In this presentation I will demonstrate how stories are elicited and told during the police interrogation and what happens in the process when the suspect's story is written up. Producing the story on paper is one of the goals of the police interrogation: to write up a report, as much as possible in the suspect's own words, about the events reported in the interrogation. It is the story in this report that matters in the further judicial process (cf. Komter, 2006; Rock, 2001) where it can serve as evidence in court. After the police officer (P) has solicited the story, there are three different ways in which the stories are told in the interrogations that I recorded. First of all, the suspect (S) produces a longer turn at talk in which his story is told from beginning to end (this is what I call the "free story"). Secondly, S starts telling a story, but is soon interrupted by P. By interrupting and asking further questions, P steers the storytelling ("supervised story"). The third way in which stories are told is that they are not told by the suspect himself, but "imposed" by the officer. Either S refuses to tell a story, or the story is not the "correct" story according to P. The officer then verbalizes a version of the story which the suspect agrees or disagrees with. The three ways in which stories are solicited, told and responded all lead to a written version of events that happened. I use a collection of 11 interrogations that I recorded in the Netherlands and their police records to look at "what happened." This data provides us with two different versions, or 'generations' (Jönsson & Linell, 1991) of the same story: a spoken and a written version. With my data I can not only show how the stories are elicited and told and what elements are transformed in the written version, but I am also able to demonstrate how these transformations take place and how the interaction plays a role in constructing the second generation story.