Main point: Threshold experiments have led to the view that visual information passes through channels selective to orientation, spatial frequency, and spatial location. We present here observations that are difficult to understand in that view. Reasoning: As the duration of a flashed grating increases, its contrast threshold decreases and its apparent contrast increases, putatively reflecting the growth of excitation in the channels sensitive to the grating. Shifting the orientation of the grating by 90í in the middle of that growth should allow excitation to decay in the previously excited channels and to rise anew in channels sensitive to the new orientation. At any given time, the excitation should lag behind the levels it would have reached if the grating's orientation had not changed (unless excitation is asymptotic). The shift should raise contrast thresholds and lower apparent contrast. We show that this is true for thresholds, but not true for the apparent contrast of suprathreshold gratings. We observe the same pattern of results when the phase of the grating is shifted by 180í, and when two different random noise patterns or two different natural images are exchanged during this period of growing excitation (thereby changing the spatial distribution of energy, the phase spectrum, and, to some extent, the energy distribution in 2D Fourier space). Conclusions: (1) growth of apparent contrast, but not contrast threshold, is independent of changes of orientation, phase, and spatial distribution during short presentations; (2) the concept of channels complicates explanation of these particular suprathreshold phenomena and gives impetus to the search a different conceptual model.