Genetic and environmental influences on adult intelligence and special mental abilities

Bouchard, Thomas J Jr

THOMAS J. BOUCHARD JR.1

Abstract I review representative biometric studies of adult twins and adoptees that have been used to generate estimates of genetic and environmental influence on intelligence (IQ) and special mental abilities. The various studies converge on a heritability estimate between 0.60 and 0.80 for IQ. Estimates of common environmental influence from the same studies are near zero. Studies of twins reared together and studies of unrelated individuals reared together yield sizable estimates of common family environmental influence in childhood but also demonstrate that this influence dissipates with age and approaches zero in adulthood. Twin studies of the major special mental abilities (verbal, spatial, perceptual speed and accuracy, memory) yield heritability estimates of about 0.50 and modest estimates of common environmental influence.

KEY WORDS: MENTAL ABILITY, IQ, INTELLIGENCE

The quantitative study of genetic and environmental influences on the enormous variation in human mental abilities begins with Sir Francis Galton, who recognized that observed relationships within biological families confounded genetic and environmental determinants of human traits. Searching for a method to disentangle these two sources of influence, Galton introduced the use of both twins and adoptees (Galton 1876, 1883). He also invented the method of correlation (Galton 1888; Stigler 1989), variants of which underlie most twin research (Neale and Cardon 1992). Galton can thus be fairly called the founder of quantitative behavioral genetics.

Theories of Mental Abilities

Since Galton’s time, enormous progress has been made in both our understanding of the nature of human mental abilities and methods for disentangling the various genetic and environmental sources of variance in these abilities (Bouchard 1996). Psychometricians agree that it is both theoretically and practically useful to conceptualize the domain of cognitive abilities as a hierarchy of abilities (Carroll 1993; Gustafsson 1984; Gustafsson and Un&eim 1996; Undheim and Gustafsson 1987) rather than as a taxonomy (Guilford 1967). There are views on the nature of intelligence other than the psychometric hierarchical one (Gardner et al. 1996), but because only psychometric theories have generated data relevant to the issue of genetic influence on intelligence, I limit this review to studies in this tradition.

Hierarchical theorists have a number of disagreements. The principal one is whether the hierarchy is truncated, thus having the appearance of a mountain range (a number of higher-order abilities with no general factor spanning them), or whether it is a true hierarchical structure with a general factor (Spearman’s g) spanning the higher-order factors (Carroll 1996). The mountain range model is actually the current version of the theory of fluidcrystallized intelligence first put forward by Cattell in the 1930s and 1940s (Cattell 1933, 1943, 1963; Hakstian and Cattell 1978) and championed by John Horn since the mid-1960s (Horn 1976, 1994, 1998; Horn and Cattell 1967). The peaked model is the enduring theory of g originally proposed by Charles Spearman (Spearman 1904, 1927) and currently championed by a number of investigators (Brand 1996a; Carroll 1993, 1996; Herrnstein and Murray 1994; Jensen 1992, 1993, 1994). This controversy remains to be fully resolved.

A number of recent studies, however, strongly favor the g view. Bickley et al. (1995) demonstrated that comparable measures (16 ability tests) administered to 8 age groups (6, 8, 10, 13, 16, 30-39, 50-59, 70-79 years) have the same organizational structure across all age groups, and that structure is a three-stratum structure that requires a g or general factor at the top. Bickley specifically tested and rejected the hypothesis that g was identical to any of the second stratum factors. Carretta and Ree (1995), using a different set of ability measure and a large military sample-269,969 applicants for US Air Force commissions between the ages of 18 and 27 years-showed that the structure of cognitive abilities is virtually identical across males and females and across ethnic groups (white, black, Hispanic, and Native American). These studies confirm Carroll’s (1993) conclusion, based on a meta-analysis of the world literature, that “there is abundant evidence for a factor of general intelligence, . . . found at the highest order. . of analysis for a given data set . . . that dominates factors or variables that emphasize the level of difficulty that can be mastered in performing induction, reasoning, visualization, and language comprehension tasks” (p. 624). Consequently, I examine genetic and environmental influences on g (the third-order factor), special mental ability factors (second-order factors), and specific tests.

The controversy generated by the publication of The Bell Curve (Herrnstein and Murray 1994), which strongly championed the g view and the argument that g was heavily influenced by genetic factors (Bouchard 1995; Dorfman 1995), should make it clear that both the issue of genetic influence on mental abilities and the construct validity and utility of mental ability tests continue to be widely debated (Fischer et al. 1996; Gottfredson 1997a; Gould 1996; Rushton 1997a). Contrary to the picture portrayed in the media, however, most knowledgeable behavioral geneticists believe that mental abilities are to a considerable degree under genetic influence (Gottfredson 1997b; Plomin and Petrill 1997; Snyderman and Rothman 1988), and most experts believe that mental ability tests have substantial validity and utility (Baker et al. 1993; Gottfredson 1996, 1997c; Hunt 1995; Lubinski and Humphreys 1997; Neisser et al. 1996; Snow 1995). Some dissenting views can be found in the special issue of the American Psychologist devoted to intelligence (Sternberg 1997). A striking example of how controversial the topic of g continues to be can be found in the de-publication (censorship) of a recent book titled The g Factor, by Christopher Brand (1996b).

It is important to state at the beginning that when I discuss genetic influences on IQ, I am speaking about genetic influences on individual differences in IQ. What I am explaining is quantitative variation within a population. It is possible to study processes that influence the mean of a population (i.e., early intervention attempts to raise performance on IQ scores or special mental abilities) separately from genetic influences. In the long run, however, the same processes underlie both individual and mean differences (Rowe 1997; Turkheimer 1991).

The behavioral genetics literature dealing with genetic and environmental influence on mental abilities is immense; consequently this review will be selective. The focus is largely on adult data and representative findings. I do not, for example, deal with the growing literature on genetic influence on intelligence in minority populations (Levin 1994, 1997; Loehlin 1992; Lynn 1994; Rushton 1997b; Waldman et al. 1994). The documentation, however, will be sufficient so that the interested reader can quickly access the larger literature.

Quantitative Behavioral Genetic Models

In the experimental sciences one of the more powerful methods of exploring causal hypotheses involves the creation of model systems. In animal behavioral genetics one such model system involves the use of inbred strains. Individual organisms with identical genotypes can be reared in a variety of environments to assess the influence of environments, and several inbred strains can be compared to assess genotypic influences on a phenotypic character. Human beings cannot ethically be subjected to breeding experiments, but they do engage in what might be called experiments of society (adoption), and when such experiments are combined with a well-known experiment of nature (twinning), we have a relatively powerful quasi-experimental design: the twins reared apart design (Bouchard and Pedersen 1998). The usefulness and generalizability of the findings from such studies depend, as they do for all research designs, on how well the assumptions of the design are met.

