We report the preliminary results of a continuing prospective evaluation of a screening programme for congenital dislocation of the hip (CDH) which uses ultrasound imaging to provide delayed selective screening to complement neonatal clinical screening. Of 26,952 births in the Southampton district, 1894 infants were referred for secondary screening because of a clinical abnormality or the presence of a predetermined risk category for CDH. Pavlik harness treatment was required for only 118 infants, giving a treatment rate of 4.4 per 1000 births. Of those referred with clinical instability, 35% did not require treatment. Dislocation or subluxation was detected in 17 of 643 infants referred only because they fell within one of three risk categories: breech presentation, foot deformity and family history. All 17 had normal clinical examinations and cases were discovered in each category. Six children presented with CDH after 12 weeks of age, giving a late presentation rate of 0.22 per 1000 births. All had normal clinical examinations within 24 hours of birth and none was in a risk category. Surgery has been required in ten children, giving a surgical treatment rate of 0.37 per 1000 births. We conclude that, in Southampton, delayed selective secondary screening with ultrasound is more effective than clinical screening alone. It targets treatment to those infants who need it, and reveals a number of dislocated and subluxed hips that would otherwise be missed.