When a mysterious odor wafted through the city this week, Mayor Michael Bloomberg quickly appeared on television to reassure unnerved New Yorkers that the smell — whatever it was — was harmless.
The pronouncement was more than guesswork.
Over the past three years, the U.S. government has deployed hundreds of air-sniffing sensors in at least 30 metropolitan areas to create an early warning system for a chemical or biological attack. In some cities, the devices test the air 24 hours a day for traces of anthrax, smallpox and other deadly germs.
When the smell came through New York on Monday, it did not set off any of the system's alarms. And that helped offer the public some reassurance.
Most of these urban monitoring networks are still in a fledgling stage, and authorities warn that they have their limitations. But the surveillance network has been steadily improving.
In some big cities, including New York, Boston and Washington, monitors have been installed in major train and subway stations to sample the air for poisonous chemicals or explosive gases. Also, environmental agencies have been given portable air sensors that can be driven around in vans or carried by hand.
Wireless technology may soon make it possible for the machines to do automatic testing and relay the results to a central monitor, eliminating the need to carry samples to a lab.
"Our objective is to make it an almost instantaneous result," said Christopher Kelly, a spokesman for the Department of Homeland Security's science and technology division.
The system has its limits. Among them, the devices are at the whim of wind patterns and can detect only substances that have already been released into the air — meaning that their primary usefulness is in getting victims treated quickly and preventing a contagion from spreading.
The sensors also cannot test for everything. The director of Homeland Security's biological countermeasures program told Congress last May that the latest generation of the BioWatch system, the part of the sniffing network that monitors for deadly germs, will test for about 20 different microbes and toxins.
None of the many air-sampling systems available to investigators was able to actually identity the rotten-egg smell that wafted across parts of New York and New Jersey on Monday, officials said, and city investigators relied mostly on traditional methods of analyzing the odor, including handheld meters long used by utility crews to check for gas leaks. Investigators have yet to identify the smell.
The ability of the terror surveillance network to spot an actual attack is still largely untested.
Some experts have warned that there are far too few monitors in place. A report by the Environmental Protecttion Agency's inspector general raised questions in 2005 about the reliability and efficiency of what was then a $129 million BioWatch program.
Criticism of the system has lessened somewhat, however, as its technology has improved.
New York City's subway system was pleased enough with the performance of the chemical and biological detectors it installed in Grand Central Terminal and Pennsylvania Station that it recently decided to spend an additional $3.9 million on the system.
When the sensors were first implemented in 2004, they had trouble distinguishing between deadly chemicals and more routine stuff, like buckets full of cleaning fluid.
But the system has since become better, in part because of the addition of cameras that allow its operators to see if anything close by might be triggering the alarm.
The drive to invent an early warning system for bioterror still has a long way to go, said Penny J. Hitchcock, a senior associate at the Center for Biosecurity at the University of Pittsburgh Medical Center.
Monitoring for airborne toxins, she said, "is an enormously difficult problem and an expensive one." But she added that the system's potential may be huge, and not just in fighting terrorism.
A good network of automated air sensors, she said, may someday be able to alert public health officials about flu outbreaks, monitor the air in disease-prone poultry markets, or detect when someone with a contagious disease like SARS walks into an emergency room.
The surveillance network's potential, she said, was hinted at when BioWatch sounded its very first warning about a possible attack in Houston in 2003. That alert came after multiple air sensors simultaneously detected microorganisms used in germ warfare to cause tularemia, a potentially fatal fever.
The bacteria turned out to have occurred naturally in the environment and no one became ill, but while authorities were still investigating, they warned hospitals to be on the lookout for flu-like symptoms.
Doctors began looking at patients more carefully, and made a surprising discovery: People were getting sick — not from tularemia, but with a strain of flu that hadn't been included in that season's flu vaccine. That kind of discovery can save thousands of lives, Hitchcock said.
"If you think about harnessing this," she said, "I actually think it has a lot of potential."