Objective: Our objective was to develop an efficient way to search for radiology information on the World-Wide Web (WWW)., Materials and Methods: With the Harvest Information Discovery and Access System and a DECstation 3000 computer, each week we gather and update WWW documents that relate to radiology. To limit the size of the gathered documents, certain data (such as image files) are omitted. To date, we have gathered from 20 main radiology sites. By accessing our server, thousands of documents can be efficiently searched for specific radiology information., Results: At no charge, individuals with access to the WWW can request a search of our server for specific radiology information. We return to the requester in HyperText Markup Language a document with results of the search included as embedded pointers. Our server address, or uniform resource locator, is http://glimpse.cs.arizona.edu/radiology.html., Conclusion: Use of a server dedicated to radiology is an efficient way to search the internet for radiology resources, such as teaching files. We provide access to many sites with one search, and we select those sites to ensure that only relevant and timely information is gathered. We also index that information.