A causal argument rests on the accumulation of evidence along five major domains: biologic plausibility, consistency, temporality, dose-response gradient, and strength of association (Last, 2001). Our review of the literature on maternal smoking and subsequent child neurocognitive functioning indicates evidence in each of these domains. As described, there are multiple biologically plausible mechanisms by which prenatal and passive tobacco smoke exposure could result in neurocognitive and behavioral alterations. At this time, the large number of epidemiologic studies and the consistency of their findings concerning increased rates of behavioral problems provide a broad base of consistency across populations and across various study designs and endpoint measures. The temporal sequence of exposure preceding outcome, while somewhat cloudy in some human studies, is clearly evident in the animal models. Numerous studies also support a dose-response relationship, i.e., the greater the exposure, the higher the likelihood that children will demonstrate adverse behaviors. Moreover, the animal data provide a strong body of evidence of diminished neurocognitive performance associated with prenatal exposure while eliminating the possibility of misattributing these alterations to confounding with social or psychological characteristics that frequently accompany parental smoking. Some remain skeptical (Ernst et al., 2001; National Cancer Institute, 1999; Olds, 1997; Ramsay & Reynolds, 2000) of research demonstrating adverse effects of smoking during pregnancy. As noted elsewhere, differences exist between smoking and nonsmoking mothers that might explain adverse outcomes among the offspring of smoking mothers, i.e., heavy and moderate smokers receive less prenatal care, recognize their pregnancies later, and report more symptoms of depression than do nonsmokers and light smokers (Sameroff & Seifer, 1983). These differences suggest that smokers are more likely to be depressed and less likely to practice health-promoting behaviors for themselves or their children. However, more than a dozen observational studies have controlled for numerous potential confounders, including depression and substance abuse, and the association between smoking and adverse child neurocognitive outcomes remained. Intervention studies and animal models also potently contradict the contention that maternal smoking simply is a proxy for other factors responsible for adverse neurocognitive and behavioral outcomes. The magnitude of the adverse behavioral and neurocognitive effects of tobacco exposure for individual children remains unclear, and some available measurements seem modest, i.e., a decrement of 4 to 5 IQ points (Olds, Henderson, and Tatelbaum, 1994b) and an odds ratio of approximately 1.5 for adverse developmental or behavioral outcomes. Yet, considerable public resources have been directed to other problems having a similar magnitude of effect and, given the vast numbers of children affected, the net effect on a societal level would be expected to be quite large. It also is essential that we recognize that an insult of this type and magnitude, when coupled with other risks that tend to cluster among a significant percentage of exposed children, may have substantial effects on functioning and quality of life across the life span. While many questions remain, both animal model and human epidemiologic data clearly suggest a causal relationship between prenatal tobacco exposure and adverse behavioral and neurocognitive effects on children.