I agree with many of Avraham Sonenthal's points (The Scientist, July 7, 1997, page 10) regarding your recent article "To Effectively Discuss Evolution, First Define Theory" (R. Lewis, The Scientist, May 12, 1997, page 13). I disagree, however, that all science is experimental or that the extrapolation of evolution into the distant past is completely conjectural. Tests of theories can be observational and employ not just predictions but also retrodictions. Retrodictions are extrapolations into the past about things that should be (but have not yet been) found in the geological, paleontological, or archeological record if a theory is correct.
Darwin, for example, retrodicted that many intermediate forms would be found in the fossil record linking human beings and other primates and that similar intermediates would be found linking modern horses with primitive mammals. We must remember that, at the time, no such fossils were known. Obviously, these retrodictions have proved ...
