology (see Chapter 11). The other
important way that evolution may influence and
perhaps supersede the role of interactions in deter-
mining large-scale patterns of community structure,
such as regional diversity gradients, is by determin-
ing the diversity and composition of regional species
pools from which communities are assembled (Rick-
lefs 2004, 2008).
It is a source of some discomfort to community
ecologists that the causes of both local and global
diversity gradients remain uncertain. It is particular-
ly troubling that the most conspicuous large-scale
diversity pattern, the latitudinal gradient in species
diversity, has no generally accepted explanation, and
is instead the subject of continuing debate. In general,
it seems possible that both local and regional diversi-
ty patterns could be the result of (1) ongoing eco-
logical interactions, including priority effects,
(2) historical evolutionary processes, the conse-
quences of speciation and adaptive radiation, and
(3) purely neutral processes, including statistical
sampling processes such as the mid-domain effect.
At local scales, it seems that a number of factors
can conspire to produce a range of diversity pat-
terns. At least four different productivity–diversity
patterns have been observed in studies of different
systems (Waideet al. 1999), including ones that are
concave-up, concave-down, increasing, or essen-
tially flat. One explanation proposed for the differ-
ence between concave-down and increasing
patterns is the scale of study (Chase and Leibold
2002). Concave-down patterns appeared at a local
scale (among nearby ponds drawing on the same
potential species pool), while increasing patterns
appeared at larger spatial scales where the species
pool might be expected to increase with productiv-
ity. However, other processes can produce a range
of productivity diversity patterns, even when
drawing on exactly the same species pool. Fukami
and Morin (2003) showed that priority effects
related to the sequence of species assembly in com-
munities arrayed along a productivity gradient
could produce a variety of productivity–diversity
patterns, including hump-shaped (concave-down),
approximately linear increasing, and concave-up
patterns. Because these patterns resulted over the194 FUTURE DIRECTIONS