Charles Murray.

WE, Homo sapiens, are about to learn how to alter human nature at roughly the same time that we finally learn for sure what that nature is.

Our ignorance about the underlying truth of human nature has not been for want of trying. Philosophers took up the question as one of the very first that human beings systematically asked about themselves. But philosophers produced answers as various as Aristotle's and Rousseau's. Since the late 1900s, behavioral and social scientists too have tried to understand human nature. But while they have illuminated many useful bits and pieces, they have failed as system-builders. What is left of Freud, out of the beliefs that were so intellectually pervasive in mid century? Psychotherapy remains, in profuse variety, but only remnants of Freudianism. What is left of B. F. Skinner? Behaviorism is still a productive branch of psychology, but the Skinnerian vision of human nature that once seemed so compelling is dead. As for Marx, does anything at all survive? For more than a century, Marxism was throughout continental Europe the leading intellectual framework for thinking about how political institutions can realize the nature of man. That edifice has collapsed utterly.

How can we have expended so much of our collective genius on understanding human nature and still know so little for certain? Because up until now, we have been able to observe only behavior. People can hold very different views of human nature-man is by nature altruistic or by nature selfish; by nature amoral or by nature endowed with a moral sense- -because we observe in the human animal, in abundance, every sort of behavior. Or to put it statistically, human nature does not consist of universal human characteristics but of distributions. Is mankind altruistic or selfish? From everyday experience, we know that some people behave selfishly and some behave altruistically. When one says that human beings are by nature altruistic or selfish, one is actually saying that a distribution of the human population on the characteristic of "underlying biological propensity to altruism" will have a certain shape and median. The implications of a distribution in which, for example, the average value is "fairly selfish" has very different implications from a bell curve in which the average value is "fairly altruistic." The implications of a curve that is narrow and steep (meaning that almost all human beings are very close to the median value) are very different from those of a shape that is wide and short (meaning that human nature for this characteristic is all over the map).

The problem is that, while scientists can measure the observed shape of these behaviors, they have been stymied by the nature/nurture problem. This is not to say that we know nothing. Just as geologists know a lot about the probability of finding oil based on rock formations on the surface, psychologists have learned to infer a lot about the heritability of observed traits. But in both cases, the observer is dealing with outcroppings and probabilities, while the exact, inarguable truth lies hidden.

This situation is about to change. No one can tell how rapidly and how completely the story will unfold. A few brave souls--brave indeed to buck the consistent lesson of the last five hundred years of science--still argue that the mysteries of the human mind will forever be mysteries. But E. O. Wilson's reading of the situation in his 1998 book, Consilience, seems much more plausible. The neuroscientists, increasingly understanding how the brain works, and the molecular biologists, increasingly understanding which genes do what, are about to link up with the social sciences, according to Wilson, in a "webwork of causal explanation" that brings human behavior within the realm of rigorous investigation previously reserved for physical phenomena. And not just individual behavior. "The explanatory network now touches on the edge of culture itself," in Wilson's words--or to put it another way, we are on the edge of understanding how human nature in individuals produces social and political institutions.

What we know now is fragmentary. But the speed with which that knowledge is expanding is so fast, and accelerating, that it is reasonable to expect that we are going to know a great deal about many, many aspects of human nature and their social implications within just a few more decades. By the end of the 21st century, we will be approaching biological truth about these topics.

It will be a winding road, with many false pronouncements that will be revised a year later, as new data come to light. Even those new findings that are solidly based will seldom be exciting individually. We will not find an aggression gene or a marriage gene or an IQ gene. Instead, we will learn about complex combinations of genes and their alleles that affect a behavior, and about how they interact with the unimaginably complicated neural and hormonal processes that affect behavior. We will learn about the interaction between biology and environment.

TRANSLATION INTO POLITICAL CODES

The practical importance of these impending discoveries lies in this: The great conflicts of the last two centuries have in large part been the story of differing views of human nature translated into political codes. Communism's use of Marxism was the paradigmatic example, explicitly and aggressively asserting that human nature is soft plastic that can be molded into any configuration by society's political and economic institutions. But the debates over social policy within the democratic West have also been underlain by conflicting understandings of human nature. Are mothers peculiarly suited to raising small children or can fathers do it just as well? Should women be in combat? The positions one takes are based on assumptions about the innate differences between men and women. The welfare state makes sense, or doesn't, depending on underlying beliefs about how human beings respond to economic incentives and, more profoundly, about how human beings achieve satisfaction in life. Should we try to deal with crime by attacking root causes? Depending on your definition of "root causes," attacking them could mean an anti-poverty program or more prisons--and your definition can ultimately be traced back to your beliefs about human nature.

It will be a cumulative process, but, as time goes on, our increasingly certain knowledge of human nature is going to shrink the wiggle-room for certain political positions. Think of the process as a scientific version of the Alger Hiss case. As of 2000, we have, analogously, already discovered the Pumpkin Papers. The scientific literature already in hand, not to mention common sense, gives us a pretty good idea of where this story is leading, just as dispassionate observers in 1948 had a pretty good idea that Hiss was guilty. Hiss's advocates defended his innocence for decades after the Papers were found, but those advocates dwindled as new evidence periodically came to light. Finally the Venona intercepts were revealed, and the debate effectively ended. So it will be with the uncovering of human nature.

The choice of analogy betrays my own expectations of the unfolding story. What we already know leads me to believe that the story of human nature as revealed by genetics and neuroscience will be Aristotelian in its philosophical shape and conservative in its political one. We will learn for certain such things as that women innately make better nurturers of small children than do men and that men innately make better soldiers than do women. Regarding these and many other human characteristics impinging on marriage, the upbringing of children, and the enforcement of social order, I am predicting that the adages of the Right will usually prove to be closer to the mark than the adages of the Left, and that many of the causes of the Left will be revealed as incompatible with the way human beings are wired.

To put it in terms of Left versus Right, however, understates the magnitude of what is likely to happen. Of all the casualties of our growing knowledge of human nature, the most politically far-reaching will be the 20th century's curious attachment to literal human equality. Let me draw on the response to The Bell Curve to illustrate what uncharted territory we are sailing into. As authors of the book, Richard Herrnstein and I thought that The Bell Curve contained powerful ammunition for the political Left. If IQ is important in determining life's outcomes and IQ is not acquired by merit, then one legitimate line of argument is that the government should intervene to make up for the unfairnesses of nature and capitalism. What we did not realize was how important the egalitarian premise is to the worldview of the Left. It is not enough that governments guarantee equal rights to all; it is not even enough that governments intervene to equalize outcomes. It must also be true that inequalities in individuals are the result of the social, economic, and political system, rather than of inherent differences in ability. I am still not sure why this premise is so important--the intellectual case for redistributionist policies does not depend on it--but it is.

In their own way, politicians of the Right are equally in thrall to the egalitarian premise. For example, no major Republican politician is willing to say in public that some of the social problems we most deplore are rooted to some degree in personal deficiencies. Try to imagine a GOP presidential candidate saying in front of the cameras, "One reason that we still have poverty in the United States is that a lot of poor people are born lazy." You cannot imagine it because that kind of thing cannot be said. And yet this unimaginable statement merely implies that when we know the complete genetic story, it will turn out that the population below the poverty line in the United States has a configuration of the relevant genetic makeup that is significantly different from the configuration of the population above the poverty line. This is not unimaginable. It is almost certainly true. It is also almost certainly true that statistically significant distributions of biological makeup separate just about any other groups that show substantially different patterns of behavior.

BEYOND RACE AND SEX

The group differences that people obsess about have to do with race and sex, but let me try to reach past that reflexive response to make a broader point: Statistically significant genetic differences beyond the self-evident ones probably separate men from women, and people who call themselves "white" from people who call themselves "black" or "Asian," but they also probably distinguish the English from the French, employed Swedes from unemployed Swedes, observant Christians from lapsed ones, and people who collect stamps from people who backpack.

None of this should be earthshaking. Often we will be talking of group differences so subtle that they can be teased out only with the most sophisticated methods. Often these differences will have nothing to do with "better" or "worse," but just vive la difference. Even when the differences are substantial, the variation between two groups will almost always be dwarfed by the variation within groups--meaning that the overlap between two groups will be great. In a free society where people are treated as individuals, "So what?" is to me the appropriate response to genetic group differences. The only political implication of group differences is that we must work hard to ensure that our society is in fact free and that people are in fact treated as individuals. And yet I can tell you from personal experience that "So what?" is not a response that many others share. Today, to suggest that genetically based group differences are even probable provokes a reaction that resembles hysteria.

Now imagine a world a few decades hence in which it has been demonstrated that biologically based differences separate individuals and groups, and that some of these differences involve characteristics that are important to success in life. What will happen when a position that is taboo in public discourse is proved to be scientifically accurate?

Nothing will happen, it might be argued. Even now, hardly anyone really believes in his heart of hearts that the strictly egalitarian line is true, so what difference will scientific proof make? But history suggests otherwise. Thomas Kuhn taught us in The Structure of Scientific Revolutions that first the old scientific paradigm begins to show cracks, then those cracks spread, and then, with remarkable speed, Kuhn's famous "paradigm shift" occurs. In just a few years, there are no more Ptolemaic astronomers, only Copernican ones; no more Aristotelian physicists, only Newtonian ones. Today we are at the stage of the spreading cracks, even if the egalitarian premise seems more politically impregnable than ever. But when the scientific debate eventually ends, it will not be merely a matter of scientists on the wrong side saying, "Oh, well," while the rest of our intellectual perspective continues unchanged; with the displacement of the old paradigm comes a new way of looking, not just at isolated bits of scientific truth, but at the way the world works. Think of the Newtonian revolution and the Enlightenment. Closer to our own time, the Darwinian and Einsteinian revolutions were central to the development of the nonscientific intellectual world of the 20th century. So too will it be with the consequences of the neurogenetic revolution that is about to unfold. It will have transforming effects that spill over into our conceptions of politics, religion, and social relationships.

A NEW CAUSE FOR THE LEFT

A chief characteristic of a paradigm shift is that its consequences are unexpected. But I can illustrate the nature of the spillover with one of the few obvious possibilities--that eugenics will become a cause of the Left. Why obvious? After all, eugenics is not only in disrepute everywhere, "eugenicist" is one of the Left's cursewords for people engaged in neurogenetic research. But eugenics is in disrepute because of Nazism, which has led us to forget that before Nazism it referred to a movement centered in Britain that was respectable and especially popular among intellectuals. To put it simply, the eugenicists made an assertion and drew from it a policy implication. Their assertion was that social problems would be greatly reduced if the lower classes had fewer children and the better classes had more. The policy implication they drew was that government policies should encourage that result.

As the biological basis for personal qualities statistically associated with social problems--low IQ, impulsiveness, short time-horizons, sociopathy, indolence--is understood, the old arguments about causality (e.g., "It's poverty and disadvantage that create the low IQ, not the other way around") will be resolved. There will still be a large role for environmental causes and solutions to social problems, but understanding the portion that is biological will permit analysts of the future to make fairly precise forecasts about the extent to which changes in fertility patterns may be expected to affect crime and poverty. The only difference will be that the old eugenicists had to rely on a rough statement ("the lower classes"), whereas eugenicists of the future will be able to be more precise ("people with the following genetic profiles").

Now turn to the eugenicists' political conclusion, that government should act to shape fertility patterns. It is not something that today's Left likes to recall, but eugenicism was predominantly a movement of the British Fabian and socialist Left, not of Tories or the old Liberals. This political affinity was no accident, for a reason expressed by Sidney Webb, one of the brightest lights of British socialism. "No consistent eugenicist can be a 'Laisser Faire' individualist," he wrote, "unless he throws up the game in despair. He must interfere, interfere, interfere!" Sidney and his wife Beatrice were joined in their enthusiasm for eugenics by the likes of George Bernard Shaw, Emma Goldman, and H. G. Wells.

As genetic engineering matures, a new, more insidious brand of eugenics will become possible--one that does not require the lower classes to stop having children, only to start having better children. This will be eugenics tailor-made for a Left constituted of people who are more squeamish about being repressive than their forebears, but who are just as ready to interfere, interfere, interfere, in a good cause. Will the Right stand firm against this ultimate intrusion of government? I pray so. But don't bet on it. In the face of temptation, mainstream Republican politicians cannot be relied upon to say It's None of the Government's Business--an unhappy fact that may have reverberations when eugenics with a smiling face is upon us.

I have no idea how the new eugenicism will play out, only a general expectation that eugenics, anathema today, will be a spinoff of the neurogenetic revolution tomorrow. My main point is that many such effects will be triggered, that most of them are now unforeseeable, and that this will turn the intellectual and political landscape topsy-turvy.

What of the broader manipulation of human nature? Putting aside government intervention and confining ourselves to the voluntary choices of individuals, should we expect that Homo sapiens will take it into its collective head to redesign itself? I confess to a certain optimism. I suppose that sex selection will be common, and that some parents will, if they can, opt to make their babies more compassionate, or more competitive, or "more" of some other personality trait that they favor. Some parents may want to grow seven-foot-tall basketball players. But one of the main reasons that couples have babies is to produce their baby, the product of their combined genes. Motivations don't get much more basic than that, and I think it unlikely that the typical parent will want to distort the process too much. The popular voluntary uses of gene manipulation are likely to be ones that avoid birth defects and ones that lead to improved overall physical and mental abilities. I find it hard to get upset about that prospect.

One may hypothesize a variety of darker sides to the ability to manipulate human tendencies. There are the unforeseeable effects of homogenization, for example. A world in which all the children are above average might be duller than we suppose. I am not confident in our ability to tweak the human sense of the rhythm of life to correspond with the extensions in lifespan that may occur. More broadly, our ability to affect the physical aspect of the human animal may run ahead of our ability to accommodate those changes to the ways in which the human psyche achieves happiness.

But these specific worries are the ruminations of a 20th-century man, destined to look as myopic a century from now as the predictions of 19th- century men about the 20th. I am confident of just one thing: Many of the people reading these words will live to see one of Thomas Kuhn's paradigm shifts--one as broad and deep, as demoralizing and inspiriting, as destructive and creative, as any that has taken place before.

Mr. Murray is the Bradley Fellow at the American Enterprise Institute.

Understanding Biological and Social Influences on Religious Affiliation, Attitudes, and Behaviors: A Behavior Genetic Perspective. Brian M. D'Onofrio, Lindon J. Eaves, Lenn Murrelle, Hermine H. Maes, Bernard Spilka. Journal of Personality Dec 1999 v67 i6 p953 View extended citation and retrieval choices

Jennifer Couzin.

Baby rats are experts at not only recognizing their mother's nursing style but imitating it when they have offspring--even if they're adopted. Strong maternal rodent behavior, like the tendency to breast-feed with an arched back or groom frequently, is passed down without any genes involved, Canadian neuroscientists reported last week. Their study, published in Science, suggests that environmental influences may explain more behaviors than previously thought.

The team divided baby rats into several groups, keeping some with their biological mothers and assigning others to different moms. They found that the pups raised by outgoing, relaxed mothers (who groom their pups frequently) were less affected by stress and strongly maternal toward their own offspring. Whether the pups had been born to outgoing or reserved mothers proved irrelevant.

Scientists are hunting for genes that control everything from happiness to aggression, but this work highlights the fact that "things are not necessarily just programmed by genotype," says Robert Bridges, a professor at Tufts University School of Veterinary Medicine in North Grafton, Mass. While similar research on humans is still sketchy, scientists know that infant animals may reprogram neurons depending on how they're treated.

GRAPHIC:%9 Picture: With baby rats, nurture wins over nature. (Srulik Haramaty--Phototake)

Susan L. Warren; Stephanie Schmitz; Robert N. Emde.

Author's Abstract:

Objective: To conduct behavioral genetic analyses of self-reported childhood anxiety at 7 years of age. Method: Three hundred twenty-six same-sex twin pairs (174 monozygotic, 152 dizygotic) completed the Revised Children's Manifest Anxiety Scale at 7 years of age. Behavioral genetic analyses were conducted on the total and subscale scores. Results: Monozygotic within-pair correlations were higher than dizygotic correlations for physiological and social anxiety symptoms, suggesting heritable influences on these aspects. These results were found to be statistically significant with structural equation modeling. Conclusion: Certain symptoms of self-reported anxiety in children 7 years of age seem to result, at least in part, from genetic factors. Physiological and social anxiety symptoms, which may be related to behavioral inhibition, appear to be genetically influenced. These results are linked to previous findings in older children and adults.

Key Words: anxiety, twin studies, behavioral genetics, development.

Research has suggested that the development of childhood anxiety disorders may be genetically mediated because anxiety disorders run in families (Biederman et al., 1991; Last et al., 1991). This, however, may be the result of environmental influences. Twin studies of panic disorder and phobic disorders in adults have supported a genetic contribution toward adult anxiety disorders (Torgersen, 1993). Research on depression, however, has suggested that the manifestation of psychiatric symptoms may result from different etiologies at different ages, as depression in adolescents may be both genetically and environmentally mediated but in children it may have primarily an environmental component (Harrington et al., 1996; Thapar and McGuffin, 1996). It is possible, therefore, that anxiety disorders are genetically influenced in adults but are the result of environmental influences in childhood and/or adolescence.

Little is known about the genetics of childhood anxiety. The aspect of temperament referred to as behavioral inhibition has been linked to childhood anxiety disorders (Biederman et al., 1993) and appears to be genetically mediated (Robinson et al., 1992).

Fearfulness has also been found to have a genetic contribution (Abe et al., 1984; Freedman and Keller, 1963; Goldsmith and Gottesman, 1981; Phillips et al., 1987; Rose and Ditto, 1983; Stevenson et al., 1992).

Direct research concerning the genetics of child and adolescent anxiety, however, has been inconclusive. Thapar and McGuffin (1995) reported that when anxiety symptoms were rated by adolescents using the Revised Children's Manifest Anxiety Scale (RCMAS), only shared environment appeared to play a role in family resemblance for the development of anxiety. In contrast, when the parents rated their twins, there appeared to be a genetic contribution (Thapar and McGuffin, 1995). Legrand and colleagues (1996) reported a genetic contribution to self-reported anxiety (using the State-Trait Anxiety Inventory) for 11-year-old female twins but not for 17-year-old female twins. Topolski and colleagues (1997) found genetic contributions to anxiety in 8- to 17-year-old children using child reports with the RCMAS. Using a semistructured diagnostic interview, they also reported a genetic contribution to overanxious or generalized anxiety symptoms but not to separation anxiety symptoms, for which shared environment appeared to play a major role (Topolski et al., 1997). Moreover, the genetic contributions appeared to be larger for girls than for boys in terms of manifest anxiety but not overanxious or separation anxiety symptoms (Topolski et al., 1997).

In summary, there is evidence for genetic and shared environmental contributions to childhood anxiety. However, depending on age and rater/informant, the results have been inconsistent. Moreover, little research has examined specific aspects of anxiety which may account for the different findings.

This study was designed to focus on the genetic contribution to self-reported anxiety in children 7 years of age because no previous research had studied such young children. Examining such young children could clarify differences found in previous studies between younger children and adolescents. In addition, a plan was developed to focus on specific aspects of anxiety (physiological symptoms, worry and social concerns) because these particular aspects had not been examined previously and could show differing genetic contributions, perhaps elucidating important patterns in the genetics of childhood anxiety.

Subjects were part of the MacArthur Longitudinal Twin Study (Emde et al., 1992). Twins were recruited through the Division of Vital Statistics in Colorado and were preferentially selected for normal birth weight ([greater than or equal to]1,700 g) and lack of medical complications. All twin pairs were raised together by their biological parents. Ninety-two percent of the sample was white, 7% was Hispanic, and 1% was African-American. In the current analyses, for monozygotic (MZ)/dizygotic (DZ) and gender distributions, there were 91 MZ female twin pairs, 83 MZ male twin pairs, 70 DZ female twin pairs, and 82 DZ male twin pairs. Zygosity was determined on the basis of 8 physical attributes (Nichols and Bilbro, 1966) by 2 independent testers on each of numerous occasions in the home and in the laboratory. This approach has been shown to be highly accurate (Goldsmith, 1991). If there was less than 90% agreement of the MZ or DZ classification across all testers, twin zygosity was considered ambiguous. In those cases, DNA testing was used.

Measures and Procedures

The children were studied in the home and laboratory at regular intervals starting at 14 months of age. The RCMAS (Reynolds and Richmond, 1978), the measure used in this research, was administered separately to each twin by different examiners during one home visit. It was read to the children when they were 7 years of age (mean = 7.21, SD = 0.12). Children were asked to indicate their responses by putting a card into the appropriate box. One box indicated that the statement described the child, and the other box indicated that the statement did not describe the child. The RCMAS has a total score and also subscales for Physiological Anxiety (an index of the child's expression of physical manifestations of anxiety), Worry/Oversensitivity (an index of worry frequency), and Social Concerns/Concentration (primarily an index of concern about social experiences and about how one is viewed by others). There is also a Lie scale composed of items which children might endorse if they were not being completely honest and instead were answering in a manner which they believed to be socially desirable.