One of the most powerful complementary model systems to the twins reared apart design is another adoption design: the unrelated but reared together design. As noted, the problem that Galton attempted to solve was the unconfounding of genetic and environmental influence on the similarity between relatives. Surprisingly, even though Galton introduced the use of both twins and adoptees, he did not recognize that taken together these two designs could yield ordinary correlations that were in principle unconfounded (Bouchard 1996). Although Galton did discuss the similarity between parents and their adopted offspring (popes and dignitaries of the Roman Catholic Church), he did not discuss the possibility of studying unrelated individuals of the same age reared together.

The easiest way to illustrate the logic of these methods is by means of path analysis and reliability theory. Figure la shows the ordinary interpretation of parallel form reliability. There are two tests, A and B. The correlation between them is caused by a latent trait called the true score, and it is represented by the sum of the cross-products of the paths connecting the scores, r^sub AB^. This correlation is computed by means of analysis of variance (Hayes 1973), as are most twin correlations, and it represents what is called the true score variance-it is not squared. Precisely the same path model underlies the correlation between monozygotic (MZ) twins reared apart (r^sub MZA^), and this model is shown in Figure 1b. By the rules of path analysis, the correlation is the sum of the cross-products of the paths or h X 1.00 X h, or in this instance r^sub MZA^ = h^sup 2^. This model makes the usual assumption of no placement with regard to trait-relevant environmental features, no systematic genotype by environment interactions, and no systematic genotype by environment correlations. If there is nonadditive genetic variance for the trait, it is captured in G, and the correlation estimates what is called the broad heritability.

To represent the correlation between dizygotic (DZ) twins reared apart (r^sub DZA^), who share half their genes in common by descent, the correlation between G’s in Figure lb would become 0.5 and the equation would read r^sub DZA^ = 0.5h^sup 2^. The correlation between DZ twins reared apart directly estimates one-half the narrow heritability. This version of the model makes the assumption that there is no nonadditive genetic variance and no assortative mating on the genetic component of the trait.

Dizygotic twins (reared together or apart) share only 0.25% of their dominance variance (one component of the nonadditive genetic variance), whereas MZ twins share all of the dominance variance; consequently, if the DZ correlation is reliably less than half the MZ correlation, this is presumptive evidence of nonadditive genetic variance. Interactions between loci (epistasis) can also create nonadditive genetic variance and cause DZ twins to be less than half as similar as MZ twins. Dominance and epistasis are often lumped together and simply called nonadditive sources of genetic variance.

Figure 1c shows the model for the correlation obtained from unrelated individuals reared together (r^sub URT^). The correlation in this instance (r^sub URT^ = c2) estimates what is called common environmental influence. Figure ld shows the components of the correlation for MZ twins reared together (r^sub MZT^). The correlation in this instance is confounded (r^sub MZT^ = h2 + c2) and estimates the sum of the genetic and common environmental variance. It simply combines the models in Figure 1b and 1c. The correlation for DZ twins reared together would be r^sub DZT^ = 0.5h^sup 2^ + c^sup 2^. The twin literature is full of heritability estimates based on the formula 2(r^sub MZT^ – r^sub DZT^). This equation, called the Falconer heritability formula (Falconer 1990), assumes that c2 is the same for MZ and DZ twins reared together-the equal trait-relevant environment assumption-and that all genetic influence is additive.

Assortative mating on the genetic component of a trait creates a greater similarity between first-degree relatives than would be found under random mating and thus would make the DZ correlation more similar than half the MZ correlation. There is considerable evidence for assortative mating for IQ and somewhat less evidence for assortative mating for special mental abilities. I do not discuss the complexities of assortative mating for IQ because of a lack of space. Jensen (1978) provided a thorough discussion of this topic.

When data from several kinships are gathered, a variety of model fitting procedures are available to efficiently estimate the underlying genetic and environmental parameters and to statistically test various hypotheses (e.g., Does the addition of a parameter for nonadditive genetic variance result in a significantly better fit or will additive variance suffice? Can the pattern in the data be explained by environmental factors, or does a genetic factor have to be postulated to generate an adequate fit?) (Neale 1995; Neale and Cardon 1992).

Genetic and Environmental Influences on General Intelligence

Direct Estimates of Heritability Using MZ and DZ Twins Reared Apart. The current findings for the IQ correlations between MZ and DZ twins reared apart are shown in Table 1.

The correlations between adult MZ twins reared apart are high (weighted mean = 0.75) and replicable across time, tests, languages, and populations in Western industrialized societies. Two of the studies in Table 1 were carried out in the United States 53 years apart. Two were carried out in the early 1960s in Denmark and Great Britain [the Juel-Nielson (1980) reference given in Table 1 is a republished version of a study originally published in 1965]. The remaining study was recently completed in Sweden.

Only two correlations for DZ twins reared apart have been published. One is from the Swedish Adoption/Twin Study of Aging (SATSA): 0.32 (95% confidence interval = 0.30-0.56). The second, 0.47 (95% confidence interval = 0.13-0.71), is from the Minnesota Study of Twins Reared Apart (MISTRA). The correlation for DZ twins reared apart is reported as a control variable in this study. The MISTRA IQ correlations have not yet been fully analyzed. We are awaiting completion of the study before conducting a full analysis. As the confidence intervals show, these values do not differ significantly from each other. The weighted mean (0.38) is close but somewhat less than the typical value reported for same-age first-degree relatives reared together-if twins are excluded (Bouchard and McGue 1981). Doubling the correlation for DZ twins reared apart yields a heritability of 0.76, a value close to the heritability estimated by the correlations for MZ twins reared apart.

The quality of the three early studies on MZ twins reared apart has been vigorously attacked on a variety of grounds (Kamin 1974; Lewontin et al. 1984; Taylor 1980). Bouchard (1982, 1983, 1993, 1997) systematically examined all the criticism and found it seriously wanting. The two more recent studies benefited from the published criticisms and were able to assess and reject the common criticisms, such as the influence of contact, age of separation, and biased testing. The most frequent criticisms of these studies include (1) failure to separate the twins immediately at birth, (2) contact between the twins while growing up and or contact after finding each other but before assessment, and (3) placement effects (e.g., both twins being placed in homes of similar socioeconomic status).

