This article addresses Paul Churchland’s attempt to identify colors with surface reflectance spectra. Of particular concern is Churchland’s novel method of approximating surface reflectance spectra. While those approximations are generated by objective means and yield a striking match with human phenomenological color space, they are not physically meaningful. The reason for this is that the method used to produce the approximations induces equivalence classes on surface reflectances that are not invariant under physically appropriate changes of measurement convention. This result undermines (...) Churchland’s proposed response to the objection from metamers commonly raised against color physicalism, as his surface reflectance approximations are supposed to provide an objective, physical unifying basis for metamers. *Received December 2007; revised July 2009. †To contact the author, please write to: Department of Logic and Philosophy of Science, School of Social Sciences, 3151 Social Science Plaza A, University of California, Irvine, CA 92697; e‐mail: [email protected]. (shrink)