Today, bibliographical information is kept in a vari
MotivationThe integration of existing digital libraries and electronic catalogues of publications is considered as one of the main tasks in the digital library community [22]. Publishers, libraries and computer scientists engage in developing digital libraries and electronic catalogues in order to make the publications themselves (digital contents) and their accompanying meta-data (bibliographical data about the publications) available. In different libraries and catalogues various services based on the provided information are offered for certain area of research, for instance the ACM and IEEE digital libraries providing meta-data and full text for computer science in general and the DBLP (Computer Science Bibliography) service provide meta-data for the field of database research. Generally, the provided services differ in their quality, for instance full text versus bibliographical data, quantity and the associated cost model. Today, the integration of meta-data from the available heterogeneous sources and the identification of objects in this global scope are major problems for the integration of digital libraries. For the latter problem current developments in the WWW technology and probabilistic approaches for identifying same objects seem to be promising. New techniques to tackle the problem of meta-data integration were developed in the area of integration of information systems, such as federated database systems and mediator based architectures. The provision of a single point of access to bibliographical meta-data for the users would allow to look up publications more efficiently. Furthermore, the integration makes it possible to derive information about implicit relationships among data spread over various sources, for instance citation linking information.For the integration of bibliographical data various projects were started, and newly developed methods and technologies from the research area of federated databases and mediator systems were applied successfully. But some problems remained unsolved, because of the complexity resulting from the high level of heterogeneity regarding data representation, contents and interfaces. All approaches have a similar multi-tier architecture in common that provides adapters and/or wrappers for certain classes of sources to deal with interface heterogeneity and, depending on the approach, various issues of data representation. Due to the interface heterogeneity and varying levels of cooperativeness of data providers all of these approaches have their limitations.In this paper our approach for integrating heterogeneous digital libraries based on federation services will be presented, whereby both cooperative and noncooperative sources are taken into consideration. The remainder of this paper is structured as follows. In section 2 we give a short overview of related work. In section 3 we introduce an exemplary scenario for the integration of digital libraries in heterogeneous environ-