No enemy escapes the weedkiller ants: A rainforest shrub enjoys the protection of some powerful little friends, says Malcolm Smith

It would be every gardener's dream. No more back- breaking hand weeding or hoeing. Just plant shrubs and watch them grow, the soil remaining weed-free naturally, year after year. A sort of biological cordon sanitaire surrounding each blossoming plant.

Gardeners may have to dream awhile yet, but, in nature, it seems that weed control around desired plants is a reality, courtesy of some incredibly dedicated ants.

In one of the most unusual symbiotic relationships yet discovered, ants living exclusively on Tococa shrubs in the Amazonian rainforests protect them from plant predators by killing off any other plant that appears nearby, by using a herbicide.

They guarantee that the Tococa shrubs survive and spread. So determined are these tiny, herbicide-happy creatures that other shrubs with the effrontery to germinate and grow near the Tococas are killed within days. Some perish within hours.

The discovery was made by three Austrian biologists, Wilfried Morawetz, Martin Henzl and Bruno Wallnofer, from the University of Vienna. They were working in the Sira Mountains in the Peruvian Amazon, where rainforest covers steep hills and ridges.

Tococa is a rainforest shrub growing no more than 3m (9.6ft) high. As it is light demanding, it only ever gets going on the tops of hills where gaps appear in the rainforest cover. A quick grower, it can gain half its eventual height in its first six months.

Ants first appear with the young shrubs, making themselves at home both inside the hollow stems and in the leaf pores. Once the population reaches about 1,500 individuals it begins its herbicidal mission.

As merciless as a garden sprayer filled with weedkiller, these tiny ants foray forth, killing any plants growing between the Tococa bushes and in a zone up to 3m wide round about them.

In trials with 40 pot-grown plants of eight different species, Dr Morawetz and his team found that none survived for more than five days near the Tococa. Several were dead within three days. One papaya plant lasted three hours before it was reduced to a soft, blackened, decaying clump.

Most impressively, the ants attack and kill trees up to 10m tall. Systematically, they tackle each leaf until the defoliated tree, injected with herbicide, gives up the ghost.

How do these ants do their gardening work? To increase the effectiveness of their herbicide, they first bite into plant leaves at their most vulnerable points; for instance where the nutrient-carrying veins meet up. Then they spray poison into the wounds from glands on their abdomens.

Damaged leaves develop brown patches which slowly rot. The patches spread and whole leaves die. Small plants succumb quickly; larger trees can take days, even weeks. As the ground around is denuded of any potentially competitive vegetation, the Tococa grove extends outwards by sending up suckers.

Dr Morawetz's team estimates that a grove of about 60 Tococa shrubs would give refuge to one or two million ants.

In this highly advanced relationship, the Tococa is guaranteed the light and space it needs to grow to produce a monoculture up to 30m in diameter. Without its lodgers, Tococas would barely get a look-in in the rough and tumble of rainforest species jockeying for life-giving light.

So what do the Tococa plants provide to repay all this help? They simply give the ants food and places to hide. The biologists observed ants cutting open small glands on the surface of Tococa leaves and feeding on the exudate, a liquid that is probably rich in nutrients.

The ants do not inject herbicide when they feed on the Tococa leaves and so do little harm to the shrubs. They also appear to eat scale insects that live in parts of the leaves. All in all, a somewhat Spartan diet.

Somehow the ants are able to recognise the particular Tococa shrubs that give them their sustenance. Dr Morawetz's team tested this recognition ability by cutting twigs from the same grove and placing them on the ground within the cordon sanitaire. The ants did not attack the cuttings. Yet twigs cut from another Tococa grove were set upon as if they were an alien shrub.

The upshot of this is that ants react only to colony-specific markers on their host plants. The markers are probably complex chemicals.

So, with a symbiotic partner like these ants fighting their corner, why haven't Tococa shrubs taken over the Amazonian rainforests?

The reason is that the ants don't appear to recognise trees growing at a certain distance from the spreading Tococa colony. As these develop larger and larger canopies high above the Tococas, they eventually completely shade the light-dependent shrubs.

The formidable ants are powerless to stop the fading conditions. Slowly, the grove dies out. The ants die, too, or move on to seek out a developing Tococa patch elsewhere.

But they leave a legacy in their wake. For several years, few plants can grow in the herbicide-contaminated soil. Heavy rainstorms may wash it away, exposing bare rock. Perhaps they are not the sort of ants we would want in our gardens after all.