Figure 9. Derived rod sensitivities in darkness and on four backgrounds, for one subject.

A, the derived rod response for subject CF, obtained as in Figs 7 and 8, has been divided by test flash intensity to yield the response per unit intensity, in units of μV Td⁻¹ s⁻¹. B, these results have been further normalized with respect to the probe-alone response R_alone, to yield the fractional response per unit intensity, in units of Td⁻¹ s⁻¹. This normalization means that the response is expressed as the fractional suppression of the existing circulating current in the steady state. The experiment was performed over two days, with three of the backgrounds studied on both days; the background intensity, test intensity and peak response amplitude corresponding to the different symbols are tabulated below. Each point was obtained from typically six test-probe presentations, with the probe-alone response obtained from typically 10 flashes. As in Fig. 8B, the points have been averaged over a measurement window from 4.6 to 7.4 ms. In B the points for the middle three sets of data (♦, ▴, ▿) were calculated as the weighted means of the points plotted in A obtained on the two days, as indicated by the groupings tabulated below. The curves plotted in B are eqn (6) with the parabolic rise of eqn (7b), using A= 4.3 s⁻² (the average amplification constant found in fitting the a-wave families) and t_d= 2.7 ms, and with the decay parameters t₀ and τ_rec in eqn (6b) set as below, with n = 8.