The application of computer graphics in archaeology is becoming more commonplace, allowing us a view of the past that might otherwise be difficult to appreciate. The potential of computer graphics as an aid to the archaeologist, both as a research tool and as a method of interpretive display, is now widely recognised. While this new perspective may enhance our understanding of past environments, if we are to avoid misleading impressions then the computer generated images should not only look ?real? but should accurately simulate the physical evidence recorded by the archaeologist. The current trend for artistic conception and photo-realism in reconstructions is not enough to benefit the archaeological community. As the world of televised archaeology and audio-visual museum displays draws the public deeper into an exploration of the past, high quality virtual scenes are demanded. For the archaeologist to use the computer-generated environments as a research tool, stricter controls are necessary. The ultimate goal is the creation of scenes that are perceptually indistinguishable from an actual scene. Physically and perceptually realistic graphics give a new perspective on past environments, allowing hypotheses to be tested in a safe and controlled manner. Realism owes much to the distribution of light in a scene, thus for the reconstruction of past environments we must take into account the illumination appropriate to the period. The bright, steady light that illuminates our world today has little resemblance to that of past environments. The correct spectral values of daylight and flame should be used to illuminate our models, allowing us to see the site or artefact under the conditions in which it was created and used. Current work at the University of Bristol involves the application of realistic lighting to reconstructions to provide models indistinguishable from a real scene, which are ultimately of greater use to the archaeological community.