The purpose of this thesis is to investigate how the metadata format Dublin Core is applied within five Swedish search engines. The objects of the investigation are Rättskällan, studera.nu, Svenska miljönätet, Svesök and the web catalogue of Uppsala University. Dublin Core is a flexible and adjustable format, which can be used in many different ways and in various projects depending on the target groups and objectives of the activity. We compare the differences and similarities between the five search engines concerning their usage of metadata in general and Dublin Core in particular. The material of the investigation consists of metadata records, which are collected by a robot and examined manually. We present the reader with an overview of the metadata concept and explain in particular the Dublin Core Metadata Element Set. The review of the investigation result begins with a presentation of the five search engines and is concluded by an exposition of the usage frequency of the fifteen Dublin Core elements within the examined search engines. The five search engines show notable differences with respect to the usage of Dublin Core. The number of Dublin Core elements in the templates of these search engines varies from five to fifteen, and furthermore different priorities have been given to the selection of mandatory elements and qualifiers. The investigation shows that the metadata producers often limit themselves to registration of the mandatory elements. The study shows further that the differences are bigger than the similarities concerning the usage of metadata within the five search engines. The search engines take advantage of the flexibility of the format by focusing on different elements and qualifiers and complementing the descriptions of the resources with different types of additions. We conclude that the search engines have to collaborate and coordinate their usage of metadata in order to improve public access to reliable and trustworthy information on the Internet.