To estimate the fatigue strength of riveted joints represents a problem of primary importance both for the re-analysis of old bridges and the assessment of relatively recent structures. Even if the riveted solution is going to be progressively substituted by bolted and welded solutions, the behaviour of riveted members remains a matter of considerable economic importance. Technical specifications available in the Standards in force give to engineers engaged in fatigue design problems a reference curve relating to only one detail and the recommended values appear to be very different. In particular, recent re-analysis reported in the literature showed that the value suggested by the Eurocode 3 for shear splices that use non pre-loaded high strength bolts is strongly non-conservative when applied to riveted connections taken from old bridges. The paper presents the results of fatigue tests carried out on 3 different riveted joints and compares them to the design curves reported by some European and American Standards in force.