According to some, the historian must for working purposes assume that nature is uniform, i.e., that miracles do not occur. For otherwise, it is suggested, he may place no confidence in the historical reliability of the records and artifacts on which he relies: such confidence can exist only where it is assumed, for example, that ink marks in the form of words do not sometimes appear spontaneously on old bits of paper. In this article I spell out this methodological thesis (...) in greater detail and determine to what extent it should be accepted. (shrink)