That's my understanding as well. But Tringa (if I understand his posts correctly) was talking about metalizing the other surfaces. To me it seems that would make no difference, given that total internal reflection would have taken place off those surfaces ... but I'd be more than glad for someone more knowledgeable to explain why that may not be so.To my knowledge only one prism surface of Schmidt-Pechan prisms has to be supplied with a metal mirror, since the light rays do not enter under an angle of perfect reflection on that surface leading to light losses.