Back in the mid-1970s the University of Washington initiated an experimental physical fitness program for its students. To evaluate its effects, all those who registered for the program at the beginning of the first year of operation were tested at registration and retested at the end of the year. Physical fitness improved remarkably. As a result, it was decided to run the program for a second year. It was not considered necessary to test new participants at the beginning of the second year, since the results from the first year were available as a benchmark. At the end of the second year, all the participants were tested. The tests showed a marked deterioration in physical fitness by comparison with the scores at the beginning of the first year.