A note on ray tracing mirages (comments and author's reply)

The author comments on an article by M. Berger et al. (ibid., vol.10, no.3, p.36-41, 1990), which describes a model for forming a mirage. In particular, he makes some points regarding their interpretation of the model and their description of their implementation of a version of it. He points out that the primary bending agent in mirage formation is total reflection, not refraction, as Berger et al. suggest. This in turn indicates that a purely reflective model, without refraction might well be sufficient. The author asserts that this would be beneficial, as dispersive ray tracing is not particularly easy or inexpensive to implement, and it would obviate the statement by Berger et al. that the use of layers allows a digitized approach to the continuous spectrum. A response by M. Berger, essentially agreeing with the author, is included.