The purpose of this thesis is to investigate norm in contemporary Russian verbal stress. In a first step the concept of norm is explored. It is shown that the criteria generally used in Russian for defining norm (correspondence to the language system, usage and authority/tradition/necessity) are not applied strictly. It is also concluded that any study of norms must take into account the distinction between the explicit norm, i.e. the codification, and the implicit norm, i.e. the usage and attitude of educated native speakers. In a second step the explicit norm is investigated. The analysis is based on the stress notation in two orthoepic dictionaries. This comparison shows that there is not, as is often suggested, one unanimous, "objectively existing", explicit stress norm. In a third step, the implicit norm is examined. This is done through a survey of reported and actual usage, carried out on 106 Russian speakers in Moscow. Subsequently, implicit norms are related to explicit norms. There is compliance between these in many cases, but the discrepancies are numerous. Furthermore, there is no direct or predictable relationship between the implicit stress norms and the labels these stresses are assigned in handbooks. A comparison with additional sources demonstrates that among the, in all, nine sources no two are perfectly alike in their notation. Sources that reflect the implicit norm better than others are identified. Finally, dictionary data and the survey results are compared with results from previous surveys (1956-1994). This shows that certain stress variants have apparently functioned as the implicit norm for several decades, but this has not yet been taken into account in codification. The general conclusions are that there is in theory an unclear definition of norm; there is in practice disagreement in codification; there is no official codex, although some sources might be considered more reliable; there is in many cases a discrepancy between explicit and implicit norms, which is most likely a result of arbitrariness or subjectivism and of conservatism. It is possible that these conclusions are valid for areas of language normativisation other than verbal stress.