The striking feature of these criticisms is that they are almost always made in isolation, that is, without reference to other relevant data that would also need to be addressed if the criticism were correct. If the criticisms were valid, then one would expect a number of other kin correlations (e.g., cousins, siblings) to be as high as if not higher than the correlations for MZ twins reared apart. The question is a quantitative one, and the most informative group is that of unrelated individuals reared together. As the path diagram in Figure 1c shows, these individuals do not share genes; the reason they would be alike is because they share a common family environment. These individuals are not just placed in homes similar on some index of socioeconomic status or parental educational background; they are placed in the same home! This comparison controls for adoption status, a control lacking when twins reared apart are compared to twins reared together. No mental ability data appear to have been collected from samples of adopted MZ and DZ twins reared together. There could be some placement bias on the background characteristics of the parents of unrelated individuals reared together, but as I will show, these factors must not be trait relevant because they do not explain the results for adults.

Direct Estimates of Environmentality Using Unrelated Individuals Reared Together. The result of the studies of unrelated individuals reared together are shown in Figure 2.

For the data gathered in childhood-the bulk of the studies-there is a considerable range of values (Burks 1928; Freeman et al. 1928; Horn et al. 1979; Leahy 1935; Scarr and Weinberg 1977; Skodak 1950). It does not matter whether the data are organized by adopted versus adopted children or adopted versus biological children (both are unrelated and reared together); the degree of similarity is essentially the same. The childhood data (age range in the studies is 4-16 + years) estimate c^sup 2^ to be 0.28. The adult data, however, show an entirely different picture (Loehlin et al. 1997; Scarr and Weinberg 1978; Scarr et al. 1993; Teasdale and Owen 1984). They suggest an estimate of c^sup 2^ of essentially 0 (0.04). There are fewer adult studies, and one study does provide an estimate of c^sup 2^ of 0.19 (Scarr et al. 1993). Two studies in this group report longitudinal data: Scarr and Weinberg (1978) and Scarr et al. (1993). Scarr and Weinberg (1978), with a sample of 108, found that the drop from childhood to adulthood was from 0.31 to 0.19. In the Texas Adoption Study both the adopted versus adopted and the adopted versus biological groups declined, from 0.20 to – 0.03 and from 0.11 to – 0.02, respectively (Horn et al. 1979; Loehlin et al. 1997). It is also important to realize that almost all the adults in these samples are between age 16 and 21. It would be desirable to have data on an older sample.

The striking age effect on c^sup 2 shown in Figure 2 was missed entirely by Bouchard and McGue (1981) in th^eir review of the world’s IQ kin correlations, and all studies on unrelated individuals reared together were pooled. Because a similar age effect can be found in the longitudinal twin data (see later section), there is convergence across methods. Consequently, the finding is probably real and there is no good reason for continued reproduction of that aggregated correlation in textbooks. These findings also explain why adoption studies using young children do indeed show common family environmental influence (Capron and Duyme 1989, 1996). The real question now is the extent to which these effects will persist into adulthood (McGue 1989).

Heritability Estimate Using the Four-Group Design. The SATSA reported an overall analysis in terms of a four-group design (MZ and DZ twins reared together and apart), and I summarize these findings in the first section of Table 2. The special mental abilities data in Table 2 are discussed later in this review.

The broad heritability estimate for the first principal component (IQ) is 0.81, almost precisely what would be inferred from only the correlation for MZ twins reared apart (0.78). Pedersen et al. (1992) reported that a model with all nonadditive genetic variance fits the data best, but such a model is biologically improbable and, as they point out, the twin design alone is not a powerful one for estimating nonadditive effects. The reason for this odd estimate (large nonadditive genetic effect) is the very low correlation for DZ twins reared together (0.22) and the modest correlation for DZ twins reared apart (0.32). The correlation for DZ twins reared together is well below values reported in numerous other samples. Consistent with the data for unrelated individuals reared together, these adult data yield an estimate of 0 of shared environmental influence and an estimate of nonshared environment plus error of 0.19. The SATSA also reported a follow-up on the data for twins reared apart and showed that both the variability and the stability in cognitive abilities in later life are largely genetic in origin (Plomin et al. 1994).

Importance of Age. The strong age effects in data on unrelated individuals reared together (the disappearance of common environmental influence) are mirrored in the ordinary twin data, where we also see an increase in genetic influence. Figure 3, based on the twin literature broken out into five age categories (McGue et al. 1993), shows these effects graphically.

The change appears to occur at about the time most individuals leave home and begin to function as adults on their own. Similar findings have been reported for the twins reared together being studied in the Minnesota Twin Study of Aging (McGue et al. 1993). A study that has combined data from the adult twins of the Minnesota Study of Aging and the adult twins (reared together) from the SATSA showed that for the age periods 27-50 years and 50-65 years the heritability of IQ was 0.81 and c^sup 2^ was essentially 0 (Finkel et al. 1995). The older Swedish twins (65-85 years), however, showed somewhat of a decline in heritability (0.54), whereas the Minnesota twins did not. More recently, the Swedish group has published data from their study of same-sex twins reared together who are 80 or more years old (McClearn et al. 1997). Only twins with no major cognitive, sensory, or memory impairment were included. For the first principal component of an assessment battery the MZ correlation was 0.75 (N = 110 pairs) and the DZ correlation was 0.38 (N = 130 pairs). The estimated heritability was 0.62. The shared environmental effect of 0.11 was not significantly different from 0 and could be dropped from the model.

Heritability Estimate from Nontwins: Siblings Reared Apart and Together. Teasdale and Owen (1984), who provided one of the adult data points for unrelated individuals in Figure 2, actually reported on five types of sibling correlations: full siblings reared apart, maternal half-siblings reared apart, paternal half-siblings reared apart, unrelated individuals reared together, and full siblings reared together. Their sample was obtained from the Danish adoption register and consisted of only males between the ages of 18 and 26 years. All Danish males, whether fit for military service or not, are required to report to the conscription board and complete a welldeveloped intelligence test (Rasch 1980). Teasdale and Owen’s cases were drawn from the conscription board files. As a consequence, their sample is perhaps one of the most representative samples that has ever been used for estimating the heritability of IQ in a population, at least for males. Their correlations and the genetic and environmental parameter estimates are given in Table 3.

Teasdale and Owen (1984) tested for but found no evidence of the effects of nonadditive genetic variance or assortative mating. It is of interest to note that their heritability of 0.96 is the highest value yet reported in the literature. It is often asserted that twin studies (both reared apart and reared together) overestimate the heritability of IQ relative to other designs. Teasdale and Owen disprove the claim. Their group of kinships was well sampled, tested with an excellent instrument, and contained no twins, yet it yielded the highest heritability estimate of IQ ever reported.

