On the analysis of nerve signals deduced from metacontrast experiments with human observers.

1. This paper reviews Alpern, Rushton & Torii's (1970a-d) derivation of the size of the inhibitory nerve signal arising from after flashes in the metacontrast experiment. 2. Their geometric argument is recast in terms of simple functional equations. This form of argument clearly displays the role of their assumptions in obtaining their main conclusion: nerve signal is linear in intensity over a range of 3-4 log units. 3. Two disadvantages of their approach are discussed. First, it is noted that in the presence of the data the assumption they employ in their analysis is logically equivalent to their conclusion. 4. Secondly, accepting their claim that the nerve signal generated by the after flash is linear over a broad range of intensities, and that this inhibitory signal simply cancels the excitatory signal of the test flash, leads to the conslusion that over this same intensity range the excitatory nerve signal is a power function with an exponent of close to two. This is incompatible with the suggestion that photoreceptor signals have been measured.