The RCMAS is widely used with children 6 years of age and older and shows adequate reliability and validity (Reynolds and Richmond, 1985). Test-retest reliability in primary school children has ranged from 0.98 over 3 weeks (Pela and Reynolds, 1982) to 0.68 over 9 months (Reynolds, 1981). The RCMAS has shown convergent validity, significantly correlating with the Trait scale of the State-Trait Anxiety Inventory for Children (r = 0.85,p [less than] .001) (Reynolds, 1980), and not with IQ (Reynolds, 1980).

Method of Analysis

To assess formally the genetic and environmental influences on childhood anxiety, structural equation modeling was used. Total phenotypic variation ([V.sub.p]) can be divided into the portion due to genetic effects ([V.sub.g]), shared environmental effects ([V.sub.c]), and nonshared environmental effects (including measurement error, [V.sub.e]) so that [V.sub.p] = [V.sub.g] + [V.sub.c] + [V.sub.e]. The proportion of the total phenotypic variance which is explained by genetic effects ([V.sub.g]/[V.sub.p]) is known as heritability ([h.sup.2]). The proportion due to shared environment ([V.sub.c]/[V.sub.p]) is [c.sup.2], and the proportion due to nonshared environment ([V.sub.e]/[V.sub.p]) is [e.sup.2].

Figure 1 shows the path diagram for the ACE model (for the genetic [A], shared environmental [C], and nonshared environmental [E] components). The model compares MZ twin covariances with DZ twin covariances assuming (1) a genotypic correlation of 1.0 for MZ twins (all genes are shared); (2) a genotypic correlation of 0.5 for DZ twins (DZ twins share half of their segregating genes on average); and (3) a correlation between twins' shared environmental influences of 1.0 for both MZ and DZ twins (shared environmental influences are assumed to be of similar magnitude for MZ and DZ twins).

Models were fit using maximum-likelihood procedures, using a structural equation modeling program, Mx (Neale, 1995). Chi-square statistics and their associated probabilities were used to assess the model fit and to test the significance of a particular parameter. Reduced models were used because the resulting change in [[Chi].sup.2] and model fit is the test statistic for the parameter excluded from the reduced model.

In addition to the ACE model described above, an ADE model was tested. ADE models contain nonadditive genetic variance which includes the possibility of dominant/recessive transmission and perhaps epistasis (an interaction between loci). ADE models are composed of additive genetic (A), nonadditive genetic (D), and environmental (E) components.

Approximately 6% of the sample showed elevated levels of anxiety (T score [greater than] 60). Correlations between the physiology, worry, and social subscale scores for the entire sample ranged from 0.55 to 0.61 (p [less than] .001). Correlations between these subscales and the total anxiety score ranged from 0.82 to 0.89 (p [less than] .001). Internal consistency of the scales, assessed using Cronbach's [Alpha], was found to be adequate (total anxiety, .84; physiology, .64; worry, .78; social, .66).

Table 1 shows the within-pair correlations for MZ and DZ twins for the total score and the subscales. Since MZ twins are genetically identical whereas DZ twins share half their segregating genes on average, correlations between MZ twins should theoretically be 1.0 and correlations between DZ twins should theoretically be 0.5 if a trait is completely genetically mediated and measured without error. Similarly, if shared environment were a predominant influence, within-pair correlations for MZ and DZ twins should be substantial and approximately equal.

When both genders were grouped together, within-pair correlations were substantially higher for MZ than DZ twins for the physiology subscale. This would suggest that the physiology subscale has a genetic component. For the worry subscale, the MZ within-pair correlation was quite similar to that for the DZ twins, indicating a high likelihood [TABULAR DATA FOR TABLE 1 OMITTED] of shared environmental influences. For the social subscale, within-pair correlations were substantially higher for MZ than DZ twins, again suggesting a genetic component. For the total anxiety score, the MZ within-pair correlation was quite similar to that for the DZ twins, indicating a high likelihood of shared environmental influences. When the genders were examined separately, there appeared to be no clear differences between them.

Table 2 shows parameter estimates for the total anxiety score and subscales obtained using the ACE model, which provides estimates of additive genetic contributions ([h.sup.2]), shared environmental contributions ([c.sup.2]), and nonshared environmental contributions ([e.sup.2]). Genetic influences accounted for approximately one third of the individual variation on the physiology subscale. Shared environment accounted for approximately one third of the individual variation on the worry subscale. Genetic influences also accounted for approximately one third of the individual variation on the social subscale and approximately one quarter of the variance for the total anxiety score. For all of the subscales and the total anxiety score, nonshared environment showed the greatest influence.

TABLE 2 Parameter Estimates for ACE Model Genetic Shared Nonshared Influences Environment Environment ([h.sup.2]) ([c.sup.2]) ([e.sup.2]) Physiology 0.30(*) 0.00 0.70 Worry 0.00 0.30 0.70 Social 0.34(*) 0.00 0.66 Total 0.25 0.11 0.64 * p [less than] .05.

After testing the full ACE model, we tested additional models, separately dropping A and C. For the physiology subscale, dropping the genetic component produced a significantly worse-fitting model ([[Chi].sup.2] difference of 4.058, p [less than] .05, df = 1), suggesting a significant genetic contribution. When shared environment was dropped, there was no significant change, suggesting no significant shared environmental influence for physiological anxiety symptoms.

For the worry subscale, dropping the genetic or shared environmental component did not significantly change the model fit. However, when they were both dropped, the model fit deteriorated significantly. This finding suggests that in a sample this size, neither shared environment nor genetic influences alone are significant determinants for the worry subscale but that together they may be influential.

For the social subscale, dropping the genetic component did produce a significantly worse-fitting model ([[Chi].sup.2] differences of 5.997, p [less than] .05, df = 1), suggesting that there is a significant genetic contribution to social anxiety symptoms. Dropping shared environment did not change the fit of the model, suggesting no significant shared environmental influences for social anxiety symptoms.

For the total anxiety score, dropping the genetic or shared environmental component did not significantly change the model fit. However, when they were both dropped, the model fit deteriorated significantly. This finding suggests that in a sample this size, neither shared environment nor genetic influences alone are significant determinants for the total anxiety score but that together they may be influential.

For all the subscales and the total anxiety score, including nonadditive genetic variance (that is, applying the ADE model) did not improve the model fit. This suggests that there are no significant nonadditive genetic influences for anxiety that we can detect. However, detecting nonadditive genetic influences requires large sample sizes, as thousands of twin pairs would be necessary to differentiate heritabilities of this moderate magnitude into additive and nonadditive effects.

There were no statistically significant differences between girls and boys, again perhaps due to the small sample size. Thus, parameter estimates for the entire sample are reported.

DISCUSSION

The most striking finding of this research is that 7-year-old child self-reports of certain anxiety symptoms are significantly genetically mediated. Genetic influences appear to contribute approximately one third of the variance for the physiological and social anxiety scores. This degree of influence is consistent with previous findings using similar statistical procedures with adults (Kendler et al., 1992, 1995; Phillips et al., 1987).

This is the first study to examine the genetics of anxiety in children younger than 8 years of age. Previous research (Legrand et al., 1996) suggested that self-reported anxiety may be genetically mediated in younger girls (11 years of age) but not in adolescents, in direct contrast to depression, which may be genetically mediated in adolescents but not in younger children (Thapar and McGuffin, 1996). The findings of the current study are consistent with previous anxiety research, suggesting that genetic influences play an important role in the development of some anxiety symptoms in young children.

This study is also the first to report results related to different aspects of anxiety. Moreover, the findings with the specific subscales are interesting because of linkages with other research.

The social subscale may be measuring an aspect related to shy behavior, which may relate to behavioral inhibition. Behavioral inhibition has been found to be genetically mediated, with heritabilities ranging from 0.51 to 0.64 (Robinson et al., 1992). Moreover, researchers have suggested that social phobia results, at least in part, from genetic influences (Fyer, 1993). Kendler and colleagues (1992) found MZ within-pair correlations of 0.31 and DZ within-pair correlations of 0.12 for social phobia in adults. The genetic variance estimated through behavioral genetic modeling was 0.31, similar to the 0.34 obtained for social fears in this study.

Analogous to social fears, these data suggest that physiological symptoms of anxiety are likely to have a genetic contribution. It is possible that greater physiological symptoms in children represent an early, undifferentiated or limited form of panic as panic disorder in children and adolescents includes many physiological symptoms (Bradley and Hood, 1993; Last, 1991). Panic disorder also appears to be genetically mediated in adults (Torgersen, 1993).

Alternatively, physiological symptoms of anxiety may be the result of genetic influences, unrelated to panic disorder. Gustavsson and colleagues (1996) found that genetic factors accounted for individual differences on scores of somatic anxiety in adult twins, some of whom were adopted and reared apart. Kendler and colleagues (1995) also reported substantial genetic effects (heritabilities 0.25-0.49) for somatic symptoms of anxiety in 2 adult twin samples.

In summary, previous research concerning behavioral inhibition, physiological symptoms, and social anxiety, in combination with these analyses, would suggest that anxiety in young children may result from a biological predisposition to experience physiological symptoms and fearfulness in new situations. Such a predisposition, in interaction with unique environmental influences such as the attachment relationship (Warren et al., 1997), may lead to the development of anxiety disorders in children and adolescents.

The results of this study are limited by the fact that anxiety in young children is extremely difficult to measure (Perrin and Last, 1992; Silverman, 1991), and only one measure of anxiety was used. Young children tend to not easily report their anxiety or negative experiences (Glasberg et al., 1982; Harter and Pike, 1984; Schwab-Stone et al., 1994). Moreover, parent reports are problematic in that they may be biased by knowledge of twin zygosity and may not reflect the child's internal feelings (Weissman et al., 1987). At this time we are limited by the lack of availability of standardized measures other than the RCMAS for assessing anxiety in children of this age. Although we are relying on one measure only, we are somewhat reassured by the fact that problems with measurement would be expected to occur in the nonshared environmental variance, not in the genetic variance, which comprises the significant findings of this research.

It is interesting to note that behavioral inhibition, a precursor of anxiety disorders (Biederman et al., 1993), shows greater heritability than we found for physiological and social symptoms of anxiety in young children (Robinson et al., 1992). It could be that children who develop anxiety disorders inherit a genetically mediated temperament, like behavioral inhibition, but then do not develop anxiety unless they encounter certain environmental experiences. Alternatively, behavioral inhibition could show higher heritability because it is a more clearly defined construct than anxiety (it is based on observable behaviors rather than on internal feeling states) and thus is measured with less error.

Although the physiological and social subscales showed a genetic contribution, worry and the total anxiety scale did not. In addition, unique or nonshared environmental influences accounted for the largest variance for all of the subscales and total anxiety score. Other research would suggest that unique environmental influences do contribute in a significant way to the development of anxiety (Warren et al., 1997). However, it is not possible to statistically test this with the currently described twin study design.

An additional limitation of this research is that the subscales used may not be clearly enough defined as separate constructs to accurately delineate differences between them, and the differences may not be truly significant. Further research is needed in children, particularly focusing on clearly defined aspects of anxiety and on specific clinical diagnoses, such as separation anxiety disorder or social phobia. Besides understanding the genetics of specific anxiety disorders, it is also important to examine genetic contributions to anxiety at different ages and to link early anxiety with anxiety disorders in later childhood and adolescence, as is being done in the Virginia Twin Study (see Topolski et al., 1997).

The current study supports the results of previous research which indicates that genetic factors may play a role in the development of certain aspects of anxiety, namely physiological symptoms and social fears in young children. Even if genetic factors were found to predominantly influence certain aspects of anxiety, research on environmental factors is still essential because environmental and experiential factors influence the expression of genetic liabilities (Kandel, 1998). Research that clarifies specific environmental influences contributing to anxiety in the face of genetic risk, as well as to protective factors, can point to effective intervention strategies. Such research is therefore now key for the prevention and treatment of anxiety in young children and adolescents.

REFERENCES

Abe K, Oda N, Hatta H (1984), Behavioral genetics of early childhood: fears, restlessness, motion sickness and enuresis. Acta Genet Med Gemellol (Roma) 33:303-306

Biederman J, Rosenbaum JE Bolduc EA, Faraone SV, Hirshfeld D (1991), A high risk study of young children of parents with panic disorder and agoraphobia with and without comorbid depression. Psychiatry Res 37:333-348

Biederman J, Rosenbaum JE Bolduc-Murphy EA et al. (1993), A 3-year follow-up of children with and without behavioral inhibition. J Am Acad Child Adolesc Psychiatry 32:814-821

Bradley SJ, Hood J (1993), Psychiatrically referred adolescents with panic attacks: presenting symptoms, stressors, and comorbidity. J Am Acad Child Adolesc Psychiatry 32:826-829

Emde RN, Plomin R, Robinson J et al. (1992), Temperament, emotion, and cognition at 14 months: the MacArthur Longitudinal Twin Study. Child Dev 63:1437-1455

Freedman DG, Keller B (1963), Inheritance of behavior in infants. Science 140:196-198

Fyer AJ (1993), Heritability of social anxiety: a brief review. J Clin Psychiatry 54:10-12

Glasberg R, Aboud F (1982), Keeping one's distance from sadness: children's self reports of emotional experience. Dev Psycho 118:287-293

Goldsmith HH (1991), A zygosity questionnaire for young twins: a research note. Behav Genet 21:257-269

Goldsmith HH, Gottesman II (1981), Origins of variation in behavioral style: a longitudinal study of temperament in young twins. Child Dev 52:91-103

Gustavsson JP, Pedersen NL, Asberg M, Schalling D (1996), Origins of individual differences in anxiety proneness: a twin/adoption study of the anxiety-related scales from the Karolinska Scales of Personality (KSP). Acta Psychiatr Scand 93:460-469

Harrington R, Rutter M, Fenbone E (1996), Developmental pathways in depression. Dev Psychopathol 8:601-616

Harter S, Pike R (1984), The pictorial scale of perceived competence and social acceptance for young children. Child Dev 55:1969-1982

Kandel ER (1998), A new intellectual framework for psychiatry. Am] Psychiatry 155:457-469

Kendler KS, Neale MC, Kessler RC, Heath AC, Eaves LJ (1992), The genetic epidemiology of phobias in women. Arch Gen Psychiatry 49:273-281

Kendler KS, Walters EE, Truett KR et al. (1995), A twin-family study of self-report symptoms of panic-phobia and somatization. Behav Genet 25:499-515

Last CG (1991), Somatic complaints in anxiety disordered children, J Anxiety Disord 5:125-138

Last CG, Hersen M, Kazdin A, Orvaschel H, Perrin S (1991), Anxiety disorders in children and their families. Arch Gen Psychiatry 48:928-934

Legrand LN, McGue M, Iacono WG (1996), A twin study of state and trait anxiety in childhood and adolescence (abstract). Behav Genet 26:590

Neale MC (1995), Mx: Statistical Modeling, 3rd ed. Richmond: Medical College of Virginia, Department of Psychiatry

Nichols RC, Bilbro WC (1966), The diagnosis of twin zygosity. Acta Genet Stat Med 16:265-275

Pela OA, Reynolds CR (1982), Cross-cultural application of the Revised Children's Manifest Anxiety Scale: normative and reliability data for Nigerian primary school children. Psychol Rep 51:1135-1138

Perrin S, Last CG (1992), Do childhood anxiety measures measure anxiety? J Abnorm Child Psychol 20: 567-578

Phillips K, Fulker DW, Rose RJ (1987), Path analysis of seven fear factors in adult twin and sibling pairs and their parents. Genet Epidemio 14:345-355

Reynolds CR (1980), Concurrent validity of what I think and feel: the Revised Children's Manifest Anxiety Scale. J Consult Clin Psycho 148:774-775

Reynolds CR (1981), Long-term stability of scores on the Revised Children's Manifest Anxiety Scale. Percept Mot Skills 53:702

Reynolds CR, Richmond BO (1978), What I think and feel: a revised measure of children's manifest anxiety. J Abnorm Child Psychol 6:271-280

Reynolds CR, Richmond BO (1985), Revised Children's Manifest Anxiety Scale (RCMAS) Manual. Los Angeles: Western Psychological Services

Robinson JL, Kagan J, Reznick JS, Corley R (1992), The heritability of inhibited and uninhibited behavior: a twin study. Dev Psychol 28:1030-1037

Rose RJ, Ditto WB (1983), A developmental-genetic analysis of common fears from early adolescence to early adulthood. Child Dev 54:361-368

Schwab-Stone M, Fallon T, Briggs M, Crowther B (1994), Reliability of diagnostic reporting for children aged 6-11 years: a test-retest study of the Diagnostic Interview Schedule for Children-Revised. Am J Psychiatry 151:1048-1054

Silverman WK (1991), Diagnostic reliability of anxiety disorders in children using structured interviews. J Anxiety Disord 5:105-124

Stevenson J, Batten N, Cherner M (1992), Fears and fearfulness in children and adolescents: a genetic analysis of twin data. J Child Psychol Psychiatry 33:977-985

Thapar A, McGuffin P (1995), Are anxiety symptoms in childhood heritable? J Child Psychol Psychiatry 36:439-447

Thapar A, McGuffin P (1996), The genetic etiology of childhood depressive symptoms: a developmental perspective. Dev Psychopathol 8:751-760

Topolski TD, Hewitt JK, Eaves LJ et al. (1997), Genetic and environmental influences on child reports of manifest anxiety, and symptoms of separation anxiety and overanxious disorders: a community-based twin study. Behav Genet 27:15-28

Torgersen S (1993), Relationship between adult and childhood anxiety disorders: genetic hypothesis. In: Anxiety Across the Lifespan: A Developmental Perspective, Last CG, ed. New York: Springer, pp 113-127

Warren SL, Huston L, Egeland B, Sroufe LA (1997), Child and adolescent anxiety disorders and early attachment. J Am Acad Child Adolesc Psychiatry 36:637-644

Weissman MM, Wickramaratne P, Warner V et al. (1987), Assessing psychiatric disorders in children. Arch Gen Psychiatry 44:747-753

Dr. Susan L. Warren was Assistant Professor, Department of Psychiatry, University of Minnesota, Minneapolis. She is currently Assistant Professor, Department of Psychiatry, George Washington University, Washington, DC. Dr. Stephanie Schmitz is a Research Associate, Institute for Behavioral Genetics, University of Colorado, Boulder. Dr. Robert N. Emde is Professor, Department of Psychiatry, University of Colorado Health Sciences Center, Denver. Article A57645112

Mark Genetic studies of alcoholism and substance abuse. (Psychiatric Genetics '99)(Statistical Data Included) Theodore Reich, Anthony Hinrichs, Robert Culverhouse, Laura Bierut. American Journal of Human Genetics Sept 1999 v65 i3 p599(7)

Mark Testing the Genetics of Behavior in Mice. Science Sept 24, 1999 v285 i5436 p2067 View extended citation and retrieval choices

Peter McGuffin; Neilson Martin.

Abstract: Quantitative genetics indicates that genes have important influences on human behaviour but that this nearly always is combined with environmental elements. A basic pitfall in studying the genetics of behaviour is to assume that familiarity necessarily implies genetic transmission. But the individual is the result of both genetic and environmental forces. Monozygotic or identical twins are naturally occurring clones, having all of their genes in common. Dizygotic or fraternal twins share about half of their genes. Twin studies show that some traits are environmental such as bulimia, whereas schizophrenia, autism, and manic depressive disorder seem to be genetic.

New discoveries in genetics seem to hit the headlines almost daily, and probably the most eyecatching and controversial are those dealing with human behaviour. Thus, there has been popular media interest in reports of genes conferring susceptibility to psychiatric diseases and a whole range of "genes for" traits such as aggression, intelligence, and neuroticism. The scope for sensationalism and oversimplification is great. Here, we outline some of the basic concepts and lines of evidence from quantitative genetics indicating that genes do have important influences in determining human behaviour but that this nearly always involves an interplay with the environment. We then look at ways in which molecular methods are being used to locate and identify genes and how such approaches may impact on clinical practice.

We tend to resemble our parents, siblings, and other close relatives not just in the way we look but in the way we behave. However, the types, patterns, and causes of familial behavioural traits are extremely varied. They range from rare single gene disorders, such as Huntington's disease, that present with dramatic changes in behaviour in adult life, to commoner but more genetically complex disorders, such as schizophrenia and manic depression, through to normal variation in traits that are usually measured quantitatively, such as personality or intelligence.[1] They can also include characteristics such as political or religious persuasion and career choice.

