Postmortem iliac crest biopsies were performed on 16 uremic patients. 3 had been treated conservatively while 13 had been entered into a maintenance dialysis program. The dialysate was treated by reverse osmosis for more than 10 years, and the aluminium concentration was consistently below the detection limit of 0.15 mol/l. 14 patients had been treated with aluminium hydroxide. Bone histomorphometry, aluminium labelling intensity, osteoid surface aluminium labelling extent (Al/OBI) and bone aluminium concentration were measured. 14 patients had significant bone aluminium deposition, including 2 who were not on dialysis of whom 1 had not received aluminium hydroxide. Bone aluminium concentration and labelling intensity were correlated to total aluminium hydroxide consumption (p less than 0.001, p less than 0.05) and present dose (p less than 0.01, p less than 0.01), while Al/OBI was not. The two patients with the highest aluminium concentrations had symptomatic osteomalacia, but 4 patients with significantly raised concentrations and mineralisation front labelling had secondary hyperparathyroidism. It is concluded that bone aluminium deposition occurs despite the use of aluminium-free dialysate and is associated with total and present aluminium hydroxide consumption; heavy aluminium deposition is associated with severe and symptomatic osteomalacia, but can also be observed in the presence of predominant hyperparathyroidism; aluminium deposition can occur in the absence of treatment with dialysis or aluminium hydroxide; bone aluminium concentration and labelling intensity are a better measure of bone deposition than Al/OBI.