OBJECTIVE Vanishing Optotype letters have a pseudo high-pass design so that the mean luminance of the target is the same as the background and the letters thus 'vanish' soon after the resolution threshold is reached. We wished to determine the variability of acuity measurements using these letters compared to conventional letters, and in particular how acuity is affected by the number of alternatives available to the subject. METHODS Acuity was measured using high contrast letters of both conventional and Vanishing Optotype design for three experienced normal subjects. Thresholds were determined for central vision in a forced choice paradigm for two alternatives (2AFC; AU and OQ), 4AFC (AQUO), 6AFC (QUANGO) and 26AFC (whole alphabet) using a QUEST procedure. Three measurements were made for each condition. RESULTS Threshold letter size was always larger for the Vanishing Optotypes than conventional letters, although the size of this difference (0.11-0.34 logMAR) depended on the number of alternatives and what they were. The effect of the number of AFC, and the individual letters employed, was smaller for the Vanishing Optotypes, implying that they are more equally legible than conventional optotypes. Variability was also lower for the Vanishing Optotype sets (0.01-0.03 logMAR) than the conventional letter sets (0.03-0.06). CONCLUSIONS The smaller effect of the number of letter alternatives, combined with more equal discriminability and lower threshold variability, implies that Vanishing Optotypes may be appropriate targets from which to design letter charts to measure small clinical changes in acuity.