Abstract
A survey of non-circulating books was conducted at Long Beach (California) Public Library and Information Center between August 1987 and February 1989. Two computer lists were compared - one for books that had not circulated between 1981-87 (generated as one list) and one for books that had not been checked out between 1987-88. Data revealed that 8 percent of the books that were supposed to be on the library shelves were missing. This report discusses the methodology used and looks at the parameters that could account for the missing books - errors (both human and machine) and theft.