we can extend our understanding of how disparities in health across these indicators of socioeconomic position may be generated by antecedent factors or mediated via subsequent factors. Several such elaborations are important in thinking about reducing other socioeconomic and racial disparities in health.

First, socioeconomic position (SEP) has to be thought of as an intergenerational as well as intragenerational phenomenon. Thus, parental socioeconomic position may importantly shape childhood well-being and hence educational and later adult socioeconomic attainment and health, as shown in Figure 1B. The work of Barker (e.g., Barker and Osmond, 1986) and others (Kaplan and Salonen, 1990; Elo and Preston, 1992; Blane et al., 1996; Kuh and Ben-Shlomo, 1997) has indicated that childhood socioeconomic position and experiences can have long-term effects on adult health (see also Paper Contribution C). This is sometimes interpreted to mean that childhood socioeconomic position is a more important determinant of health than adult socioeconomic position. However, Figure 1B suggests that most such effects are likely to be channeled through and reinforced by later socioeconomic attainment, and the unique impact of childhood SEP or its sequelae must be evaluated net of later socioeconomic or other experiences. When this is done, the unique effects of childhood SEP on adult health are often found to be small or even nonexistent relative to the effects of later adult socioeconomic attainment and experiences (e.g., Lynch et al., 1994). *Thus, although the impact of socioeconomic position on childhood health and well-being is a very important problem in its own right, it cannot and should not be viewed as a major explanation of adult socioeconomic or racial/ethnic disparities in health or hence as the major, preferred, or necessary route for reducing such adult disparities.

However, Figure 1B is also highly simplified, neglecting the changing socioeconomic position of the families of many children. Thus, the socioeconomic position of a child often changes from preschool to elementary school to secondary school and onward through adulthood. Socioeconomic advantage and disadvantage may be viewed as ebbing and flowing or cascading over a person's life course. Although recent socioeconomic position is usually the best predictor of future outcomes, sustained socioeconomic deprivation over time is likely to be even more damaging (Wolfson et al., 1993, Lynch et al., 1997), and uncertainty or variability in socioeconomic position may be deleterious even to those of generally solid middle- or higher-level SEP (McDonough et al., 1997). Thus, knowledge of the full life course of socioeconomic position is ideally desirable for understanding socioeconomic disparities in health and a target for efforts to alleviate such disparities.

Finally, Figures 1A and 1B must be further elaborated, as in Figure 1C, to take account of the impact of more ascribed and relatively fixed social