One of the basic pitfalls in studying the genetics of behaviour (or indeed any other type of trait) is to assume that familiarity necessarily implies genetic transmission or that strong evidence of familial clustering implies single gene effects. Geneticists have long been aware that mendelian inheritance can be simulated by other mechanisms.[2] To illustrate this, McGuffin and Huckle studied the educational histories of the families of a cohort of medical students: they found that attending medical school was roughly 80 times more common in the students' first degree relatives than in the general population and that when complex segregation analysis (a computerised method of exploring the mode of transmission) was performed the "trait" closely resembled an autosomal recessive one).[3]

A more satisfactory basic model for complex traits is that the phenotype results from a combination of the genotype and file environment to which it is exposed. Environment can be split into two broad types--shared environment, which acts on all family members to make them similar, and non-shared experiences that are specific to individuals, which would be expected to cause differences in how family members behave.

Two types of "natural experiment" enable us to estimate the extent to which complex traits are familial because of shared genes, shared environment, or a combination of both. The first is afforded by twinning and the second by adoption.

Twin studies

The basis of twin studies is that monozygotic or identical twins are naturally occurring clones, having all of their genes in common. On the other hand, dizygotic or fraternal twins share on average half of their genes. If we assume that the environment shared by twins is roughly the same for monozygotic and dizygotic pairs, then any greater observed similarities in monozygotic than dizygotic pairs should reflect the action of genes. This "equal environments assumption" is open to criticism since there is some evidence that monozygotic pairs report greater environmental similarities (such as dressing the same and sharing friends) than do dizygotic pairs, especially during childhood.[4] Nevertheless, the assumption can be tested by incorporating environmental measures in twin studies or by looking at the effects of mistaken zygosity (the twins and their parents believe that they are monozygotic when they are actually dizygotic or vice versa). Sometimes it is also possible to compare the resemblance of monozygotic twins reared together with those reared apart. In practice, these checks suggest that the equal environments assumption is generally valid).[5]

Figure 1 summarises results from twin studies on various disorders and traits including psychological "symptoms" in normal twins. There are clear differences between monozygotic and dizygotic twins for some phenotypes--schizophrenia,[6] manic depressive disorder,[7] unipolar depression,[4] childhood autism,[8] attention deficit hyperactivity disorder,[9] and cognitive ability as measured by IQ test[1]--suggesting genetic effects. For bulimic behaviour[10] and disabling fatigue in childhood,[11] there is evidence of familiality (there are positive correlations) but less clear genetic effects (modest differences in the monozygotic and dizygotic correlations).

Adoption studies

Adoption studies have been used less extensively than twin studies but have been crucial in studying some conditions, such as schizophrenia. Heston studied 47 index adoptees separated from their schizophrenic mothers within 72 hours of birth and compared them in their mid-30s with 50 control adoptees who did not have schizophrenic parents).[12] Five of the index adoptees (11%) became schizophrenic (roughly the rate expected in non-adopted offspring of schizophrenics) compared with none of the controls. Subsequently, a series of studies was carried out in Denmark; the most recent study showed that the frequency of schizophrenia was 16% in the biological relatives of schizophrenics adopted early in life, compared with 1.8% in adoptive relatives and the relatives of control adoptees.[13]

Major advances in quantitative psychiatric genetics have occurred in the past two decades with the increased availability of high speed computing. This has enabled researchers to go beyond merely estimating heritability--that is, the proportion of total phenotypic variance explained by additive genetic affects. It is now possible to accurately quantify both genetic and environmental effects and assess the extent to which a reduced model (such as one with no additive genetic or no common environmental effects) explains the data compared with a full model. A recent application to schizophrenia showed that the heritability may be as high as 80%, and, although this leaves 20% of variance to be explained by the environment, this seems to be entirely of the non-shared type.[6] The table shows examples of behavioural disorders or traits that are thought to have a genetic component.

Some behavioural disorders and traits, their pattern of inheritance, and the status of gene mapping studies

Behavioural trait Pattern of inheritance Huntington's disease Rare autosomal dominant Familial Alzheimer's disease Rare autosomal dominant Late onset Alzheimer's disease Common complex Dyslexia Common complex Schizophrenia Common complex Aggression Common complex Homosexuality Common complex Behavioural trait Gene mapping Huntington's disease Gene identified (huntingtin) Familial Alzheimer's Three distinct genes identified disease (for presenilins 1 and 2 and amyloid precursor protein) Late onset Alzheimer's Increase risk with applipoprotein e4 disease allele firmly established Dyslexia Two contributory loci on chromosomes 6 and 15. Findings replicated Schizophrenia Several reported linkages, including chromosomes 6, 13, and 22, but no consensus. Promising candidate genes include that for [5-HT.sub.2A] Aggression X linked mutation reported in monoamine oxidase gene in a single family, but no evidence of broader relevance Homosexuality X linked locus reported in a study of affected sib pairs but not replicated

It is also possible to carry out multivariate model fitting that examines two or more disorders or traits at a time.[14] For example, life events are generally thought to have a causal relation with depression, but a recent study suggested that at least part of the co-occurrence of life events and depressive symptoms results from genetic and shared environmental influences on the extent to which both are reported.[15] A similar approach can be applied to psychiatric syndromes that often occur together. It had been shown that the same genes influence both anxiety and depressive disorder but that the environmental influences on the two types of disorder are apparently distinct.[16]

Such findings are not an end in themselves. Demonstrating that a disorder is substantially heritable or that a pair of disorders overlap genetically but have different environmental influences leads to two other types of study. The first is to identify the specific environmental factors that co-act or interact with genes, and the second is to locate and identify the genes themselves. Here we consider only the latter.

Finding genes

Mapping and positional cloning

Figure 2 shows the general process of positional cloning. A chromosomal region containing a gene that confers susceptibility to the disorder of interest is identified by linkage mapping. The region is then narrowed down by means of various methods until the gene itself is identified. Subsequently, the mutations (or variations) that confer susceptibility to disease are identified. The distribution, level of expression, and functions of the gene products can then be studied. The process of positional cloning carries the possibility of incremental benefits for clinical practice in the forms of predictive testing, refined diagnosis, and, eventually, the development of targeted specific treatments.

[Figure 2 ILLUSTRATION OMITTED]

However, conventional linkage analysis requires several assumptions: that there are major gene effects to be detected, that the sample is genetically homogeneous (that is, all of the affected individuals have the same cause of disorder), and that the mode of transmission of the disorder is known. A study of large, multiply affected pedigrees with early onset of Alzheimer's disease has identified mendelian subforms of the disease: three distinct forms have been mapped, and the mutations identified.[17] This has not been the case with other psychiatric disorders such as schizophrenia or manic depression. Here a bewildering and apparently contradictory array of positive linkage study results has been reported,[18 19] suggesting that single genes of large effect are rare or non-existent and that these disorders are oligogenic (resulting from the combined effects of several genes) or perhaps polygenic (resulting from many genes).

Sib pair analysis

An alternative approach that is useful in oligogenic or polygenic diseases is sib pair analysis. It is possible to take several hundred DNA markers roughly evenly spaced along the 23 pairs of human chromosomes and carry out genotyping in a series of sibling pairs of which both have the same disorder. The probability that siblings share 0, 1, or 2 alleles at any marker locus are, respectively, 0.25, 0.5, and 0.25. However, if a marker locus is close to (and therefore linked with) a locus conferring susceptibility to the disease this will be detectable as increased allele sharing at the marker. This approach has been successful in identifying susceptibility loci for disorders such as type 1 diabetes, and a variant of the method has been used to map genes involved in reading disability (or dyslexia) on chromosomes 6 and 15.[20]

The main drawback is that susceptibility loci of very small effect (such as conferring a relative risk of less than 2) may require large numbers of sib pairs--in the region of 600 to 800--to be detected.[19] In a disorder such as schizophrenia the relative risk in a sibling of an affected individual is about 10; thus, if several additive genes are involved, none may individually have a relative risk of more than 2.

Allelic association

Just as it is possible to search through the entire genome for disease susceptibility genes using classic linkage or sib pair methods, it is now becoming feasible to do the same thing using allelic association to detect linkage disequilibrium. Linkage disequilibrium occurs when a marker allele and the locus for susceptibility to a disease are so closely linked that their association is preserved over many generations of recombination. The advantage of mapping genes using linkage disequilibrium is that it can detect very small effects[1] and might the only way of finding the genes involved in polygenic disorders. The disadvantage is that, because linkage disequilibrium takes place only over very short distances, several thousand markers are required to complete a genome search. New methods of high throughput genotyping are now being developed that can accomplish this based on DNA pooling[21] or studying single nucleotide polymorphisms on microarrays.[22] The first "top to bottom" search of a whole chromosome by means of DNA pooling has now been published,[23] and such methods should allow whole genome scans for linkage disequilibrium to be completed in the foreseeable future.

Predicting the clinical impact of mapping genes

The most immediate benefit of identifying genes contributing to common familial psychiatric diseases will be in our understanding of the basic neurobiology of disease, but this should lead on to the discovery of new and more specific drug treatments. It is as much commercial farsightedness as scientific altruism that has led the major pharmaceutical companies into serious investment in genotyping technology and the development of detailed maps of single nucleotide polymorphisms.[24] Safer, more effective treatments will obviously be to the benefit of patients, but it is also possible to envisage DNA testing being used to predict which patients will respond to different types of antipsychotic or antidepressant treatment or who will be susceptible to particular side effects. Preliminary evidence suggests that these approaches will work. For example, response to antipsychotic drugs seems to be influenced by genotype at the gene for serotonin receptor 5-[HT.sub.2A].[25]

What about predicting whether someone will develop a psychiatric disorder and attempting to prevent it? Prediction is already possible for rare dementias of early onset, such as Huntington's disease, or single gene forms of Alzheimer's disease. Understanding the molecular basis of these conditions should lead on to the development of effective treatments and preventive methods, but none exists yet. Somewhat paradoxically, genetic predictive testing is much more difficult for commoner disorders such as depression or schizophrenia, where effective treatments exist but were the genetic basis is complex. Psychiatric genetic counselling is already available in some specialist centres, but at present it is possible only to offer empirical figures for those at high genetic risk. For example, the child of a schizophrenic parent has about 10 times the risk of developing the disorder compared with a member of the general population, where the lifetime risk is 1%. The child's risk increases to 16 time if a sibling is already affected and is over 40 times file population risk when both parents have schizophrenia.[26]

It is likely that individual risk prediction will become more precise once the molecular genetic basis of schizophrenia is better understood, but it is unlikely that risk prediction will ever be better than about 50% accurate since monozygotic twins are discordant for schizophrenia 50% of the time. In our view, this means that DNA based population screening for complex psychiatric disorders (including Alzheimer's disease of late onset) will never become a reality but that screening for high risk relatives probably will. This could have obvious benefits in advising relatives on avoiding risk factors (such as the recreational use of cannabis in the case of schizophrenia) or, more controversially, in attempting prevention with a low dose antipsychotic agents.

There will also be less immediately tangible benefits to sufferers from psychiatric disorders. It has sometimes been feared that "geneticisation" could contribute to the stigma of mental disorder. So far the experience has been just the opposite. Alzheimer's disease is now widely perceived as a "real" disorder with a rapidly unfolding molecular aetiology. We predict that this is the start of a trend and that identifying the genes involved and understanding causation will do much to improve public perception and acceptance of other psychiatric disorders.

Funding: PMcG is director of an MRC funded research centre. NM holds an MRC studentship.

Competing interests: None declared.

Predicted developments

Identification of relevant genes will improve understanding of the molecular neurobiology of psychiatric disorders

This will lead to the development of more efficacious and more specific drugs

DNA testing may be used in predicting response to treatment and susceptibility to side effects

DNA testing will be useful in counselling patients' relatives at high risk of heritable disease but not for population screening

Public perception of psychiatric disorder will change: improved understanding of the causes and mechanisms of disease is likely to reduce stigma

[1] Plomin R, Owen MJ, McGuffin P. The genetic basis of complex human behaviours. Science 1994;264:1733-9.

[2] Edwards JH. The simulation of mendelism. Acta Genet 1960;10:63-70.

[3] McGuffin P, Huckle P. Simulation of mendelism revisited: the recessive gene for attending medical school. Am. J Hum Genetics 1990;46:994-9

[4] McGuffin P, Katz R, Rutherford J, Watkins S. The heritability of DSM-IV unipolar depression. A hospital based twin register study. Arch Gen Psychiatry 1996;53:129-36.

[5] Plomin R, DeFries JC, McClearn GE, Rutter M. Behevioral genetics. New York: WH Freeman, 1997.

[6] Cardno AG, Marshall EJ, Macdonald AM, Coid B, Ribchester TR, Davies NJ, et al. Concordance rates and biometrical model fitting for operational diagnoses in the Maudsley twin psychosis series. Arch Gen Psychiatry 1999;56:162-8.

[7] Bertelsen A, Harvald B, Gauge M. A Danish twin study of manic-depressive disorders. Br J Psychiatry 1977;130:330-51.

[8] Folstein S, Rutter M. Infantile autism: a genetic study of 21 twin pairs. J Child Psychol Psychiatry 1977;18:297-321.

[9] Thapar A, Holmes J, Poulton K, Harrington R. Genetic basis of attention deficit and hyperactivity. Br J Psychiatry, 1999;174:105-11.

[10] Rutherford J, McGuffin P, Katz R, Murray RM. Genetic influences in eating attitudes in a normal female twin population. Psychol Med 1993;23:425-36.

[11] Farmer A, Scourfield J, Martin N, Cardno A, McGuffin P. Is disabling fatigue in childhood influenced by genes? Psychol Med 1999;29:279-82.

[12] Heston LL. Psychiatric disorders in foster home reared children of schizophrenic mothers. Br J Psychiatry, 1966;112:819-25.

[13] Kety SS, Wender PH, Jacobsen B, Ingraham LJ, Jansson L, Faber B, et al. Mental illness in the biological and adoptive relatives of schizophrenic adoptees. Replication of the Copenhagen study in the rest of Denmark. Arch Gen Psychiatry 1994;51:442-55.

[14] Neale MC, Cardon LR. Methodology for genetic studies of twins and families. Dordrecht: Kluwer Academic Publishers, 1992.

[15] Thapar A, Harold G, McGutfin P. Life events and depressive symptoms--shared genes or shared adversity? A research note. J Child Psychol Psychiatry 1998;39:1153-8.

[16] Kendler KS, Heath A, Martin NG, Eaves LJ. Symptoms of anxiety and depression in a volunteer twin population. The etiologic role of genetic and environmental factors. Arch Gen Psychiatry 1986;43:213-21.

[17] McGuffin P, Owen MJ, O'Donovan MC, Thapar A, Gottesman II. Seminars in psychiatric genetics. London: Royal College of Psychiatrists, 1994.

[18] Risch N, Botstein D. A manic depressive history. Nature Genet 1996;12:351-3.

[19] Moldin SO. The maddening hunt for madness genes. Nature Genet 1997;17:127-9.

[20] Grigorenko EL, Wood FB, Meyer MS, Hart LA, Speed WC, Shuster A, et al. Susceptibility loci for distinct components for developmental dyslexia on chromosomes 6 and 15. Am J Hum Genet 1997;60:27-39.

[21] Daniels J, Holmans P, Williams N, Turic D, McGuffin P, Plomin R, et al. A simple method for analysing microsatellite allele image patterns generated from DNA pools and its application to allelic association studies. Am J Hum Genet 1998;62:1189-97.

[22] Asherson P, Curran S, McGuffin P. Molecular genetics--approaches to gene mapping. CNS 1998;1(4):18-22.

[23] Fisher PJ, Turic D, McGuffin P, Asherson P, Ball D, Craig I, et al. DNA pooling identifies QTLs for general cognitive ability in children on chromosome 4. Hum Mol Genetics 1999;8:915-22.

[24] Woodman R. Wellcome Trust and drug giants fund gene marker database. BMJ 1999;318:1093.

[25] Saikh S, Makoff A, Collier D, Kerwin R. Dopamine D4 receptors potential therapeutic implications in the treatment of schizophrenia. CNS Drugs 1997;8(1):1-11.

[26] Gottesman II, Aston SJ, Moldin SJ. Schizophrenia genetic risks: a guide to genetic counselling for consumers, their families, and mental health workers. 2nd ed. Arlington, VA: NAMI, 1999.

Social Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, Kings College London, London SE5 8AF

Peter McGuffin, director

Division of Psychological Medicine, University of Wales College of Medicine, Cardiff CF4 4XN

Neilson Martin, MRC PhD student

Correspondence to: P McGuffin p.mcguflin@iop.kcl. ac.uk Article A55249521

My four-year-old son asked for a Barbie this year. His blue eyes were hopeful, his small face angelic. His mother was suspicious.

Between this child and his older brother, our house is a Toys R Us warehouse of heavily muscled action figures, dinosaurs with jagged teeth, light-up swords and leaking water pistols. Complaint is constant--Oh, Mom, you're no fun--over my refusal to buy more additions to the arsenal. My older son at one point began to see weapons in household objects the way adults dream up phallic symbols. "Shoot her with the toothbrush," he once shouted to a companion as they chased the cat around the house.

"Why do you want the Barbie, honey?" I asked.

"I wanna chop her head off."

There I was again, standing at the edge of the great gender divide, the place and the moment where one becomes absolutely sure that the opposite sex is, in fact, opposite. I know of no way for women of my generation, raised to believe in gender neutrality, to reach this edge faster than through trying to raise children.

"I did not do this," a friend insisted on the day her son started carefully biting his toast into the shape of a gun. "I think my daughter has a pink gene," a British journalist confided recently, as she confessed that her daughter has not only a Barbie collection but all the matched plastic purses and tiny high-heeled shoes. I don't think in pastels myself. I think jungle-green, blood-red. Most of all, I think there's a reason--a reasonable biology--to the differences we see in little boys and girls, men and women, males and females.

We are, I hope, moving past the old politically correct notion that we are pure culture, that children are born blank slates to be influenced--or, worse, manipulated--by the adults around them. There's a straightforward reason why we are a male-female species: Reproductively, it works. We are all born with bodies designed to be the same (breathe, circulate blood) and to be different (produce sperm, produce eggs, produce milk, produce none). There's an internal biology--structural and behavioral--that supports those differences. It's not all of who we are, but it's a part. When is biology the primary influence? Where does culture overtake it, and at what point, in the startlingly fluid landscape of human behavior, does one alter the other?

One of my favorite illustrations of the way culture fine-tunes us for gender roles has to do with the Barbie versus Godzilla effect. It turns out that lots of little boys ask for dolls and other so-called girl toys. They aren't encouraged, though; parents really hesitate to buy their children "gender inappropriate" toys. In a study involving almost 300 children, researchers found that if little boys asked for a soldier equipped with battle cannons for their birthday, they got it some 70 percent of the time. If they asked for a Barbie doll, or any of her plastic peers, the success rate was 40 percent or less. Can you think of a child who wouldn't figure out in, oh, a day, how to work that system?

Marc Breedlove, a neuroscientist at the University of California, Berkeley, points out that splitting apart biology and culture is analogous to splitting hairs. But scientists try to separate the strands anyway, exhaustively exploring early development. A few ambitious scientists have even looked for prebirth differences, arguing that it's difficult to slap too much cultural attitude onto a fetus. It turns out that boy fetuses are a little more active, more restless, than girl fetuses. And in the first year after birth, toy preferences already seem distinct: Boy infants rapidly engage with more mechanical or structural toys; little girls of a few months gravitate toward toys with faces, toys that can be cuddled.

The world of play--the toys we gravitate to, how we play with them, how we play in general--has now become serious business to scientists. Today's hottest theory of play is that it's a practice run at the challenges of adult life. Through games, the experts tell us, we learn the art of measuring the competition, how to win and lose gracefully (we hope), which leads pretty directly into how to build friendships. In scientific terms, we learn socialization.

"Play offers a non-life-threatening way of asserting yourself," says Christine Drea, a researcher. "By playing, you learn skills of managing competition and aggression." We are a social species. We find isolation destructive, and we establish patterns of childhood play that reflect adult social structures. In humans, our patterns tend to conform to our chemistry: Human males are likely to produce seven to 10 times more testosterone, for example, than females. (Hyenas are quite a different story; see box, page 52.) And so, you would correctly predict, little boys tend to be more rough-and-tumble than little girls. That's true, in fact, for the entire realm of primates (monkeys, apes, man).

Back in the late 1970s, Robert Goy, a psychologist at the University of Wisconsin, first documented that young male monkeys consistently played much more roughly than juvenile females. Goy then went on to show that if you manipulate testosterone level--raising it in females, cutting it off in males--you reverse those effects, creating sweet little boy monkeys and rough-and-tumble girls. We don't experiment with human development this way, obviously. But there are naturally occurring genetic variations that make closely comparable points. As mentioned earlier, human males circulate higher levels of testosterone. There's a well-known exception, however, called congenital adrenal hyperplasia (CAH), in which a baby girl's adrenal gland inadvertently boosts testosterone levels. Researchers have found that CAH girls, in general, prefer trucks and cars and aggressive play. That doesn't mean they don't join in more traditional girl games with friends--but if left to choose, they prefer to play on the rowdy side of the street.

