1. Exercise-induced haemolysis has been implicated in the sub-optimal iron status of endurance-trained athletes. Accordingly, erythrocyte survival studies using 51Cr were performed on male and female distance runners (n = 20) and sedentary control subjects (n = 10) in order to determine whether the rate of erythrocyte destruction was altered as a consequence of repetitive exercise training. 2. The chromium half-disappearance time of the male (25.4 +/- 3.6 days, mean +/- SD) but not the female (28.3 +/- 4.6 days) athletes was significantly lower than that of the male (33.1 +/- 4.5 days) and female (32.3 +/- 2.6 days) control subjects (P less than 0.01). The mean erythrocyte lifespan of the male and female distance runners (67.2 +/- 22.2 and 72.4 +/- 26.0 days, respectively) was significantly shorter than that of the non-exercising male and female subjects (113.4 +/- 31.0 and 114.1 +/- 29.0 days, respectively) (P less than 0.01). 3. There was no correlation between the mean erythrocyte lifespan and the haemoglobin concentration, serum ferritin levels, body mass, weekly training distance, number of years running or daily protein intake. The mean cell volume and reticulocyte count measured in the same athletes before and after completing a standard 42 km marathon race were within the normal range, whereas the plasma haemoglobin levels were elevated (77.0 +/- 50.5 mg/l) and the serum haptoglobin levels were decreased (0.89 +/- 0.4 g/l) at rest, with a further significant decrease after running (0.69 +/- 0.4 g/l) in the latter measurement (P less than 0.05). 4. It is concluded that the demonstrated increase in erythrocyte turnover may be sufficient to precipitate an iron deficiency in endurance athletes when dietary intake or absorption does not meet the accelerated erythropoietic demands.