Although birth weight, gestational age, and IUGR are important risk factors for perinatal mortality, we did not include them in our final models. Studies have shown that these conditions are also associated with exposure to air pollution (Ritz and Yu 1999; Ritz et al. 2000), so it is very likely that they lie in the causal pathway between traffic-related air pollution and perinatal deaths. Other pregnancy complications such as hypertension and diabetes may also make the fetus more susceptible to the additional insult of air pollution, but it is less likely that they are in the causal pathway because, in our study, these are conditions developed during pregnancy and thus essentially are associated with the pregnancy itself. In all these cases, these variables might modify the effect of exposure. However, we did not have sufficient statistical power to evaluate this hypothesis.