Average time required to determine whether an alphanumeric character was presented in its normal version or as its mirror image increased from 500 msec to 1,000 msec as its angular departure from upright increased from 0 to 180 deg. However, when Ss already knew the identity of the upcoming character and when advance information as to its orientation was available for 1,000 msec, this reaction time was reduced to about 400 msec regardless of the orientation of the test stimulus. In this case, Ss claimed that they could prepare for the rotated stimulus by imagining the normal version of the designated character rotated into the indicated orientation and that they could then rapidly test for a match against the ensuing stimulus.