Trench fever, cause by Bartonella quintana, disappeared decades ago. Between 1992 and 1994 about 20 cases were reported in industrialized countries [6, 7]. Some of these cases were associated with HIV infection [5], but most affected the homeless. The epidemiology of B. quintana is unknown in Europe, but infection and transmission are favoured by poor housing and hygiene. This paper describes the results of a seroprevalence survey among the homeless living in Marseille. The eight major non-profit organizations caring for the homeless in the city participated for the four winter months. Four of these organizations also provide health care. Each subject who came to be housed for more than one day were asked to complete a questionnaire and give a small blood sample by micropuncture. They were clearly informed about the aims of the research and were allowed to refuse participation freely. All blood samples were tested by the WHO collaborative center for ricketssiae using a previously tested and validated immunofluorescence technique. The subjects were scored positive if the IgG titer was above 1/100. Two hundred and twenty one subjects agreed to participate in the study. Most were of French origin (51.4%) and 40% under 40 years old. The seroprevalence was 1.8% (IC 95%: 0.05-3.55). None of the four positive subjects presented symptoms and clinical examination was normal. All four were over 40 years old; two were of French origin, 1 from Algeria and 1 from the Comoro Islands. One had an IgG titer above 1/3, 200, possibly reflecting recent contamination and/or severe infection. Statistical analysis revealed no significant association between seropositivity and housing or hygiene conditions. No socio-demographic data are available about the homeless in Marseille and indeed, there is no list of such individuals. It is therefore not possible to estimate the representativity of our sample. Nevertheless, we show that the homeless include infected individuals. Physicians who care for such subjects should prescribe B. quintana serological investigations, and organizations that take in these people should implement preventive action and health education programs. Any such policy must be designed so as to be acceptable to homeless people. This work should also encourage public health professionals to investigate more extensively infectious diseases among the homeless.