Higher testosterone levels are also responsible for another characteristic: competitiveness. In fact, testosterone is almost predictable in this regard. It shoots up before a competition; that's been measured in everything from chess matches to soccer games to courtroom battles to brawls. It stays up if you win, drops if you lose. Its role, scientists think, is to get you up and running and right on the competitive edge.

Even in preschool, boys and girls fall into very different play patterns. Boys tend to gather in larger, competitive groups. They play games that have clear winners and losers and bluster through them, boasting about their skills. Girls, early on, gather in small groups, playing theatrical games that don't feature hierarchy or winners. One study of children aged three to four found they were already resolving conflict in separate ways--boys resorting to threats, girls negotiating verbally and often reaching a compromise.

There are some provocative new insights into that verbal difference. Recently, researchers at Emory University have found that little female monkeys are much quicker to pick up "verbal" skills than little boy monkeys. Sound familiar? The small female monkeys do more contact calling (cooing affectionately) than their male counterparts. And it appears, again, that this is related to their mothers' prenatal hormones. Some very preliminary tests suggest that females exposed to androgens early in their fetal development become more like male monkeys: They are less likely to use language to express themselves.

In humans, too, we look for natural biological variations. In general, girls have sharper hearing than boys--the tiny hair cells that register sound waves vibrate more forcefully. These are ears tuned for intense communication. (The rare exception tends to be in boy-girl twin pairs. Those girls are more likely to have ears built a little more like their brothers'--less active hair cells, notched-down response. Researchers looking at this suspect a higher exposure to androgens in utero.) There's something about the biology of the egg-producing sex that seems to demand more acute communication abilities.

Of course, there's a whole range of personalities and behaviors that don't fall into any of the obvious stereotypes. What about tomboys, those exuberant girls who prefer softball to tea parties? What about the affectionate sweetness of little boys, who--away from the battle zone of their friends and brothers--turn out to be surprisingly cuddly and clingy? What about the female stiff, the chatty male, and so on, into infinity? The quick answer: Sex differences are group differences, overall patterns.

The complex of genes and hormones and neurotransmitters and internal chemistry that may influence our behaviors varies from person to person and is designed to be flexible. There's nothing in average, everyday biology that forbids either the truck-loving girl or the boy who likes to play house, the aggressive, competitive adult woman or the nurturing, stay-at-home man. Human biology makes room for every possible type of personality and sexuality in the range between those stereotypes.

And finally, the way we behave can actually influence our biology. The link between testosterone and competition makes this point perfectly. Yes, corporate lawyers tend to have higher testosterone levels than ministers. But there's a chicken-or-egg aspect to this. Is the lawyer someone born with a high testosterone level? Or is it the profession that pushed it up? Or some combination of both? It's worth noting that the parallel works in men and in women; women in competitive jobs have more testosterone; men who stay home with their children have less.

Nothing in biology labels behaviors as right or wrong, normal or abnormal. Any stereotypes we impose on children--and, by extension, adults--are purely cultural, not biological. For example: Little boys are noisy and rambunctious; we tend to equate that with being emotionally tough. But what science actually tells us is the exact opposite. Little boys, we're learning, need a lot of emotional support. One revealing study of children of depressed and withdrawn mothers, done at U.C. Berkeley, found that a lack of affection actually lowered the IQ of little boys. Laura Allen, a neuroscientist at the University of California, Los Angeles, explains it like this: "I think boys need more one-on-one attention. I think affection may change the sex hormone level in the brain, which then affects brain development." Both the Berkeley study and a more recent federal daycare study find a different pattern in girls. They're emotionally sturdier--I think most of us have already figured this out--and their healthy development seems most harmed by being restricted. It's confinement that seems to drive down IQ in our daughters.

What's the real difference between boys and girls? More, and less, than we thought. With rare exceptions, the anatomy of gender is straightforward, separate. But the chemistry of gender is more complex. It's a continuum, I think, and we can each find a place within the wide band of "normal." What's more, we can change our place. And we can influence our children's places--not by force but by guidance.

And so, if you're wondering, I did not buy my son the doll. I'm too grown up these days to approve of dismembering pricey toys. I did let him pick out a scaled-down Barbie, instead of a toy car, in one of those fast-food kid's meal promotions. It turned out to be cream and gold in appearance, annoyingly indestructible, and he lost interest. These days, he likes to make books and draw pictures of blood-dripping dinosaurs. Me? I pass him the red crayons.

[BOX]

Further reading: Among the best of the recent books that explore the shifting sands of gender identity are Deborah Blum's Sex on the Brain: The Biological Differences Between Men and Women (Penguin) and Natalie Angier's Woman: An Intimate Geography (Houghton Mifflin).

[BOX]

Boy Meets Girl

Timing is everything, especially when it comes to sex. So, forget Biology 101: Gender depends on a lot more than simple X's and Y's.

We know that each egg cell contains one X chromosome, and sperm cells carry either an X or a Y. When an egg is fertilized, it's the sex chromosome in the sperm that determines the potential gender of the baby. But the chromosomal sex (XX or XY) won't always match the baby's physical appearance.

During the crucial first trimester, the embryo's sex-related genes go to work. A gene on the Y chromosome triggers a flood of androgens from the embryo's developing testes. Androgens, such as testosterone, are steroid (sex) hormones--chemical messengers that eliminate potential femininity and ensure that an XY baby is born looking and functioning like a "normal" male. (Later on in life, testosterone will also give him a deep voice and a beard.)

A genetically female embryo's adrenal glands (which produce hormones) also secrete androgens, but in smaller quantities. Instead, the embryo (with help from Mom) produces--and is bathed in--loads of estrogen. By the ninth week, her ovaries will be neatly tucked into place; by the end of the 12th, an ultrasound will reveal a female. But if too many androgens are floating about, an XX baby can end up with male equipment. That's where the timing comes in.

The process usually works amazingly well. Indeed, to call what happens the miracle of birth is almost an understatement.

[BOX]

The Mean, Lean (Female) Fighting Machine

Christine Drea, currently a researcher at U.C. Berkeley, studies the fast-moving, quick-witted African spotted hyena. (The hyena colony at Berkeley is so extensive that Disney illustrators went there, rather than to Africa, to study up for The Lion King. The scientists tried to convince Disney that hyenas should be the heroes.) Hyenas are the natural world's demonstration of the effects of androgens, the steroid hormones that include testosterone.

The hyenas' unique biology steeps the female fetuses in an androgen-rich soup; the placenta pumps testosterone directly to the young. In the male young, this creates normal masculinity. But for the females, it's transforming. Female hyenas are literally born to fight--teeth already in place, biting their siblings as soon as they tumble out. The females are bulkier, meaner and noticeably more aggressive than their brothers.

"The females have aggression down pat at birth," Drea says. Only later do they manifest such behaviors as play and social bonding. Researchers found that females--apparently revved on androgens--are the group rowdies; males are the low-key players. But all their androgens, Drea notes, do not interfere with the females' ability to mother.

The more hostility people show in their behavior and attitudes, the more likely they are to have calcium deposits in the arteries of their hearts, a new study finds. Such calcium deposits are early signs of atherosclerosis, also called hardened arteries, which can lead to high blood pressure and heart attacks.

Researchers led by Carlos Iribarren of Kaiser Permanente's research division in Oakland, Calif., divided 376 men and women aged 18 to 30 into two groups according to assessments of their hostility. Volunteers were characterized as hostile if they had a cynical view of the world and aggressive responses to stressful situations.

The more hostile participants were two and a half times more likely than the less hostile participants to have calcium deposits in their heart arteries. The hostile subjects were also more than nine times as likely to have enough calcium deposits in their arteries to indicate the beginnings of fatty plaques, hallmarks of atherosclerosis, says Iribarren.

The link between hostility and the early signs of heart disease held true even after the researchers took into account differences in education, weight, smoking, and blood pressure. Since the study participants were relatively young, few showed even early signs of heart disease, so larger and longer studies are needed to strengthen the link, Iribarren says. Article A54525603

D. Wahlsten.

As traditional behavioral genetics analysis merges with neurogenetics, the field of neurobehavioral genetics, focusing on single-gene effects, comes into being. New biotechnology has greatly accelerated gene discovery and the study of gene function in relation to brain and behavior. More than 7,000 genes in mice and 10,000 in humans have now been documented, and extensive information about the genetics of several species is readily available on the World Wide Web. Based on knowledge of the DNA sequence of a gene, a targeted mutation with the capacity to disable it can be created. These knockouts--also called null mutants-- are employed in the study of a wide range of phenotypes, including learning and memory, appetite and obesity, and circadian rhythms. The era of examining single-gene effects from a reductionistic perspective is waning, and research with interacting arrays of genes in various environmental contexts is demonstrating a need for systems-oriented theory.

KEY WORDS: Human Genome Project, quantitative trait locus, reductionism, World Wide Web, targeted mutation

INTRODUCTION

Unlike previous behavioral genetics reviews in this series, which divided the field by species (human and nonhuman) (Wimer & Wimer 1985, Rose 1995), this review divides the field according to single-gene and biometrical methodologies. It focuses on remarkable progress and prospects in the discovery and understanding of specific gene effects in several species, including humans. Many neurological mutations exhibiting large effects have had their DNA sequences decoded and their protein products identified, and much has been learned about how gene expression is regulated by the environment. Behavioral genetics researchers have advanced to a new stage, and have now begun to examine interacting pairs of genes and to identify viable genetic variants that exert more subtle effects on behavior. As the field of neurobehavioral genetics emerges, genetic tools are becoming central to research in physiological psychology.

Typically, the biometrical or quantitative genetic approach is applied when many unknown genes, each with presumably small effect, are believed to be involved. Instead of identifying specific genes, this methodology seeks to partition variance among several components attributable to genetic and environmental variation. There has been a tension between the two approaches since the early days of genetics, as reflected in Johannsen's (1911:138) opinion of the correlational methods employed by Francis Galton and Karl Pearson: "They have nothing at all to do with genetics--or general biology! Their premises are inadequate for insight into the nature of heredity." This tension continues. Although the mathematical models of biometrics have undoubtedly become more sophisticated and are being applied to both nervous system and behavioral analysis (Rijsdijk & Boomsma 1997), fundamental disagreements abound concerning the basic formulation and assumptions of the quantitative models (Devlin et al 1997, Schonemann 1997, Wahlsten 1990, 1994), and many practitioners of quantitative genetics are being drawn to the study of single-gene effects (Boomsma et al 1997, McClearn et al 1991). In the view of Plomin and associates (1994), "additional quantitative genetic studies are no longer needed to document the importance of genetic influence" (i.e. heritability) on intelligence, and researchers should instead attempt to identify specific genes.

The ultimate goal of behavioral and neural genetics is a comprehensive understanding of the identities, functions, and multifarious relations of genes relevant to the behavior of organisms. In this regard, it is important to know how many distinct genes a species possesses, how many of these have already been identified, and how many are likely to be important for behavior.

The Compleat Genome

The Human Genome Project seeks to determine the entire sequence of the nucleotide bases (A, C, G, T, or adenine, cytosine, guanine, thymine) in the DNA of the chromosomes of several species. The overall size of the genome in terms of millions of bases (Mb) is listed for several species in Table 1. Once the entire DNA sequence is known, molecular biologists can identify every unique gene by noting the telltale signatures of base sequences that indicate where to start and stop the transcription of DNA into messenger RNA (mRNA). Each mRNA molecule is translated into a protein molecule. If one knows the DNA sequence of a gene, the structure of its corresponding protein can be readily deduced from the genetic code. This has been accomplished in several unicellular organisms, including a yeast with 6297 genes.

Table 1 Size of the genome in relation to the number of genes, proteins, and neurons in several species that are intensively studied in behavior genetics(a)

Genome Known Nerve Species (Mb) Genes proteins cells Yeast (Saccharomyces 13.50 6,297 6,297 1 cell cerevisiae) Nematode worm (Cae- 100 14,000 11,274 302 norhabditis elegans) Fruit fly (Drosophila 165 12,000 1,566 250,000 melanogaster) House mouse (Mus 3,300 70,000 7,161 40 million domesticus) Human being 3,300 70,000 11,060 85 billion (Homo sapiens) Species Web sites Yeast (Saccharomyces genome-www. cerevisiae) stanford.edu Nematode worm (Cae- elegans.swmed.edu norhabditis elegans) Fruit fly (Drosophila flybase.bio.indiana.edu melanogaster) House mouse (Mus www.informatics.jax.org; domesticus) biomednet.com Human being bioinfo.weizmann.ac.il/ (Homo sapiens) cards; gdbwww.gdb.org; www3.ncbi.nlm.nih.gov/ Omim

1. Sources: Miklos & Rubin (1996), Henikoff et al (1997), Gottlieb (1998).

The task of sequencing is immensely more tedious in vertebrates because the segments of DNA (exons) that are transcribed into mRNA and translated into protein are interspersed by numerous and large segments (introns) that do not code for protein. It is estimated that in humans and mice the informative exons comprise a paltry 2% of the total DNA; a fabulously expensive effort to date has completed the sequencing for only 2% of the human and 0.2% of the mouse genomes (Rowen et al 1997). Once the numbers of genes in DNA that has already been sequenced are known, the total number can be estimated (Table 1). The process of gene identification in vertebrates can be greatly accelerated by studying the mRNA expressed in a variety of tissues from different age groups. This mRNA can be reverse-transcribed into complementary DNA (cDNA), which consists entirely of exons that can be analyzed to yield expressed sequence tags (ESTs). This has been accomplished on a large scale for the human genome, and ESTs exist for one or more exons of some 40,000 of the expected total of 70,000 human genes (Rowen et al 1997). Progress toward a relatively complete accounting of expressed genes can be assessed by the number of protein structures known in a species (Table 1).

Sequencing the genome of convenient, nonhuman species has major benefits for gene discovery in humans because many genes and proteins are homologous owing to descent from a common but remote ancestor. For example, at least 1914 of the 6297 proteins of the yeast S. cerevisiae have homologs in mammals (Botstein et al 1997), and homology is substantially greater in more closely related species such as mice and humans (see www.informatics.jax.org/reports.html for an extensive list of homologies).

Of the tens of thousands of genes in a mammal, how many might be relevant for understanding nervous system development and behavior? This question can be approached directly. Once the DNA sequence of an exon of a gene is known, a custom DNA probe can be constructed and then inserted into that specific gene (Joyner 1993). This procedure creates a targeted mutation that usually prevents synthesis of the corresponding protein (called a knockout, or null mutation), but it is also possible to change only one specific amino acid in a protein (Giese et al 1998). The mouse is the preferred subject for this technique, and the 129 inbred strain is commonly the source of cells that are genetically altered. Because one common substrain (129/SvJ) has been genetically contaminated (Simpson et al 1997, Threadgill et al 1997) and the 129 strain, like all inbred strains, has a number of neural and behavioral abnormalities, interpretation of results is sometimes clouded (Crawley et al 1997, Gerlai 1996, Wahlsten & Sparks 1995, Wolfer et al 1997). Nevertheless, the knockout technique is invaluable and can be refined to address earlier shortcomings. Hundreds of kinds of mice have been created that lack a specific protein (such as the estrogen receptor from the Estra gene), and numerous mouse models of human genetic diseases have been created by altering the relevant gene (e.g. the Fmr1 knockout model of Fragile X mental retardation).

The null mutation is a relatively blunt instrument, but in many Instances researchers have been surprised to obtain viable animals that experienced only minor damaging effects or showed no perceptible effects at all. For example, mice with a disabled dopamine [Beta]-hydroxylase gene (Dbh) are unable to synthesize norepinephrine and have motor difficulties, but are otherwise able to learn reasonably well (Thomas & Palmiter 1997). Because the use of small sample sizes is common in work with knockout mice, most such experiments lack statistical power to detect small or moderate effects and make it risky to proclaim the genesis of a completely normal mouse. Furthermore, researchers usually focus on one phenotype and target genes of particular interest, implying that the extant sample of mutations is not at all representative of the mouse genome. A functional scan of the entire genome by knocking out one gene at a time is now feasible for yeast; in the near future the scan may also be applicable to nematode worms, but not to more complex animals.

Another approach is to create random mutations (many of which will occur in unknown genes) and record how many of these then impair development of an organ. Although a precise number cannot be ascertained at present, available data suggest that thousands of genes--perhaps as many as 70% of all genes--are required for the normal development of a complex organ such as the eye (Miklos & Rubin 1996).

Linkage and Chromosome Mapping

DNA-based technology can reveal all genes, whether or not they have alternate forms (alleles) that create protein polymorphism, or individual differences in behavior in a population. Consequently, much of this genetic information is of less interest to psychologists for whom the relatively few genes pertinent to behavioral disorders provide more relevant information. The classical approach to genetics begins with a noteworthy difference in phenotype and then asks whether inheritance follows Mendelian rules and whether the hypothetical gene is linked to a marker at a known location on a chromosome. The search for linkage has been greatly facilitated by the discovery of thousands of phenotypically neutral and highly polymorphic DNA markers scattered widely across the genome of mammals (Dietrich et al 1995). If a mutation in an unknown gene with major effects on brain or behavior occurs, it is now possible to detect it quickly and locate it accurately on a chromosome map.

A good example is provided in mice by the barrelless (brl) mutation that obliterates the normal barrel-shaped pattern of neuron assemblies in somatosensory cortex. The first description of the phenotype was published recently (Welker et al 1996), and fewer than 2 years later it was mapped to a narrow zone on chromosome 11 (Abdel-Majid et al 1998) that was already known to contain six other genes. These six became plausible candidates for brl, and the search quickly narrowed to focus on the gene Adcy1, which codes for the enzyme adenylyl cyclase type I, an important part of an intracellular signalling pathway involving cyclic AMP (cAMP) in neurons. As it turned out, that enzyme had reduced activity in the mutant mice, and an unrelated knockout strain that lacked a functional Adcy1 gene was found to lack the brain barrels. The mutant mice also suffered memory deficits. The causes of the barrel structures were also clarified. This distinct pattern is impressed on the cerebral cortex by neural input from the vibrissae in the animal's snout, and alteration of the cAMP pathway by a mutation prevents the anatomical imprint of sensory experience. Now that the gene is better understood, it is properly referred to as the adenylyl cyclase type I gene, and the mutation becomes the loss of function allelle [Adcy1.sup.brl].

Genetic mapping with neutral markers also works well in human subjects and has recently been used to detect genes pertinent to many rare neurological disorders. Most cases of unequivocally successful mapping of disease genes have involved dichotomous phenotypes that differ distinctly between normal and abnormal individuals, and where the mutation has a large effect. Tremendous efforts have been made to detect linkage with hypothetical genes pertaining to some of the more common psychiatric disorders that fall into rather arbitrary diagnostic categories, such as manic depression and schizophrenia. Several published claims of linkage have proven to be false positives, and the most recent evidence for linkage of schizophrenia with markers on chromosomes 6 and 8 remains only weakly suggestive (Kidd 1997, Moldin 1997, Moldin & Gottesman 1997). Given the many studies done on this topic, it is reasonable to conclude that no single gene contributes in a major way to the etiology of schizophrenia. Any genetic influence most likely involves the "nonlinear interaction of multiple genetic and environmental factors" (Cloninger 1997), and these effects will be very difficult to identify with conventional linkage analysis that assumes all effects to be independent (Kidd 1997).

Quantitative Trait Loci

Most behavioral variation is continuous and most genetic effects are probably not very large in the normal range of variation. A moderate effect size of a quantitative trait locus (QTL) can be detected by its linkage with neutral DNA markers (Belknap et al 1997, Lander & Schork 1994, McClearn et al 1991). The results are most readily interpreted when the experiment begins with two inbred strains because there can only be two alleles, and the marker alleles will be known in both strains. In an [F.sub.2] hybrid cross, genotype frequencies will have Mendelian ratios at the marker. The closer the marker locus is to the hypothesized gene, the lower will be their recombination probability. Thus, if the QTL has an appreciable effect on behavior, there should be a statistically significant difference in mean behavioral test scores of individuals with different marker genotypes. By examining several markers on the same chromosome and using the MAPMAKER computer program, the QTL can be localized within a confidence interval.

Two major difficulties challenge the users of QTL methodology. (a) A scan of the whole genome typically involves several markers on each of 20 independent chromosomes in mice, and in humans, 23. Thus, there is an appreciable risk of a false positive association when the conventional Type I error probability [Alpha] = 0.05 is used for each test, so most of the QTL harvest will probably be spurious. Lander & Krugylak (1995) argued persuasively that researchers should use [Alpha]= 0.0001 for each test to keep Type I error at 5% for the entire linkage study. (b) Even if the evidence for existence of a QTL is compelling, the width of the confidence interval along the chromosome may still be too great to allow for rapid gene identification and sequencing. A 1% recombination frequency corresponds to a distance along the chromosome of about 1 centiMorgan (cM), that in mice contains about 2 Mb of DNA and about 65 genes. A review of 22 QTLs believed to be important for alcohol and drug sensitivity (Crabbe et al 1998) found that the interval in most cases was more than 15 cm. If the QTL can be localized only within a 15 cM interval, it could be any one of about 1000 genes (Belknap et al 1997).

