Background: Many studies have found that there is a low-to-moderate association between exercise during adolescence and exercise habits in adulthood. A question that arises from these earlier studies, with a follow-up period of about five to 20 years, is how children's and adolescents' physical activity affects their inclination to exercise later in life, after the age of 50, e.g. after a follow-up period of almost 40 years. Purpose: The main objective of this study is to illustrate whether, and in that case how, differences in sports experiences during childhood and adolescence and differences in cultural capital are reflected in exercise habits in middle age. The analysis takes its starting point in Pierre Bourdieu's conceptual framework. The specific research questions are as follows. (1) How does "sport habitus" formed during childhood, in physical education lessons at school and through leisure time sports during leisure time, relate to exercise habits in middle age? (2) How does the "cultural capital" acquired during childhood, expressed in social position and success at school, relate to exercise habits in middle age? (3) How do "sport habitus" and "cultural capital" assign in childhood relate to exercise habits in middle age if, at the same time, the influence of the individual's level of education is controlled for? Research, design and data collection: The individuals included in this follow-up study were first contacted in 1968 when they were 15 years old. I then contacted 91 randomly selected school classes in Year 8 from four counties in Sweden. Follow-up contact and information gathering were conducted on six additional occasions, primarily via questionnaires sent by post. In 2007 the highest possible number of responses was 1979. A total of 1518 responses were received, which corresponds to 77% of the total reachable group. Findings: Neither membership of a sports club nor the amount of time spent on sports activities at the age of 15 had any significant association with the exercise habits displayed in middle age when the individual's breadth of sport experience was used as a simultaneous control. Sporting breadth was significantly related to later exercise habits, however. As indicators of sport habitus, both the "breadth of sport experience" during leisure time and "grades in physical education" had a significant association with later exercise habits. There was also an almost five times greater chance that an individual with a very high cultural capital at the age of 15, as defined by their social background and grades in theoretical subjects, was still an active exerciser 38 years later in comparison with an individual with a very low culture capital. Conclusions: A middle-aged individual's level of exercise is closely linked to that person's social position and, accordingly, to his or her educational capital. The children and adolescents with the greatest chance of achieving this middle-class position were those from backgrounds with a relatively high social positions and/or high grades in school. If they had a strong sport habitus as well, their inclination to exercise was strengthened. (Contains 10 tables, 5 figures and 4 notes.)