Even if a water sample has been collected with proper trace-metal clean procedures, it's not completely clear what, for example, "10 parts per billion of lead" means. Is that lead dissolved in the water? Is it attached to small particles of soil? Is it a colloid?
A mixture of two substances is called a colloid if the mixture remains stable, without separating, over a long period of time. Some familiar examples are milk, blood, ink, and paint. In milk, tiny globules of fat are suspended in water, apparently defying the rule that "oil and water don't mix." The fat certainly doesn't dissolve, but it doesn't separate out, either (at least for a few days).
Since the 1950s, environmental chemists have tacitly ignored colloids. Anything that passes through a filter with .45-micron pores was considered to be "dissolved." It was merely a definition of convenience, as .45 microns was the size of the finest filter available in the 1950s. Chemists now have finer filters available to them, but the old standard has become entrenched in tradition and environmental regulations. "The point where a particle becomes dissolved has never been objectively defined," says Arthur Horowitz, a chemist for the U.S. Geological Survey.
Charles Alpers of the U.S. Geological Survey has experimented with filters with pore sizes down to .005 microns, and found that the finer filters trap many colloidal particles that were previously reported as "dissolved." In this sense, measuring concentrations of a chemical isn't like measuring the weight of a rock: as the equipment used to do the measurement changes, so does the measurement. So what's a chemist to do? Stick with the old procedure, arbitrary as it is? Set new standards, which might force the government to rewrite all its regulations?
Ultimately, Horowitz thinks, we need to understand which particles are absorbed by the human body, or by the animals and plants that use the water. "That's the really hard question, and one that most individuals, such as Flegal, never really address," Horowitz says. "Chemical concentration per se is not what we're concerned about."
But in fact, Flegal's and Creasey's work was a spin-off from research on colloids. Creasey planned to compare their measurements of trace metal concentrations with different filters to the results of the commercial consultants. But she realized that there was a discrepancy even when they used the normal .45-micron filters. That's when she and Flegal decided to publish their finding.