Accepted for publication September 1995 There is an extensive number of yearly publications relevant to socioeconomic inequalities in health. It is not easy to gain a complete overview of this literature if relying solely on available computerised literature databases such as Medline. Firstly, a comprehensive overview requires the use of many databases with different specialisations, given the multidisciplinary character of this research area. Secondly, a general database does not allow for a detailed reference system, which necessarily limits the possibilities for a good selection of relevant literature. For this reason, a specialised documentation centre was set up in The Netherlands. The centre now offers a collection of more than four thousand publications, which includes not only journal articles but also book and reports. The coverage of literature published since 1985 is almost complete, and a selection of major pre-1985 publications is also available. A computerised system contains bibliographic data on all the literature. The emphasis is on publications written in English, but the database also contains material written in Dutch, French, and German. New titles are identified by regularly reviewing several computerized databases, scientific journals, bibliographies, congress proceedings etc. Supplements of accessions are issued quarterly. The centre operates a specially developed automatic reference system. A minimum of eight keywords are assigned to each publication. These keywords can be used to identify the character of the publication (for example, a report on a longitudinal study) and the way socioeconomic status and health is indicated (such as occupational status and mortality from cancer). They can also be used to identify the explanatory mechanism studied (such as smoking), as well as the country/period in which the publication is set. By means of these keywords, the compilation of a bibliography on a certain subject is relatively easy, and may be done on request. Moreover, the database can be used for a quantitative description of scientific developments. This may be illustrated by the following example. The number ofpublications on material/structural and behavioural/cultural explanations respectively was compared, using a selection of publications relating to empirical studies in the United Kingdom, published since 1985. This selection elicited 40 publications on the socioeconomic distribution of structural factors, whereas for behavioural factors the corresponding number was 89. These kind of descriptions may be helpful to identify gaps in scientific knowledge.' We hope the documentation centre will be useful to researchers, policy makers and other parties interested in this subject throughout Europe, and we welcome any request for information, as well as the forwarding of relevant publications.