Heritability Estimate from Numerous Kinships: Texas Adoption Study. The Texas Adoption Study (Horn et al. 1982) is a longitudinal adoption study that makes use of 14 kinship pairings, some of which involve adoptive relations and some of which involve biological relations, because adoptive families often have their own biological children as well as adopted children. The Texas study provided some of the data points for the adult unrelated individuals reared together (see Figure 2). The most recent data collected used the Beta test, which also had been given to the parents 10 years earlier and to all the children in the follow-up study. The earlier analysis of data from the Texas Adoption Study, collected when the children were young and when different tests were used for both parents and children, yielded several estimates: a true score variance (corrected for unreliability) of 0.38 for additive variance (narrow heritability), a shared environmental influence of 0.19, a geneenvironment correlation of about 0.10, and a nonshared environmental effect of 0.33 (Loehlin 1979; Loehlin et al.1989). The more recent analysis (Loehlin et al. 1997) yielded a narrow heritability for true score variance (0.776), no gene by environment correlation, and virtually no shared family environmental influence.

The Texas Adoption Study as a whole shows the same decrease in shared environmental influence with age as the studies for unrelated individuals reared together and the studies for MZ and DZ twins reared together.

Sex Effects. Turner (1996) attracted considerable attention by suggesting that genes coding for intelligence evolved on the X chromosome. The evidence that there are some genes on the X chromosome that influence intelligence is indeed growing, and it is well known that the fragile X site and other specific genes (Thapar, Gottesman et al. 1994) increase the proportion of males in the lower tail of the IQ distribution. The suspicion that there are more males in the upper tail of the IQ distribution has some support as well (Halpern 1997; Hedges and Nowell 1995), but no genetic mechanism to explain it has been proposed. Nevertheless, for the ordinary range of intelligence there is virtually no difference in the magnitude of IQ correlations for samesex and opposite-sex familial correlations (Bouchard and McGue 1981), demonstrating that most of the quantitative trait loci (previously called polygenes) lie on the autosomes, not on the X chromosomes.

Maternal Environmental Effects. Recently, Devlin et al. (1997) reanalyzed a subset of the kin correlations originally compiled by Bouchard and McGue (1981). Devlin claimed that twin maternal effects account for 20% and sibling maternal effects account for 5% of the IQ variance in this database. With this maternal parameter in place, Devlin’s model estimated the additive genetic variance at 34% and the nonadditive genetic variance at 15% (48% total genetic variance). This model estimated the shared environmental parameter to be 17%. Devlin’s analysis challenges the view expounded here that there are strong age effects and that the heritability of IQ increases with age, with common environmental influence fading out in adulthood. Devlin et al. (1997) purported to have tested this hypothesis, and although their data were not inconsistent with the hypothesis, they claimed that their maternal hypothesis fitted the data better. Curiously, the database did not contain unrelated individuals reared together. Devlin’s preferred model would predict that this group should have a correlation of 0.17 both in childhood and in adulthood, a prediction falsified by the data presented in Figure 2. The correlation for unrelated individuals reared together of 0.32 (based on childhood data) has been cited for years to support important common environmental influences on IQ. The Devlin model requires that this evidence be sacrificed in favor of maternal effects. The question is, Which is more probable?

Devlin et al. (1997) claimed that the age effect was tested by dividing twins into three categories, youths (18 years). McGue et al. (1993) formed their adult group with twins older than 20 years (see Figure 3) because so many young adults still live at home. A different breakdown of the data may yield a different result, especially because Devlin et al. (1997) apparently did not include all the adult twin data available. Indeed, their database as a whole is quite old. Contrary to the Devlin claim that “it is not difficult to garner evidence for the importance of maternal environment” (p. 470), the evidence actually leans the other way.

Almost 50 years ago Price (1950, 1978) argued from an extremely comprehensive review of the evidence that almost all MZ twin prenatal biases were difference producing rather than similarity producing, leading twin studies to underestimate genetic influences on various traits. At that time the literature was so large that the entire bibliography (Price 1950) was not published. Price’s complete bibliography was published in 1978 with an additional 260 references, at which time Price reiterated his conclusions. Research since that time largely reinforces rather than challenges the hypothesis that genetic influences are underestimated (Bryan 1993; Hall and Lopez-Rangel 1996; Macdonald et al. 1993). None of the research cited by Devlin et al. (1997) regarding possible in utero effects on IQ is unimportant; it simply does not support their narrow argument that maternal effects create excessive similarity in twins.

This is an empirical matter, however, and data gathered on large representative samples of adult twins and other adult kinships will help to resolve the question. The Teasdale and Owen (1984) sibling study and the Texas Adoption Study, both discussed earlier and not included in Devlin et al.’s analysis, yield findings that contradict Devlin et al.’s conclusions. The Colorado Adoption Project (Plomin et al. 1997) recently reported its IQ correlations between adoptive parents and their offspring, between biological parents and their offspring, and between a control sample of biological parents and their offspring. They demonstrated that “genetic influence increases monotonically from infancy to childhood to adolescence” (p. 446), a finding consistent with the conclusion of this review. In addition, they report a narrow heritability of 0.56 at age 16, a value clearly much larger that the 0.34 estimate put forward by Devlin et al. (1997). Additional studies of the relationship between adult kin of various types is sorely needed.

Genetic and Environmental Influences on Special Mental Abilities

Heritability Estimates Using the Four-Group Design. The SATSA reported the overall analysis of their data in terms of a four-group design (MZ and DZ twins reared together and apart), and I have summarized their findings in Table 2. The results for the first principal component (IQ) have already been discussed.

The special mental abilities all yield somewhat more modest heritabilities, ranging from 0.38 to 0.58, with only the verbal crystallized abilities suggesting nonadditive genetic variance. It should be noted that, despite the fact that the sample size of MZ twins reared apart is by far the smallest of the four groups, the correlations for these twins closely approximate the broad heritability estimate for all special mental abilities.

The special mental ability findings from MISTRA based on MZ and DZ twins reared apart are shown in Table 4, where they are analyzed as a fourgroup design using data from MZ and DZ twins reared together. The MZ and DZ twin data for twins reared together are from an update of the meta-analysis initially carried out by Nichols (1978) and updated by Bouchard, Segal et al. (1990). The analysis was carried out on the weighted mean of the correlations using the mean sample size of the studies [see Bouchard, Segal et al. (1990, Tables 5-8)]. There were 43-45 samples (some studies contributed more than 1 sample) for verbal ability, 54 or 55 samples for spatial ability, 20 or 21 samples for perceptual speed and accuracy, and 16 samples for memory ability.

