Lindeman (1942). Fig. 1.1c shows the quantification
of the patterns of energy flow in a salt marsh in
North America among the major functional compo-
nents (Teal 1962). Owing to measurement con-
straints and the assumed functional equivalence of
different species involved with respect to nutrient
and energy transformation, these networks typical-
ly have low taxonomic resolution, and have mostly
been developed as a way to understand and
describe energy fluxes and material in ecosystems,
and the associated ecosystem functions and ser-
vices (such as primary and secondary productivity,
carbon sink capacity, etc.). That is, the focus is the
dynamics of energy and material, rather than
the dynamics of species and their interaction struc-
ture. An exception is the work by Ulanowicz (1997),
who is one of the rare authors who has extensively
considered origins and consequences of different
topological patterns of energy and nutrient flows
in food webs. In his work, he stresses the impor-
tance of positive feedback among species, resulting
in autocatalytic loops within subsystems of whole
webs, which may even lead to ecosystem-level
competition between such alternative loops. Also,
the work on soil foods webs by de Ruiter, Moore
and others (e.g. de Ruiteret al.1998) has tried to
map the flows of nutrients and energy from a much
more topological perspective than is done in classic
ecosystem science. The outcome of this work will be
addressed further in Chapter 2.
The second broad type of information repre-
sented by a link is theeffectof one entity on another.
For example, if individuals of two species each have
a positive effect on the growth and reproduction of
individuals of the other, they would be linked. This
positive–positive interaction is termed mutualism
and a range of biological mechanisms might under-
lie this link. Pollination is an example of a mecha-
nism that can result in a positive–positive
interaction. Associational resistance, in which a
species gains protection from its consumers by spa-
tially associating with its potential competitors, is
another example that can lead to a net positive
interaction between species (Olffet al.1999). Facili-
tation, such as often found among grazing herbi-
vores (Huisman and Olff 1998; Arsenault and
Owen-Smith 2002), also belongs to this class of in-
teractions. A negative–negative interaction is
termed competition, and can result from consump-
tion (a biological process) by two species of a shared
resource, as well as from a number of other non-
consumptive mechanisms, such as direct beha-
vioural interference among species (Schoener 1983).
When links representeffectsof one species on
another, they may result from direct effects, indirect
effects or net effects. Competition via consumption
of a shared resource can be represented as a
negative–negative link arising from an indirect
interaction between two competing species
mediated by their joint consumption of a third
resource species; that is, there need not be a transfer
of anything tangible from one competitor to the
other, yet changes in the abundance of one will
alter the abundance of the other. Other indirect
effects (or interactions) include apparent competi-
tion (negative–negative effects caused by the pres-
ence of a shared consumer; e.g. Holt 1977) and
apparent mutualism (positive–positive effects, e.g.
in a trophic cascade; Paceet al.1999).
Ecological networks in which links represent the
net effect of one species on another (the outcome of
all direct and indirect effects) are often, and per-
haps confusingly, termedinteraction webs. As this
also describes the general class of ecological net-
works in which species interact (irrespective of the
nature of their interactions) they should be more
correctly calledinteraction-sign networks. They de-
pict the sign (positive, negative, zero) and some-
times the magnitude of the net impact of changes
in the abundance of one species on another. These
net impacts can be the result of all kinds of direct
and indirect interspecific interactions. Interaction
webs are less commonly depicted in the literature,
perhaps because of the practical difficulty of
determining the sign of net interactions, especially
within trophic levels, for which elaborate field ex-
periments are often required. However, sometimes
the nature of interactions can be inferred from
large-scale natural disturbances to ecosystems. A
classic example is the effect of rinderpest on the
food web structure and ecosystem functioning of
the Serengeti (Sinclair 1979). In this work, abiotic
components and processes, such as fire, are also
taken into account as ‘nodes’ in the food web, be-
cause of the strong effectbioticfeedbacks that they
receive. One good example of an interaction-sign10 SHAPE AND STRUCTURE