1. The importance of synchronous activation in maintaining cortical binocularity was studied physiologically in kittens that had been reared under different regimens of alternating monocular deprivation. 2. Three different techniques were employed to provide alternate monocular stimulation: a) mechanical shutters placed before the animals' eyes; b) goggles fitted with complementary colored cutoff filters, which restricted visual input to one eye at a time; and c) two rotating gratings that were 90 degrees out of phase. In the third technique, the gratings were always orthogonal to one another and viewed separately through cutoff filters. This allowed us to exploit the orientation selectivity of cortical cells and thereby stimulate them alternately through each eye without simultaneously affecting activity in the dorsal lateral geniculate nucleus (dLGN). 3. We based our conclusions on a sample of 691 neurons, which we recorded in 21 animals. Results with all techniques were remarkably consistent. Binocular cortical inputs predominated at normal or nearly normal levels, even when a number of seconds elapsed between successive exposures of each eye. 4. An interonset interval of at least 10 s was required to make a substantial reduction in binocularity. This interval can be separated into two parts--the duration of exclusive monocular stimulation and the time when neither channel receives input. Of these, the latter appeared to be less important. Blanking times of 0.15--1.0 s did not affect binocularity if the interonset interval was 1 or 10 s; and in one experiment where the blanking time was 9 s, the resulting disruption in binocularity was less than that found with shorter blanking times and the same interonset interval. 5. Our results imply that mechanisms responsible for the disappearance of binocular cortical inputs require independent stimulation of each eye for periods of at least a few seconds; this stimulation must be of a kind that is known to excite cortical cells. Our results with the rotating grafting show, in addition, that the mechanisms whose timing we have measured are intrinsic to the cortex.