Research on situated multiagent systems (situated MAS) in- vestigates decentralized architectures for software systems that have to deal with highly dynamic operating conditions. To re- alize the system requirements, the agents of a situated MAS have to coordinate their behavior. The agent environment pro- vides a means to mediate (i.e., enable and constrain) agents' activities in the system. Laws embedded in the agent environ- ment allow to define application specific constrains on agents activities. In this paper, we declaratively specify the semantics of laws for perception, action, and communication in situated MAS. We illustrate the laws with concrete examples in an au- tomated transportation system that we have developed. Medi- ation of agents' activities via the agent environment improves separation of concerns in MAS and helps to manage complex- ity, especially in open and pervasive environments.