Observers looked monocularly into a tunnel, with gratings on the left
and right sides drifting toward the head. An exposure period was follo
wed by a test with fixed gratings. With fixation points, left and righ
t retinal fields could be stimulated selectively. When exposure and te
st were on the same retinal fields, but fixation was on opposite sides
of the tunnel during exposure and test periods, aftereffects of retin
al sweep and of perceived looming were in opposite directions. The two
effects tended to cancel, yielding no perceived aftereffect. When the
y did occur, aftereffects in the retinal and the looming directions we
re equally likely. Cancellation was significantly more likely in the e
xperimental conditions than in the control, when fixation always remai
ned on the same side. When areas of retinal stimulation in the exposur
e and test periods did not overlap, cancellation was less frequent and
aftereffects of looming were more frequent. Results were not signific
antly different for left and right visual fields, indicating that cort
ical vs. subcortical OKN pathways do not influence the illusion. Vecti
on resulted for 16 of 20 observers under one or another of our conditi
ons.