Demographers usually study population and environment in preindustrial settings where "environment" means food, forest, or land. California, in contrast, is an advanced industrial state with rapid population growth and complex environmental stresses. In this paper I examine the effects of population growth on carbon monoxide (CO) and ozone, the principal ingredient of smog. Ozone and CO are monitored at numerous local sites throughout California. Wind currents are strong, so the level of ozone or CO at a site may depend on population size and other factors upwind as well as at that site. I use longitudinal data for a sample of sites to estimate panel models of trends in ozone and CO. Population growth is measured at three levels: site, county, and upwind; and trends in per capita income and air pollution regulations are controlled. Local population growth has a substantial impact on CO; in contrast, population growth at any level has a very small or even negative impact on ozone. The methodological and policy implications of this implausible finding are discussed.