(1972) was followed in all essential aspects; incubation was at 37°C in an atmosphere of H2 90% and COZ 10%. A slope of Simmons citrate medium seeded with Pseudomonas aeruginosa was included in each jar as a control.Specimens. Sterile toothpicks were used to collect samples of subgingival plaque from 20 normal healthy adults. The subjects were medical students aged 19-20 years; none had any gross oral or dental pathology and none was receiving antibiotic therapy.Isolation of Bacteroides. The methods and media were derived from those of Baird-Parker (1957), Loesche, Hockett and Syed (1971), Syed and Loesche (1973) and Williams et al. (1975); they were evaluated by Holbrook (1976) and Holbrook et al. (1978). The samples of plaque were seeded directly on to a sector of a plate of pre-reduced BM agar (see Williams et al., 1975) with kanamycin 75 pg/ml and vancomycin 2.5 pg/ml and the inoculum was streaked over the remainder of the plate. The plates were examined after anaerobic incubation for 48 h; all colony types were noted and representative colonies were subcultured from each specimen on to plain BM agar. After incubation for a further 48 h, any additional colony types were noted and further representative colonies were subcultured to a total of 10 from each specimen. The colony types were selected in approximate proportions to their comparative numbers on the primary isolation plate. The isolates were suspended in a freezing medium of Nutrient Broth No. 2 (Oxoid) with inactivated horse serum (Wellcome) 10% and glucose 1% and held at -70°C in a liquid-nitrogen container.IdentiJcation of isolates. The isolates were identified by the methods of Duerden et af. (1976, 1980). The tests were: colony morphology after 48-h incubation on blood agar; cell morphology in gram-stained smears from 48-h cultures on blood agar and in BM broth with cooked-meat particles (see Deacon, Duerden and Holbrook, 1978); pigment production on BM agar; haemolysis on human blood agar; motility in BM broth; antibiotic-disk resistance tests with neomycin 1000 pg, kanamycin 1000 pg, penicillin 2 units and rifampicin 15 pg disks; tolerance tests with taurocholate, deoxycholate, Victoria blue 4R and ethyl violet; biochemical tests for the production of indole, digestion of gelatin and hydrolysis of aesculin; fermentation tests with glucose, lactose, sucrose, rhamnose, trehalose, mannitol and xylose.

The need for chemoprophylaxis for bacterial endocarditis is partly dependent on the risk of bacteraemia associated with the procedure, which has not been adequately defined for skin surgery. The incidence of postoperative bacteraemia in 149 immunocompetent out-patients with non-infected lesions was 0.7% (95% CI 0.3-3.8%). Procedures included excisions, flaps, grafts and micrographically controlled surgery. Coagulase-negative staphylococcus was the most common skin isolate at the site of surgery, present in 68.5% of patients. The most effective chemoprophylaxis would be intravenous vancomycin, which is inconvenient and has an inherent risk of morbidity. Given the low incidence of bacteraemia and the disadvantages of the optimum chemoprophylaxis, surgery on non-infected lesions does not warrant prophylactic antibiotics to prevent the very low risk of bacterial endocarditis.

A Sudanese girl became desperately ill with liver and kidney abscesses due to Nocardia asteroides. She did not have pulmonary or cutaneous infection. She recovered after surgical drainage of the abscesses and prolonged treatment with intravenous amikacin and high dosage cotrimoxazole and sulphadimidine. After recovery normal neutrophil function, cell-mediated and humoral immunity were demonstrated.

The need for chemoprophylaxis for bacterial endocarditis is partly dependent on the risk of bacteraemia associated with the procedure, which has not been adequately defined for skin surgery. The incidence of postoperative bacteraemia in 149 immunocompetent out-patients with non-infected lesions was 0.7% (95% CI 0.3-3.8%). Procedures included excisions, flaps, grafts and micrographically controlled surgery. Coagulase-negative staphylococcus was the most common skin isolate at the site of surgery, present in 68.5% of patients. The most effective chemoprophylaxis would be intravenous vancomycin, which is inconvenient and has an inherent risk of morbidity. Given the low incidence of bacteraemia and the disadvantages of the optimum chemoprophylaxis, surgery on non-infected lesions does not warrant prophylactic antibiotics to prevent the very low risk of bacterial endocarditis.