The background and goals of a national study to determine the feasibility of blind proficiency testing in U.S. forensic DNA laboratories are discussed. Part of the project involved designing and executing a series of fifteen blind proficiency tests. Execution included biological specimen donor recruitment and case evidence manufacturing. Simulated cases were submitted to DNA laboratories by law enforcement agencies and in some cases by other forensic-science laboratories. Replicate-manufactured evidence was submitted to reference laboratories to simulate the workings of a larger-scale program. Ten tests were straightforward, and essentially tested analytical ability. Five tests involved selecting on the basis of case facts appropriate bloodstains for typing from a bloodstain pattern. We describe in detail our experience in designing and conducting these blind proficiency test trials, and relate those experiences to the overall issue of blind proficiency testing as a quality-assurance tool in forensic DNA laboratories. In this feasibility test series, one blind test was detected by a laboratory, a second one was shown to the lab by law enforcement, and a third was never completed because of lapses in communication. Turnaround times were relatively fast in the independent/commercial labs and relatively slow in the larger public laboratories. Two cross-state case-to-case CODIS "hits" were "planted" among the first series of ten blind tests. One pair was detected. One member of the second pair went to a lab that was not CODIS-ready.