Subjects responded in accordance with the apparent distance of a target, alternating between normal vision and minification. Changes in performance induced by changing the visual condition were as marked after repeated exposure as initially, provided that no current feedback was supplied. This was taken as lack of support for Taylor''s (1962) prediction that effects and after-effects should decline with alternation of visual conditions. Contrary to the differential conditioning model of adaptation, their elimination apparently continues to depend on response-produced feedback.