I disagree that all evidence must be reproducible, which is the standard for experimental evidence (simple physics comes to mind). Just off the top of my head,

I disagree that all evidence must be reproducible, which is the standard for experimental evidence (simple physics comes to mind). Just off the top of my head, there is historical evidence and geological evidence, for example. Historical evidence (whether documentary or archaeological or otherwise) is certainly producible, but not reproducible in the sense that you can provide the exact same thing again. Geological evidence (again, very simply, like rock strata) is not reproducible. Even saying that this kind of formation always accompanies another kind isn't reproducible. (The "all ravens are black" problem.") There are other logical kinds of evidence that don't meet the "scientific" standard, and yet are quite valid. My objection really is that the original statement was far too narrow.