JET progressively established elements of an environmental monitoring programme well in advance of tritium operation in order to determine baseline levels. Prior to the first JET tritium experiment (PTE) in 1991, an extensive programme was in place and agreed with the regulatory authority. This consisted of tritium in air, rain, ground and river water, and crops on and off the JET site. Air is sampled continuously and averaged monthly. Other samples are taken quarterly or, at an appropriate point in the growing season. The discharges of tritium from the JET stacks are monitored by on-line instrumentation and silica gel-based samplers. The performance of these is described and improvements arising from the PTE experience are discussed. In particular the implications of tritiated methane on sampling and analysis are considered. The results of environmental monitoring are presented and comparisons made with predictions from discharges made during the PTE and subsequent operations. The implications of washout on the site liquid discharge authorisation is considered. From a comparison of observed and predicted concentrations, routine releases of tritium from JET will have insignificant environmental impact.