Both the equidirection (nonius) and the equidistance horopter criteria were used on a group of subjects with and without the application of a 5% overall magnifier before the right eye. For three of six subjects, shifts in the equidirection horopter induced by this afocal lens could not be predicted on the basis of the magnification properties of the lens and the assumption that corresponding points are fixed. Equidistance horopter plots showed reasonably systematic reductions in slope between the predicted and the measured values for all subjects. Irregularities in all data were accounted for by shifts in the vergence angles during the experiment.