The testing for truth or falsity of a clinical psychoanalytical theory, as distinct from seeking its confirmation and from applying it to non-clinical material, is illustrated by an examination of Forster's adaptation of Melville's novel for the opera Billy Budd. Forster's defensive response to the novel is treated as being formally equivalent to a homosexual analysand's response to the interpretation of his envious hatred for his homosexual love-objects; as such, his response of reversal, of hate into love and love into hate, is itself interpretable in terms of that interpretation. Wisdom's criterion for testing a clinical interpretation in the session in which it is made is extended to cover the use of non-clinical material for the testing of a clinical theory.