than mean differences between communities. Many studies of the epidemiology of low-level lead toxicity fall into this category. Tooth lead levels in children have been related to intelligence and attention span deficits since the pioneering work of Needleman et al. (1972). Other studies have linked blood lead levels to children's stature (Schwartz et al., 1986), umbilical cord lead levels to congenital anomalies (Needleman et al., 1984), and identified bone demineralization as an internal source of lead exposure (Silbergeld et al., 1988).

It is clear that whenever one can use an individual marker of exposure, the power of such studies is greatly augmented. Monster and Smolders (1984) did an imaginative analysis of exhaled air for tetrachlorethane in teachers and their five-year-old pupils at a kindergarten near a factory with fugitive emissions. They then compared these levels to those found in a control group. Exhaled tetrachlorethane was then measured in a group of residents of an old people's home situated near a chemical-waste dump. Significant differences were found between the two groups of children, with those closer to the factory having higher levels. It was also shown that residents living on the first floor of the home (which is closer to the waste site) had significantly higher levels than did residents who lived on the second floor.

There are three categories of studies of acute morbidity: diary studies, population-based studies, and analysis of large data banks.

Diary studies of respiratory symptoms were first used in the 1950s as indicators of the pulmonary effects of air pollution (Lawther et al., 1970). These involved analyses of daily subjective records of respiratory symptoms, as they relate to pollution. There has recently been a resurgence of interest in this method, because it reduces the possibility of confounding by other factors that are difficult to control in cross-sectional studies, which compare rates of illness or symptoms across areas with different pollutant concentrations. In the longitudinal diary study, the population acts as its own control. Hence, variations between subjects in reporting rates due to subjective factors, differences in susceptibility, and passive smoke exposure are normalized and do not confound the pollution relationship.

Despite the inherent subjectivity of self-recorded data on symp-