Deception research has neglected the fact that legal-workers often have to try to detect deceit on the basis of statements derived from pairs of suspects, each having been interrogated repeatedly. To remedy this shortcoming we conducted a study where each memeber of 10 truth-telling pairs and 10 lying pairs was iterrogated twice about an alibi. One hundred and twenty undergraduate students were enrolled as lie-catchers. The main fidings were that (a) overall deception detection accuracy was modest; (b) lie-catchers given access to a large number of statements did not outperform lie-catchers given access to a lesser number of statements; (c) when asked to justify their veracity assessments the most frequently reported cue was 'consistency within pairs of suspects': (d) all cues to deception were of low diagnostic value. Psycho-legal aspects of integrating sequential information in deception detection context are discussed.