A subchronic oral toxicity study with potassium nitrite (KNO2) was carried out in rats. Groups of ten male and ten female 6-wk-old rats received KNO2 in the drinking-water (tap-water) at levels of 0, 100, 300, 1000 and 3000 mg/litre for a period of 13 wk. The potassium concentration in the nitrite solutions was equalized by adding potassium chloride (KCl) up to the potassium level of the 3000-mg KNO2/litre solution. An additional group of ten males and ten females received drinking-water supplemented with KCl only, at an amount resulting in a potassium concentration equivalent to that of the 3000-mg KNO2/litre solution. Body weight, food intake and food efficiency were decreased at 3000-mg/litre level in males, while liquid intake was decreased in males given 1000 and 3000 mg/litre and in females given 3000 mg/litre. There was significant increase in the methaemoglobin concentration in animals given 3000 mg/litre, while slight decreases in red blood cell variables occurred at the 1000- and 3000-mg/litre dose. No impaired renal function was observed in any of the test groups, although the relative weight of the kidneys and the plasma urea nitrogen level was increased at 3000 mg/litre. There was a slight decrease in plasma alkaline phosphatase activity at 3000 mg/litre. A small amount of nitrite was present in the saliva of the rats receiving 3000 mg/litre but there was no evidence of increased mutagenic activity in the urine of these rats. Interestingly, hypertrophy of the adrenal zone glomerulosa was observed in all test groups, the incidence and degree being dose related. It was concluded that in the study reported here the no-effect level is lower than 100 mg KNO2/litre in the drinking-water, which is equivalent to a level lower than 10 mg KNO2/kg body weight/day.