There are several asymmetries between the S-cones and the L- and M-cones (e.g. S-cones are much rarer; are absent from the central foveola; and have a peak spectral sensitivity well separated from the L and M peak sensitivities). But perhaps the most significant asymmetry is that, while the L- and M-cones have access to a variety of post-receptoral channels, the S-cone signal is largely confined to S-opponent pathways (e.g. Dacey & Lee, 1994; Sun et al., 2004). Only under special conditions is a small S-cone input to luminance observed psychophysically.

We exploited the functional asymmetries between S-cone vision and L- and M-cone vision to determine psychophysically the lights required to isolate the S-cone response. We evaluated three ways of finding the tritan line that depend on differences in: (i) spatial resolution; (ii) recovery from extinction of a yellow (-S) adapting field, and (iii) sensitivity changes caused by a blue (+S) adapting field. All three methods, if properly applied, converge to a common estimate of the tritan line for a given observer and retinal eccentricity. Using reaction time measurements and temporal masking, we estimated the latency of response to such tritan stimuli to be no more than 20-30 msec greater than latency of response to L/M-opponent stimuli, a result that agrees with earlier temporal phase measures (Stromeyer et al., 1991). Other evidence for larger differences may reflect luminance contamination in the L/M stimulus.

Tritan stimuli are particularly attractive in teasing apart visual processing, for they minimise contributions of the magnocellular (and parvocellular) pathways. Furthermore, since the S-cones exhibit little or no projection to superior colliculus, a brain area thought to be important for shifts of attention, tritan stimuli can be a powerful tool in psychological research (Sumner, Adamjee & Mollon, 2002).