Signs of magnesium toxicity appear at serum magnesium concentrations of 1.5 mmol/L. The reference range is 0.75 to 0.93 mmol/L and intraindividual variation in healthy subjects is negligible. Magnesium deficiency and marginal deficit are widespread, the latter being reflected in high intraindividual variations in serum magnesium concentration, and the former in a consistent decrease in serum levels. The distribution of magnesium can be described by a 2-compartment system. The purpose of this study was to estimate the time required to achieve steady-state concentrations (i.e. saturation of metabolism) after oral ingestion, which is safer than parenteral administration unless frequent monitoring is possible. In an open study, 10 hypomagnesaemic patients consecutively admitted to the Back Rehabilitation Unit, Ichilov Hospital, with musculoskeletal pain of nonmalignant origin were given 2 to 6 oral magnesium hydroxide tablets (125 to 750mg of Mg++) daily depending on tolerance, for 6 months, followed by 2 tablets daily until a steady-state serum concentration was achieved. The initial serum magnesium concentration was 0.703 +/- 0.008 mmol/L, measured by atomic absorption spectrophotometry, and at steady-state was 0.910 +/- 0.006. The mean 24-hour urinary magnesium was initially 3.80 +/- 2.34 mmol/24h, increasing to 5.70 +/- 1.14 mmol/24h. Mean time to achieve steady-state was 12.4 +/- 6.2 (range 3.5 to 20) months. The only adverse effect, apart from diarrhoea at high doses, was a reduction in serum phosphate concentration after 9 months in 1 patient, and after 12 months in a further 2 patients. It is suggested that the range of 0.82 to 1.06 mmol/L serum magnesium concentration observed on saturation of metabolism should serve as a basis for defining the target range of serum magnesium concentrations in magnesium therapy.