The data suggest that the heritability of the major mental abilities hovers between 0.50 and 0.55. Both studies of twins reared apart, however, suggest somewhat lower heritabilities for memory abilities. Different memory tests appear to yield different heritabilities [see Thapar, Petrill et al. (1994) for a brief review]. A detailed analysis of four Wechsler Adult Intelligence Scale (WAIS) subtests, including digit span using the Minnesota Study of Aging twin sample and the SATSA twins reared together, also suggests a lower heritability for memory relative to other special mental abilities (Finkel and McGue 1993; Finkel et al. 1995). More research is clearly needed in this area, given the theoretical importance of memory for understanding g (Kyllonen 1996; Kyllonen and Christal 1990).

Cross-Cultural Studies of Special Mental Abilities. A recent study of twins, aged 15-19 years old, in Croatia suggests that the heritability of verbal and spatial abilities is similar to that found in the United States (Bratko 1996). A comparable study in Egypt, however, reports lower heritabilities (AbdelRahim et al. 1990). Cross-cultural work in this area is still in its infancy. Other relevant studies include Park et al. (1978), Wilson and Vandenberg (1978), DeFries et al. (1974) and Johnson et al. (1976).

Genetic Influences on Special Mental Abilities As a Function of Their g Loading

Inbreeding Depression. There is now considerable evidence that, as predicted from genetic theory, inbreeding depresses the mean scores of inbred individuals on mental ability tests (Afzal 1988; Agrawa et al. 1984; Badaruddoza and Afzal 1993; Bashi 1977). Neel and his colleagues have also shown that inbreeding effects are as strong for mental abilities as they are for physical traits (Neel et al. 1970; Schull 1995; Schull and Neel 1965, 1972). A number of reports show that there is a significant correlation between the degree to which a cognitive ability loads on g and the degree of inbreeding depression that it shows (Jensen 1983, 1987; Nagoshi and Johnson 1986; Vernon 1989). Kamin (1980) severely criticized the early literature on inbreeding and IQ, and the newer studies are far from perfect (Bouchard 1993), but they are consistent.

Resemblance of Collateral Relatives on Special Mental Abilities As a Function of g Loadings. In line with the inbreeding findings, Johnson and Nagoshi (1990), using the large Hawaii Family Study of Cognition (HFSC) sample (DeFries et al. 1976, 1979; DeFries and Plomin 1978), showed that resemblance of biological collateral relatives (uncles and aunts and nephews and nieces with a one-fourth additive genetic relationship) is correlated with the g loading of special mental abilities, whereas there is no significant correlation for the same unrelated collateral relatives. This same study also showed that the correlations for cousins (who have only a one-eighth additive genetic relationship) are somewhat higher (mean of 0.16 vs. 0.20). Johnson and Nagoshi (1990) suggested that this latter finding is perhaps due to a strong environmental cohort effect.

Is There G beyond g? An interesting question that has not received a great deal of attention is, How much genetic influence remains in special mental abilities after the genetic influence on g has been removed? The answer is that most special mental abilities still show genetic influence after the genetic influence on g has been removed, but the amount remains to be determined. The SA study (Pedersen et al. 1994), which used twins reared both together and apart, suggested that most of the genetic variance in special mental abilities was due to genetic variance shared with g. Finkel et al. (1995), who used only twins reared together from both the Minnesota Twins Study of Aging and SATSA, found that a little more than half the genetic variance was shared with g and that the remainder was specific to each special mental ability.

Summary

This overview is not an exhaustive review of the relevant studies. An exhaustive review would require a comprehensive meta-analysis (Hunter and Schmidt 1990). Nevertheless, the major studies that use powerful designs all converge on much the same conclusion: Genetic factors strongly influence the expression of general intelligence in adult populations. In addition, although common environmental influence on general intelligence clearly expresses itself in childhood, it appears to dissipate in adulthood. There are fewer comprehensive studies of special mental abilities, but here also the findings from various studies converge. Genetic factors strongly influence special mental abilities but less than for general intelligence. General intelligence, however, mediates much of the genetic variance found in special mental abilities and other genetic phenomena, such as inbreeding depression and the correlations between collateral relatives.

These findings apply to the broad middle class in industrialized Western societies (the group from which most of the samples have been drawn). They should not be generalized beyond that population.

Acknowledgments This research was supported by grants to the Minnesota Study of Twins Reared Apart from the Pioneer Fund and the Koch Charitable Foundation. Received 17 January 1997; revision received 9 October 1997.

Literature Cited

Abdel-Rahim, A.R., C.T. Nagoshi, and S.G. Vandenberg. 1990. Twin resemblance in cognitive abilities in an Egyptian sample. Behav. Genet. 20:33-43. Afzal, M. 1988. Consequences of consanguinity on cognitive development. Behav. Genet. 18:583-594.

Agrawa, N., S.N. Sinha, and A.R. Jensen. 1984. Effects of inbreeding on Raven matrices. Behav. Genet. 14:579-585.

Badaruddoza, and M. Afzal. 1993. Inbreeding depression and intelligence quotient among North

Indian children. Behav. Genet. 23:343-347.

Baker, L., J. Asendorf, D. Bishop et al. 1993. Intelligence and its inheritance: A diversity of views. In What Are the Mechanisms Mediating Genetic and Environmental Determinants of Behavior? Twins as a Tool of Behavior Genetics, T.J. Bouchard and P. Propping, eds. Chichester, England: Wiley, 85-108.

Bashi, J. 1977. Effects of inbreeding on cognitive performance. Nature 266:440-442. Bickley, P.G., T.Z. Keith, and L.E. Wolfle. 1995. The three-stratum theory of cognitive abilities: Test of the structure of intelligence across the life span. Intelligence 20:309-328. Bouchard, T.J., Jr. 1982. Review of Identical Twins Reared Apart: A Reanalysis. Contemp. Psychol. 27:190-191.

Bouchard, T.J., Jr. 1983. Do environmental similarities explain the similarity in intelligence of identical twins reared apart? Intelligence 7:175-184.

Bouchard, T.J., Jr. 1993. The genetic architecture of human intelligence. In Biological Approaches to the Study of Human Intelligence, P.A. Vernon, eds. Norwood, NJ: Ablex, 33-93.

Bouchard, T.J., Jr. 1995. Breaking the last taboo: Review of The Bell Curve. Contemp. Psychol. 40:415-421.

Bouchard, T.J., Jr. 1996. Behavior genetic studies of intelligence, yesterday and today: The long journey from plausibility to proof-The Galton Lecture. J. Biosoc. Sci. 28:527555.

Bouchard, T.J., Jr. 1997. IQ similarity in twins reared apart: Findings and response to critics. In Intelligence: Heredity and Environment, R.J. Steinberg and E.L. Grigorenko, eds. New York: Cambridge University Press, 126-160.

Bouchard, T.J., Jr., and M. McGue. 1981. Familial studies of intelligence: A review. Science

212:1055-1059.

