They hypothesized that when arsenic comes in contact with this modified E. coli, it will activate the arsenic switch, whose gene in turn will activate the breakdown of lactic acid. French says that a simple color litmus test can then pick up the resulting change in pH levels, also indicating the presence of arsenic. "It works under laboratory conditions, but we haven't tested it with real water samples," he says.

So far, the team has worked with relatively clean lab-controlled solutions of water and inorganic arsenic mixed with modified E. coli and a pH indicator in a sealed glass tube. After a while, the water turns red or yellow, depending on the presence or absence of arsenic. A potential drawback is the test's lag time: it takes about five hours for the biological interactions to complete.

Ultimately, the group hopes to design a cheap and simple kit of powder--freeze-dried bacteria and pH indicator--in a disposable, screw-top plastic tube. In the near future, French's group may test its design on actual water samples from Bangladesh that contain other biological and chemical materials. For now, the researchers have added their device to Endy's "bio-bricks" registry for other scientists who might want to build a better arsenic biosensor.