Spectral sensitivity in the red-green spectral range typically reflects the joint influence of the middle-wavelength-sensitive cones (the M or green cones) and long-wavelength-sensitive cones (the L or red cones). The balance of M- and L-cone influence can be altered by presenting the test lights superimposed upon steady background fields of long or short wavelength. We find that presenting test stimuli just after an abrupt exchange between two colored backgrounds permits an easier and closer approach to cone isolation than presenting them either on a steady background or following an intense bleach. Background exchange drives the flicker detection or flicker photometric spectral sensitivities measured at 17 Hz to a limiting condition at lower intensities than do steady backgrounds. This condition is consistent with either M- or L-cone isolation. Steady backgrounds do not produce complete cone isolation: even on backgrounds that push spectral sensitivity closest to M or L, there are substantial phase differences between flickering lights of different color. In contrast, no phase differences remain following background exchange. The improvement in cone isolation produced by the exchange procedure is not confined to flicker measurements: the spectral range over which subjects are temporarily monochromatic is more extended following background exchange than on steady fields.