This dissertation explores political representation and its manifestation within the European Union (EU). The main purpose is to examine the representative roles of Members of the European Parliament (MEP) in the context of cross–pressure between the national level and the EU level. This involves an analysis of how the MEPs under-stand their roles, how they organize their work, and how they have voted in the European Parliament (EP) in 1999-2002. It also includes a study of how national party organizations adapt to the EU environment and how this influences the MEPs link to the national arena. The study is based on various sources, such as interviews, formal documents and voting data. The most under-researched part of the cross-pressure has been the national link and the empirical focus of the thesis is on that link. It is a comparative study of parties in Ireland, the Netherlands and Sweden. In each country, three parties were selected (social democratic, right-wing and green parties). By using a focused comparative method, and by controlling for certain independent variables, the ambition is to go beyond description and identify explanations for why MEPs adopt certain roles. The overall picture that emerges is of a relatively weak link between MEPs and the national level. To a certain extent, MEPs express frustration over their limited role in the national arena and over the lack of input from the national arena in their work at the European level. Most of the parties struggle to include MEPs in their organizational set-up, and the MEPs experience a growing hostility within the parties toward them. In general, the lack of interest and knowledge in the national arena, concerning the EU in general and specifically the work of the MEPs, obscures the role of the MEPs. They become EU ambassadors at the national level, rather than elected representatives at the EU level. The dissertation also tests variables that are thought to influence MEPs’ roles: the type of electoral system, popular opinion on EU issues, whether their party is in government, the party’s ideological heritage, and if the party organizes more advanced coordination mechanisms. The main result is that the working assumption that MEPs are influenced by characteristics in the national arena is shown to be largely correct. That is, some of the identified aspects of the national political context do influence how the MEPs understand their roles. For example, the character of the electoral system influences attitudes among the MEPs. However, that relationship is not as simple and straightforward as much of the literature suggests. Rather, the results in this study suggest that the most important aspect of the relationship between the national level and the MEPs is whether parties or others (such as national parliamentarians) actively engage in the work of the MEPs. It matters how parties design the relationship between the levels, especially for how and where MEPs direct their main attention, but also in terms of how MEPs vote in the EP. The conclusion emphasizes the importance of further research into how parties facilitate the link between the national and the EU level.