Heterochromatic flicker photometry (HFP) is often regarded as setting the standard with which other methods of determining macular pigment optical density (MPOD) should be compared. We have developed a reflectometry method employing a retinal camera modified with multi-bandpass filters that provides the reflectance of the retina at a number of specific wavelengths. Based on a simple model of sequential absorbers (MP, cone and rod photopigments, and melanin), image analysis software is employed to derive the optical density distribution of the MP. We have tested the procedure on a group of 15 subjects in the age range 18 to 24, whose MPOD was also determined by HFP. We found a less than perfect correlation (R2 = 0.65, P < .001) between MPODs obtained by the two methods that led us to re-examine one of the basic assumptions of HFP. It is assumed that photoreceptor relative spectral sensitivity is the same for the foveal and parafoveal locations where measurements are made. Preliminary data indicate that this may not be a valid assumption. We modified the flicker photometer to determine the complete MPOD spectrum from 410 to 680 nm. The spectrum should be essentially flat with an OD of zero above ~ 540 nm, yet we found for the 3 subjects tested an apparent OD increasing with wavelength from zero at ~ 560 nm and, in one of the subjects, reaching a value of ~ 0.3 at 680 nm. This suggests that the LWS/MWS cone ratio may be significantly lower, at least in some subjects, in the fovea than in the parafovea. The result would be a decreased sensitivity to longer wavelengths in the fovea, equivalent to the effects of an absorbing pigment. It is possible that, from these measurements, we could determine the LWS/MWS cone ratio in the fovea relative to the parafovea and make appropriate corrections to the MPOD spectrum.