Many claims of QTLs assigned to map locations have now been published in the behavioral genetics literature, and in some cases provisional gene symbols have been proposed. In many cases, the validity of these claims is suspect and the field would benefit from greater circumspection and rigor. Crusio (1998) remarks that "on closer examination, as yet the promise of the QTL method has not been fulfilled at all." It makes good sense to reduce Type II errors by casting a wide net in the first phase of a study, but it seems unwise to claim something has been mapped or provisionally mapped merely because there is statistically significant evidence of linkage. Further confirmatory testing should be mandatory to cull the false positives and substantially narrow the confidence interval (Darvasi 1998). Real success should be recognized not in long lists of weakly substantiated QTLs but in one or two conclusive discoveries of genetic variants with moderate effects.

Several fruitful strategies for confirming hypothetical QTLs are available (see Crabbe et al 1998, Darvasi 1998). Buck and coworkers (1997) studied severity of alcohol withdrawal symptoms in mice derived from the strains C57BL/6J and DBA/2J. An initial screening against 1522 genetic markers in 21 recombinant inbred (RI) strains yielded seven chromosomal regions appropriately designated as showing "potential linkage" with a "putative QTL." In a sample of 451 [F.sub.2] hybrid mice evaluated only at regions implicated in the first phase, three of these were clearly supported and another weakly supported (see Belknap et al 1996). The researchers then selectively bred two lines of mice for high and low withdrawal severity, and the allele frequencies at three marker loci diverged rapidly and significantly, thereby confirming the existence of three QTLs with independent evidence in the predicted direction. Although map locations suggested plausible candidate genes--including several GABA receptor subunits on chromosome 11--95% confidence intervals were more than 10 cM wide.

A similar approach has been employed to study the acute response (loss of righting reflex) to a high dose of ethanol. When 124 markers were tested in 27 RI strains derived from the long sleep (LS) and short sleep (SS) lines, 11 "provisional QTLs" were located with a very lenient ([Alpha] = 0.05) criterion (Markel et al 1996). A study of the mice with the most extreme scores in a sample of 1072 [F.sub.2] hybrids supported only two of these QTLs, which were tentatively localized within intervals of about 16 cM (Markel et al 1997). As a further test, [F.sub.2] hybrid mice of known genotype at the marker loci flanking the hypothetical QTL were then mated and their offspring tested for ethanol-induced sleeping (Bennett et al 1997). Although the sample sizes were too small to yield conclusive results, this application of marker-assisted selection (Ruane & Colleau 1995) holds great promise for confirming the presence of a QTL, localizing it to a narrower interval, and studying interlocus interactions.

Once the presence of a QTL has been adequately confirmed, its precise identity must be demonstrated. This is probably feasible only for a gene already documented at the biochemical level. From a chromosome map, researchers can locate plausible candidate genes in the confidence interval for their QTL. For example, Buck and colleagues (1997) noted that a QTL on chromosome 11 was near genes for three subunits of the GABA receptor (Gabra1, Gabra6, Gabrg2). The full DNA sequences of the exons of a gene in the two strains might reveal a polymorphism that gives rise to different forms of the protein. If only one of several candidate genes differs between the strains, it will become the object of intense scrutiny, whereas the presence of several polymorphic genes in the interval will confound progress. Further evidence could be obtained by knocking out the gene in question, but this evidence could also be misleading. The QTL itself might involve a rather minor difference in viable alleles, whereas a total knockout of another nearby gene might very well have major pleiotropic effects on that behavior. Thus, the knockout could implicate one gene without proving that gene to be the source of the QTL.

The task of identifying genes of moderate effect will benefit from a comprehensive effort in a wide variety of common mouse strains to determine the DNA sequence of exons for genes known to code for many nervous system proteins. Genes already proven with the knockout method to be relevant for brain development or behavior would provide a good starting place. To date, the knockout method has taught us much about development but not about individual differences. The classical era of mouse behavioral genetics documented large variations among common inbred strains for a wide range of behavioral phenotypes. We need to know whether the genes targeted by molecular biologists are indeed the genes that gave rise to these ubiquitous strain differences. If researchers would assess the possible relevance for behavior of definite protein polymorphisms rather than search for the proverbial needle in the haystack using the QTL method, answers would come more easily and be less prone to error.

Allele Association Studies

The allele association approach is being used to assess the relevance of well-known nervous system proteins to behavioral variation in humans. In the first step, several alleles of a gene that lead to altered forms of a protein are identified. For example, a 48-base sequence coding for a string of 16 amino acids in the dopamine type 4 receptor (the DRD4 gene) is often repeated, and a world wide survey identified 9 alleles with 2 to 10 repeats (Chang et al 1996). Many relatively common alleles in the dopamine D2 receptor (DRD2) gene are also well documented (Kidd et al 1996). Of critical importance is the observation that allele frequencies generally differ markedly from one geographic population to another (Kidd 1996).

The second step is to establish a correlation between specific alleles and behavioral differences. Claims have been made--but doubts persist--that the A1 allele of the DRD2 gene leads to higher risk of alcoholism. If the study sample is ethnically diverse, an allele that is more common in a group which has a higher rate of alcoholism could result in a spurious correlation. The best recourse is to examine allele associations within a more homogeneous population. For example, in three populations in Taiwan, there is no association of alcoholism with alleles in either the DRD2 or DRD4 gene (Chang et al 1997, Lu et al 1996). In surveying the literature, Kidd (1996) concluded that "the better designed studies have been consistently negative on association" with alcoholism.

Genetic polymorphisms also speak to the chronically vexatious issue of race in behavioral genetics. A comprehensive assessment of allele frequencies around the world by Cavalli-Sforza and colleagues (1994) found little support for racial categories. More recent data prompted Kidd (1996) to comment: "It is my belief that racial classifications of humans are scientifically indefensible since there are essentially no boundaries of qualitative genetic difference and the vast majority of genetic variation shows a continuous pattern around the world."

Considerable publicity has been given to two studies published in 1996 that claimed an association between the personality trait of novelty seeking and the long 7 repeat allele of the DRD4 gene. As revealed in Figure 1, eight subsequent studies from several countries have obtained mixed results. A meta-analysis of these data suggests that scores on the novelty seeking questionnaire have a standard deviation roughly d = 0.06 higher in people with longer repeat alleles (95% confidence interval from -0.03 to 0.16). Because there is significant heterogeneity among the samples (Q = 32.9, df = 9, P = 0.0001), it is possible that epistatic interaction with the genetic background or interaction with test situations or local environments could yield a significant association in certain populations but not others.

[Figure 1 ILLUSTRATION OMITTED]

This exercise with meta-analysis and the history of false positive linkage results for schizophrenia teach important lessons. When a new claim is made of weak allele association or linkage with some other measure (such as IQ), experience should caution us against premature enthusiasm until the result is replicated adequately and survives meta-analysis. Otherwise, there arises a serious risk that false claims will mislead public discourse, as allegations of sex-based differences in the human brain (Bishop & Wahlsten 1997) and an alleged relation between serotonin metabolism and impulsive violence (Balaban et al 1996) have already done.

Presuming the allele association method does eventually point to a genetic variant with reliable behavioral correlates, proof that the connection is causal does not follow automatically. The one gene might be linked to another locus that is actually responsible for the observed difference, and hence it would be wise to assess several nearby genes rather than to restrict the scope of the search too early on. The knockout method will not, of course, be available for confirmatory studies in humans, but highly specific DNA-based drugs might be used to substantiate an effect on behavior for the locus in question.

Linkage and allele association methods are entirely adequate for detecting genes with large phenotypic effects, but these kinds of genetic variants tend to be uncommon in the human population. Nevertheless, the work can be justified because of the potential benefit new knowledge may provide for the prevention or alleviation of suffering. Meanwhile, the hunt for ubiquitous polygenes pursues an elusive quarry. Detecting small effects requires extraordinarily large samples, even if the best available research designs are employed (Risch & Merikangas 1996). There is a profound conflict inherent in this enterprise. In terms of a social calculus of the cost-benefit ratio, the smaller the potential good that might result from a new discovery, the more expensive the purchase of that knowledge will be.

Databases of Genetic Information

Despite my reservations about research on genes of small effect, impressive progress has been made in the detection of genes affecting the nervous system and behavior; and even a cursory account of the present state of knowledge exceeds the scope of this review. Fortunately, a vast reservoir of current genetic information is now readily available on the World Wide Web at species-specific sites (Table 1). Investigators can search these databases for long lists of genes residing on a specific chromosome, detailed information about a specific gene, or lists of genes with possible relevance to a specified phenotype or syndrome. A formal course on skills for the Web would be a very useful addition to the neurobehavioral genetics curriculum.

The Mouse Genome Database (MGD) can be reached via the Jackson Laboratory (jax) site. In February of 1998, a search for the phenotype "obesity" yielded 13 relevant genes. The gene symbol Lep, or its name, leptin (formerly obese), yielded the precise map location on chromosome 6, a lengthy abstract, a current bibliography, and other useful information. It also provided links to the homologous gene LEP in humans and the DNA sequence of several ESTs. If one does not know the official gene symbol, it is best to start with a search for a closely related keyword. For example, the calmodulin kinase II [Alpha] subunit gene symbol is Camk2a. If the protein symbol CaMKII (often cited in the neuroscience literature) is entered in a search of the MGD, nothing is found, whereas a search using the keyword "calmodulin" successfully calls up information on the gene in question and several others. The MGD currently lists 20,080 genetic markers that have been placed on the mouse chromosome map and 8911 genes, of which 6171 have been mapped and 6396 have at least partial DNA sequences available. One must exercise caution when searching for genes affecting phenotypes, because many in the database are poorly validated and may be ancient apparitions. One such is the gene absent corpus callosum (ac), originally reported by Keeler (1933) but not seen by anyone in the past 60 years. Once a gene is listed in the catalog--no matter how flimsy the case for its existence--it tends to remain there. Any mouse gene for which an accurate map position is lacking should not be taken seriously.

After completing a brief registration procedure on the BioMedNet Web site, one can access the Mouse Knockout Database, which provides extensive data on targeted mutations. A search detected over 300 articles on over 100 single-gene knockouts that yield viable animals with alterations in the nervous system and/or behavior.

The Weizmann Institute of Science Gene Cards facility is especially recommended for accessing human genetics information. It allows searches for gene symbols or keywords involving phenotypes, yields chromosome map locations, protein characteristics, and homologies with mice, and offers convenient connections to the Genome Database (GDB) or Medline literature search. The GDB is presently the most authoritative source on human genetics, but it may soon cease operations because funding by the US Department of Energy is being discontinued (Letovsky 1998). Online Mendelian Inheritance in Man(OMIM) provides a lengthy abstract and bibliography for each gene and can also be accessed by entering phenotypic keywords. Searching OMIM for "dyslexia" in February 1998 yielded three gene symbols (DYX1, DYX2, THRB). The GDB listing for DYX1 is based on a single entry from July 19, 1996, and the existence of this gene is far from certain; no information is cited on chromosome map location, and the fine print reveals that it is merely a "reserved symbol," meaning that this will be its official designation if the gene is ever confirmed. DYX2 yields a map location on chromosome 6 that has been supported by an independent group of researchers (Grigorenko et al 1997) but only for one of five reading-related phenotypes (phonological awareness) and only with nonparametric (rather than parametric) methods. The confidence interval for gene location is more than 10 cM wide, and no protein or DNA sequence information is known. A search on "schizzophrenia" yielded 40 entries including the gene symbol SCZD1 assigned on April 12, 1989, to a region on chromosome 5 that is now recognized as not harboring a gene influencing schizophrenia (Moldin & Gottesman 1997). Any gene name returned by a search of OMIM should be carefully checked against more authoritative sources, especially the GDB, where a history of the SCZD1 symbol reveals it was "unassigned" on October 16, 1991. If a protein, DNA sequence, or homology with a mouse gene is listed, one can be confident that the gene is real, but a map location by itself provides no guarantee. Several symbols included in the catalog represent false positives that have not been culled.

The only unequivocal evidence for a gene is elucidation of its DNA sequence and associated protein structure. The quality control for this kind of biochemical information on the Web is good, in part because of facilities provided by the Human Genome Project. Unfortunately, quality control for weaker claims about genes relevant to phenotypes is inadequate, and speculative assertions in the mass media about genetic determination of socially significant behaviors (Colt & Hollister 1998) all too often are based on hasty proclamations from behavioral geneticists who should know better.

GENE FUNCTION

A cornucopia of genes relevant to the nervous system and behavior is now available for research on function. The question of how genes influence behavior and how the activities of genes are themselves regulated is of prime concern for psychology. Function can be understood at different levels.

Natural Polymorphisms

A mutation that seriously impairs the function of an important gene typically is rare in a breeding population, but not all major gene effects on behavior are grossly aberrant misfits. Two remarkable behavioral polymorphisms found in wild fruit flies seem to persist because they aid a species to exploit a wider variety of environments. The foraging locus influences activity of larvae in the presence of food; the dominant rover allele ([for.sup.R]) leads to longer forays into the environment, whereas the recessive sitter allele ([for.sup.s]) results in more localized feeding. Both alleles are common in wild fruit flies living in an urban habitat (Toronto). An exemplary series of studies demonstrated that the sitter mutation occurs in a previously documented gene, dg2, that codes for a cyclic GMP-dependent protein kinase and causes a small change in activity of the enzyme that is sufficient to alter foraging behavior (Osborne et al 1997). The rover phenotype predominates in crowded living conditions, whereas the sitter allele increases in lower population densities where the food supply is not so readily exhausted (Sokolowski et al 1997). A more subtle polymorphism occurs in the circadian clock gene period, where the allele which is more common in northern Europe leads to more efficient adaptation of the 24-h activity rhythm to temperature changes than the allele more common near the Mediterranean (Sawyer et al 1997). By combining carefully controlled genetic analysis in the laboratory with studies further afield, the science of individual differences thus advances our understanding of behavioral ecology and evolution.

Genetic Dissection

Most psychologically interesting behaviors are multifactorial, involving numerous genes whose actions are influenced by diverse features of the environment. Although individual studies usually concentrate attention on one specific gene, it is generally understood that many genes are relevant. As emphasized by Tully (1997), "Single-gene mutant analysis can be informative only when pursued within the framework of interacting polygenes." Powerful techniques to create mutations have spawned new possibilities for genetic dissection of complex processes. No satisfying account of genetic involvement in any complex behavior has yet been achieved, but significant progress has been made in several domains. Olfactory learning and memory in fruit flies is a process in which certain mutations exert their effects primarily on a specific component (see Figure 2), but the famous flow diagram does not imply that one or two genes provide a sufficient explanation for a biologically distinct component of memory; yet the diagram is a useful device for integrating a large corpus of experimental data. The fact that different mutations result in flies with different temporal profiles of memory loss and different interactions with drugs that block protein synthesis proves the multifactorial nature of the memory process. By examining flies affected simultaneously by two different mutations, a scheme for parts arranged in series or in parallel may be perceived. Numerous other genes are undoubtedly involved, and pleiotropy, the occurrence of multiple phenotypic effects of one gene, is to be expected. For example, the turnip mutation reduces motor activity and sensitivity to shock while also impairing learning (Mihalek et al 1997). To some readers, this may render it less interesting because its effects are not restricted to the memory process, but genetic dissection clearly reveals it to be an integral part of the process.

[Figure 2 ILLUSTRATION OMITTED]

Targeted mutations have led to a resurgence of interest in learning and memory in mice, and the list of genes known to be important is rapidly growing (see Table 2). Although admirable efforts have been made to comprehend the interconnections of gene-derived proteins involved in memory formation within a synapse (e.g. Abel et al 1998), the horizons of this metabolic landscape are rapidly expanding, with no limit in sight. Many of these genes have pleiotropic effects as well, such as Camk2a, which is important in spatial memory but also impinges on numerous other behaviors (Chen et al 1994), and Creb, which also reduces symptoms of morphine withdrawal (Maldonado et al 1996).

Table 2 Mouse genes on specified chromosomes that are important for psychological processes(a)

Chromosome Learning and memory 1 Creb1 Sele 2 dbh Grin 1 3 Gria2 112 4 Pde4b 5 Ache En2 Hdh 6 Kcna1 7 Pkcc 8 Pkaca 9 Ncam Rasgrf1 10 Fyn 11 Adcy1 Cbx2 12 Fos 16 App 18 Camk2a X Fmr1 Chromosome Appetite and obesity 1 2 [A.sup.Y] anx Mc3r 3 Ap2 4 [Lepr.sup.db] 5 6 [Lep.sup.ob] 7 Ad Gys1 tub 8 [Cpe.sup.fat] Insr Mclr Mt1 9 Cck 10 Adn 11 Slc2a4 12 Pomc1 16 18 X Htr2c

1. Genes included in the table must have an accurate map location in the Mouse Genome Database and be implicated by at least one study in the process. Most were demonstrated by the targeted mutation method, although a few were spontaneous mutations. Superscript symbols refer to the neurological mutations known prior to the discovery of the specific protein for which the gene codes. Human homologues have been verified for all but four genes in the table.

The organism's genes are of course present from conception, and many participate in formative processes as well as in dynamic adult functions. Embryonic effects can be notably different from involvement in the mature brain, and to distinguish between developmental and current effects of a gene knockout is challenging indeed. Several clever techniques have been employed to overcome this problem. Tsien and coworkers (1996) deleted the NMDA receptor (Grin1 gene) selectively from the CA1 region of the hippocampus and obtained memory deficits similar to those from a nonspecific gene knockout; Mayford and colleagues (1996) were able to limit the expression of a Camk2a mutation to the forebrain of adult mice and still obtained memory deficits. Guzowski & McGaugh (1997) altered spatial memory by injecting synthetic DNA directly into the hippocampus of adult rats to specifically modify the action of the Creb gene. These sophisticated methods confer unprecedented clarity on results for psychopharmacology. Molecular genetics is thus becoming a tool in the kit of physiological psychology.

Appetite and obesity in mice are proving to be physiologically and genetically complex (Table 2), and conceptual schemes for synthesizing this knowledge are still lagging behind the burgeoning data. This is happening with regard to circadian rhythms as well, where newly discovered genes are revealing previously unimagined elements of a larger picture (Albrecht et al 1997).

Investigations of obesity provide an emerging portrait of diverse organs connected in feedback loops involving the environment (Figure 3). Under normal conditions, overeating leads to growth of white fat cells (adipocytes) which in turn synthesize the protein leptin and secrete it into the bloodstream. One of leptin's effects occurs in the hypothalamus, where it binds to the leptin receptor and decreases appetite by inhibiting the synthesis of neuropeptide Y (Npy gene)--a neurotransmitter that tends to increase appetite. The obese mutation ([Lep.sup.ob/ob]) prevents the synthesis of leptin in white fat, thereby increasing appetite when NPY levels rise unchecked, and gene therapy to restore leptin in the [Lep.sup.ob/ob] mice prevents both obesity and diabetes (Muzzin et al 1996). The diabetes mutation ([Lepr.sup.db/db]) disables the leptin receptor and renders mice insensitive to high levels of leptin in the blood, which again leads to overeating (Caro et al 1996). By using a double mutant combining [Lep.sup.ob/ob] with the gene knockout [Npy.sup.-/-], it was shown that there are parallel pathways for leptin-related appetite control in the hypothalamus (Erickson et al 1996). The [Lep.sup.ob/ob] plus lethal yellow ([A.sup.Y]/a) double mutant revealed another parallel pathway that acts via the melanocortin-4 receptor (Mc4r gene) where normal stimulation of MC4-R decreases appetite but the [A.sup.y] gene product antagonizes it (Boston et al 1997). Not all obesity is mediated by the leptin loop (Schonfeld-Warden & Warden 1997), not all leptin effects are mediated by appetite changes (Yu et al 1997), and diabetic symptoms are not joined inexorably with obesity (Hotamisligil et al 1996). Although Table 2 suggests that separate sets of genes impact learning and appetite, this inference may not be warranted because those working with obesity typically do not assess a wide range of behavioral phenotypes. It seems highly likely that variations in appetite would indeed influence the acquisition of certain kinds of tasks, but little recent work with obesity mutations has been done by psychologists interested in motivation.

[Figure 3 ILLUSTRATION OMITTED]

Genetic dissection thus proceeds through several stages. (a) Research projects in the early stages seek to discover a single mutation and explore its phenotypic effects. (b) After several genes are known to be important parts of the system, work begins with double mutants and factorial gene-environment or gene-drug interaction studies to elucidate serial and parallel processes, each study focusing on a limited sector of the larger system. (c) Eventually, attempts are made to integrate this knowledge into comprehensive models that can be tested with multifactorial experiments. Most research in neurobehavioral genetics is presently entering the second stage, and none has yet reached the third.

