AIMS--To assess the best medium for primary isolation of enteric pathogens; to determine the need for a second primary culture medium; and to gather information on stool culture media used in 20 other randomly selected laboratories. METHODS--Specimens were cultured on desoxycholate citrate agar (DCA), Hektoen enteric agar (Hektoen), and xylose lysine deoxycholate agar (XLD). Non-lactose fermenters were screened with Rapidec Z (bio-Mérieux) and identified with API 10S (bio-Mérieux) where appropriate. Shigellas were identified with API 20E (bio-Mérieux) and serology, and salmonellas biochemically and by serology. A telephone survey was carried out to enquire into different culture practices and whether they had been evaluated for cost effectiveness. RESULTS--The isolation rate of enteric pathogens on primary stool culture media was 97% on DCA, 88% on XLD, and 76% on Hektoen. Seventeen of 18 shigellas grew on DCA, 13 of 18 on XLD, and 14 of 18 on Hektoen. DCA missed one Salmonella, XLD three, and Hektoen 13. XLD and Hektoen both missed Yersinia enterocolitica. The telephone survey revealed a diverse range of both primary and subculture plates. There was little evidence of evaluation of stool media, but firm personal convictions concerning the advantages and disadvantages of each type of medium at each stage of culture. CONCLUSIONS--DCA performed best and was the most cost effective of the three media. Neither XLD nor Hektoen were satisfactory as primary culture media because they grew fewer pathogens than DCA.