An experiment by Colin 
Pitblado and Lloyd Kaufman, 
then at the Sperry Rand 
Research Center, supports the 
theory that the Poggendorff 
illusion results from 
displacement of angle contours. 
They used a stereoscope to 
present a rectangle against a 
background of converging 
lines, shown in the stereogram 
at left. The rectangle appeared 
in a vertical plane, separate 
from the converging lines, 
which appeared in another 
plane even though the retinal 
image still contained the 
crossing of the rectangle’s 
vertical contours and the 
converging line contours. 
Observers saw a distorted 
rectangle. The top edge looked 
longer than the bottom edge, 
presumably because the sides 
appeared to slope outward, due 
to overestimation of the angles 
at the intersections of the 
rectangle’s sides with the 
converging lines. Earlier, the 
same investigators did a 
variation of this experiment. 
They found that the Ponzo 
illusion disappeared with 
stereo viewing, suggesting that 
the original Ponzo effect, 
unlike the intersecting-lines 
variation, was the outcome 
solely of inappropriate depth 
processing.