Science: The sensible host stays cool: Bumble bees have found a way to fight against parasites. Chris Barnard reports

Chilling out may not seem the obvious way to deal with a nasty infection, but the arms race between hosts and their unwelcome invaders can lead to desperate measures. Bumble bees, it seems, take a cool approach.

As part of their armoury, many parasites exploit or modify their host's behaviour, usurping it for their own ends. Certain helminth (worm) infections of fish, for example, cause their hosts to be less wary of predators, so increasing their chances of being eaten and absorbed into the system of a new host in the interests of completing their life cycle. The incidence of such behaviour among many species suggests it is a strategy hosts find hard to counter. Bumble bees, however, appear to have turned the behavioural weapon back on their tormentors.

Workers of the bumble bee Bombus terrestris are commonly parasitised by conopid flies which inject an egg into the bee's abdomen. The larva hatches, developing for 10 to 12 days before killing the host. Up to 70 per cent of worker bees can be parasitised in this way. While the flies seem to set a time bomb ticking inside their hosts, the bees appear to have found a way of defusing it, or at least of slowing it down.

Christine Miller and Paul Schmid-Hempel, working at the Institute of Terrestrial Ecology in Zurich, have found that bees parasitised by the flies tend to stay out of the colony both during daylight foraging hours and, importantly, during the cool hours of night. Their absence from the colony seems bizarre as workers among social insects can reproduce only by proxy, by helping their mother-colony rear offspring. Time out from the colony is therefore time out from reproduction.

Miller and Schmid-Hempel's study, however, suggests that colony desertion could benefit parasitised bees by exposing the parasite to a hostile thermal environment. Larval development in the flies is dependent on a warm temperature. By staying out of the warm colony, especially at night when temperatures are cooler, parasitised bees might be able to slow down the development of the larvae and so prolong their life. Miller and Schmid-Hempel tested the idea by setting up bumble bee colonies with marked workers in the field. By monitoring individual bees through the summer, they were able to show parasite-free workers spent on average less than 10 per cent of their time outside their colony at night. When the same workers became parasitised, however, this rose to more than 70 per cent. Other effects, such as age, were carefully ruled out.

To check that bees were staying out in order to be cold, Miller and Schmid-Hempel gave parasitised and non-parasitised bees the opportunity to wander between a cold and a warm compartment in a choice chamber, where temperatures were matched to those outside at night or inside the colony. While non-parasitised bees spent around 50 per cent of their time in the cold compartment, parasitised bees chose to spend over 80 per cent of their time there. The bees therefore seemed to choose where to be on the basis of temperature.

The experimenters also checked that the overnight temperatures experienced by stop- outs had the required effect on their parasites and thus on their lifespan. Infected bees subjected to a controlled temperature of 19 C lived an average of two to three days longer than those subjected to a temperature of 28 C. In keeping with this, nearly 50 per cent of parasites recovered from dead bees which had been subjected to the lower temperature had failed to complete their development.

Their cool anti-parasite response gives one explanation as to why bumble bees are so often seen sluggishly lumbering about outside early on summer mornings.

Increasing body temperature is a well-known strategy for combating parasitic infections among vertebrates and some insects. Miller and Schmid-Hempel's bees appear to be the first example of a comparable strategy based on reducing body temperature, in this case by behavioural means.

The author works in the Behaviour and Ecology Research Group at the Department of Life Sciences, University of Nottingham.