The clinical course of 135 patients with lupus nephritis was examined long-term as part of a prospective study. Biopsies were classified according to modified WHO criteria and showed 17 per cent of patients had mild mesangial lesions, 10 per cent focal lesions, 21 per cent mild diffuse proliferative lesions, 37 per cent severe diffuse proliferative lesions and 15 per cent membranous lesions. Overall patient survival of 83 and 65 per cent at five and 10 years respectively from onset of nephritis was similar to other recently published series. In contrast to the latter, the severe proliferative group had a significantly worse outcome than the other proliferative groups (p less than 0.01) and only patients in this group progressed to end-stage renal failure. Haematuria was more common (p less than 0.05) in the severe group and there was a striking correlation between histologic activity assessed semiquantitatively (Table 1) and urinary red cell count (p less than 0.001). There was no correlation between serum creatinine, proteinuria or chronic lesions with urinary red cell count. In contrast to a previous study there was no correlation between the presence of hyaline thrombi on initial biopsy and subsequent development of glomerular sclerosis. Although the value of renal biopsy has been questioned, we suggest that it remains a most important investigation in the management of lupus nephritis. Determination of urinary red cell count provides a most useful monitor of disease activity and response to treatment.