This study evaluated the repeatability both within one session and between two sessions of a new technique of axial length measurement using 'A' scan ultrasound through the eyelids. The validity of the technique was assessed by comparing the results to those taken through the cornea. The new technique was very reliable, the mean difference between two readings at the first visit being 0.01 mm and between the two visits being 0.01 mm with 95% of differences between two readings in both cases being less than two standard deviations. The axial length measurements through the lid were a mean of 0.05 mm longer than the corneal measurements. The standard deviation of the difference between the two forms of measurement was 0.16 mm and 95% of axial length readings through the eyelid were within two standard deviations (+/- = 0.32 mm) of the corneal measurement. Providing strict criteria are applied to the quality of the 'A' scan images we believe that satisfactory axial length values may be obtained using this technique.