Six strains of Haemophilus influenzae were distributed to 417 United Kingdom laboratories who were asked to test susceptibility of the strains to ampicillin, augmentin, tetracycline, chloramphenicol, and trimethoprim and to test for beta lactamase production. Laboratories were also asked to provide details of their methods by completing a questionnaire. The incidence of reports recording sensitive strains as resistant was 8% (ampicillin), 7% (augmentin), 3% (tetracycline), 1% (chloramphenicol), and 12% (trimethoprim). The incidence of reports recording resistant strains as sensitive was 9% (ampicillin), (2% with beta lactamase producing strains, 24% with non-beta lactamase producing strains), 51% (augmentin), 10% (tetracycline), 20% (chloramphenicol), and 3% (trimethoprim). High error rates were associated with several methods or practices. These included use of general purpose growth media rather than susceptibility testing media and failure to add lysed blood to the media when testing trimethoprim susceptibility; standardise the inoculum; use suitable control strains; and the use of high content discs for testing chloramphenicol, tetracycline, and ampicillin.