Measurements of chemical tracers whose spatial gradients are primarily due to the time dependence of sources and/or sinks are often used to define tracer ages in an effort to diagnose transport. However, a major problem with interpreting these tracer ages is that different tracers can yield different ages, and at present, it is not clear what aspects of the transport are measured by the different tracers. We use the concept of a distribution of transit times to compare the timescales derived from different tracers, including CFCs, tritium-helium, and radioactive tracers. By performing a systematic study over a range of transit time distributions we examine under what conditions two tracers yield similar or different ages. It is shown that there can be significant differences in tracer ages and that in general, tracer ages are not fundamental timescales of the flow. Furthermore, even if ages from two tracers are similar these ages can be very different from the mean (ideal) age or the age of a third tracer. It is also shown that significant temporal variations in tracer ages can occur for steady transport and that these changes are of similar magnitude to the changes in CFC and tritium-helium ages observed in the North Atlantic and North Pacific over the 1980s and 1990s. Accounting for the changes in tracer ages caused by steady transport is necessary before attributing changes in tracer ages to changes in transport. The possibility of using the differences in ages from different tracers to infer information about the transit time distribution is also examined. It is shown that two tracer ages can constrain the first two moments (mean age and width) of the distribution, but how tightly these are constrained depends on the tracers used, the certainty of the age calculations, and the flow characteristics.