Gog, this is true, however they are talking pure metals here. @kemal try with a white background, no sun and a single omni light instead. This way you are insured that there are no variables other than the refraction. I have tried this in the past and could see no difference between refraction indexes on solid objects with no transparency. As I stated earlier, bryce refraction functions solely on transparent objects. Bryce does not support reflecting light or true caustics because it lacks any significant photon simulation engine. Reflecting light and caustics are created by reflective and refractive properties. Let's start with caustics. When a light goes through an object it is bent due to the refractive index, we've been over this. This is only incident on the surface of the object unless the refraction changes throughout the object itself. While a beam of light is inside the object and otuside the object it remains straight. Let's take a sphere for example. The light enters at say 45 degrees.. it is then bent (for example's sake) 20 more degrees. (this is why a straw looks like it's bent when you put it in a glass of water) anyway, now the beam of light exits the sphere at a new angle due to the incident with both the front and the back face of the sphere. The light has not passed straight through but has instead been redirected, in the case of a sphere, all light entering is slightly bent towards the center upon exit, This is why the caustics effect of a transparent sphere with relatively high refraction is a bright dot. When bouncing off of the surface of a pool light is gathered in a different effect. But anyway, all caustics is is a non uniform distribution of light brought about by refractive surfaces. Light bouncing off a mirror is like a simplified version of the above. Unfortunately bryce only supports direct illumination (disregarding ta, which really has nothing to do with illumination at all, but with the proximity of ambient surfaces) This effect would again be predicted with photons. Think of photons as light particles. they bounce off of reflective surfaces and stop on opaque surfaces, tracing a line behind them. Now take a million of these and throw them out from a light.. this is what photon simulation does. The photons then would bounce off of the reflective surface at the opposite interior angle to the incident beam. once they have reached the opaque surface that stops them they illuminate it. Bryce lacks such an engine, thus anything like reflecting light off of mirrors or creating caustics (without faking it) is indeed impossible.