Some security experts said the system is unlikely to catch a bioterrorism attack in time to save many lives. And they said it is powerless to spot an attack in an enclosed area, like an airport terminal or subway line, and unable to detect attacks unless they are big enough to scatter over several blocks.
"If you saw planes going over and releasing major clouds of this stuff, there's a chance that people would get suspicious a long time before anybody checked the filters," said Jacqueline Cattani, director of the Center for Biological Defense at the University of South Florida.
The sensors have been in place since early spring, and while the government won't say exactly where, regional health officials confirmed the list includes Philadelphia, New York, Washington, San Diego, Boston, Chicago, San Francisco and St. Louis.
The White House in January said the "Biowatch" monitoring system would cost about $1 million annually per city.
In participating cities, filters within the machines are removed daily and immediately analyzed for spores and chemicals that could have been dumped from a plane or building and left to drift in the air.
If an attack was close enough to a sensor, authorities could know about it within 12 hours, according to Bob Bostock, homeland security chief for the Environmental Protection Agency. That is much faster than it would take people exposed to anthrax to develop symptoms, he said.
"The main advantage or having a system like Biowatch is that prior to it being rolled out, the only real way to tell if a biological agent had been released was to see if people started turning up sick or worse," Bostock said. "By knowing in advance that a contaminant has been released, you can start treating it before symptoms develop."
Much about the system is being kept secret; the government won't say who makes the detectors, how much they cost, or what they look like. Officials also won't say which labs are analyzing the detectors' filters, other than to say that some are operated by state health departments.
If a filter tests positive for a particle, scientists can estimate where it came from, based on its physical properties, Bostock said.
The system, though, has plenty of critics.
Calvin Chue, a researcher at the Center for Civilian Biodefense Strategies at Johns Hopkins University, said the cost of testing and replacing the filters daily will be high and the probability of spotting a contaminant low. He also said the results will be difficult to confirm, especially in polluted cities or places where natural organisms found in the air can give false-positive results.
Researchers who studied what would happen if someone dropped 2.2 pounds of anthrax from a tall building in New York said the sensors could save lives, but only if officials detected an attack immediately and instantly began distributing medication.
Stanford Business School professor Lawrence Wein said in that scenario, the number of dead could be cut from 120,000 to 70,000 if sensors detected an attack within six hours. He said deaths could be cut to 50,000 if the government allowed people to stockpile antidotes ahead of time.
Bostock, of the EPA, wouldn't discuss how the detectors have performed so far, but he said the monitors have neither detected an attack, nor produced the type of false-positive reading that triggers an emergency response.
"We have a high degree of confidence in the results we have been getting," he said.
The EPA, he noted, has used similar sensors to monitor air pollution for decades.
"It's not like, with a sensor like this, you have to have one on every street corner," he said. "We have quite a bit of experience in terms of how things move in the air."