from short-term exposure, such as that resulting from a single episode of exposure that occurs at the critical time in pregnancy, and other windows of vulnerability exist for neurodevelopmental effects (NRC, 1989). Such time- and dose-dependent exposure also could be involved in the development of other chronic diseases, such as learning and behavioral disabilities caused by exposure to lead.
For studies of hazardous-waste sites where common diseases with multiple causes are found, the finding of a dose-response relationship can be obscured by the operation of multiple causal factors and the absence of valid estimates of exposure. As Neutra (1990) notes, we can rarely reconstruct the individual exposures accurately, especially because they can stem from periods several decades past. In the absence of detailed measurements of exposure, we are forced to assume that people in a given neighborhood endured comparable and uniform exposures, even where that is not likely to have been the case.
Where an assumed causal relationship exists, removal of the suspected cause in individuals should reduce or eliminate the suspected effect, unless the effect is irreversible. Thus, those who stop smoking reduce their risk of contracting lung cancer. At the population level, reductions in cigarette smoking among men and women in the United Kingdom and among men in the U.S. have resulted in reduced rates of lung cancer (NCI, 1990). Where different causes contribute to a single disease, this principle will be relevant only for the specific causal factor removed.
The above considerations explain why it is not easy to evaluate the effects of removal of an exposure at many hazardous-waste sites. Despite these problems, after the fact analyses have been produced that permit some causal inferences regarding a few studies. Thus, allowing for a five-year latency, no new cases of leukemia have occurred in families in Woburn, Massachusetts, since those families stopped using contaminated wells (R. Clapp, Center for Environmental Health Studies, JSI, personal communication, 1991). Similarly, in the Lipari Landfill study by the state of New Jersey (NJDOH, 1989) and at Love Canal, after exposure from hazardous materials declined, birth weights returned to normal (Goldman et al., 1985; Vianna and Polan, 1984).
Does the association make sense in terms of the current understanding of basic human biology? Animal studies or other experi-