The bias and precision associated with five new in/face respirator sampling procedures was evaluated. The procedures evaluated consisted of a continuous, low sampling rate flush on the respirator midnose and mouth probing (CLF) procedure; a continuous, high sampling rate deep front of mouth probing (CHD) procedure; a pushed exhalation, deep front of mouth probing (PED) procedure; an exhalation valve discharge (EVD) procedure; and a pulsed inhalation, deep front of mouth pushing (PID) procedure. The CLF procedure was the recommended methodology for sampling respirator facepieces in the United States. The procedures were evaluated using nine full facepieces and five half facepieces. All facepieces were normally equipped with organic vapor cartridges and were NIOSH certified. The full facepieces did not have nose cups. The facepieces were punctured at three sites to generate leaks corresponding to penetrations of 1 or 0.1% and mounted on a specially designed respirator/simulated breathing test system. Acetone vapor was used as the test agent. The amounts of acetone vapor passing through the leaks were measured using the CLF, CHD, PED, EVD, and PID. The bias for each procedure was determined as the difference between the true acetone concentration and the concentrations measured inside the facepieces. Precision was estimated using a modified analysis of variance technique. The average sampling biases on the full facepieces obtained with the CLF, CHD, PED, EVD, and PID procedures were -21, -3, 0.7, -14, and -12.3%, respectively. The corresponding biases on the half facepieces were -26, -13, -4, -2, and -24%. The precision of the measurements obtained on the full facepieces by the CLF, CHD, PED, EVD, and PID procedures were 14.0, 11.8, 5.0, 11.9, and 21.0%, respectively. The values obtained on the half facepieces were 14.0, 11.7, 5.8, 3.0, and 29.2%, respectively. The authors conclude that based on the results obtained with the full facepieces the PED is the best procedure and the CLF procedure the worst.