Food Web Ecology: Playing Jenga and Beyond

See allHide authors and affiliations

Science  01 Jul 2005:
Vol. 309, Issue 5731, pp. 68-71
DOI: 10.1126/science.1096112

Naturalists have long noted that the distribution, abundance, and behavior of organisms are influenced by interactions with other species (1). Motivated in part by Paine's (2) work in the rocky intertidal zone and May's (3) theoretical work on the relationship between the complexity and the stability of ecosystems, the study of food webs gained momentum in the late 1970s and early 1980s (4). These studies precipitated a convergence of different approaches—mathematical treatments, descriptive work, manipulative field studies, and a formal treatment of energy flow and matter. This in turn allowed mapping of the interrelationships among the structure of an ecological community, its stability, and the processes occurring within the ecosystem—that is, construction of a food web (5).

Over the past decade, new issues arising in ecology, such as environmental change, spatial ecology, and the functional implications of biodiversity, require a different view of ecosystems and ecological research (6). The food web approach, with its focus on static structure and reliance on stability or persistence of species, seemed ill-equipped for analyzing these more dynamic topics.

Indeed, the often-used metaphor for the relationship among species, community structure, and stability was that of a stone arch with the loading forces among stones (species) representing interactions among species, and the “keystone” representing the species that had the dominant role in regulating structure and stability of the community. But many ecologists now view such a static representation of biological communities as inappropriate. Moreover, food web descriptions have been criticized for incompleteness because they do not fully account for all the species and links that are present, and because they generally ignore spatial and temporal variability. For these reasons, food web approaches have rarely been applied to current environmental issues.

The metaphor of the static arch might better be replaced with the metaphor of the structures built during a variation of the game Jenga (see the first figure, caption). Simple rules of balance and energetics govern the stability of both arch and Jenga structures, but unlike an arch, a Jenga structure is constantly changing, with additions and deletions of stones, and its stability at any moment depends on the importance of a given ingoing or outgoing stone's contribution to the structure. By realizing that dynamics are key to understanding complex structures, we can see stable food webs not as static entities, but as open and flexible Jenga-like systems that can change in species attributes, composition, and dynamics. Recent food web studies have incorporated data on spatial and temporal dynamics. Here, we highlight a few examples of such studies and discuss their implications for environmental management.


In a game of Jenga, players successively take away parts and place them on top until the structure becomes unstable and crashes. Each part can thus be a keystone. When parts are replaced at other positions, the stability of the Jenga structure can be maintained.

Over time, food webs change in species composition and in population life history parameters and abundances, and individual organisms within the web change in growth, size, and behavior. Dynamic relationships among different levels of the biological hierarchy govern food web structure and stability. Field observations and theoretical models show that environmental heterogeneity creates subsystems (compartments) of interacting species within food webs, especially at the lower trophic levels (7, 8). Organisms at the higher trophic levels act as integrators, linking the lower pathways in space and time, and stabilize the dynamics of their resources (prey) via density-dependent foraging (9, 10). This relationship explicitly supports MacArthur's idea (1) that community complexity buffers against perturbations and thereby overrides the inherent constraints on system stability imposed by complexity (3). Remarkably, this mechanism is similar to the way in which dynamics at the level of individual behavior influence food web structure and stability. For example, a predator switching to new prey affects population dynamics, because such dietary shifts inhibit rapid population growth of abundant prey while allowing rare prey to increase (11). Such shifts can be rapid; hence, when food web architecture changes (by changes in species composition or through fluctuating population abundances), web structure may quickly stabilize and may even result in a positive complexity-stability relationship.

The food web of Tuesday Lake, 1984.

The width of the horizontal bars shows the body mass (log10 kg), number (log10 individuals per m3), and biomass (log10 kg/m3), respectively, of each species. The vertical positions of the species show trophic height (20). Despite a major change in species composition, following a manipulation, this energetic setup of the food web remained roughly the same (19).

Life history processes in structured populations also can influence community dynamics in extraordinary and even counterintuitive ways. Size-structured food web models, in which growth in body size is density-dependent, predict that size-selective predators decrease the overall biomass of the prey as expected, but change the prey size distribution. This alters competition among prey individuals to such an extent that it actually leads to an increase in the abundance of the preferred prey stages. This positive feedback allows for large populations of predators to persist under conditions where small populations are likely to become extinct (12).

The trophic position of species in dynamic food webs may influence the risk of loss of that species, with possible consequences for ecosystem functioning (13). Experiments on pond food webs show that the effects of species on ecosystem processes depend on the interplay between environmental factors (such as productivity) and trophic position, whereby species at higher trophic levels tend to have larger effects (14). The risk of a given species' extinction and its consequences (in terms of secondary extinctions and ecosystem functioning) will be different in different ecosystems and will vary within ecosystems over space and time. Hence, a keystone species in one setting may have relatively little influence on community dynamics in another setting.

What does this research tell us? First, the performance and persistence of species and the role of species in biological communities should be examined in the context of a dynamic food web. The studies also show how to go beyond playing Jenga by revealing the general rules of demography and energetics that tend to stabilize biological communities (15, 16). The stabilizing effect of flexible substructures imposed by environmental change and variability (9) echoes the notion of stabilizing trophic pyramids (17) occurring in food webs and food web compartments (10). When compartments are viewed as trophic interaction loops, we can even understand mathematically the stabilizing effect of pyramidal structures within loops on food web structure (18).

Dynamics and stability can be seen in the intensively studied food web of Tuesday Lake in Michigan, USA (see the second figure). Removal of three planktivorous fish species and addition of one piscivorous fish species changed the lake's community structure remarkably. Although this manipulation had almost no effect on species richness (56 in 1984, 57 in 1986), about 50% of the species were replaced by new incoming species within less than 2 years (19). The energetic setup of the food web, in terms of distributions of body sizes and abundances over trophic levels, unexpectedly remained roughly the same. Apparently, the new web structure allowed the community to conserve key ecological features in the face of a major disturbance.

The notion of the ecosystem as a static arch has restricted our vision. In contrast, viewing food webs as open and flexible Jenga-like structures that accommodate changes in species composition, attributes, and dynamics reveals the features of the ecosystem that are critical to our understanding of community resistance and resilience to environmental change and disturbance. Recent theoretical advances in food web research must be accompanied by rigorous experiments and detailed empirical studies of food web modules in a variety of ecosystems. As food web science continues to develop, it surely will contribute new tools and new perspectives for the management of both natural and human-affected ecosystems.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article