Background. There is pressure on acute admission services in inner-city areas. Two deprived London districts with markedly different acute bed ratios but similar sociodemographic backgrounds were compared to test the hypothesis that more facilities mean better service. Method. An instrument for auditing the use of short-stay hospital beds was constructed to collect information concerning admissions to, and short-stay patients in, the chosen districts during a three-month period. Results. There was a higher admission rate and substantially greater use of beds per unit population in south Southwark than in Hammersmith and Fulham. Much of the difference was attributable to a higher rate of admission of patients with affective disorders in south Southwark. Conclusions. The results are not explained by Variations in population need, longer in-patient stay, or poorer aftercare leading to early relapse. The question of whether there is over-provision of services compared with real need in south Southwark, or under-provision (particularly for people with affective disorders) in Hammersmith and Fulham, is considered but left open for discussion following a study of ethnic issues and the reasons for admission.