An experiment was performed to test the capability of measures of gamut area to predict the color discrimination capability of lamps with highly structured spectra. Sixteen subjects completed a Farnsworth Munsell 100 Hue Test under four different lighting conditions of similar chromaticity: light emitting diodes (LEDs) that used red, green, and blue primaries; two types of linear fluorescent; and tungsten halogen. Total Error Score (TES) was significantly worse under the LEDs than the other three conditions, which produced statistically equivalent scores. No measure considered in this work, including Color Rendering Index, Color-Discrimination Index, or Farnsworth Munsell Gamut Area, was able to order correctly, let alone correlate with, the color discrimination capability of the four lamps, as characterized by TES. This calls into question the concept of using a measure of gamut area to characterize color discrimination as part of a multi-index tool for quantifying color rendition, especially since lamps with highly structured spectral output are becoming more prevalent for general illumination. (c) 2011 Wiley Periodicals, Inc. Col Res Appl, 2011;