There is limited evidence that beryllium is a lung carcinogen to man, and several compounds of beryllium are carcinogenic to the lungs of the rat, rabbit and monkey. One such compound is beryllium sulphate (BeSO4.4H2O). This soluble salt has been evaluated in a range of genotoxicity tests. It was non-mutagenic to Salmonella typhimurium (strains TA1535, 1537, 1538, 98 and 100) when evaluated in the plate-incorporation assay at dose levels up to 5 mg/plate (+/- induced rat-liver S9 mix). It was also non-clastogenic to Chinese hamster lung (CHL) cells cultured in vitro. When dosed to male CBA mice via oral gavage at dose levels of 80% and 50% of the medium lethal dose (2.3 and 1.4 g/kg, respectively) it failed to increase the incidence of micronucleated polychromatic erythrocytes in the bone marrow (sampled at 24, 48 and 72 h post-dosing). However, a marked depression of erythropoiesis was evident 24 h after dosing suggestive of beryllium-mediated bone-marrow toxicity. When tested in the strain A mouse lung tumour bioassay, BeSO4 induced a significant increase in the number of tumour-bearing animals but not in the number of lung tumours per animal. These findings are discussed within the contexts of other genotoxicity data published for BeSO4, and of current strategies for the detection of possible human carcinogens.