The new version 2.0 of the Unified Modeling Language (UML) was targeted at improving expressiveness and semantic precision. These developments are particularly evident in activity diagrams which have not only acquired many new features, but a completely new metamodel and semantic foundation. The UML contains some hints that Petri-nets are the inspirational source for the new semantics. In this paper we will investigate how strong the alignment of UML's activity diagrams to Petri-nets really is. We start by providing a mapping of the basic elements of activity diagrams to Petrinets and discuss the problems arising when trying to extend this approach to some of the advanced features of activity diagrams, namely exceptions, traverse-to-completion, and streaming. This examination raises several syntactic and semantic questions concerning activities. We conclude that for basic activities, the analogy works pretty well, but for higher-level constructs, no such intuitive connection exists.