With the evolving use of ultrasonography, there has been an increasing emphasis on the technically oriented diagnosis of developmental hip dysplasia. Graf's alpha and beta angles and Harcke's bony rim percentage are two of the more common criteria, but their repeatability and reliability have not been well documented. Fifty-eight hip ultrasounds of 29 infants were measured independently by three physicians, and statistical analysis was performed. It was concluded that Graf's and Harcke's methods could be repeatable with training and frequent use of ultrasonography. Harcke's guidelines have a large equivocal zone, which limited their clinical usefulness. Graf's angles gave a better classification, but their clinical relevance has not been proven.