For several months in 1986, an outbreak of Streptococcus viridans pseudobacteremia occurred at a large teaching hospital. All sources of laboratory blood culture contamination were excluded. A retrospective epidemiological study indicated that one phlebotomist, “P,” collected a disproportionate number of the positive blood cultures. Further comparison of the paired blood culture results from the three months when the incidence was highest revealed a good concordance of results among all other phlebotomists (Kappa = 0.5), while P's results concurred with others less frequently than would be expected even by chance (Kappa < 0.0). Clinical follow-up showed that P did not routinely wear gloves while drawing blood and had eczema of the hands. Skin scrapings from the hands, right index finger/fingernail grew predominantly S viridans species that were compatible with those recovered from contaminated blood cultures. This epidemic demonstrated the need for early detection of this source as a cause of nosocomial pseudobacteremia.