have been made regarding fish, their populations being supported by zoobenthos
thus allowing increased predation on zooplankton (Schindler & Scheuerell,
2002 ; Vander Zandenet al., 2005), and the communities of islands in the Sea of
Cortez, where subsidies from the marine ecosystem support a cascade through
the terrestrial community (Polis & Strong,1996). If these suggestions of external
subsidies fuelling increased predation are applicable to trophic cascades gener-
ally it is clear that the trophic cascade is nothing more than an indirect effect of
alternative prey.
It is clear also that longevity increases with body size (Brown, Allen & Gillooly,
this volume) and, as a consequence, the populations of large-body-size predators
high in the food web are damped in their response to fluctuations in the popula-
tions of small-body-size basal prey; dependent upon their susceptibility to
resource depletion individuals may be able to survive periods of reduced food
availabilityas they drivetheirprey’s populationstolowlevels, even in the absence
of alternative prey. Such longevity is reflected in cohort dynamics and periodic
release of prey populations when top predators pass through a reproductive cycle
(Persson & De Roos, this volume); evidence of five- to six-year cyclicity has been
found in fossil pigments in the sediments of lakes, suggesting that zooplankton,
and thus phytoplankton, are released from predation as the dominant cohort of
zooplanktivorous fish is replaced (Carpenter & Leavitt,1991 ). In comparison, the
generation time for phytoplankton may be as short as a few hours or days, with
zooplankton taking several days to complete a generation cycle.
As biomass abundance is the product of body size and numerical abundance,
high densities of large individuals in the higher trophic levels lead to a deviation
in the biomass distribution within the web from the typically assumed trian-
gular trophic pyramid to a more square or inverted triangle pattern. The distri-
bution of biomass within trophic levels has been linked to certain characteristics
of food-web structure (Neutel, Heesterbeek & de Ruiter, 2002 ). Complex, highly
connected webs where the majority of links are weak typically have a triangular
distribution, and may be more stable than those that deviate from this pattern
(Raffaelli,2002). It is appealing to suggest that trophic cascades are a con-
sequence of a certain pattern of biomass distribution (square or inverted tri-
angle). However, whilst the slope of the trophic pyramid may help to identify
communities where we would expect cascades to occur, this would not advance
our understanding of the mechanism that leads to cascade dynamics. It merely
leaves us with another question ‘Why does the distribution of biomass vary
among communities?’Productivity
The Oksanenet al.(1981) hypothesis regarding trophic cascades, where top-
down and bottom-up forces regulate community structure alternately dependent
on the number of links in the food web, is an over-simplification due to122 J. I. JONES AND E. JEPPESEN