We identify and discuss the phenomenon of narrative stability, in the context of current methodological literature on interviewing. It derives from two independent studies, undertaken fifteen years apart, of members of the same genetics research group who were interviewed by different researchers. The first ('Discovery') interviews were collected very soon after the breakthrough was first published. The second ('Legacy') interviews were based on informants looking back at those events. Some strikingly similar narrative episodes across those accounts suggest strong narrative stability. In the course of interviews, informants reproduce biographical stories that are well sedimented. Not all interview materials, therefore, should be thought of exclusively in terms of co-production between interviewer and informant.