Access keys

Skip to content Accessibility Home News, events and publications Site map Search Privacy policy Help Contact us Terms of use

Untangling nature’s secrets

Untangling nature’s secrets - 15 August 2012. PEaCE Lab

Ecologists studying food webs reveal the importance, and fragility, of pollination networks.

15 August 2012

Charles Darwin once described food webs as 'entangled banks' (ref 1) to convey the complex network of species and breadth of interactions in an ecosystem.

More than 150 years later, ecologists funded by BBSRC are teasing apart the links between living organisms to find out just what makes the system tick. It's a massive task, but focusing on farmland pollinating insects is revealing just how fragile some communities are to changes in their environment.

Video to accompany the feature 'Unravelling nature's food webs'  

You need to have JavaScript enabled to view this video.

Video transcript: Untangling nature’s food webs - Video and audio help - Watch video on YouTube

"Food webs provide a means of studying the interactions between species," says Professor Jane Memmott of the University of Bristol. "The ways that species interact can have a profound impact on a community's response to species loss, stress and ecological restoration."

In fact, the researchers are as interested in how food webs respond to stress as they are in the natural stability of the network, so it's work that can reveal how rare or important species might react to conservation or agro-environment schemes, and how pollinating insects fit into the agricultural food webs that are critical for ensuring food security.

Web heads

Jane Memmott has been studying food webs – a feeding network of who eats who within an ecosystem – for more than 25 years. Much of the research by others in that time was somewhat theoretical, as biologists accumulated data and used mathematical models and computer simulations to look for large scale underlying patterns, both within and between different food webs. For example, similar patterns in network architecture have been found between very different types of habitat, such as between salty marine ecosystems and the driest deserts (ref 2), and that body size not only has a notable effect on who eats who – unsurprisingly, bigger fish tend to eat smaller fish – but this then partly determines wider network structure (ref 2, ref 3).

Food webs: the yellow dots represent groups of species; green lines show species linked by predation. Image: PEaCE Lab

Food webs: the yellow dots represent groups of species; green lines show species linked by predation. Image: PEaCE Lab

But after decades of data accumulation and mathematical modelling in purely natural habitats, some scientists began to ask if food web research was too theoretical and less relevant to the problems facing ecosystems already under stress from human activities. Memmott summed up the quandary ably with the question – Food webs: a ladder for picking strawberries or a practical tool for practical problems? (ref 4)

In the paper, Memmott outlined areas where food webs have been used as an applied tool, such as restoration ecology, managing alien species, biological control, conservation ecology and habitat management. She then defined areas in which the food web approach could prove rewarding, including urban ecology, agroecology, studying habitat fragmentation and, importantly, ecosystem services such as pollination.

Flowers for food security

Insects are an essential part of ecosystems and insect pollination has been estimated to value EUR158Bn per year (ref 5), so maintaining pollinator biodiversity is essential for food security. Many crops, such as apples, beans, raspberries and tomatoes, are entirely reliant on pollinating insects.

Memmott says if we are to conserve and manage pollination, we need to understand the structure of these networks. "They will tell us how resilient the community is likely to be in the face of environmental change. For example, if honey bees disappear or if climate change leads to changes in flowering times of plants."

Memmott and colleagues are one of the first groups to take a close look at pollination networks in relation to other food webs. "The driver here was the clear reality that if you stand in a wood, a heathland, a farm or a rain forest, all the different networks are clearly linked," Memmott explains. "For example if you don't have a pollination network there won't be any seeds for a seed dispersal network."

Bees, such as this bumble bee, as essential pollinators of crop plants. Image: D. Harding

Bees, such as this bumble bee, as essential pollinators of crop plants. Image: D. Harding

But another reality is that the way biologists do their research makes them look like separate webs: pollination biologists make pollination networks, ornithologists make seed dispersal networks and entomologists make parasitoid networks.

So Memmott's team took on the huge task of looking at multiple food webs across one farm. But the benefits would be clear: data on whether perturbations and patterns in one web would affect another. Can species extinctions, climate change or habitat stress that impact on one part of the network cascade across the whole web?

