The aim of this study was to examine the degree of concordance of two ultrasound-classifications of hip morphology by three investigators with different levels of experience, with two different probe-frequencies and two different methods of measurement. In 207 consecutive newborns (101 boys; 106 girls) the following parameters were assessed: bony roof angle (alpha-angle) and cartilage roof angle (beta-angle) according to Graf´s method and "femoral head coverage" (FHC) as described by Terjesen (modified Morin). Both hips were measured by each investigator with a 7.5 MHz linear transducer and a 12.0 MHz linear transducer. The images were measured twice by each investigator. First, they were measured with ruler and setsquare on a paper-print and later with the digital angle- and distance-function of the ultrasound system. Investigators were double blind to their own and the results of the other two investigators. Best results were achieved by the most experienced investigator for the alpha-angle with a variance of 7,1958 and for the beta-angle with a variance of 18,02. The lowest variance for the FHC achieved the least experienced investigator with 22,7529. While comparing the results of all three investigators, the results were only statistically significant for the alpha-angle and the FHC. The comparison of the different probe-frequencies revealed a significantly higher variance for the 12 MHz probe while calculating the mean over all three investigators (confidence interval: alpha-angle 1,07-1,27; beta-angle 1,07-1,28; FHC 1,12-1,34). But regarding the individual investigators the variances of the most experienced and the one of the least experienced investigator were significantly lower with the 12 MHz probe (confidence interval: pediatric orthopedic surgeon 0,55-0,81; medical student 0,5-0,74). Concerning the method of measurement a lower variance was achieved by all investigators for all parameters by using the paper-print images. But statistical significance regarding the mean over all observers was only achieved for the FHC (confidence interval 1,27-1,51). The slightly lower variances regarding the alpha- and the beta-angle over all investigators were not statistically significant. In summary it has been shown that although more experienced investigators tend to achieve lower variances, investigators with less experience achieve comparable results. There was no proven benefit from using probes with higher frequencies in general. But a higher frequency might be useful for measuring the beta-angle. The evaluation of paper-print images seems to be superior to the digital measurement technique.