PURPOSE. To evaluate the accuracy and repeatability of size lens induced aniseikonia measurement with the Aniseikonia Inspector Version 1 and a newer customized version of the Aniseikonia Inspector, Version 2. METHODS. Aniseikonia was measured on 27 subjects with both versions of the Aniseikonia Inspector in normal room illumination. Measurements of induced aniseikonia were made using size lenses in a randomized order. Twenty-five subjects were further tested in the dark using target sizes of equal visual angle for both tests. Repeatability of the intrinsic aniseikonia measurement was assessed on five subjects using randomized testing order for instrument and light and dark measurements. RESULTS. In normal illumination, the mean slopes for plots of induced aniseikonia vs. size lens magnification for Version 1 were 0.883 and 0.838 for the vertical and horizontal meridians, respectively. For Version 2, the corresponding slopes were 1.162 and 1.043. In the dark and using targets of the same size for both tests, the slopes for Version 1 were 1.038 in the vertical meridian and 0.866 in the horizontal meridian whereas for Version 2, the slopes were 1.195 in the vertical meridian and 1.127 in the horizontal meridian. The amount of underestimation or overestimation within any given testing condition showed considerable intersubject variation. Version 1 was more repeatable than 2, particularly in the vertical meridian. CONCLUSIONS. On average, the most accurate and repeatable measurement of aniseikonia was found with Version 1 in the vertical meridian in the dark. Measurement of aniseikonia in the horizontal meridian appears to be less reliable. Version 2 overestimates size lens-induced aniseikonia under all testing conditions. Intersubject variation in slopes of induced aniseikonia vs. size lens magnification should be further addressed. © 2007 American Academy of Optometry.