Effects of delay of instrumental reinforcement on human eyelid conditioning were investigated in 5 groups of 20 Ss each with delays of 0, 0.5, 1.0, 2.0, and 4.0 sec. between response and reinforcement. The reinforcement was illumination of a small lamp and operation of an electric counter, and S was told only that if he did the right thing the reinforcement would occur. Acquisition appeared to be maximal when the reinforcement was delayed from .5 to 2.0 sec. The function was a flat-topped inverted U with 0 delay and 4.0 sec. delays producing less conditioning. There was some indication of greater resistance to extinction in those Ss who received the 4.0sec. delay of reinforcement.