Bouchard, T.J., Jr., and N. Pedersen. 1998. Twins reared apart: Nature’s double experiment. In On the Way to Individuality: Current Methodological Issues in Behavioral Genetics, E.L. Grigorenko and S. Scarr, eds. New Jersey: Nova (in press). Bouchard, T.J., Jr., N.L. Segal, and D.T. Lykken. 1990. Genetic and environmental influences on special mental abilities in a sample of twins reared apart. Acta Genet. Med. Gemellol. 39:193-206.

Bouchard, T.J., Jr., D.T. Lykken, M. McGue et al. 1990. Sources of human psychological differences: The Minnesota study of twins reared apart. Science 250:223-228. Brand, C. 1996a. Doing something about g. Intelligence 22:311-326. Brand, C. 1996b. The g Factor: General Intelligence and Its Implications. Chichester, England: Wiley.

Bratko, D. 1996. Twin study of verbal and spatial abilities. Pers. Indiv. Diff 21:621-624. Bryan, E.M. 1993. Prenatal and perinatal influences on twin children: Implications for behavioral studies. In Twins as a Tool of Behavior Genetics, T.J. Bouchard and P. Propping, eds. Chichester, England: Wiley, 217-225.

Burks, B.S. 1928. The relative influence of nature and nurture upon mental development: A comparative study of foster parent-offspring child resemblance and true parent-true child resemblance. Yrbk. Natl. Soc. Stud. Educ. 27:219-316. Capron, C., and M. Duyme. 1989. Assessment of effects of socioeconomic status on IQ in a full cross-fostering study. Nature 340:552-554.

Capron, C., and M. Duyme. 1996. Effect of socioeconomic status of biological and adoptive parents on WISC-R subtest scores of their French adopted children. Intelligence 22:259275.

Carretta, T.R., and M.J. Ree. 1995. Near identity of cognitive structure in sex and ethnic groups. Pers. Indiv. Diff 19:149-156.

Carroll, J.B. 1993. Human Cognitive Abilities: A Survey of Factor-Analytic Studies. New York: Cambridge University Press.

Carroll, J.B. 1996. A three-stratum theory of intelligence: Spearman’s contribution. In Human Abilities: Their Nature and Measurement, I. Denis and P. Tapsfield, eds. Mahwah, NJ: Erlbaum Associates, 1-18.

Cattell, R.B. 1933. The Cattell Intelligence Test: Scales 1, 2, and 3. London, England: Harrap. Cattell, R.B. 1943. The measurement of adult intelligence. Psychol. Bull. 40:153-193. Cattell, R.B. 1963. Theory of fluid and crystallized intelligence: A critical experiment. J. Educ. Psychol. 54(1):1-22.

DeFries, J.C., and R. Plomin. 1978. Behavior genetics. Annu. Rev. Psychol. 29:473-515. DeFries, J.C., S.G. Vandenberg, and G.E. McClearn. 1976. Genetics of specific cognitive abilities. Annu. Rev. Genet. 10:179-207.

DeFries, J.C., R.C. Johnson, A.R. Kuse et al. 1979. Familial resemblance for specific cognitive abilities. Behav. Genet. 9:23a3.

DeFries, J.C., S.G. Vandenberg, G.E. McClearn et al. 1974. Near identity of cognitive structure in two ethnic groups. Science 183:338-339.

Devlin, B., M. Daniels, and K. Roeder. 1997. The heritability of IQ. Nature 388:468471. Dorfman, D.D. 1995. Soft science with a neoconservative agenda: Review of The Bell Curve. Contemp. Psychol. 40:418-421.

Falconer, D.S. 1990. Introduction to Quantitative Genetics, 3d ed. New York: Longman Group. Finkel, D., and M. McGue. 1993. The origins of individual differences in memory among the elderly: A behavior genetics analysis. Psychol. Aging 8:527-537. Finkel, D., N.L. Pedersen, M. McGue et al. 1995. Heritability of cognitive abilities in adult twins: Comparison of Minnesota and Swedish data. Behav. Genet. 25:421431. Fischer, C.S., M. Hout, M.S. Jankowski et al. 1996. Inequality by Design: Cracking the Bell

Curve Myth. Princeton, NJ: Princeton University Press.

Freeman, F.N., K.J. Holzinger, and B.C. Mitchell. 1928. The influence of environment on the intelligence, school achievement, and conduct of foster children. Yrbk. Natl. Soc. Stud. Educ. 27:219-316.

Galton, F. 1876. The history of twin, as a criterion of the relative powers of nature and nurture. J. Anthropol. Inst. G. Br. Ir. 5:391-406.

Galton, F. 1883. Inquiries into Human Faculty and Its Development. London, England: Macmillan.

Galton, F. 1888. Co-relations and their measurements, chiefly from anthropometric data. Proc. R. Soc. Lond. 45:135-145.

Gardner, H., M.L. Kornhaber, and W.K. Wake. 1996. Intelligence: Multiple Perspectives. New York: Harcourt Brace.

Gottfredson, L.S. 1996. What do we know about intelligence? Am. Scholar Winter: 15-30. Gottfredson, L.S., ed. 1997a. Intelligence and social policy: Special issue. Intelligence 24 (whole issue).

Gottfredson, L.S. 1997b. Mainstream science on intelligence: An editorial with 52 signatories, history, and bibliography. Intelligence 24:13-24.

Gottfredson, L.S. 1997c. Why g matters: The complexity of everyday life. Intelligence 24:79132.

Gould, S.J. 1996. The Mismeasure of Man, 2d ed. New York: Norton. Guilford, J.P. 1967. The Nature of Human Intelligence. New York: McGraw-Hill. Gustafsson, J. 1984. A unifying model for the structure of intellectual abilities. Intelligence

8:179-203.

Gustafsson, J., and J.O. Undheim. 1996. Individual differences in cognitive functions. In Handbook of Educational Psychology, D.C. Berliner and R.C. Calfee, eds. New York: Simon & Schuster Macmillan, 186-242.

Hakstian, A.R., and R.B. Cattell. 1978. Higher-stratum ability structures on a basis of twenty primary mental abilities. J. Educ. Psychol. 70:657669. Hall, J.G., and E. Lopez-Rangel. 1996. Twins and twinning. In Emery and Rimoin’s Principles and Practice of Medical Genetics, D.L. Rimoin, J.M. Connor, and R.E. Pyeritz, eds. New York: Churchill Livingston, 395-404.

Halpern, D.F. 1997. Sex differences in intelligence: Implications for education. Am. Psychol. 52:1091-1102.

Hayes, W.L. 1973. Statistics for the Social Sciences, 2d ed. New York: Holt, Rinehart & Winston.

