The purpose of this study was to determine inter- and intra-observer agreement in assessing hip morphology and stability by ultrasound. Three groups of infants, of 206, 74 and 78 newborns respectively, were examined. Morphology was classified into four categories (normal, immature, minor dysplastic and major dysplastic) according to subjective assessment, objective measurement (of the acetabular inclination angle alpha) or a combination of the two. Inter- and intra-observer agreement was determined for reading of recorded ultrasound scans, and for examination (recording plus reading of the scans). Hip stability was subjectively classified as stable, unstable, dislocatable or dislocated, and inter-observer agreement was determined. There was a high degree of agreement for morphological classification based on repeated readings of recorded scans by the same observer (206 infants, kappa = 0.7 and 0.8 for the two observers, respectively) while the degree of agreement between observers was moderate (kappa = 0.5). The agreement between observers for repeated readings and recordings was moderate when based on a subjective classification (kappa = 0.5). Adding the alpha angle did not improve agreement. There was a moderate inter-observer agreement in determining hip stability (70 infants, kappa = 0.4). The authors concluded that a high degree of inter- and intra-observer agreement in classifying hip morphology may be obtained for the reading of recorded ultrasound scans. Inter- and intra-observer agreement in producing the scans is poorer than for reading. To obtain a high degree of inter-observer agreement in assessing hip morphology and stability in the newborn, substantial training, attention to details in the technique, and evaluation of results are necessary.