In a finitely repeated principal agent relationship with adverse selection I study (exogenous) interim information that is revealed during a long term relationship. Interim information mitigates adverse selection. Verifiability, measured by the cost of signal manipulation, and the signal's informativeness determine the use and effectiveness of interim information: Less precise and more manipulable signals are used in a forward looking way exclusively. More precise and less manipulable signals are also used in a backward looking way and extract all information rents. Highly precise signals with a high degree of verifiability yield the first best. Moreover, verifiability and informativeness are substitutes.