Hedges, L.V., and A. Nowell. 1995. Sex differences in mental test scores, variability, and numbers of high-scoring individuals. Science 269:41-45. Herrnstein, R.J., and C. Murray. 1994. The Bell Curve: Intelligence and Class Structure in American Life. New York: Free Press.

Hom, J.L. 1976. Human abilities: A review of research and theory in the early 1970’s. Ann.

Rev. Psychol. 27:437485.

Horn, J.L. 1994. Theory of fluid and crystallized intelligence. In Encyclopedia of Human Intelligence, R.J. Sternberg, ed. New York: Macmillan, 443-451. Horn, J.[L.] 1998. A basis for research on age differences in cognitive abilities. In Human Cognitive Abilities in Theory and Practice, J.J. McArdle and R.W. Woodcock, eds. Mahwah, NJ: Erlbaum Associates.

Horn, J.L., and R.B. Cattell. 1967. Age differences in fluid and crystallized intelligence. Acta PsychoL 26:107-129.

Horn, J.M., J.C. Loehlin, and L. Willerman. 1979. Intellectual resemblance among adoptive and biological relatives: The Texas adoption project. Behav. Genet. 9:177-207. Horn, J.M., J.C. Loehlin, and L. Willerman. 1982. Aspects of the inheritance of intellectual abilities. Behav. Genet. 12:479-516.

Hunt, E. 1995. Will We Be Smart Enough? A Cognitive Analysis of the Coming Workforce. New York: Russell Sage Foundation.

Hunter, J.E., and F.L. Schmidt. 1990. Methods of Meta-Analysis: Correcting Error and Bias in

Research Findings. Newbury Park, CA: Sage Publications. Jensen, A.R. 1978. Genetic and behavioral effects of nonrandom mating. In Human Variation: The Biopsychology of Age, Race, and Sex, R.T. Osborne, C.E. Noble, and N. Weyl, eds. New York: Academic Press, 51-105.

Jensen, A.R. 1983. Effects of inbreeding on mental-ability factors. Pers. Indiv. Diff 4:71-87. Jensen, A.R. 1987. The g beyond factor analysis. In The Influence of Cognitive Psychology on Testing, R.R. Ronning, J.A. Grover, J.C. Conoley et al., eds. Hillsdale, NJ: Erlbaum Associates, 87-142.

Jensen, A.R. 1992. Understanding g in terms of information processing. Educ. Psychol. Rev.

4:271-308.

Jensen, A.R. 1993. Spearman’s g: Links between psychometrics and biology. Ann. NY Acad. Sci. 702:103-129.

Jensen, A.R. 1994. What is a good g? Intelligence 18:231-258. Johnson, R.C., and C.T. Nagoshi. 1990. Resemblance of collateral relatives on cognitive measures as related to g loading of the measures. Pers. Indiv. Diff. 11:871-873.

Johnson, R.C., J.C. DeFries, J.R. Wilson et al. 1976. Assortative marriage for specific cognitive abilities in two ethnic groups. Hum. Biol. 48:343-352.

Juel-Nielsen, N. 1980. Individual and Environment: Monozygotic Twins Reared Apart, rev. ed.

New York: International Universities Press.

Kamin, L.J. 1974. The Science and Politics of IQ. Potomac, MD: Erlbaum Associates. Kamin, L.J. 1980. Inbreeding depression and IQ. Psychol. Bull. 87:469-478. Kyllonen, P.C. 1996. Is working memory capacity Spearman’s g? In Human Abilities: Their Nature and Measurement, I. Dennis and P. Tapsfield, eds. Mahawa, NJ: Erlbaum Associates, 49-76.

Kyllonen, P.C., and R.E. Christal. 1990. Reasoning ability is (little more than) working memory

capacity?! Intelligence 14:389-433.

Leahy, A.M. 1935. Nature-nurture and intelligence. Genet. PsychoL Monogr. 17:241-305. Levin, M. 1994. Comment on the Minnesota Transracial Adoption Study. Intelligence 19:13

20.

Levin, M. 1997. Why Race Matters: Race Differences and What They Mean. Westport, CT: Praeger.

Lewontin, R.C., S. Rose, and L.J. Kamin. 1984. Not in Our Genes: Biology, Ideology, and

Human Nature. New York: Pantheon.

Loehlin, J.C. 1979. Combining data from different groups in human behavior genetics. In Theoretical Advances in Behavior Genetics, J.R. Royce and L.P. Mos, eds. Alphen aan den Rijn, The Netherlands: Sijthoff & Noordhoff, 303-334.

Loehlin, J.C. 1992. Should we do research on race differences in intelligence? Intelligence 16:14.

Loehlin, J.C., J.M. Horn, and L. Willerman. 1989. Modeling IQ change: Evidence from the

Texas Adoption Project. Child Dev. 60:993-1004.

Loehlin, J.C., J.M. Horn, and L. Willerman. 1997. Heredity, environment, and IQ in the Texas adoption study. In Heredity, Environment, and Intelligence, R.J. Sternberg and E.L. Grigorenko, eds. New York: Cambridge University Press, 105-125. Lubinski, D., and L.G. Humphreys. 1997. Incorporating general intelligence into epidemiology and the social sciences. Intelligence 24:159-202.

Lynn, R. 1994. Some reinterpretations of the Minnesota Transracial Adoption Study. Intelli

gence 19:21-28.

Macdonald, A.M., E.M. Bryan, R. Derom et al. 1993. Group report: What can twin studies contribute to the understanding of childhood behavior disorders? In Twins As a Tool of Behavior Genetics, T.J. Bouchard and P. Propping, eds. Chichester, England: Wiley, 85108.

McClearn, G.E., B. Johansen, S. Berg et al. 1997. Substantial genetic influence on cognitive

abilities in twins 80 or more years old. Science 276:1580-1585. McGue, M. 1989. Nature-nurture and intelligence. Nature 340:507-508. McGue, M., T.J. Bouchard, Jr., W.G. Iacono et al. 1993. Behavior genetics of cognitive ability: A life-span perspective. In Nature, Nurture, and Psychology, R. Plomin and G.E. McClearn, eds. Washington, DC: American Psychological Association, 59-76. Nagoshi, C.T., and R.C. Johnson. 1986. The ubiquity of g. Pers. Indiv. Diff 7:201-207. Neale, M.C. 1995. MX: Statistical Modeling. Richmond, VA: Department of Human Genetics

University of Virginia.

Neale, M.C., and L.R. Cardon, eds. 1992. Methodology for Genetic Studies of Twins and Families. Dordrecht, The Netherlands: Kluwer Academic.

