THIS paper examines the contribution of egocentric cues and visual con
text to manual localization of remembered targets. Subjects pointed in
the dark to the remembered position of a target previously viewed wit
hout or within a structured visual scene. Without a remembered visual
context, subjects pointed to within 2 degrees of the target. The prese
nce of a visual context with cues of straight ahead enhanced pointing
performance to the remembered location of central but not off-center t
argets. Thus, visual context provides strong visual cues of target pos
ition and the relationship of body position to target location. Withou
t a visual context, egocentric cues provide sufficient input for accur
ate pointing to remembered targets.