The aim of this study was to find out if hypertext structure has any relevance for the efficiency of information retrieval. An experiment, in which three variants of hypertext structure were compared, was carried out in the autumn of 1998. Three websites with different structures were constructed for this purpose: a sequential version, a hierarchical version and a webcluster version. 44 subjects, all students at the University of Lund, were randomly divided into three groups. Subjects were requiered to answer a number of questions using one of the three versions.