Though there were good reasons for adopting the two main features of the design of the original Exeter polished stem, these reasons did not turn out to be those for which the device functioned as it did. Thus, in one sense, the evolution of the device has been to some extent serendipitous. However, as a consequence of continued clinical follow-up, analysis of failures and retrievals, and basic investigation of the viscoelastic behaviour of acrylic cement, by 1988 sufficient had been learned to be able to say that the Exeter Universal stem had been ‘designed to subside’38. It is interesting to observe that if the early recommendations77,78 for screening newly developed implants using migration measurements had been in place in 1970, the Exeter polished stem would never have survived beyond its first year of use.