Additivity failures in heterochromatic brightness matching

The often poor correlation between luminance and perceived brightness has been attributed to the facts that first, the CIE standard luminous efficiency function V(Λ) does not represent adequately heterochromatic brightness matching of monochromatic stimuli, and second, the additivity law does not always hold in heterochromatic brightness matching. By means of a variation of the traditional experimental procedures, Guth found that two types of additivity failure can occur: additivity failures of the enhancement type and of the cancellation type. It is shown here that the two types of additivity failure can also be deduced by calculation from brightness‐matching data obtained by the traditional experimental procedure. However, the magnitude of the additivity failures and sometimes the type of failure seem to depend strongly upon the individual observers involved. New experimental data confirm Guth's results except in the middle region of the spectrum. Whereas Guth's observers find additivity failures of the cancellation type between 500 and 600 nm, the new data suggest that in that region additivity holds fairly well.