Some enzymes, such as those from extreme thermophiles, have significant half-lives above 100°C. The differences in structure and function between these very stable and less stable enzymes are relatively small and are comparable with those differences found among enzymes of similar stability. Recent evidence suggests that protein degradative reactions at high temperatures (>80°C) occur only slowly in conformationally intact proteins, so that conformational stability may still dictate the upper temperature limit for enzyme activity. The interrelationship of both conformational stability and enzymatic activity with protein flexibility suggests that in naturally occurring enzymes, we cannot expect to find stability at temperatures far above those which are optimum for the growth of the organism. Genetic and enzyme engineering studies are promising in terms of enhancing conformational stability, but are likely to require case-by-case knowledge of the enzyme concerned, and stability enhancements achieved so far are relatively small. Furthermore, engineered increases in stability may well be accompanied by a decrease in specific activity. Nevertheless, it should be feasible to engineer enzymes to be substantially more stable than any of those found so far in nature where useful half-lives above 120°C are already available.