The extensive mammalian toxicity studies performed to support the safety of Bacillus thuringiensis-containing pesticides clearly demonstrate that the tested isolates are not toxic or pathogenic. Toxicity studies submitted to the US Environmental Protection Agency to support the registration of B. thuringiensis subspecies have failed to show any significant adverse effects in body weight gain, clinical observations, or upon necropsy. Infectivity/pathogenicity studies have shown that the intact rodent immune system responds as expected to eliminate B. thuringiensis gradually from the body after oral, pulmonary or intravenous challenge. Similar clearance patterns were observed with B. subtilis and B. sphaericus, which also were non-toxic to test animals. The results also indicate that the currently used protocols for toxicity/pathogenicity evaluations of micro-organisms in laboratory animals have provided useful and necessary information for risk assessment. The study results raise the issue, however, of whether some reduced set of toxicily/pathogenicity studies could be required for new B. thuringiensis isolates of already registered subspecies, and whether this reduced data set would also apply to subspecies not yet registered as pesticidal active ingredients.