An experiment was conducted based on an information service for planners run in conjunction with Geo Abstracts Limited. Forth-eight styles of abstracting journals were produced which, although based on the same source material, differed in depth of coverage, frequency of publication, content of entry, and whether an index was included. One style of journal was produced with each possible combination of these parameters. Each style was sent for a trial period to four users who marked items not noted and followed up from each issue and completed a questionnaire at the end of the experiment. The results show a distinction between expressed preferences on the parameters studied and the styles which proved most useful in practice. In particular, although most users said they would prefer abstracts, the citation-only styles were most useful. The bulk of a publication appears to be an important factor in determining usability. The experiment also demonstrated a demand for a much wider range of information from researchers and teachers than from practitioners. (Author/SL)