The format for the analysis is adopted from Hawkins and Tiedman (1975), who have described the process by which acts or attributes come to be recognized and possibly reacted to as deviance. The act is first observed and then recognized as rule violation; causes are imputed to the act and motives to the actor; the observer then considers potential reactions to the violation, and in the context of his/her interests, chooses one of the possible reactions. Finally, this choice has a particular impact on the rule-violator. Since this model asserts the primacy of recognition over reaction in the deviance-defining process, the model is used in the discussion of book theft and mutilation in academic libraries, where the norm violation is dealt with 'by avoiding formal, unambiguously single-minded social control activities' (Katz, 1977). It appears that library security, i.e., the responsibility for protecting the collection and ascribing deviance to users, ultimately falls on the librarian. Wherever possible, unobtrusive measures are used to protect the library's holdings so as not to make security measures a dominant factor in the libraries procedures and environment. When prevention fails, librarians are loath to pursue suspected offenders because of the difficulty in proving malicious intent and the various consequences of attempting to do so. The bureaucratic context of the librarian/user interaction is concluded to minimize the chances of open conflict, and aware of the fact that the public takes a casual attitude toward institutional theft, the librarian is unlikely to do more than classify the act as 'theft assumed.' The analysis concludes that this act of classifying the incident constitutes both its explanation and its resolution. Sixty bibliographic entries are provided.