Research productivity was examined over the careers of a sample of eminent Australian academics now aged 70-90 years. Overall number of publications increased to a peak at 50-59 years of age and then declined. However, different peaks were observed for various types of publication. Productivity increased to a peak age of 40-49 years for journal articles, 50-59 years for new books and cross-disciplinary publications, and 60-69 years for revised and edited books, technical publications and non-technical books. The age of peak productivity for low and high publishers coincided. These effects could be either a consequence of ageing or due to cohort effects in publication practices.
KEY worms-Research productivity, age.
Many investigators have considered age-related changes in research productivity and creativity (Lehman, 1953; Over, 1989; Simonton, 1988a,b). Research quantity (productivity) is said to increase rapidly up to a definite peak productive age, and then to decline gradually. The high point is usually attained somewhere in the late thirties or early forties (Lehman, 1953; Simonton, 1988a; Horner et al., 1986). Research quality (creativity), on the other hand, is said to remain constant over the lifespan (Over, 1989). Simonton formalized the relationship between creativity and productivity as the 'constant probability of success model' which 'simply holds that the odds of making a contribution are a straightforward probabilistic consequence of total output ' (1988a' ( , p. 86, also Simonton, 1988b' ( , 1989)).
Although age-related changes in research productivity and creativity have been studied extensively, further investigation of the issue seems warranted. Much of the research has concentrated on tabulating the number of notable contributions (contributions cited in written histories of the sci-