A 5'-nuclease (TaqMan) assay was evaluated for its capability to recover and detect stressed Yersinia enterocolitica. Sensitivity studies of a 5'-nuclease assay for detecting Y. enterocolitica 0:8 in a pure culture system and spiked ground pork samples demonstrated that the assay has reliable sensitivity with a detection limit of 3 to 4 log CFU/ml or CFU/g. The PCR 5'-nuclease (TaqMan) assay was evaluated with the Thin Agar Layer Oxyrase method (TALO, overlaying 14 ml of Trypticase soy agar with a 1:30 dilution of "Oxyrase for Agar" onto a prepoured pathogen-specific, selective medium), and it was compared against the selective medium cefsulodin-irgasan-novobiocin (CIN) for recovering and detecting Y. enterocolitica from inoculated nonfrozen and frozen (-15 degrees C, 2 days) ground pork samples. The TALO method showed more sensitivity (detection limit, 2 log CFU/ml), and it has greater recovery capability (0.5 to 1 log CFU/ml) than CIN (P < 0.05). The 5'-nuclease assay provided rapid detection processing (5 versus 24 h after an 18-h enrichment). The sensitivity per PCR was calculated to as low as 0 to 1 log CFU per PCR reaction; however, in the method's current developmental stage, target pathogens should be enriched to 3 to 4 log CFU/ml or CFU/g to show consistent results. In a survey of 100 ground pork samples using TALO, CIN, and PCR methods, no Y. enterocolitica was recovered. A combined cultivation and an automated PCR TaqMan could be used as a presumptive screening test for detecting Y. enterocolitica in food samples.