An improved laboratory system for testing the efficacy of potential broiler carcass disinfectants against attached Salmonella was developed. Breast skin from freshly processed chilled broilers was cut into 10 cm diameter pieces, irradiated to inactivate vegetative microflora and served as the attachment surface. Salmonella typhimurium and Salmonella montevideo were grown in brain heart infusion (BHI) broth at 23 or 37 degrees C, diluted in 0.1 M potassium phosphate buffer (KPB), and inoculated onto skin samples as a coarse aerosol spray. At 10, 20 or 30 min post inoculation, skins were vigorously rinsed twice with phosphate buffer (PB) to remove ''loose'' cells. To enumerate salmonellae remaining on the skin (''attached'' cells), skins were blended with fresh PB and plated onto tryptic soy agar (TSA). Applying 10(2), 10(3) or 10(4) cells/skin consistently resulted in attachment of 23 to 44% of the applied inoculum. Culturing temperature and serotype did not affect attachment. The described skin attachment model (SAM) provides a reliable and repeatable procedure for studying the efficacy of various treatments for removing or inactivating enteropathogens that are attached to poultry skin. The SAM allows for consistency in testing of antimicrobial agents against attached bacteria and enables utilization of specific bacteria without interference from background microflora using non-selective recovery techniques.