Air mail not so easy Steven Schultz Since starting his air analysis 10 years ago at the University of Rhode Island, Michael Bender and his colleagues have developed an elaborate system for collecting his samples. "The logistics of it can be difficult," says lab technician Robert Mika. Mika manages the flow of 72 flasks of air a month that come into Bender's lab from remote points across the Pacific Ocean. Nine research stations contribute samples -- one aboard a merchant ship that makes a circuit between Los Angeles and Australia and eight land-based stations in such locations as Tasmania and American Samoa. The collection sites are remote so that the air is a thoroughly mixed sample of that region's atmosphere without local disturbances such as pollution. The Princeton researchers supplied each collection location with a device that slowly sucks air through a glass flask then seals the flask. Air being sucked in passes through a reservoir cooled by a slurry of liquid nitrogen and alcohol, which freezes all moisture out of the sample. The most sophisticated of these setups is on the ship, where a satellite-driven global positioning system alerts the crew that they are approaching a collection point. "They are not researchers, they are sailors. But they are glad to cooperate," says Mika. The collectors send the samples back to Princeton in cardboard boxes and Styrofoam packing that Mika supplies to them. "There are lot of issues involved in international shipping," says Mika. "Customs must be involved and make sure there is no commercial value." Once in Princeton, Mika loads the samples onto an automated device, which, over a period of about 16 hours, draws off some air and compares its makeup to that of "laboratory" air of known content. The device reports the ratio of oxygen to nitrogen in each sample. Over the years, Bender has developed several tricks for ensuring the reliability of his data. "You try to imagine all the things you might be doing that would harm the integrity of the sample -- and you do them more," says Bender. "Then you check to see if the numbers changed." "Also, we assume we are making mistakes, but we do everything identically, so we don't lose any information," he says. The measurements reveal not only long-term trends in global climate, but also a steady cycle of seasonal variations. Each spring, as trees in the Northern Hemisphere put out leaves, the amount of oxygen starts to go up, then dips down again in the winter. Given the subtlety of the measurements and the great potential for contamination of samples, it often seems impossible that clear trends will show up, says post-doctoral researcher Melissa Hendricks. "But after about six months, when you see the curves developing on the graph, it's really exciting." "I love working here," says Mika. "Sometimes when I'm running an experiment I can't wait to come to work in the morning and see what the results will be." Related article Message in a bottle		[an error occurred while processing this directive]