Systems of Genes

The question remains of how many genes are involved in memory, appetite, or circadian rhythm. A first approximation can be achieved by examining the array of genes expressed in mRNA under specified conditions. A sensitive and rapid method is now available to assess simultaneously the expression of hundreds of genes in mice (Figure 4) and over 1000 in humans, and customized arrays for screening any desired subset of genes may be anticipated (see Web sites www.resgen.com and atlas.clontech.com). One might contrast brains of trained and untrained mice to assess memory, or brains at midnight and at high noon under a normal light cycle or constant darkness to reveal circadian mechanisms. Furthermore, tissue from mutants and normal siblings tested under the same circumstances could be used to assess pleiotropic effects.

[Figure 4 ILLUSTRATION OMITTED]

An extraordinary glimpse of complex gene action has been obtained recently for yeast, an organism best known to psychologists for its vital role in synthesizing ethanol from sugar. As the sugar in the yeast's environment is consumed, its metabolism shifts from anerobic fermentation to aerobic respiration. Researchers were able to attach DNA sequences of almost all the 6297 yeast genes to a single glass plate 18 mm by 18 mm and record the abundance of all mynas at different stages of the metabolic process (DeRisi et al 1997). During the transition from anerobic to aerobic metabolism, the expression of 1740 genes increased or decreased at least twofold. About half of these genes were new to science, had not yet been named, and had no recognized functions. A mutation in a single gene (tup1) altered the expression of 355 other genes. In a remarkable understatement, the authors observed: "The large number of genes whose expression is altered and the diversity of temporal expression profiles highlighted the challenge of understanding the underlying regulatory mechanisms."

The one-celled yeast, of course, is a relatively simple creature that has been thoroughly studied since the time of Pasteur. It seems likely that the complete picture of gene activity during mammalian learning and memory will be even more complex. The molecular tools are close at hand but the prevailing conceptual framework in biological psychology may not be equal to the task of integrating so vast an array of data.

DNA analysis and gene discovery have been dominated by a very successful reductionistic perspective (Beckwith 1996), but research on gene function reveals the necessity of systems-oriented thinking (Gottlieb et al 1998, Strohman 1997). The idea that a gene determines a specific component of a behavioral phenotype is losing scientific credibility. Chenchik and coworkers (1998) forsee that new methods "will lead researchers away from reductionistic approaches which focus on single genes, and towards more systemic approaches that involve the simultaneous, parallel analysis of hundreds or thousands of genes." It must be acknowledged that almost every gene has widespread pleiotropic effects (Miklos & Rubin 1996), that actions of genes are commonly altered by the organism's environment (Gottlieb 1998), and that the consequences of a specific mutation often depend on genotypes at other loci (Varnam et al 1996) and the genetic background (de Belle & Heisenberg 1996, Kelly et al 1998, Miklos & Rubin 1996) as well as on epigenetic effects (Wolf 1997). Strohman (1997) concludes that the origins of complex systems "are not to be found in the matter itself, but in its interactions."

Spectacular advances in genetic analysis have captured the imagination of the public and drawn legions of students into molecular biology. The effect has not been to impoverish psychology but to renew interest in the psychology of behavioral testing--especially regarding lab mice--as testified to by the 1996 Society for Neuroscience short course entitled "What's wrong with my mouse?" (Takahashi 1996). Specialists in behavioral genetics have been both impressed by the enthusiasm and appalled at the naivete of molecular geneticists who believe psychology can provide an off-the-shelf device to measure a specific construct in mice and model its human counterpart. The new molecular genetics has created a need for a wider variety of behavioral testing and for improved test construction and standardization. The skills of psychologists are uniquely suited to this task.

Interactions with Test Situation and Environment

There is a rich variety of tests available for use with mice (Crawley et al 1997, Crawley & Paylor 1997), and it is important to know whether these are likely to yield the same results in the hands of most investigators working with the same strain or mutation. The test situation and the pretest environment are virtually never the same in different laboratories. The central issue is therefore whether genetic and environmental effects are additive or interactive. If additive, then differences among labs will merely change the overall average score but will not alter the pattern of results or rank orders of genotypes, and most tests should yield valid results even in the hands of amateurs.

A clear answer can be provided to this question. Seemingly minor task-specific factors interact strongly with genotype, and reversals of rank orders of strains are commonplace when comparing results across labs. Recent work has emphasized the importance of relatively subtle variations in protocols. Poderycki and coworkers (1998) evaluated hybrid crosses of mice for seizures induced by 5-15 repetitions of gentle tossing. Genetic analysis revealed strong evidence of linkage with a marker on chromosome 9 after 6 tests but not after 15 tests, whereas another gene on chromosome 2 was not apparent after 6 tests but showed clear signs of linkage after 15 tests. Maxson (1992) reported that the effects of the Y chromosome on agonistic behavior in his congenic strains disappeared when the colony was moved to a cleaner environment where the drinking water was acidified to suppress bacteria. Certain Y chromosome effects were most pronounced when males were reared in isolation and tested against males of the same genotype, rather than reared with a sister and tested against a standard opponent strain (Guillot et al 1995). Peeler (1995) conducted avoidance training at different times of day, all during the light phase of the cycle, and found substantial effects on some strains but not others.

Apparatus design and testing protocol are crucially important. Roullet and colleagues (1993) found that BALB/c mice used odor cues and C57BL/6 mice used spatial cues to learn a radial maze, whereas [F.sub.1] hybrids could utilize either cue. Crusio and coworkers (1993) found large changes in strain rank orders on spatial versus nonspatial versions of a radial maze; only the spatial version revealed genetic correlations with hippocampal mossy fiber anatomy. Peeler (1995) noted differences in strain rank order depending on whether the mice had to run through a hole or slot or jump a barrier to avoid shock. These and other findings show clearly that a single test configuration and procedure cannot define a single psychological construct, although two tests differing in a specific element may indicate a change in a specific construct.

Growing Need for Test Standardization

In view of these findings, there are grounds for concern about the almost universal lack of standardized apparatuses, protocols, and lab environments in psychological testing of animals. In contrast, test construction and standardization are taken more seriously in evaluation of humans. Reviews among labs of the plus-maze, a popular means for assessing anxiety, reveal numerous idiosyncratic variations that are potentially important (Hogg 1996; Rodgers & Dalvi 1997). Tests known to be valid for rats are often used inappropriately with mice. The Morris swimming pool is particularly problematic when used with mice (Whishaw & Tomie 1996). Some common inbred strains (BALB/c, 129) respond quite badly in this device (Francis et al 1995) and often resort to floating after becoming exhausted (Wolfer et al 1997). When mice can locate a submerged platform, they must be using spatial cues, but failure to learn does not necessarily signal a lack of spatial memory (unless only the cues are manipulated in different versions of the task, and a strain can learn one cue but not the other). Additional difficulties are present when investigators submit an individual animal to a battery of tests, each of which was designed and validated for use by itself: Order of testing can markedly alter results when such tests are combined.

Whereas homology among flies, mice, and humans at the molecular genetic level is undeniable, homology--and even analogy--at the level of the behaving organism is not so clear. Certain genes important for memory in flies (Figure 2) are involved in memory in mammals (e.g. the fly gene rutabaga, the mouse gene barrelless, and the human gene ADCY1 all encode an adenylate cyclase), yet the complex nature of metabolic and developmental systems (characterized by pleiotropic and epistatic gene actions plus gene-environment interactions) implies that the function of a particular gene depends on its context. Homology of behaviors must be demonstrated, not assumed. In the domain of agonistic behavior, for example, mice engage in offensive and defensive attacks that appear to be adaptive under appropriate circumstances, but do these provide good models of human violence? According to Maxson (1998), offensive behavior of male mice is not a good model of impulsive aggression in humans, although several recent publications assume uncritically that the two are essentially the same. Balaban and colleagues (1996) also stress the importance of careful definition of behaviors and contexts when seeking to establish the relevance to humans of animal models, and they question the similarity of rodent attack behavior to human crime. A good case can be made for valid mouse models of several severe medical disorders caused by single-gene mutations, but the validity of mouse or fly models for the normal range of variation in human social behavior requires convincing evidence that is generally lacking.

In his review of human behavioral genetics, Rose (1995) foresaw that "Future reviews of the field are likely to read very differently than this one." The field has indeed changed direction and is advancing like a sailboat with spinnaker unfurled, rather than tacking and making little headway. Many outstanding contributions to neurobehavioral genetics are now published in leading scientific journals with a broad readership rather than in specialty journals. Not long ago, the field was trammeled by crude techniques for detecting the presence and activities of single genes, whereas today we have a panoply of molecular methods and a rich factual base of knowledge about specific genes in relation to brain and behavior. Long lists of human attributes, each accompanied by a terse summary of the latest findings from twin or adoption studies, have become passe. The challenge of keeping aware of current developments in this field is now quite formidable, even with the aid of marvelous Internet and bibliographic search programs. As the individual research project probes ever more deeply into an ever-narrower domain of knowledge, there is a growing need to synthesize existing knowledge and make connections among the isolated parts of an expanding discipline. The next major advance must come in the domain of theory.

ACKNOWLEDGMENTS

I am grateful to John Crabbe, Gilbert Gottlieb, and Pierre Roubertoux for their critical comments on this article and to Sharon Doerksen for assistance with Web searches. Supported by a grant from the Natural Sciences and Engineering Research Council of Canada.

Abdel-Majid RM, Leong WL, Schalkwyk LC, Smallman DS, Wong ST, et al. 1998. Loss of adenylyl cyclase type I activity disrupts patterning of mouse somatosensory cortex. Nat. Genet. 19:289-91

Abel T, Martin KC, Bartsch D, Kandel ER. 1998. Memory suppressor genes: inhibitory constraints on the storage of long-term memory. Science 279:338-41

Albrecht U, Sun ZS, Eichele G, Lee CC. 1997. A differential response of two putative mammalian circadian regulators, mper1 and mper2, to light. Cell 91:1055-64

Balaban E, Alper JS, Kasamon YL. 1996. Mean genes and the biology of aggression: a critical review of recent animal and human research, J. Neurogenet. 11:1-43

Beckwith J. 1996. The hegemony of the gene: reductionism in molecular biology. In The Philosophy and History of Molecular Biology: New Perspectives, ed. S Sarkar, pp. 171-83. Netherlands: Kluwer Acad.

Belknap JK, Dubay C, Crabbe JC, Buck KJ. 1997. Mapping quantitative trait loci for behavioral traits in the mouse. In Handbook of Psychiatric Genetics, ed. K Blum, EP Noble, RS Sparkes, THJ Chen, JG Cull, pp. 435-53. New York: CRC Press

Belknap JK, Mitchell SR, O'Toole LA, Helms ML, Crabbe JC. 1996. Type I and type II error rates for quantitative trait loci (QTL) mapping studies using recombinant inbred mouse strains. Behav. Genet. 26:149-60

Benjamin J, Li L, Patterson C, Greenberg BD, Murphy DL, Hamer DH. 1996. Population and familial association between the D4 dopamine receptor gene and measures of novelty seeking. Nat. Genet. 12:81-84

Bennett B, Beeson M, Gordon L, Johnson TE. 1997. Quick method for confirmation of quantitative trait loci. Alcohol Clin. Exp. Res. 21:1-6

Bishop K, Wahlsten D. 1997. Sex differences in the human corpus callosum: myth or reality? Neurosci. Biobehav. Rev. 21: 581-601

Boomsma D, Anokhin A, de Geus E. 1997. Genetics of electrophysiology: linking genes, brain, and behavior. Curr. Dir. Psychol. Sci. 6:106-10

Boston BA, Blaydon KM, Varnerin J, Cone RD. 1997. Independent and additive effects of central POMC and leptin pathways on murine obesity. Science 278:1641-44

Botstein D, Chervitz SA, Cherry JM. 1997. Yeast as a model organism. Science 277: 1259-60

Buck KJ, Metten P, Belknap JK, Crabbe JC. 1997. Quantitative trait loci involved in genetic predisposition to acute alcohol withdrawal in mice. J. Neurosci. 17: 3946-55

Caro JF, Sinha MK, Kolaczynski JW, Zhang PL, Considine RV. 1996. Leptin: the tale of an obesity gene. Diabetes 45:1455-62

Cavalli-Sforza LL, Menozzi P, Piazza A. 1994. The History and Geography of Human Genes. Princeton, NJ: Princeton Univ. Press

Chang F, Kidd JR, Livak K J, Pakstis AJ, Kidd KK. 1996. The world-wide distribution of allele frequencies at the human dopamine D4 receptor locus. Hum. Genet. 98:91-101

Chang F, Ko H, Lu R, Pakstis A J, Kidd KK. 1997. The dopamine D4 receptor gene (DRD4) is not associated with alcoholism in three Taiwanese populations: six polymorphisms tested separately and as haplotypes. Biol. Psychiatry 41:394-405

Chen C, Rainnie DG, Greene RW, Tonegawa S. 1994. Abnormal fear response and aggressive behavior in mutant mice deficient for alpha-calcium-calmodulin kinase II. Science 266:291-94

Chenchik A, Chen S, Makhanov M, Siebert P. 1998. Profiling of gene expression in a human glioblastoma cell line using the Atlas Human cDNA Expression Array I. CLON-TECHniques 13:16-17

Cloninger CR. 1997. Multilocus genetics of schizophrenia. Curr. Opin. Psychiatry 10: 5-10

Colt GH, Hollister A. 1998. Were you born that way? Life Apr:39-50

Crabbe JC, Phillips T J, Buck K J, Cunningham CL, Belknap JK. 1998. Identifying genes for alcohol and drug sensitivity: recent progress and future direction. Trends Neurosci. In press

Crawley JN, Belknap JK, Collins A, Crabbe JC, Frankel W, et al. 1997. Behavioral phenotypes of inbred mouse strains: implications and recommmendations for molecular studies. Psychopharmacology 132: 107-24

Crawley JN, Paylor R. 1997. A proposed test battery and constellations of specific behavioral paradigms to investigate the behavioral phenotypes of transgenic and knockout mice. Horm. Behav. 27:201-10

Crusio WE. 1998. The genetic dissection of brain-behaviour relationships: an introduction to neurobehavioral genetics. Behav. Brain Res. 95:1-2

Crusio WE, Schwegler H, Brust I. 1993. Covariations between hippocampal mossy fibres and working and reference memory in spatial and non-spatial radial maze tasks in mice. Eur. J. Neurosci. 5:1413-20

Darvasi A. 1998. Experimental strategies for the genetic dissection of complex traits in animal models. Nat. Genet. 18:19-24

de Belle JS, Heisenberg M. 1996. Expression of Drosophila mushroom body mutations in alternative genetic backgrounds: a case study of the mushroom body miniature gene (mbm). Proc. Natl. Acad. Sci. USA 93:9875-80

DeRisi JL, Iyer VR, Brown PO. 1997. Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278:680-86

Devlin B, Daniels M, Roeder K. 1997. The heritability of IQ. Nature 388:468-71

Dietrich WF, Copeland NG, Gilbert DJ, Miller JC, Jenkins NA. 1995. Mapping the mouse genome: current status and future prospects. Proc. Natl. Acad. Sci. USA 92: 10849-53

Ebstein RP, Nemanov L, Klotz I, Gritsenko I, Belmaker RH. 1997. Additional evidence for an association between the dopamine D4 receptor (D4DR) exon III repeat polymorphism and the human personality trait of Novelty Seeking. Mol. Psychiatry 2: 472-74

Ebstein RP, Novick O, Umansky R, Priel B, Osher Y, et al. 1996. Dopamine D4 receptor (D4DR) exon III polymorphism associated with the human personality trait of novelty seeking. Nat. Genet. 12:78-80

Erickson JC, Hollopeter G, Palmiter RD. 1996. Attenuation of the obesity syndrome of ob/ob mice by the loss of neuropeptide Y. Science 274:1704-7

Francis DD, Zaharia MD, Shanks N, Anisman H. 1995. Stress-induced disturbances in Morris water-maze performance: interstrain variability. Physiol. Behav. 58: 57-65

Gelernter J, Kranzler H, Coccaro E, Siever L, New A, Mulgrew CL. 1997. D4 dopamine-receptor (DRD4) alleles and novelty seeking in substance-dependent, personality-disorder, and control subjects. Am. J. Hum. Genet. 61:1144-52

Gerlai R. 1996. Gene-targeting studies of mammalian behavior: Is it the mutation or the background genotype? Trends Neurosci. 19:177-81

Giese KP, Fedorov NB, Filipkowski RK, Silva AJ. 1998. Autophosphorylation at [Thr.sup.286] of the alpha calcium-calmodulin kinase II in LTP and learning. Science 279:870-73

Gottlieb G. 1998. Normally occurring environmental and behavioral influences on gene activity: from central dogma to probabilistic epigenesis. Psychol. Rev. In press

Gottlieb G, Wahlsten D, Lickliter R. 1998. The significance of biology for human development: a developmental psychobiological systems view. In Handbook of Child Psychology. Theoretical Models of Human Development, ed. RM Lerner, 1:233-73. New York: Wiley. 5th ed.

Grigorenko EL, Wood FB, Meyer MS, Hart LA, Speed WC, et al. 1997. Susceptibility loci for distinct components of developmental dyslexia on chromosomes 6 and 15. Am. J. Hum. Genet. 60:27-39

Guillot P, Carlier M, Maxson SC, Roubertoux PL. 1995. Intermale aggression tested in two procedures, using four inbred strains of mice and their reciprocal congenics: Y chromosomal implications. Behav. Genet. 25:357-60

Guzowski JF, McGaugh JL. 1997. Antisense oligodeoxynucleotide-mediated disruption of hippocampal cAMP response element binding protein levels impairs consolidation of memory for water maze training. Proc. Natl. Acad. Sci. USA 94: 2693-98

Henikoff S, Greene EA, Pietrokovski S, Bork P, Attwood TK, Hood L. 1997. Gene families: the taxonomy of protein paralogs and chimeras. Science 278:609-14

Hogg S. 1996. A review of the validity and variability of the elevated plus-maze as an animal model of anxiety. Pharmacol. Biochem. Behav. 54:21-30

Hotamisligil GS, Johnson RS, Distel R J, Ellis R, Papaioannou VE, Spiegelman BM. 1996. Uncoupling of obesity from insulin resistance through a targeted mutation in aP2, the adipocyte fatty acid binding protein. Science 274:1377-79

Johannsen W. 1911. The genotype conception of heredity. Am. Nat. 45:129-59

Jonsson EG, Nothen MM, Gustavsson JP, Neidt H, Brene S, et al. 1997. Lack of evidence for allelic association between personality traits and the dopamine D4 receptor gene polymorphisms. Am. J. Psychiatry 154:697-99

Joyner AL. 1993. Gene Targeting: A Practical Approach. New York: Oxford Univ. Press

Keeler CE. 1933. Absence of the corpus callosum as a Mendelizing character in the house mouse. Proc. Natl. Acad. Sci. USA 19:609-11

Kelly MA, Rubinstein M, Phillips T J, Lessov CN, Burkhart-Kasch S, et al. 1998. Locomoter activity in D2 dopamine receptor-deficient mice is determined by gene dosage, genetic background, and developmental adaptations, J. Neurosci. 18:3470-79

Kidd KK. 1996. Human genetic diversity and neuropsychiatric disorders. Presented at Wenner-Gren Cent. Int. Symp. Genet. Psychiatr. Disord., Stockholm

Kidd KK. 1997. Can we find genes for schizophrenia? Am. J. Med. Genet. 74: 104-11

Kidd KK, Pakstis A J, Castiglione CM, Kidd JR, Speed WC, et al. 1996. DRD2 halotypes containing the TaqI A1 allele: implications for alcoholism research. Alcohol Clin. Exp. Res. 20:697-705

Lander E, Kruglyak L. 1995. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat. Genet. 11:241-47

Lander ES, Schork NJ. 1994. Genetic dissection of complex traits. Science 265: 2037-48

Letovsky S. 1998. Termination of GDB project. http://www.gdb.org/shutdown/notice.html

Lu R, Ko H, Chang F, Castiglione CM, Schoolfield G, et al. 1996. No association between alcoholism and multiple polymorphisms at the dopamine D2 receptor gene (DRD2) in three distinct Taiwanese populations. Biol. Psychiatry 39:419-29

Maldonado R, Blendy JA, Tzavara E, Gass P, Roques BP, et al. 1996. Reduction of morphine abstinence in mice with a mutation in the gene encoding CREB. Science 273: 657-59

Malhotra AK, Virkkunen M, Rooney W, Eggert M, Linnoila M, Goldman D. 1996. The association between the dopamine D4 receptor (D4DR) 16 amino acid repeat polymorphism and Novelty Seeking. Mol. Psychiatry 1: 388-91

Markel PD, Bennett B, Beeson M, Gordon L, Johnson TE. 1997. Confirmation of quantitative trait loci for ethanol sensitivity in long-sleep and short-sleep mice. Genome Res. 7:92-99

