find that children with hard neurologic findings (cerebral palsy, seizure disorders, head injury accompanied by significant coma) or definite EEG or imaging evidence of damage showed greater aggressivity and delinquency than controls. It would be of considerable interest to have had neuropsychologic-cognitive data on these children. It may be that aggressive disorders in brain-injured children are seen only in the presence of cognitive impairments (i.e., leading to communication difficulties). In any event, the data are of interest and must temper any conclusions about obligatory relationship between head injury and externalizing behaviors. It would also be well to conduct follow-up studies of this cohort.

The second theme in these studies, which is not necessarily independent of the first, is that violence, abuse, and deprivation of a loving relationship in childhood predispose the individual to later delinquency, including homicide (Lewis et al., 1983; Tarter et al., 1984; Walsh and Beyer, 1987). The first theme is, in fact, consonant with that developed in earlier sections, namely that aggressive/violent behavior may stem from involvement of the limbic system (and related regions of the brain); however, the mediating variable is neuropsychologic deficit in communication, attention, and other cognitive skills. It is thus not the case, according to this view, that there is a direct relationship between limbic system abnormality and aggression. Rather, the link appears to be through the effects of altered cognitive capacities (pursuant to limbic system damage) of the individual. The point that perhaps should be made in this context is that while the limbic system may subserve emotional behavior, the structures of which it is comprised also serve a major role in the support of cognitive behaviors. For example, the role of the hippocampus in the support of memory is well known (Milner, 1969). However, questions remain as to how such injury to the developing brain could occur and whether there are social-environmental conditions that could foster such developmental defects. In fact, there appears to be a host of pernicious influences on the developing nervous system that could, in the perinatal period (or later), account for serious compromise of the developing brain; these influences are more likely to be present in lower socioeconomic circumstances.

Genetics We discuss only two recent contributions to this area. These are summarized in Table 14. In the study by Mednick

and Kandel (1988) of 173 recidivistically violent offenders (from the Danish health records) it was found that early commission of violent crimes predicted later commissions (a point also made strongly by Farrington, 1989). Moreover, when the offenses of adopted-away children were compared with those of biologic and adoptive parents, it was found that biologic parent-child relationships predicted property convictions, but not violent offenses. This finding supports the view that the circumstances leading to the commission of violent acts are more likely due to nongenetic (i.e., environmental) than genetic determinants.

The other study, by Walker et al. (1989), explored the relation between schizophrenic diathesis and the stress of parental maltreatment in a group of acting-out children with significant aggressive and delinquency problems. The design (summarized in Table 14) allowed the authors to conclude that the combination of parental schizophrenia and maltreatment at home was associated with a progressive increase in delinquency in both sexes over time.

Lead The possible etiologic role of small amounts of lead in the environment, in contributing to cognitive and social failure, has long been the primary research interest of Needleman and collaborators (Table 15). Not only does environmental lead (in air contaminated by leaded fuels, from paint in older houses, from plumbing systems) contribute to poorer cognitive capacities (including attention) but in follow-up studies in a Boston cohort recently reported (Needleman et al., 1989), it is also associated with significant academic failure: a sevenfold increase in the failure to graduate high school—"a serious impairment in life success." The study by Thomson et al. (1989) confirms Needleman, with an independent sample from Edinburgh. This work also implicated lead levels more directly in deviant antisocial and hyperactive behavior in school children. The remaining paper in this group (Lansdown, 1986) suggests that the relationship between lead levels and behavior may not exist in children from higher socioeconomic strata. If this were to be the case, then cofactors (some of which are discussed below) may be involved.

Alcohol Although considerable data have been reported on the later effects on the child of maternal ingestion of alcohol during pregnancy, the most complete and systematic data are probably provided by the work of Streissguth and colleagues (Table 16). This is a continuing investigation of a group of 500 Seattle children