The usefulness of measuring plasma digoxin concentrations in the diagnosis of digoxin toxicity has been assessed in 83 in-patients. The mean plasma digoxin concentration in clinically toxic patients was significantly higher than the mean concentration in non-toxic patients. The overlap between the groups, however, was extensive and could partly be accounted for by hypokalaemia in those toxic patients whose plasma digoxin concentration was less than 3 ng/ml. There was, in addition, a higher incidence of hyperkalaemia, without obvious cause, in toxic patients than in non-toxic patients. Consideration of the incidence of various non-cardiac factors, specifically plasma potassium concentration greater than 5.0 mmol/l, plasma creatinine concentration greater than 150 mumol/l, daily maintenance dose greater than 6 microgram/kg, and age greater than 60 years, led to the development of guidelines to aid in the diagnosis of digoxin toxicity. Patients with plasma digoxin concentration greater than 3 ng/ml or with hypokalaemia should be considered probably toxic and those with plasma digoxin concentration greater than or equal to 3 ng/ml in the absence of hypokalaemia should only be considered toxic if they have at least two of the non-cardiac factors outlined above. Plasma digoxin concentrations could not be predicted with more than 31 per cent certainty by considering the magnitude of those non-cardiac factors.