ABSENCE OF RELATIONAL DETERMINATION IN THE ROD-AND-FRAME EFFECT

Relational features between a rod and frame that might facilitate an angle-detection process were manipulated in 2 ways: first, by varying rod lengths with retinal eccentricities of 3.8.degree., 7.5.degree. and 14.8.degree. in a frame with inner dimensions of 35.3.degree. .times. 35.3.degree., and second, by eliminating the sides, leaving only the corners of the frame. The converse case, in which corners were eliminated but sides left intact, represented the control condition. In neither case did the opportunity to use relational features enhance the rod-and frame effect (RFE). In a frame with inner dimensions of 52.5.degree. .times. 52.5.degree., when line length had an eccentricity of about 25.degree., reducing it to 3.degree. did significantly reduce the RFE, but only for those who scored in the upper 50% as measured on the full-frame RFE. An explanation is proposed based upon visual-vestibular interactions between receptive fields in peripheral retina and orientation detectors in the vestibular nucleus.