scarce (McClanahanet al. 1996; see also Hay and
Taylor 1985). That is, the presence of two effective
herbivore groups (fishes and urchins) buffers the
system against algal overgrowth if one group is lost.
While these examples involve community (or,
more accurately, guild) stability at the expense of
population fluctuations, experiments show that in-
teractions within diverse natural assemblages may
also foster stability at the population level by
imposing density dependence in demographic
rates (Hixon and Carr 1997; Carret al. 2002). Field
experiments showed that per capita mortality rates
of recruiting coral-reef fishes in the Bahamas were
strongly density dependent in the presence of mul-
tiple predators (Hixon and Carr 1997) or predators
and competitors together (Carret al. 2002), whereas
mortality was independent of density when com-
petitors alone, or only one type of predator, was
present. These experiments provide an intriguing
contrast with the pattern found in a meta-analysis
of predator–prey experiments, in which the pres-
ence of predators tended to destabilize temporal
dynamics of their prey (Halpernet al. 2005). It
would be quite interesting to know whether the
destabilizing effects of predators in the meta-analy-
sis might be an artefact of experiments including
only one or a few species of predators.
8.4.2.2 Comparisons across space
Several prominent spatial patterns are consistent
with the hypothesis that diversity enhances stabili-
ty in marine systems. Across geographic regions,
the least stable ecosystems are those low in fish
diversity, with a few strong predator–prey links
and little capacity for prey switching (Jennings
and Kaiser 1998). For example, the dramatic cas-
cades emanating from sea otters to kelp in the rela-
tively low-diversity community of Alaska were not
observed in the more diverse kelp beds of southern
California (Daytonet al. 1998), even though both
regions lost sea otters by the early 19th century. A
possible explanation is that, in southern California,
spiny lobsters and sheephead, which also feed on
urchins, compensated for the reduced impacts of
sea otter predation; indeed, when lobsters and
sheephead were heavily exploited in the 1950s,
kelps did decline in southern California (Dayton
et al. 1998). Evidence consistent with such an expla-
nation again comes from a coral reef in Kenya,
where experimental reductions of sea urchins al-
lowed algae to proliferate to about twice the level
on fished reefs as on unfished reefs (McClanahan
et al. 1996).
A stabilizing role of biodiversity in exploited ma-
rine communities is also suggested by regional
comparisons of trophic dynamics in northwest At-
lantic continental shelf ecosystems (Frank et al.
2006). These authors analysed time-series data for
several trophic levels, from phytoplankton to har-
vested fishes, at nine heavily fished sites to assess
the strength and direction of trophic control. Corre-
lations between adjacent trophic levels varied
among sites, being predominantly negative at
higher latitudes, indicating top-down control, but
positive at lower latitudes, indicating bottom-up
control. Interestingly, the strength and sign of tro-
phic control varied systematically with species rich-
ness, with stronger top-down control in the more
depauperate northern sites and bottom-up control
at the more diverse southern sites. This weakening
of top-down control with prey species richness is
consistent with results of experiments (Steiner 2001;
Hillebrand and Cardinale 2004; Duffyet al. 2005).
However, the link to diversity in the northwest
Atlantic data is confounded by strong correlations
of species richness with latitude and temperature,
which are also expected to influence the strength of
top-down control through effects on demographic
rates (Franket al. 2006). Disentangling these influ-
ences is not possible at present.
Finally, Wormet al. (2006) conducted a compre-
hensive analysis of the links between marine biodi-
versity and response to fishing (Fig. 8.6). They
analysed relationships between species richness
and fishery production for the world’s 64 Large
Marine Ecosystems (www.fishbase.org). Regions
with naturally low fish diversity supported lower
average fishery productivity, and had more fre-
quent ‘collapses’ (strong reductions in fishery
yield) and lower resilience (degree of recovery
after overfishing) than naturally species-rich sys-
tems. Wormet al. (2006) suggested that the greater
resilience of more diverse ecosystems may be ex-
plained by the greater ability of fishers to switch
among target species in diverse ecosystems; when
abundance of a species declines to a low level, it isSTRUCTURE AND FUNCTIONING OF EMERGING MARINE COMMUNITIES 107