Abstract
The purpose of this study was to investigate the influence of axis length on center of mass perception. Participants estimated the center of mass in three-body displays by moving a small judgment dot with a mouse. The displays consisted of three black-filled dots viewed against a white background. The dots were located at the vertices of virtual right triangles. We varied triangle position (direction the hypotenuse faced), orientation (vertical and horizontal) and axis length (ratio of smallest side to intermediate-length side from 1:1 to 1:5). The center of mass or centroid in these displays is at the intersection of the medians, the lines that join a vertex to the midpoint of the opposite side. Axis length and median lengths are correlated. An increase in axis length produces a corresponding increase in the medians, which ought to make center estimation more difficult. The results confirmed our prediction. Mean error, measured as the difference between estimates and the true center, increased linearly with an increase in axis length. Our second dependent measure was response direction, the angular deviation of the estimate from the true center. Mean response direction always pointed downward but was shifted toward the orientation of the nearest downward median. Direction thus seems to be an accommodation between the influence of gravity and the orientation of the nearest median.