We review recent analyses of tracer observations within the Arctic vortex, showing anomalous tracer-tracer relationships which indicate the mixing of outside air into the vortex. Similar anomalous relationships have previously been observed in vortex fragments, following breakdown of the Arctic vortex, and interpreted as indicative of mixing within the filamentary fragments. In the light of other observations we conclude that this interpretation is incorrect: the mixing took place into the vortex prior to breakdown. We discuss the effects of mixing on tracer relationships within the vortex, illustrated by results both from a simple conceptual model and from two three-dimensional chemical transport models. These models indicate that continuous mixing across the vortex edge produces entirely changed tracer relations inside the vortex and that separate, compact vortex relationships can develop. In general, tracer relationships should not be used to estimate denitrification or chemical ozone loss unless the effects of mixing can be quantified; this may be possible using information from other tracers, but one cannot assume that the effects of anomalous mixing are manifested as simple mixing lines on tracer-tracer diagrams. ¿ 2000 American Geophysical Union