An interactive recreation of the discrimination paradigm used to screen for candidate tetrachromacy in Jordan & Mollon's studies of women heterozygous for anomalous cone photopigments — read the caveat at the bottom before drawing any conclusions from your own result.
The visible spectrum from 400–700 nm, with the three wavelengths this test uses marked: green (546), yellow (590), and red (670).
The excluded swatch (labeled below) is identical RGB to the mixture, a screen limitation — real 546/590/670 nm light wouldn't collapse this way.
Three flashes per trial — two are 590 nm yellow, one is this mixture. Pick the one that looked different. 24 trials, under 2 minutes.
24 trials here is 6 brightness levels × 4 repetitions, swept around the crossover ratio — a demo-sized simplification. The original study tested 10 ratios × 9 luminances × 3 blocks, 270 trials per participant, after practice runs. The level that would land exactly on the crossover is deliberately left out: at that one brightness, this screen renders the mixture and the yellow as the literal same RGB value, so no one — trichromat, tetrachromat, anyone — could ever get it right. That's not a metameric match the way a real colorimeter produces it, just one stimulus rendered twice, so including it would only test whether you can find a nonexistent difference. Because that level is excluded, both a trichromat and a real tetrachromat are predicted to score near 100% on the remaining six — the one condition that would actually separate them isn't part of this sweep. The ratio itself is fixed at this system's one true crossover point, since only that ratio makes the swatches above nearly identical — at any other ratio, no amount of brightness adjustment would close the gap, because the mixture and the yellow trace different lines through color space. Flashes run 150 ms each with a 300 ms gap, matching the original study's timing. This all runs on an RGB screen, which is itself a three-primary system — it cannot present the independent four-dimensional stimuli a real tetrachromacy test requires, which is why the original studies used a custom colorimeter with genuine spectral lights instead of any display. It's built to make the logic of the discrimination task easier to follow, not to diagnose anyone. In the original work, only one of the 25 obligate carriers tested performed the way a true tetrachromat was predicted to.
Inspired by "What Would It Mean to See a New Color?", The New Yorker.