Bigger ants can get along peacefully with smaller species but there is likely to be fierce competition when similar-sized species vie for resources.

A new study carried out by ecologists at Imperial College London suggests that fighting is most likely to break out where there is a narrow size difference between the ants in two colonies. This is probably because they are competing for the same resources, such as the same kinds of food, or the same places to nest. Ants of different sizes, in contrast, are more likely to be able to co-exist by using different resources. The competition between species of similar sizes scales up to affect the structure of the whole ant community. The results are published today in Ecology Letters.

The question of how species organise themselves into communities is an important one for ecologists, but it is not easy to test because so many different factors - not all of which can be observed - can influence the result.

The Imperial researchers worked in partnership with ecologists from the University of Cambridge, the Natural History Museum and Universiti Malaysia Sabah. They studied colonies of tropical ants living in ferns high up in the rainforest canopy in Borneo to try to find realistic ways of simulating how ant communities organise themselves.

Using a mixture of field observation, laboratory experiments and computer modelling, the team showed that very similar-sized ants tended not to be found together. In fact, where ants of similar sizes were introduced to each other during the laboratory experiments, the resident species would become aggressive - even attempting to push the non-resident ants off the edge of the ferns.

Software developed by the team enabled them to further analyse the ants' behaviour to predict how similar in size ants needed to be before they started to compete. They used this information to model a series of rules for ant community assembly. Finally, they used the data they had collected to set up a computer-simulated rainforest to see if, by applying these rules, they were able to generate a realistic rainforest ant community.

Lead researcher Tom Fayle is a researcher in Imperial's Department of Life Sciences, and affiliated to the Biology Centre of the Czech Academy of Sciences. "It's quite difficult to show competition among species in the wild, partly because many species are so mobile," he explains. "Ants are particularly useful for this because ant colonies tend to stay in the same place and so it's easier to study the community as a whole.

"The methods that we developed - using experimental observations to set rules for computer simulations - could be useful in many different areas of ecology and conservation. Our experiments showed how ants form communities and that ants of similar size find it difficult to co-habit. This may be one explanation for the extraordinarily high diversity of ants in tropical rainforest canopy: there are more ant species living in this one kind of fern than there are in the whole of the UK."

Dr Paul Eggleton, a researcher at the Natural History Museum, said: "Ants are fantastic groups of animals to do this sort of experimental research with as they are relatively easy to identify, and their social structures mean that they congregate into nests which can be moved around. This has allowed us to undertake some ground-breaking work on how ant communities are structured and how they evolve."

Further research by the team will look at species interaction in communities which have been disturbed by logging and conversion to oil palm plantation to see whether the same rules of interaction and competition will apply. The methods used by the team could also be used to look at other patterns of species interaction, for example in plant communities.