Markel PD, Fulker DW, Bennett B, Corley RP, DeFries JC, et al 1996. Quantitative trait loci for ethanol sensitivity in the LS x SS recombinant inbred strains: interval mapping. Behav. Genet. 26:447-58

Maxson SC. 1998. Homologous genes, aggression, and animal models. Dev. Neuropsychol. 14:143-56

Mayford M, Bach ME, Huang Y, Wang L, Hawkins RD, Kandel ER. 1996. Control of memory formation through regulated expression of a CaMKII transgene. Science 274:1678-83

McClearn GE, Plomin R, Gora-Maslak G, Crabbe JC. 1991. The gene chase in behavioral science. Psychol. Sci. 2:222-29

Mihalek RM, Jones C J, Tully T. 1997. The Drosophila mutation turnip has pleiotropic effects and does not specifically affect learning. Neurobiol. Learn. Mem. 3: 425-44

Miklos GL, Rubin GM. 1996. The role of the genome project in determining gene function: insights from model organisms. Cell 86:521-29

Moldin SO. 1997. The maddening hunt for madness genes. Nat. Genet. 17:127-29

Moldin SO, Gottesman II. 1997. At issue: genes, experience, and chance in schizophrenia-positioning for the 21st century. Schizophr. Bull. 23:547-61

Muzzin P, Eisensmith RC, Copeland KC, Woo SLC. 1996. Correction of obesity and diabetes in genetically obese mice by leptin gene therapy. Proc. Natl. Acad. Sci. USA 93:14804-8

Ono Y, Manki H, Yoshimura K, Muramatsu T, Mizushima H, et al. 1997. Association between dopamine D4 receptor (D4DR) exon III polymorphism and novelty seeking in Japanese subjects. Am. J. Med. Genet. Neuropsychiatr. Genet. 74:501-3

Osborne KA, Robichon A, Burgess E, Butland S, Shaw RA, et al. 1997. Natural behavior polymorphism due to a cGMP-dependent protein kinase of Drosophila. Science 277: 834-36

Peeler DF. 1995. Shuttlebox performance in BALB/cByJ, C57BL/6ByJ, and CXB recombinant inbred mice: environmental and genetic determinants and constraints. Psychobiology 23:161-70

Plomin R, McClearn GE, Smith DL, Vignetti S, Chorney M J, et al. 1994. DNA markers associated with high versus low IQ: the IQ quantitative trait loci (QTL) project. Behav. Genet. 24:107-18

Poderycki M J, Simoes JM, Todorova MT, Neumann PE, Seyfried TN. 1998. Environmental influences on epilepsy gene mapping in EL mice. J. Neurogenet. 12: 67-86

Rijsdijk FV, Boomsma DI. 1997. Genetic mediation of the correlation between peripheral nerve conduction velocity and IQ. Behav. Genet. 27:87-98

Risch N, Merikangas K. 1996. The future of genetic studies of complex human diseases. Science 273:1516-17

Rodgers RJ, Dalvi A. 1997. Anxiety, defense and the elevated plus-maze. Neurosci. Biobehav. Rev. 21:801-10

Rose RJ. 1995. Genes and human behavior. Annu. Rev. Psychol. 46:625-54

Roullet P, Lassalle JM, Jegat R. 1993. A study of behavioral and sensorial bases of radial maze learning in mice. Behav. Neural Biol. 59:173-79

Rowen L, Mahairas G, Hood L. 1997. Sequencing the human genome. Science 278: 605-7

Ruane J, Colleau JJ. 1995. Marker assisted selection for genetic improvement of animal populations when a single QTL is marked. Genet. Res. 66:71-83

Sander T, Harms H, Dufeu P, Kuhn S, Rommelspacher H, Schmidt LG. 1997. Dopamine D4 receptor exon III alleles and variation of novelty seeking in alcoholics. Am. J. Med. Genet. Neuropsychiatr. Genet. 74:483-87

Sawyer LA, Hennessy JM, Peixoto AA, Rosato E, Parkinson H, et al. 1997. Natural variation in a Drosophila clock gene and temperature compensation. Science 278: 2117-20

Schonemann PH. 1997. On models and muddles of heritability. Genetica 99:97-108

Schonfeld-Warden N, Warden CH. 1997. Pediatric obesity: an overview of etiology and treatment. Pediatr. Endocrinol. 44: 339-61

Simpson EM, Linder CC, Sargent EE, Davisson MT, Mobraaten LE, Sharp JJ. 1997. Genetic variation among 129 substrains and its importance for targeted mutagenesis in mice. Nat. Genet. 16:19-27

Sokolowski MB, Pereira HS, Hughes K. 1997. Evolution of foraging behavior in Drosophila by density-dependent selection. Proc. Natl. Acad. Sci. USA 94:7373-77

Strohman RC. 1997. The coming Kuhnian revolution in biology. Nat. Biotechnol. 15: 194-200

Sullivan PF, Fifield W J, Kennedy MA, Mulder RT, Sellman JD, Joyce PR. 1998. No association between novelty seeking and the type 4 dopamine receptor gene (DRD4) in two New Zealand samples. Am. J. Psychiatry 155:98-101

Takahashi JS. 1996. What's Wrong with My Mouse? New Interplays between Mouse Genetics and Behavior. Washington, DC: Soc. Neurosci.

Thomas SA, Palmiter RD. 1997. Disruption of the dopamine B-hydroxylase gene in mice suggests roles for norepinephrine in motor function, learning, and memory. Behav. Neurosci. 111:579-89

Threadgill DW, Yee D, Matin A, Nadeau JH, Magnuson T. 1997. Genealogy of the 129 inbred strains: 129/SvJ is a contaminated inbred strain. Mamm. Genome 8: 441-42

Tsien JA, Huerta PT, Tonegawa S. 1996. The essential role of hippocampal CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell 87:1327-38

Tully T. 1997. Regulation of gene expression and its role in long-term memory and synaptic plasticity. Proc. Natl. Acad. Sci. USA 94:4239-41

Tully T, Bolwig G, Christensen J, Connolly J, DelVecchio M, et al. 1996. A return to genetic dissection of memory in Drosophila. Cold Spring Harbor Symp. Quant. Biol. 61:207-18

Vandenbergh DJ, Zonderman AB, Wang J, Uhl GR, Costa PT Jr. 1997. No association between novelty seeking and dopamine D4 receptor (D4DR) exon III seven repeat alleles in Baltimore longitudinal study of aging participants. Mol. Psychiatry 2:417-19

Varnam C J, Strauss R, De Belle JS, Sokolowski MB. 1996. Larval behavior of Drosophila central complex mutants: interactions between no bridge, foraging, and chaser, J. Neurogenet. 11:99-115

Wahlsten D. 1990. Insensitivity of the analysis of variance to heredity-environment interaction. Behav. Brain Sci. 13:109-61

Wahlsten D. 1994. The intelligence of heritability. Can. Psychol. 35:244-58

Wahlsten D, Sparks V. 1995. New recombinant inbred strains expressing 100% total absence of the corpus callosum. Soc. Neurosci. Abstr. 21, p.796

Welker E, Armstrong-James M, Bronchti G, Ourednik W, Gheorghita-Baechler F, et al. 1996. Altered sensory processing in the somatosensory cortex of the mouse mutant barrelless. Science 271:1864-67

Whishaw IQ, Tomie J. 1996. Of mice and mazes: similarities between mice and rats on dry land but not water mazes. Physiol. Behav. 60:1191-97

Wimer RE, Wimer CC. 1985. Animal behavior genetics: a search for the biological foundations of behavior. Annu. Rev. Psychol. 36:171-218

Wolf U. 1997. Identical mutations and phenotypic variation. Hum. Genet. 100:305-21

Wolfer DP, Muller U, Staglier M, Lipp H. 1997. Assessing the effects of the 129/Sv genetic background on swimming navigation learning in transgenic mutants: a study using mice with a modified [Beta]-amyloid precursor protein gene. Brain Res. 771:1-13

Yu WH, Kimura M, Walczewska A, Karanth S, McCann SM. 1997. Role of leptin in hypothalamic-pituitary function. Proc. Natl. Acad. Sci. USA 94:1023-28

D. Wahlsten

Department of Psychology, University of Alberta, Edmonton, Alberta, Canada T6G 2E9; e-mail: wahlsten@psych.ualberta.ca

(genes) Meredith F. Small.

Abstract: Biologists believe people are drawn to opposite sexes because of the need to reproduce. Scientists are unable to explain, however, why people are attracted to specific members of the opposite sex. Researchers of genetics and behavior suggest people avoid prospective mates who have similar genes driving their immune systems.

Some of the genes that drive the immune system may also exert an influence in matters of the heart.

In a crowded bar during happy hour, people mill about sipping drinks and eating canapes. For no apparent reason, a man locks eyes with a woman on the other side of the room. Pulled by an invisible magnet, he slowly moves toward her. She acknowledges his gaze by slightly lifting her glass and then begins to navigate through the crowd toward him. Once together, they stand within touching distance, awkwardly demonstrating with body language that something special is going on. They exchange pleasantries, then phone numbers, and a year later, they are married.

What draws people of the opposite sex together? Biologists have a general answer--we are members of a sexually reproducing species and need to join up with someone of the opposite sex to mate, conceive offspring, and pass our genes on to the next generation. But what scientists cannot yet fully explain is why people are attracted to particular members of the opposite sex. Why that man and not another; why that woman and not her friend?

Recent research combining genetics and behavior suggests there may be a hidden agenda at work: it looks as if we may avoid prospective mates who are too similar to us in the genes that drive the immune system. And data that may explain some of the basics of human attraction and mating are coming from a very special group of people--the Hutterites, a closed religious sect in which people marry young, stay married for life, and value fidelity.

In a nation that is commonly likened to an ethnic melting pot, the Hutterite communities have remained remarkably homogeneous. Every Hutterite can trace his or her ancestry to the 400 or so members of the Anabaptist sect who migrated from Europe to the United States in the 1870s to escape religious persecution. The founders settled on three communal farms in South Dakota. Today, the Hutterites, more than 35,000 strong, are spread across the northern United States, as well as western Canada, in 350 colonies.

In the late 1950s, biologist Arthur Steinberg and his colleagues at Case Western Reserve University began visiting the Hutterites for an extensive project that included measuring people, checking their health (administering EKGs, for example, and testing cholesterol levels), drawing blood for genetic studies, and, most important for analysis, recording genealogies. They visited most of the colonies and developed a priceless bank of information about the population. In the 1970s, these data were handed down to Alice Martin, of Northwestern University Medical School, who in turn passed them on to Carole Ober, a geneticist with anthropological training now working in the Department of Human Genetics at the University of Chicago. In 1982, Ober began a long-term project on one of the original lineages, focusing on thirty-one of the forty-four colonies in South Dakota.

From a population geneticist's point of view, the Hutterites are ideal research subjects. The 400 original settlers themselves were closely related; they are believed to be descended from about 90 individuals. Written genealogies go back nearly sixteen generations. The rules of Hutterite culture, which include within-group marriage as well as lifelong fidelity, prevent the usual shenanigans responsible for the exchange of genes across geographic, ethnic, racial, and religious borders. As a result, the Hutterites are relatively inbred. In addition, they are highly fertile (with an average family size of nine) and do not practice birth control, allowing a geneticist to get a picture of how genes are passed along under "natural" conditions.

Ober's main interest is a genetic system found in all vertebrates, the major histocompatibility complex, or MHC (known in humans as human leukocyte antigen system, or HLA), and its effect on fertility and mating. The MHC is a set of closely linked genes (found on chromosome number 6 in humans) and is best known for its role in the immune system. In concert with the immune system's T-cells and B-cells, HLA genes help guide the production of antibodies and are also responsible for rejecting foreign tissue or infectious agents.

Each of the genes so far identified in the HLA system--referred to by the letters A, B, C, DR, DQ, DP, E, F, and G--comes in several forms, or alleles. For example, there are at least 60 different forms of the A gene, 150 of the B, 180 of the DR, and so on. At least 80 million combinations are possible, which means that any one set of HLA genes, or haplotype, is shared by only a small number of people. (Organ transplants are difficult, in part, because the recipient's HLA system tells the body to reject tissue that does not match its own--in other words, to spurn tissue from just about every unrelated individual.)

Imagine that your body contains two strands of multicolored lights. Each strand represents a haplotype. Each bulb is a gene, the different colors representing the various possible forms of a particular gene. If you placed your two strands next to each other, the bulbs would line up in pairs. A pair might match (a red next to a red) or differ (a red next to a blue). The number of combinations is enormous. Where you have a pair of red bulbs, for example, someone else may have one teal and one magenta bulb.

Each of us gets one set of HLA genes from each parent, as if a strand of lights were passed down intact from each side of the family. These sets can be traced back through the generations because, unlike most genes, individual HLA genes do not shuffle about during replication and fertilization. Instead, they remain linked together as they are passed on, providing a genetic road map into the past.

That the genes at each position come in so many different forms intrigues evolutionary biologists, who wonder why such variety evolved in the first place and how it might be maintained. One possibility is that the variety is a product of the ongoing arms race between pathogens and antibodies. Having a multitude of combinations in the HLA system would increase the chances that whenever a new pathogen appears on the scene, the immune system will already have some ammunition with which to fight it.

Another possibility is that some HLA genes may provide protection against particular diseases and thus be maintained in a population by natural selection. Some HLA haplotypes, for example, are associated with resistance to tuberculosis, leprosy, and malaria. Other haplotypes, however, may increase the chances of developing a disease. Celiac disease, ankylosing spondylitis, multiple sclerosis, and type 1 diabetes have all been tagged to specific HLA genes. Recently, HLA genes have been implicated in the progress of diseases such as AIDS and in the rate at which tumors grow.

Researchers have also discovered--and this is where Ober's work reenters the story--an intriguing relationship between HLA and pregnancy. Although half of its genes come from its mother, a fetus is still essentially foreign to the mother's body. Since one role of the HLA system is to differentiate self from nonself and then reject any tissue that is not an exact match, you might think "foreignness" would spell trouble for the fetus. But the opposite appears to be true: several studies have shown that trouble--in the form of higher risk for miscarriage--arises when a man and a woman with similar HLA haplotypes conceive. If the best immune system is one able to produce a wide variety of antibodies, natural selection may favor fetuses with HLA genes different from their mothers'.

Ober found that for the Hutterites, fetal loss increased significantly when parents shared all the genes of the HLA system or just genes in the HLA-B region. Ober is quick to say that, at this stage in the research, we can/not declare with certainty that the HLA genes themselves--and not some other genes found in the same region on the chromosome--interfere with fertility, but since pregnancy is essentially an intrusion by foreign tissue, HLA is probably involved. If natural selection is operating so strongly during reproduction, organisms should be "designed" to avoid the fetal wastage that comes with shared HLA genes. But how?

In the late 1970s, Kunio Yamazaki and colleagues observed that inbred laboratory mice were more likely to mate with mice having dissimilar MHC genes. Working with mice in a more natural setting (outdoor pens that gave the animals considerably more freedom to move about and interact), Wayne Potts, now at the University of Utah, confirmed Yamazaki's findings: the first generation of offspring in his mouse colony produced far fewer pups carrying matching MHC regions than might be expected if the parents were mating at random.

In some of his experiments, Yamazaki, now at the Monell Chemical Senses Center in Philadelphia, placed newborn male mice with foster mothers of a different MHC strain. He noted that when the time came for these males to choose mates, they avoided females whose MHC genes were similar to their foster mothers', even when that meant winding up in the position of mating with a female whose MHC genes were similar to their own. In other words, exposure during development, rather than some innate MHC mate-choice program, appears to be at work.

These startling results suggested that animals might somehow be able to detect prospective mates whose MHC genes are more or less similar to their own and then to choose mates accordingly. They may have evolved such an ability because of MHC's importance in the immune system or simply because MHC sensibility is a way to distinguish relatives from nonrelatives and thus to avoid inbreeding. To find out, scientists have shifted their studies of mate choice from the more traditional variables of outward appearance to the realm of genetics.

Inspired by the mouse studies, Ober launched a project on mate choice in the Hutterites in 1992. Hutterites go about finding mates as most people do--by socializing. After graduating from high school, young adults travel with their parents to other colonies to help with farmwork and household chores, often when a new baby arrives. These social gatherings are opportunities for young people to interact and are the way that most Hutterites meet their spouses. Most marry in their early twenties, after a year of courtship. The same colonies tend to exchange visits over and over, resulting in a history of marriages between particular colonies and even particular families. In fact, 20 percent of Ober's sample were "double" marriages, with two or more siblings from one family married to siblings from another.

After genetically typing 411 couples, Ober discovered that, in this inbred population, where there is a much higher likelihood of falling in love with someone carrying a similar haplotype, spouses shared fewer HLA genes than expected. Ober traced HLA inheritance further and found that when a husband and wife did have a matching strand, the set of genes in question most often came from the paternal side, for both the wife and the husband. In other words, men and women appear to have avoided finding a mate with HLA genes that resembled their mother's; it did not seem to matter if their partner had genes like their father's. As in mice, people may somehow imprint on their mother's HLA and then be drawn to someone with genes different from hers.

But how do people, or mice for that matter, know who is who in the first place? Research on mice, rats, and even humans has implicated smell as a possible cue for distinguishing MHC/HLA genes. Mice and rats, researchers have shown, can easily distinguish the urine and body odor of fellow creatures caged in the same way, fed the same food, and bred to be genetically identical in all but their MHC genes. Remarkably, research has also shown that humans can smell the difference in body odor, urine, and feces among mice of various MHC types.

A controversial finding, published in 1995, suggests that the sense of smell also plays a role in human mate choice. Claus Wedekind and colleagues, at the University of Bern, presented women with T-shirts imbued with the body scent of men with various HLA haplotypes. The women (who were not taking birth control pills and were in the middle of their menstrual cycle, and thus at their most fertile) preferred the T-shirts of men with HLA haplotypes unlike their own. The women also tagged these T-shirts as smelling like their current partners, suggesting that odor has something to do with their real-world mate choices. (Interestingly, women who were taking the pill--which simulates pregnancy--preferred T-shirts of men with HLAs like their own.)

Rachel Herz and Elizabeth Cahill, also at Monell Chemical Senses Center, would not be surprised to find that women attempt to sniff out a genetically proper mate. Surveying both sexes about which senses are important, they found that "for females, how someone smells is the single most important variable" in choosing a lover. If odor does provide information about the immune system, it makes evolutionary sense for women to pay attention to smell: they have much to lose if they mate with an inappropriate male and give birth to a baby with a reduced ability to fight off disease.

But unlike mice, humans--as primates--rely mostly on vision, not smell, to navigate through life. Perhaps other sensory cues--embedded in facial features or body shape, for example--alert us to the hidden genes in potential partners. If these cues are there, however, Ober and other researchers have yet to discover them.

During our ancient hunter-gatherer days, when humans lived in small bands separated by great distances, finding a mate with a different set of HLA genes and producing children with immunologic advantages may have been critical to survival. At this point in human history, with our restless, globe-trotting populations and relatively unlimited number of possible mates, we are unlikely, even across a crowded room, to spy someone with the same HLA haplotype. But even if we did, as the Hutterites have demonstrated, our genes would probably compel us to look the other way.

Mag.Coll.: 95E0467 Article A21084296

(morality and genetics)(Editorial)

Charles Colson; Nancy Pearcey.

The strategy of evolutionary psychology is to debunk traditional morality by reducing it to genetic self-interest.

Richard Dawkins, the flamboyant British biologist who gave us the phrase "selfish genes," now offers a genetic explanation for President Clinton's alleged foibles with les femmes. Our evolutionary ancestors were harem builders (like seals), Dawkins explains, instead of monogamous (like Canadian geese). Any male monopolizing power and wealth also monopolized the females, thus ensuring the survival of his genes. Clinton's behavior is simply a fossilized remnant from our genetic past.

Well, Dawkins's Just-So story may be good for a chuckle. But reducing human behavior to genetics is serious business these days. The latest fad is an updated version of sociobiology known as evolutionary psychology, which seeks explanations for human behavior in our genes. Surprising numbers of liberals and conservatives are eagerly claiming it as support for their own political philosophies.

Why is evolutionary psychology so popular? It promises to fill a gap in the Darwinist world-view: the need for a workable morality. Ever since Darwin, many have recognized that evolution leads to moral nihilism. For example, Cornell biologist William Provine (himself a loyal Darwinist) acknowledges that Darwinism implies "no free will" and "no ultimate foundation for ethics."

But we all experience the angst of facing moral choices, so evolutionists keep trying new ways to fit morality into the picture. Evolutionary psychology claims that by examining our evolutionary history, we can identify which behaviors have been selected for their adaptive value. These provide the basis for a genuinely scientific morality.

Does this new theory succeed in rescuing evolution from moral nihilism? No. The first problem is that any behavior practiced anywhere can be judged to have survival value--after all, it has survived--including behavior widely considered immoral.

