Research on classical conditioning of human pupillary responses is reviewed and evaluated. Es have usually employed an auditory CS with illumination change as the UCS. While positive findings were reported in early studies, later investigators generally met with little success both in replications and in more sophisticated designs. Crude methodology, lack of proper control group data, and the subjective nature of response measurement were characteristics of earlier studies which contributed to the presumably erroneous findings. However, Young's (1958) contention that pupillary responses cannot be conditioned is not supported by results of mote recent investigations using shock as the UCS. Implications of these findings are discussed and suggestions for future research are offered.