The excretion of urinary hydroxyproline has been measured before mastectomy in 342 patients presenting with breast cancer for the first time to Guy's Hospital. The first 106 women were maintained on a gelatine-free diet whilst the remainder were on unrestricted diet. In both dietary groups hydroxyproline levels or the ratio of hydroxyproline to urinary creatinine were not related to pathological stage or histological grade. The time between initial presentation and subsequent bone metastases was negatively and significantly associated with hydroxyproline excretion (P less than 0.05) and the ratio of hydroxyproline to creatinine (P less than 0.01) in women on a gelatine-free diet. A similar, but not significant, trend was observed in patients on unrestricted diet. Although hydroxyproline excretion was related to the time to onset of bone metastases the amount of hydroxyproline excreted by these patients was not significantly different from patients who had recurrences at sites other than bone or patients who were disease-free up to 5 years after initial diagnosis. The conclusion is that hydroxyproline is of little value in the early detection of bone metastases.