If you go down to the woods today, you'll see a patchwork of different trees. Now Alissa Packer and Keith Clay, of Indiana University, announce that one reason for this diversity is that trees attract disease, making the ground around them uninhabitable for their offspring.

Trees face a battery of herbivores, parasites, fungi and diseases anxious to do them down, many of which are specialized to attack only one species. As you would expect, these attackers to congregate around their target species. A mature tree is usually up to the challenge, but this is bad news for any young and vulnerable offspring nearby. Other species, though, will find the area more hospitable. This could explain why many tree species coexist in tropical rainforests, where there are many tree pathogens.

Packer and Clay came at this idea from a new angle: they found that, in the eastern United States, a disease-causing fungus prevents the seeds of a black cherry tree (Prunus serotina) from growing next to it, as they report in Nature1.

They charted the fate of seeds germinating beneath six cherry trees over three years. Most seedlings sprout less than 10 metres from the base of their parent trees, but these stay-at-homes do badly -- survivors tended to germinate further afield. "The new aspect is that this is happening in temperate forest," says Wim van der Putten of the Netherlands Institute of Ecology.

Next, they needed to be sure that it was disease, and not overcrowding, that was killing seedlings. Experiments showed that, sure enough, soil is more important than density: seedlings did better growing in soil gathered far from the tree or in sterilized soil. This was a bigger influence on survival than the number of seedlings in the pot.

The offending fungus is an as-yet-unidentified member of a group called Pythium -- a bane of gardeners that causes root-rot in many plants. The fungus forms "a shadow underneath the canopy of the tree," says van der Putten. It must have a taste for black cherry, as other trees germinate far more successfully in the infested ground.

'Distance-dependent' effects could make themselves felt in many ways. They could help to explain why trees go to so much trouble to spread their seeds far and wide -- wrapping them in tempting fruit, for example.

"Our work suggests that dispersal through space may be more effective than dispersal through time -- that is, dormancy," says Packer. But, she adds, if a species is common, this may not be good enough. "It may disperse away from the parent, but still germinate close to another member of the same species," preventing any one tree from dominating a woodland.

And, paradoxically, the more disease-ridden a forest, the more diverse it should be. "Temperate forests are less diverse because they lack the host-specific predators and pathogens found in tropical forests," says Packer.

As a general rule, forests contain fewer tree species the further you go from the equator: in Canada or Sweden, for example, a few species of conifer can spread over huge areas. Packer says that these harsh environments "may be more stressful and limiting to soil pathogens".

But, however powerful any one idea, biodiversity has many causes. "The problem in ecology," says van der Putten, "is that hardly any hypothesis holds for 100 per cent of cases, because of the tremendous variation we are dealing with."