Recently it has been claimed that some early Renaissance painters used concave mirrors to project real inverted images onto their supports (paper, canvas, oak panel, …) which they then traced or painted over, and that this was a key source of an apparent increase in naturalism and realism in European painting around 1420. This bold theory makes implicit and explicit assumptions about the illumination and associated optical technology used for such projections. We compute and experimentally verify that the illumination requirements of the projection method are quite severe, and that in most cases subjects would have had to have been illuminated by direct sunlight, which seems unlikely for many specific paintings. We show how modern “re-enactments” of the theory's procedure in this regard are sometimes misleading or flawed, generally biased in favor of the theory. In certain versions, and for certain paintings, the theory also has testable implications for the color in final paintings. Through computer manipulation of digital images of key Renaissance paintings, we test informally whether it is faithful reproduction of form and contour or instead of color that best explains the naturalism in early Renaissance paintings, and conclude that it is subtleties in color. We demonstrate how the optical projection technique never aids in the accurate rendering of color, and in certain implementations severely impedes in the accurate rendering of color. Our analysis of color and illumination argues against the projection theory generally, and further supports conclusions from image analyses of specific paintings in the debate.