How safe are the Gulf oil dispersants?

reveals that the agency’s assumptions about dispersant toxicity are based on unreliable data.


When it demanded that BP find less toxic dispersants, the EPA referred to a table summarizing experiments in which a fish called Menidia beryllina and Mysidopsis bahia shrimp were exposed to a mixture of dispersant and diesel fuel oil in a ratio of 1:10.

These are standard experiments that must be submitted for inclusion on the EPA’s National Contingency Plan Product Schedule, which lists products authorized for use on an oil spill. They determine the concentrations in parts per million (ppm) required to kill half of the animals in a given time.

They are only part of a series of tests that are required by the EPA: the dispersant must also be tested alone, the fuel oil must be tested alone, and finally the lab must run controls testing a “reference toxicant” known as dodecyl sodium sulphate (DSS).

Aldhous notes that looking across all the listed dispersants, test results for the fuel oil alone are highly variable, with the concentrations required to kill half of the fish varying from 5.95 to 201.8 ppm. This could indicate that some of the tested oil samples had lost their most toxic volatile components.

This is a major problem. The tests reveal varying toxicity for the different dispersant-oil mixes. Given the significant differences between results for samples of oil alone, however, it becomes very difficult to compare the results for the mixes. “It’s absurd,” says Joannie Docter, president of GlobeMark Resources in Atlanta, Georgia, which makes a dispersant called JD-2000.

Incomplete data

The DSS control experiments pose even bigger problems, says Mitchelmore. The concentrations of DSS required to kill half of the fish, for example, vary from 1.14 ppm for the tests submitted for Sea Brat #4, to 159.6 ppm for the controls submitted for Nokomis 3-F4,made by Mar-Len Supply of Hayward, California.


Aldhous says that, presumably, this huge variation reflects inconsistencies in the testing procedures used by the labs hired by manufacturers to run the tests. If so, it throws the EPA’s conclusions about the products’ relative toxicities into serious doubt.

Docter has complained to the EPA about its reliance on “incomplete and misleading” data; Nalco has similarly drawn the agency’s attention to the huge variability in the results supplied by different dispersant manufacturers. “They acknowledged that there are some shortcomings,” says David Horsup, vice-president for research and development with Nalco Energy Services.

Mitchelmore argues that the EPA should also run tests on the growth and reproduction of shrimp and fish, to judge the dispersants’ longer-term effects. Those tests are not yet being run.

The EPA says it has commissioned a single laboratory to retest all the authorized dispersants, running the standard tests but using Louisiana crude rather than fuel oil. Establishing uniform conditions and ensuring the animals are all of the same age takes time, the agency told New Scientist. “These tests take more than a few days to run.”