The depth effect induced by motion of two dots was investigated in two experiments. Subjects were asked to estimate the length and the angle of inclination of a ‘perceived rod’ that moves in a 3-D space as if the two stimulus dots were its end points. In the first experiment the dots moved on parallel straight and elliptical paths. For the elliptical paths two positions were of special interest: when both dots were on the inner halves of paths and when they both were on the outer halves. The results are set against predictions derived from the frontoparallel principle (Johansson and Jansson, 1968 Perception & Psychophysics 4 165 – 170). This principle was confirmed for the inner halves, the results were inconclusive for the straight paths, and the principle was not confirmed for the outer halves of elliptical paths. This implies that the frontoparallel principle might be a special case of some more general principle, suggesting that subjects used some additional cues to estimate length in 3-D space. The results also indicate some compensatory strategies—while the length of the 3-D rod tends to be overestimated, its angle of inclination is underestimated. With these outcomes in mind a geometrical model of the percept is proposed. In the second experiment, the angle and the length of the “perceived rod” were varied. The dots moved on semi-elliptical paths (inner and outer halves). The results revealed that the angle of inclination that the perceived rod ought to have in the 3-D space is a factor in the estimation of its length. The geometrical model also provides an insight into the type of motion of the rod; specifically, rotation for the inner halves and oscillation for the outer halves of the elliptical paths.
