The method of constant stimuli was used to study the relative contributions of two factors (frame aspect ratio and pictorial depth) on the horizontal-vertical illusion. Four different contexts were created by combining vertically versus horizontally oriented frames with horizontally versus vertically slanted texture gradients. Points of subjective equality and just-noticeable differences were estimated from the psychophysical functions for each observer. The data revealed illusions ranging in size from 10.2% to 4.0% for the various frame-slant combinations, with no associated changes in response variability. The most important finding was that the effects of frame and depth were additive, suggesting that the illusion is determined by at least two different sets of mechanisms.