If any single test of a scientific hypothesis contradicts it, the hypothesis must be revised. (Assume that you've ruled out errors in the testing process; that is, the test result really does contradict the hypothesis.)