Photopigment regeneration after bleaching can reveal much about retinal function in health and disease. Recently a “rate-limited” model has been proposed, whereby 11-cis retinal diffuses into photoreceptor outer segments through a resistive barrier from a constant pool in the pigment epithelium, resulting in regeneration proceeding linearly with time rather than as an exponential (Lamb & Pugh, 2004; Mahroo & Lamb, 2004). We tested this hypothesis for cone pigment regeneration following very intense bleaches, posing two questions. Does regeneration follow a linear rate after intense bleaches? If so, is the rate the same as for smaller bleaches, as the model predicts? We used a conductive fibre electrode to record electroretinogram photopic a-wave responses to red flashes (0.4 photopic cd m^-2 s) following one-minute bleaching exposures (11 000 – 130 000 photopic cd m^-2) in two normal subjects with dilated pupils. A blue background (40 scotopic cd m^-2), present throughout, eliminated rod signals. Post-bleach response amplitudes were used to estimate pigment regeneration using a transformation published previously (Mahroo & Lamb 2004). Recoveries proceeded according to a linear rate: first order exponentials gave strikingly poorer fits than the “rate-limited” model (see Fig. 1). However, the rate was around 30% slower than that obtained previously from the same subjects following less intense bleaches (Mahroo & Lamb, 2004), suggesting that the model needs modification. Cone pigment appears to regenerate more slowly following very intense bleaches. This may indicate a reduction in the 11-cis retinal pool available, shedding new light on retinal mechanisms after exposure to intense illumination. Figure 1. Cone pigment recovery following intense bleach Fifteen dim red flashes (0.40 cd m^-2 s at 0.5 s intervals) were presented every 10 s after 1 min exposure to yellow light of 40,000 photopic cd m^-2. Response amplitudes, measured 14-15 ms after each flash, were used to estimate pigment level (Mahroo & Lamb, 2004, Eqn (9) and (10)). Points plot mean ± S.E.M, over 20 s windows, with results from six exposures, so each point averages c.180 flash presentations. Panels fit recovery for a 100% bleach with different models: points differ slightly between panels as fits give different estimated dark-adapted amplitudes, affecting normalization. A, Curve shows best-fitting form of recovery as a single exponential, with time constant of 2.2 min. B, Solid curve: best-fitting recovery according to rate-limited model, giving parameters K_m = 0.2 and rate v = 0.33 min^-1. Dashed curve: expected recovery if v is 0.5 min^-1, which fitted recoveries in this subject from less intense exposures (Mahroo & Lamb, 2004).