Neel, J.V., W.J. Schull, M. Yamamoto et al. 1970. The effects of parental consanguinity and inbreeding in Hirado, Japan. II. Physical development, tapping rate, blood pressure, intelligence quotient, and school performance. Am. J. Hum. Genet. 22:263-286. Neisser, U., G. Boodoo, T.J. Bouchard Jr. et al. 1996. Intelligence: Knowns and unknowns. Am. Psychol. 51:77-101.

Newman, D., A. Tellegen, and T.J. Bouchard Jr. 1998. Individual differences in adult ego development: Sources of influence in twins reared apart. J. Pers. Soc. Psychol. (in press). Newman, H.H., F.N. Freeman, and K.J. Holzinger. 1937. Twins: A Study of Heredity and Environment. Chicago, IL: University of Chicago Press. Nichols, R.C. 1978. Twin studies of ability, personality, and interests. Homo 29:158-173. Park, J., R.C. Johnson, J.C. DeFries et al. 1978. Parent-offspring resemblance for specific cognitive abilities in Korea. Behav. Genet. 8:43-52.

Pedersen, N.L., R. Plomin, and G.E. McClearn. 1994. Is there G beyond g? (Is there genetic influence on specific cognitive abilities independent of genetic influence on general cognitive ability?) Intelligence 18:133-143.

Pedersen, N.L., R. Plomin, J.R. Nesselroade et al. 1992. A quantitative genetic analysis of cognitive abilities during the second half of the life span. Psychol. Sci. 3:346-353. Plomin, R., and S.A. Petrill. 1997. Genetics and intelligence: What’s new? Intelligence 24:5378.

Plomin, R., D.W. Fulker, R. Corley et al. 1997. Nature, nurture, and cognitive development from 1 to 16 years: A parent-offspring adoption study. Psychol. Sci. 8:442-447. Plomin, R., N.L. Pedersen, P. Lichtenstein et al. 1994. Variability and stability in cognitive

abilities are largely genetic later in life. Behav. Genet. 24:207-215. Price, B. 1950. Primary biases in twin studies: A review of prenatal and natal differenceproducing factors in monozygotic pairs. Am. J. Hum. Genet. 2:293-352. Price, B. 1978. Bibliography on prenatal and natal influences in twins. Acta Genet. Med. Gemellol. 27:97-113.

Rasch, G. 1980. Probabilistic Models for Some Intelligence and Attainment Tests, 2d ed. Chicago, IL: University of Chicago Press.

Rowe, D.C. 1997. A place at the policy table? Behavior genetics and estimates of family environmental effects on IQ. Intelligence 24:133-158.

Rushton, J.P. 1997a. Race, intelligence, and the brain: The errors and omissions of the “revised” edition of S.J. Gould’s The Mismeasure of Man (1996). Pers. Indiv. Diff. 23:169-180. Rushton, J.P. 1997b. Why The Bell Curve didn’t go far enough on race. In Intelligence, Political

Inequity, and Public Policy, E. White, ed. Westport, CT: Prager, 119-140. Scarr, S., and R.A. Weinberg. 1977. Intellectual similarities within families of both adopted and biological children. Intelligence 1:170-191.

Scarr, S., and R.A. Weinberg. 1978. The influence of family background on intellectual attainment. Am. Sociol. Rev. 43:674-692.

Scarr, S., R.A. Weinberg, and I.D. Waldman. 1993. IQ correlations in transcracial adoptive families. Intelligence 17:541-555.

Schull, W.J. 1995. Effects of Atomic Radiation: A Half-Century of Studies from Hiroshima and Nagasaki. New York: Wiley-Liss.

Schull, W.J., and J.V. Neel. 1965. The Effects of Inbreeding on Japanese Children. New York: Harper & Row.

Schull, W.J., and J.V. Neel. 1972. The effects of parental consanguinity and inbreeding in

Hirado, Japan. V. Summary and interpretation. Am. J. Hum. Genet. 24:425-453. Shields, J. 1962. Monozygotic Twins: Brought Up Apart and Brought Up Together. London, England: Oxford University Press.

Skodak, M. 1950. Mental growth of adopted children in the same family. J. Genet. Psychol. 77:3-9.

Snow, R.E. 1995. Validity of IQ as a measure of cognitive ability. In Workshop on IQ Testing and Educational Decision Making. San Diego, CA.

Snyderman, M., and S. Rothman. 1988. The IQ Controversy: The Media and Public Policy. New Brunswick, NJ: Transaction Books.

Spearman, C. 1904. General intelligence: Objectively determined and measured. Am. J. Psychol. 15:201-293.

Spearman, C. 1927. The Abilities of Man: Their Nature and Measurement. New York: Macmillan.

Sternberg, R.J. 1997. Introduction to the special issue on intelligence and lifelong learning. Am. Psychol. 52:1029.

Stigler, S.M. 1989. Francis Galton’s account of the invention of correlation. Stat. Sci. 4:73-86. Taylor, H.F. 1980. The IQ Game: A Methodological Inquiry into the Heredity-Environment Controversy. New Brunswick, NJ: Rutgers University Press. Teasdale, T.W., and D.R. Owen. 1984. Heritability and familial environment in intelligence and educational level: A sibling study. Nature 309:620(622. Thapar, A., S.A. Petrill, and L.A. Thompson. 1994. The heritability of memory in the Western Reserve twin project. Behav. Genet. 24:155-160.

Thapar, A., I.I. Gottesman, M.J. Owen et al. 1994. The genetics of mental retardation. Br. J. Psychiatr. 164:747-758.

Turkheimer, E. 1991. Individual and group differences in adoption studies of IQ. Psychol. Bull.

110:392-405.

Turner, G. 1996. Intelligence and the X chromosome. Lancet 347:1814-1815. Undheim, J.O., and J.E. Gustafsson. 1987. The hierarchical organization of cognitive abilities: Restoring general intelligence through the use of linear structural relations (LISREL). Multivar. Behav. Res. 22:149-171.

Vernon, P.A. 1989. The generality of g. Pers. Indiv. Diff 10:803-804. Waldman, I.D., R.A. Weinberg, and S. Scarr. 1994. Racial-group differences in IQ in the Minnesota Transcracial Adoption Study: A reply to Levin and Lynn. Intelligence 19:29-44. Wilson, J.R., and S.G. Vandenberg. 1978. Sex differences in cognition: Evidence from the Hawaii family study. In Sex and Behavior, T.E. McGill and D.A. Dewsbury, eds. New York: Plenum, 317-335.

1 Department of Psychology and Institute of Human Genetics, University of Minnesota, Minneapolis, MN 55455-0344.

Multivariate analysis of the Italian population

Institutional pharmacy to create 75+ jobs in Toledo