The time-order effect (TOE), commonly referred to as time-order error, has been the subject of much investigation and has been extensively discussed in the literature.l It has both practical and theoretical importance. One of the most valuable practical results of work on TOE has been the knowledge that comparisons must be made in two time-orders if errors are to be minimized. In the absence of a quantitative theory, it has been customary to assume, however, that the TOE for standard first, variable second (S1V2) will be exactly equal to TOE for variable first, standard second (V1S,).2 If this assumption is incorrect, serious errors may creep into psychophysical measurements designed to establish psychological equality between physically different stimuli; for example; in the establishment of the luminosity function of the C.I.E. standard observer.3 The theoretical importance of TOE lies in its implications for psychological scaling and for theories of perception and judgment. The successful application of rating-scale techniques and adaptationlevel (AL) theory to comparative judgments makes possible the development of a treatment of TOE which is quantitative and which integrates TOE with other psychological phenomena.4 The present study was undertaken to test the validity of an extension of adaptation-level theory to timeorder effects.