The U.S. Environmental Protection Agency (EPA) has struggled in the past two months to come to grips with dispersants, the chemical cocktails being used to break up the Gulf of Mexico oil spill into tiny droplets that are easier for microbes to eat. It now appears that the EPA failed to require adequate controls for dispersant toxicity testing, despite the fact that the agency set very explicit criteria for how such chemicals should be tested. This follows the agency's call for BP to use less deadly dispersants (pdf) with toxicities below a certain threshold when, in toxicology, it's smaller figures that indicate a concentration that is more dangerous.

When testing a given chemical's toxic impact, using the same methods on the same chemical—a so-called reference toxicant—should give consistent results no matter who carries out the tests. That way the toxicity of various products can be rated against each other, and against a certain benchmark. In addition to setting such a benchmark, the EPA also sets rules for how the chemical test solutions should be made, the concentrations required, and the organisms to be tested (in the case of dispersants, a fish, Menidia beryllina, and a shrimp, Mysidopsis bahia), among other specifications.

The reference toxicant provides scientific assurance that the tests were carried out in the manner intended, and helps to legitimize data that is, after all, provided by the makers of the chemicals themselves. In the case of the dispersants pre-approved for use by the EPA in the event of an oil spill, that reference toxicant is a surfactant called dodecyl sodium sulfate, a chemical that helps liquids spread out. Assuming all tests were carried out in a similar fashion, they should deliver similar results on the toxicity of SDS.

But toxicologist Carys Mitchelmore of the University of Maryland says that industry-provided testing values on the toxicity of SDS relative to various dispersants on the market vary "up to 300-fold." The cause, she says: "They could have had bad water, bad animals, bad reference toxicant or they could have just run it badly."

And NOKOMIS 3-F4 (purported by EPA's list to be the least toxic dispersant) didn't even use SDS as its reference toxicant, relying instead on copper sulfate. "That one's the real outlier," Mitchelmore says. As she testified to the U.S. Senate last week, "If you do not have consistent [lethal concentration data] for your reference toxicant then you cannot assume any of the other tests are reporting accurate data."

That means the EPA does not know the relative toxicity of various dispersants available on the market and pre-approved for use, and never accurately knew their toxicity prior to the ongoing disaster in the Gulf of Mexico. So far, more than five million liters of dispersants have been used there. Which begs the question: Did anybody at EPA check the industry's dispersant homework? "What does concern me is that there clearly is no quality control. How do they make the decision that these things are approved then?" Mitchelmore asks. "How could they have accepted this data?"

The agency is currently carrying out its own toxicity tests on the various dispersant options, an effort that should take no more than six weeks and could be done in as few as two. In the meantime, the impacts of the dispersed oil from BP's Macondo well will provide answers. "We're going to be filling in some of those data gaps," Mitchelmore says, with data from the real world.

Image: A U.S. Air Force chemical dispersing C-130 aircraft from the 910th Airlift Wing at Youngstown-Warren Air Reserve Station, Ohio, drops an oil dispersing chemical into the Gulf of Mexico as part of the Deepwater Horizon Response effort, May 5, 2010. The 910th AW specializes in aerial spray and is the Department of Defense's only large area fixed wing aerial spray unit. U.S. Air Force Photo by Tech. Sgt. Adrian Cadiz.