Catching an oil leak in its earliest stages is critical for directing appropriate cleanup efforts, says Van der Meer. A spill may not leave a visible trace, in the form of tar, until long after its most toxic effects have come and gone. By allowing for quick and easy detection of spills very soon after they occur, biosensor bacteria may make possible an earlier, more effective intervention.

Chemical testing will still likely be necessary, however. The bacterial sensors can give a rough estimate of the relative amounts of each chemical class, but only rigorous chemical analysis can determine exactly how much of each substance is present. "We tried to develop this method to be relatively quick, and to give you an overview," says Van der Meer, adding that biosensors could perhaps identify areas where more-extensive testing is warranted.

Van der Meer ultimately hopes to incorporate the glowing bacteria into buoy-based devices, which would continuously monitor seawater for hints of an oil spill and relay pertinent information back to a laboratory. His group is developing microfluidic systems that could maintain a constant, contained population of sensor bacteria to periodically test the waters.

Such a device would be subject to the vagaries of living organisms: its usefulness would be entirely dependent on whether the bacteria were alive and thriving. A negative reading could mean that no toxins are present, but it could also mean that the bacteria have died. "If they're not healthy," says Richardson, "the system is broken." Deploying living sensors also raises the risk of releasing genetically altered organisms into the environment. In this case, the chemical-sensing bacteria are theoretically harmless and unlikely to survive long in the harsh open environment.

Beyond detecting oil spills, Van der Meer's group has developed and tested a bacterial strain that detects arsenic in rice. Other potential applications include testing for pollutants in soil and groundwater, and for antibiotics in meat and milk. But for now, his vision for the future of biosensor bacteria remains largely aquatic.

"Why not have a robotic fish that swims through the water," he speculates, "and if it detects something, it could send out a signal by GPS? Technically, I think these things are possible."