All programming was performed using the commercially available Quintus Prolog system (Release 2.4.2). General details of the Prolog language may be found in the book Programming in Prolog (Clocksin, W. F. and Mellish, C. S, Springer-Verlag, FRG, 1981). Briefly, Prolog stores facts and deductive rules in a database, where the facts are equivalent to a collection of relations in a relational database. The Prolog interpreter when presented with a query (or goal as it is often called) searches the database of facts and rules to see if any of the facts and rules match the query. If there are facts that match, then the variables in the query are unified with the constant symbols in the fact(s). If there is a rule whose consequent (the head of the rule) matches the query, then the interpreter proceeds by attempting to recursively prove the subgoals of the rule. Specific examples relevant to the protein structure information stored in our database are described below. The database system was originally developed on a Sun 3/50 workstation with 4MBytes of memory. The current installation is running on a Sun SPARCstation 1 with 12MBytes of memory.