"A loss of bumblebees is likely to affect the pollination of the weeds that feed seed-feeding [granivorous] birds. This then affects the birds of prey that feed on granivorous birds," says Memmott. "The entomologist studying the bees and the ornithologist studying the birds would be very unlikely to see these links."

Small worlds

Over two years, the BBSRC-funded researchers looked at 1501 unique interactions between a total of 560 organisms on the 125-hectare Norwood Farm in Somerset. They constructed different webs, such as between aphids and their parasitoids; seeds, rodents and ectoparasites, and between flowers and butterflies for example. They then simulated extinctions in ecosystems by running more than 20,000 computer simulations that removed species from the web to see whether it would collapse or remain stable (ref 6).

The scientists found that different food web networks varied in their robustness to species loss, and networks including pollinators appeared to be particularly fragile. This ran against the researchers' initial presumptions. "Our expectation was actually that these [pollination] networks would be rather robust to species loss as pollinators are not very specialised," says Memmott.

Why were the pollinator networks so fragile? Memmott notes that Norwood Farm had been organic for a long time, theoretically good for pollinating insects such as bees and butterflies, and was species rich in comparison to an intensively managed farm, but flowers were not actually that common. In addition, the pastures are eaten by livestock which reduces the number of flowers, the field margins are full of grasses rather than flowering plants, and habitats such as the woodlands on the farm are only good for pollinators for a short period in the spring.

A team of ecologists sampled pollinator networks for more than two years. Image: J. Memmott

A team of ecologists sampled pollinator networks for more than two years. Image: J. Memmott

The researchers also found that some plants such as thistles, cow-parsley, clover and buttercups were disproportionately well linked to animals through the food web. In fact, the research showed that overall the food web networks did not strongly co-vary in their robustness – some groups of animals were more sensitive to the loss of plants than others, and what was bad for one group of animals was not necessarily bad for others.

The findings, co funded by Defra, have clear practical benefits and policy implications. Ecological restoration programmes and techniques (ref 7), through agri-environment schemes for example, which benefit one group of organisms will not inevitably benefit others. In addition, because some individual plant species were disproportionately well linked to many other species, this type of information can be used in restoration management because it identifies the plant groups that can potentially lead to above average gains in biodiversity.

From farms to urban habitats

Although the study only took in one farm, previous research by Memmott's group has compared organic and conventional farming across multiple sites. This too, showed counter-intuitive findings, namely that the provision of natural pest control was similar between the two farming methods, even though biodiversity of natural predators was higher on the organic farms, as expected (ref 8, ref 9).

Following farms and fields, Memmott is now setting her sights on the city. Her current big project concerns the ecology and conservation of pollinators in urban habitats and is funded under the BBSRC Insect Pollinators Initiative (IPI).

Memmott says that although urban webs sound very different to farm webs, the sampling is very similar. "But instead of divvying up a farm into hedgerows, crops, woodland and field margins, we are divvying cities up into parks, allotments, cemeteries and car parks!" she says. "We want to find out how urban habitats compare to nature reserves and farmland in terms of biodiversity and network structure; where are the hotspots of pollinators in cities; and what can we do to conserve urban pollinators."

The Insect Pollinators Initiative

The Insect Pollinators Initiative (IPI) is a £9.6M, 4-year programme joint-funded by BBSRC, Defra, the Natural Environment Research Council (NERC), the Scottish Government, the Wellcome Trust and is managed under the auspices of the Living with Environmental Change (LWEC) partnership.


  1. Darwin Online (external link)
  2. Simple rules yield complex food webs (external link)
  3. Phylogeny versus body size as determinants of food web structure (external link)
  4. Food webs: a ladder for picking strawberries or a practical tool for practical problems? (external link)
  5. Economic valuation of the vulnerability of world agriculture confronted with pollinator decline
  6. The robustness and restoration of a network of ecological networks
  7. New tools to boost butterfly habitat quality in existing grass buffer strips (external link)
  8. Do differences in food web structure between organic and conventional farms affect the ecosystem service of pest control? (external link)
  9. Parasitoid diversity reduces the variability in pest control services across time on farms (external link)

Tags: feature


You may need to download additional plug-ins to open