Using a binocular search coil technique, we measured oculomotor behavior during gentle pressing with a finger on the outer canthus of one eye. With a steadily pressed viewing eye, and the fellow eye occluded, the occluded eye deviates while the pressed eye does not. If the eye is rapidly pressed and released, the compensation of rotational force in the pressed eye becomes incomplete, so that both eyes move. At high frequencies of press (greater than 1 Hz) the pressed eye is deviated and the contralateral eye no longer moves. In darkness the pressed eye always rotates but the contralateral eye never does, demonstrating that any inflow from proprioceptors sensing rotation of the pressed eye does not affect oculomotor posture as measured in the fellow eye. With binocular viewing the results are more variable. On some trials neither eye moves, while on others both move. The results can be interpreted as a Hering's law controlled attempting to reconcile disparate inputs from the two eyes. The results confirm and extend, with objective measures, earlier conclusions from subjective experiments.