In a recent article in the New York Times, Steven Pinker, a telegenic science popularizer, urges us to "understand" teenage girls who kill their newborns, arguing that "the emotional circuitry of mothers has evolved" by natural selection to include "a capacity for neonaticide." Infanticide is built into our "biological design," and we can't blame people for doing it.

Pinker's article triggered outraged responses, and even he backs away from his own conclusions: In the same article he unexpectedly writes, "Killing a baby is an immoral act." So which is it? Either evolution provides a moral guide or it doesn't.

The same contradiction runs through Dawkins's article, which explains harem building as a product of natural selection. Sensing his readers might take this as justification for immorality, Dawkins confides that he has made the "un-Darwinian personal decision" to be "deliberately monogamous." But if the Darwinian process provides a moral guide, why should anyone need to make "un-Darwinian" decisions? For that matter, how could anyone?

The second problem with evolutionary psychology is its genetic determinism. It claims to base morality on the genes that govern our behavior. But if genes really do that, moral choice is an illusion.

In The Moral Animal, Robert Wright spends hundreds of pages denouncing freedom as an illusion and describing human beings as "robots," "machines," and "puppets" of our genes. Then he turns 180 degrees, arguing that we are free to choose moral ideals contrary to the "values" of natural selection.

These fatal contradictions make hash of every effort to derive morality from biology--and Christians need to press the point in our classrooms and living rooms. Evolutionary psychology is being dished up to an anxious public via magazine and newspaper articles. In an era dogged by declining morality and social decay, it offers the soothing promise of a morality buttressed by the certitude of science.

Make no mistake. The goal of evolutionary psychology is utterly radical: to replace traditional religious morality with a new scientific morality. The strategy is to debunk traditional morality by reducing it to genetic self-interest. Wright is typical: He unmasks all "thoughts and feelings," all "moral values," as "stratagems of the genes." Even Jesus' teachings are nothing but ideologies serving his "evolutionarily ingrained interests." Evolutionary psychology proposes to base morality frankly on "selfish genes."

Some Christians have hoped to make peace with Darwinism as long as it is restricted to biology. But evolutionary psychology demonstrates that there is an inexpungable imperialism in Darwinianism--a compulsion to reduce all society to material mechanisms. Just as Darwinist theory in biology aims to replace divine design with natural processes, so in ethics it aims to replace revealed morality with a naturalistic morality. Sociologist Howard Kaye observes evolutionary psychology is nothing less than a secularized natural theology--an attempt to use nature to justify a secular world-view.

Exactly 100 years ago in his Princeton lectures, Dutch statesman Abraham Kuyper argued that Christians are not up against individual theories but comprehensive world-views. The only sure defense is to frame Christianity as an equally comprehensive world-view. Only then will genetic Just-So stories be relegated to the storybooks, where they belong.

Mag.Coll.: 95F4053

an analysis based on twins reared apart.

Thomas J. Bouchard Jr.; Yoon-Mi Hur.

The Myers-Briggs Type Indicator (MBTI) is a widely used personality

assessment inventory. It was designed to measure four bipolar

dimensions of personal preference derived from Jungian personality

theory (Myers & McCaulley, 1985). The dimensions are Extraversion

versus Introversion (EI), Sensing versus Intuition (SN), Thinking

versus Feeling (TF), and Judging versus Perceiving (JP).

To our knowledge, the MBTI has been used in only one behavior genetic

study. Vandenberg (1967) admininistered the MBTI to 27 same-sex

dizygotic twins (DZ) and 40 monozygotic twins (MZ) in the Louisville

twin study. He reported a Holzinger heritability of .46 for EI and

heritabilities of zero for TF, JP, and SN. No twin correlations or details

of the sample were presented in his brief report. His results are highly

discrepant with the larger literature dealing with genetic influence on

personality measures (see below), and in addition his twins were young

in age. The appropriateness of the inventory for a young sample is open

to question.

There are no adult behavior genetic studies of the MBTI. Consequently,

we will compare our findings with the results of behavior genetic

findings reported for the big five, the dominant model in American

psychology (Goldberg, 1993) and the model with which the MBTI is

most often compared (Furnham, 1996). The EI and SN scales of the

MBTI are good markers of Extraversion and Openness as measured by

the NEO-PI. Correlations between EI and Extroversion and SN and

Openness were, respectively, -.74 and .72 in a male sample (N = 267)

and -.69 and .69 in a female sample (N = 20 1). We can, therefore,

directly compare our findings with those reported for Extraversion and

Openness in the literature. TF and JP have only a modest relationship

with Agreeableness and Conscientiousness, yielding correlations of .44

and -.49 (males) and .46 and -.46 (females) (McCrae & Costa, 1989).

Consequently, comparison of our findings with behavior genetic results

in the literature for Agreeableness and Conscientiousness will be much

less informative.

Loehlin (1992) summarized all of the kinship literature dealing with the

genetics of personality under the rubric of the big five. He found that the

average heritability was .42, with slightly lower heritabilities for

Agreeableness (.35) and Conscientiousness (.38) than for the other three

traits. Bouchard (1994, 1997) reanalyzed the best big five markers on the

Multidimensional Personality Questionnaire (MPQ) in data sets obtained

from the Minnesota Study of Twins Reared Apart (MISTRA) and the

Minnesota Twin Registry. The average heritability was .41.

Agreeableness yielded the lowest heritability (.30), and

Conscientiousness yielded an average heritability (.42). Jang, Livesley,

and Vernon (1996) have reported heritabilities for the big five, as well as

for the big five facet scales. The heritabilities were derived from a single

design, a large sample of MZ and dizygotic DZ twins reared together

who completed the NEO-PI-R (Costa & McCrae, 1992). Heritabilities

ranged from.41 (Neuroticism and agreeableness) to .61 (Openness),

with a mean of .48. On the basis of the extensive literature demonstrating

genetic influences on personality, we expect to find significant genetic

influence on all four continuous scales of the MBTI.

METHOD

The sample includes sets of monozygotic and dizygotic twins and triplets

reared apart (MZA and DZA twins) and other individuals (mostly

spouses of twins), over 17 years of age, who have participated in

MISTRA since 1979. MISTRA is an ongoing longitudinal study of twin

and triplet sets separated in infancy, reared in different homes during

their formative years, and reunited as adults. Participants engage in

approximately 50 hours of comprehensive medical and psychological

assessments over a 1-week time span at the University of Minnesota.

Zygosity determination has varied over the years as the Minneapolis

Memorial Blood Bank has changed the genetic systems assessed for

determination of twin type. Until 1994, we never used fewer than 9

markers (blood groups, serum proteins, and enzymes). We currently use

6 blood group markers and 3 DNA fragment length polymorphisms.

Physical resemblance is always assessed by at least one experienced twin

researcher. The probability that a DZA pair would be concordant on all

markers, and thus misclassified as an MZA pair, is probably less than

.001 (Lykken, 1978). The study protocol and recruitment procedures are

described in greater detail elsewhere (see Bouchard, Lykken, McGue,

Segal, & Tellegen, 1990).

Table 1 presents descriptive characteristics of the twin sample used in the

present investigation. As Table I shows. the sample was predominantly

female (62%) and adult (mean age of 42.13 years, standard deviation of

13.10 years). Although the age at separation varied, 82% of the twin

pairs were separated during the first year of life. Twins had varying

degrees of contact prior to assessment; total amount of contact time

ranged from 1 week to 1,233 weeks.

Table 1 Means, Standard Deviations, and Ranges for Age at Testing and

Measures of Contact for MZA and DZA Twins

Age at Total contact Age % separation time (in years) Females (in years) (in weeks)

MZA (N = 61) Mean 40.71 62.3 .48 93.22 SD 12.71 .76 194.38 Range 18 to 68 .01 to 4.00 1 to 1233

DZA (N = 49) Mean 43.90 62.2 .78 44.76 SD 13.44 1.09 56.13 Range 22 to 77 .01 to 4.50 1 to 235

Total (N = 110) Mean 42.13 62.3 .61 71.43 SD 13.10 .93 150.71 Range 18 to 77 .01 to 4.50 1 to 1233

Note. Two triplet sets were each entered as three sets. DZA twins

included 13 opposite-sex pairs.

Form F (166 items) of the MBTI (Myers & McCaulley, 1985) was

administered to the sample. There are a variety of scoring systems for the

MBTI (Harvey & Murry, 1994; Myers & McCaulley, 1985; Saunders,

1987). In this research the standard continuous scores (Myers &

McCaulley, 1985) were chosen because of their widespread use by

others in the research context (McCrae & Costa, 1989) and the world of

work Furnham, 1992; Furnham & Stringfield, 1993).

Analysis of the influence of separation and contact. As indicated in Table

1, these twins were separated at various ages and had varying degrees of

contact prior to participation in the MISTRA assessment. The influence

of these variables was estimated by correlating them with twin absolute

differences on the MBTI scales. None of the correlations was statistically

significant, suggesting that amount of separation and contact had little

effect on twin similarity for the MBTI.

Age and sex correction. To determine genetic and environmental

influences on the four MBTI scales, two types of data analyses were

conducted: twin intraclass correlations and model-fitting analyses. As the

existence of age and sex effects can bias the estimates of twin similarity,

prior to these analyses the raw scores on the MBTI scales were corrected

for gender, age, age(2) and age x gender interactions using a regression

procedure described in McGue and Bouchard (1984).

Twin intraclass correlation. Twin intraclass correlations were computed

for the MZA and DZA twins using the standard formula, r = (MSB -

MSW) /(MSB + MSW), where MSB and MSW are, respectively, mean

squares between and within twin pairs derived from a one-way ANOVA.

Resemblance for the MZA twins includes both additive and nonadditive

genetic effects and thus provides the best estimate of total genetic

influence (broad heritability).

Nonadditive genetic effects involve interactions among alleles at a single

locus (dominance-recessiveness) as well as interactions among alleles at

different loci (epistasis). Whereas monozygotic twins share all

nonadditive genetic effects, dizygotic twins share approximately a quarter

of genetic effect due to dominance and relatively little genetic effect due

to epistasis. Thus, if nonadditive genetic effects are important for a

particular trait, the correlation for dizygotic twin correlations will be less

than half the correlation for monozygotic twins (Plomin, DeFries, &

McClearn, 1990). The comparison of the MZA and DZA twin

correlations in the present design will be used to test for evidence of

nonadditive genetic effects on the MBTI scales.

Model-fitting. The quantitative genetic model assumes that observed

phenotypic variance ([V.sub.p]) is a linear additive function of genetic

([V.sub.g]) and environmental ([V.sub.e]) variances. Symbolically,

[V.sub.p] = [V.sub.g] + [V.sub.e].

Assuming a nonadditive component, [V.sub.g] can be further

decomposed as

[V.sub.p] = ([V.sub.a] + [V.sub.a]) + [V.sub.e]

where [V.sub.a] refers to variance due to additive genetic effects,

[V.sub.d] refers to variance due to nonadditive genetic effects,(1) and

[V.sub.e] refers to variance associated with environmental effects. As

our twins have been reared apart, the environmental variance in the

model represents residual variance not explained by hereditary influence,

that is, nonshared environmental variance confounded with measurement

error.

From quantitative genetic theory, we can derive the expected covariance

between any two relatives as a function of the variance components given

above. The expected covariances between the MZA and DZA twin pairs

will be,

[COV.sub.(MZA)] = [V.sub.a] + [V.sub.d], and

[COV.sub.(DZA.] = 0.5[V.sub.a] + 0.25[V.sub.d]

The general assumptions in the model are: (a) there is no

genotype-environment correlation or interaction, (b) all nonadditive

genetic effects are dominance effects, (c) mating is random with regard to

the traits under study, and (d) no selective placement on trait-relevant

factors has occurred. We test the assumption of random mating below.

Extensive discussions of the other assumptions in behavioral genetic

designs are available elsewhere (Bouchard et al., 1990; Bouchard &

McGue, 1990; Plomin et al., 1990).

Expected variances and covariances based on the model for the MZA and

DZA twins were applied to the observed variances and covariances using

a maximum likelihood estimation procedure in Mx (Neale, 1995). One of

the advantages of the model-fitting technique is that it provides the

opportunity to test competing theoretical models. We fit a full model and

three reduced models to the twin covariances. The full model consisted

of additive and nonadditive genetic variance parameters and an

environmental plus measurement error variance parameter; the three

reduced models entailed a reduction of one or more of these parameters.

To measure overall fit of the model, we used the chi-square test statistic

with the criterion for rejection of models at the .05 probability level. In

evaluating the relative fits of the various models, particularly those that

could not be rejected on the basis of the chi-square test statistic alone, the

Akaike Information Criterion (AIC = [X.sup.2] - 2[df]) was examined.

AIC quantifies the information content of a model in terms of the joint

criteria fit and parsimony. In general, small chi-square values from

models with few free parameters lead to small AICs, representing

maximum parsimony, whereas large chi-square values from models with

many parameters yield large AICs, representing lack of parsimony

(Akaike, 1987).

Descriptive Statistics

Means and standard deviations for MBTI scales by zygosity are shown

in Table 2. Two-tailed t-tests and variance ratio tests were carried out to

test for differences between MZA and DZA means and variances. Except

for the JP scale, t-tests yielded no significant mean differences. On the

JP scale, the mean was significantly higher for the MZA twins than for

the DZA twins (p [is less than] .05). There were no significant variance

differences by zygosity on any of the scales. The SN means were low

for both the MZA and DZA groups compared to normative group means

reported in the manual.

Table 2 Means and Standard Deviations for the Four Myers-Briggs Type

Indicator Continuous Scales for MZA and DZA Twins

Scale MZA (N=61) DZA (N=49) Mean SD Mean SD

Extraversion-Introversion (EI) 100.07 26.72 102.86 25.70

Sensing-Intuition (SN) 89.52 25.63 83.12 26.11

Thinking-Feeling (TF) 108.30 21.04 105.57 20.90

Judgment-Perception (JP) 99.48(*) 28.74 90.33(*) 27.93

Note. (*) means differ at the p [is less than to] .05 level (two-tailed test).

Twin Intraclass Correlations

The twin and spouse intraclass correlations for the MBTI scales are

shown in Table 3.

The spouse correlations for EI, TF, and JP were essentially zero,

supporting the model assumption of random mating. The spouse

correlation of .23 for SN is too modest to cause any serious estimation

errors in the model fitting.

As discussed earlier, in the absence of selective placement on

trait-relevant factors, the MZA correlation is a direct estimate of genetic

influence. The MZA correlations in Table 3 show that the heritability

estimates for the MBTI scales ranged from.40 to.60, With a mean of

.50. For three scales -- EI, TF, and JP -- the DZA correlations were less

than half the MZA correlations and not significantly different from zero.

This pattern of twin correlations strongly suggests the possibility of

nonadditive genetic influence on EI, TF, and JP.

Table 3 Intraclass Correlations and 95% Confidence Intervals for the

Four Myers-Briggs Type Indicator Continuous Scales for MZA and DZA

Twins and Spouses

Scale MZA (N=61) DZA (N=49) Spouse (N=92)

Extraversion- Introversion (EI) .60 (.41-.74) .02 (-26-.30) .10 (-.11-.30)

Sensing- Intuition (SN) .40 (.17-.59) .34 (.07-.56) .23 (.03-.41)

Thinking- Feeling (TF) .58 (.39-.72) .00 (-28-.28) -.01 (-.21-.19)

Judgment- Perception (JP) .41 (.18-.60) .15 (-.13-.41) -.15 (-.34-.05)

Chi-square values and their probability estimates for the full and reduced

models are summarized in Table 4. The full model (ADE) included

additive genetic variance, nonadditive genetic variance, and

environmental variance combined with measurement error. Three

reduced models were tested in order to assess the significance of the

nonadditive effects (AE model), the additive genetic effects (DE model),

and total genetic effects (E model). Model-fitting produced results similar

to those based on an examination of the twin correlations. As the test

statistics in Table 4 show, the full model and its two submodels, AE and

DE, fit the data well for all four of the MBTI scales. The E model,

however, failed to fit for any of the four scales. Thus none of the data

can be explained by environmental factors alone.

[TABULAR DATA 4 NOT REPRODUCIBLE IN ASCII]

To arrive at the most parsimonious explanation of the data, we examined

the AIC values for each of the models. For the SN and JP scales, the AE

model yielded the lowest AIC values, showing that the genetic influences

on these scales are primarily additive. In contrast, for the EI and the TF

scales, the DE model produced the lowest AIC values, suggesting that

genetic influence on these scales is predominantly nonadditive.

Table 5 contains standardized maximum-likelihood parameter estimates

and their 95% confidence intervals derived for the ADE model for each

scale. The model-fitting heritability estimate (the sum of A and D) was

.57 for EI, .46 for SN, .60 for TF, and .39 for JP. The average of these

four estimates was .50. These results corresponded closely with the

estimates derived from the MZA correlations alone.

[TABULAR DATA 5 NOT REPRODUCIBLE IN ASCII]

DISCUSSION

Overall, these estimates of genetic influence on the MBTI scales are very

much like those found for most other personality measures -- in the

.40-.60 range. Extroversion has been the personality variable most

studied by psychologists as well as behavior geneticists. The majority of

behavior genetic studies report significant nonadditive genetic variance

for this trait, regardless of the measuring instrument used, and

heritabilities of about .54 (Bouchard, 1997, Tables 16.2 to 16.5).

Consequently, the high heritability ([h.sup.2] = .57) and evidence of

nonadditive genetic variance demonstrated for EI are very consistent with

previous findings. Jang et al. (1996), on the other hand, report a

heritability of .55 but no nonadditive genetic variance. Three of the six

facets of Extraversion do, however, show strong evidence of

nonadditive genetic variance. This heterogeneity of mode of genetic

transmission is characteristic of the facets underlying all of the big five

except Neuroticism, for which only one of the six facets deviates from

strict additivity.

On the basis of his comprehensive analysis of Openness, Loehlin (1992)

reported a heritability of .45, no nonadditive genetic variance, and a very

small common environmental component (.06). The results reported here

are very similar: a heritability estimate of .45 and little evidence for

nonadditive genetic variance. The MPQ Absorption scale, a reasonable

marker of Openness, yields estimates of .29 for additive variance, .15

for nonadditive variance, and .07 for common environmental variance

based on the MZA and DZA samples in this study as well as a large

reared-together twin sample (Bouchard, 1997, Table 16.9). Jang et al.

(1996) report a heritability of .61 consisting largely of nonadditive

genetic variance. Three of the six facets of Openness, however, show

primarily additive variance.

TF corresponds only modestly with Agreeableness and carries mostly its

own specific variance. The findings for TF include a high heritability

([h.sup.2] =.60) and strong evidence for nonadditive genetic variance.

These findings diverge from the results reported for Agreeableness by

others, findings that themselves are not very consistent. Loehlin (1992)

reported estimates of .24 for additive genetic variance and .11 for

nonadditive variance, yielding a broad heritability of only .35. Again

using the same sample as in this study and a large reared-together twin

sample, the MPQ Aggression scale, a reasonable marker of

Agreeableness, yields estimates of .05 for additive variance and .25 for

nonadditive variance, yielding a broad heritability of only .30 (Bouchard,

1997, Table 16.9). Jang et al. (1996) report a heritability of .41

consisting largely of additive genetic variance. Two of the six facet

scales, however, show nonadditive genetic variance.

JP corresponds only modestly with Conscientiousness and carries a great

deal of its own specific variance. Our heritability estimate of .38

corresponds exactly with Loehlin's estimate. Jang et al. (1996) report a

heritability of .44 consisting largely of additive genetic variance. Unlike

the other big five measures, three of the six Conscientiousness facets

showed significant common environmental influence.

The heterogeneity of the purported genetic and environmental

mechanisms underlying the big five, as assessed by different

instruments, as well as the heterogeneity of mechanisms underlying their

facets, shown in the Jang et al. (1996) study, throw into question the

view that the big five represent unity factors, at least at the biological

level (i.e., the parts are influenced by quite different underlying genetic

mechanisms).

The existence of large amounts of nonadditive genetic variance for some

traits (EI and TF in this study) is an important finding, if it is replicable,

as it predicts very modest correlations between first-degree relatives,

especially between parents and offspring, even though the trait itself is

under a considerable degree of genetic influence (Lykken, McGue,

Tellegen, & Bouchard, 1992; Waller, Bouchard, Lykken, Tellegen, &

Blacker, 1993).

The spouse correlations are either zero or very modest, and are similar to

those reported by others in the behavior genetic literature (Eaves,

Eysenck, & Martin, 1989). They add the findings from another

personality inventory to a large body of literature that shows, contrary to

widespread belief among both laypersons and professionals, that spouse

similarity in personality is quite modest (Lykken & Tellegen, 1993).

Nevertheless, the implications of the present findings are limited because the small number of kinships and modest sample sizes does not allow us to resolve D from A clearly. The classical twin design (MZ and DZ twins

reared together) has relatively low power to detect nonadditive genetic

variance. The twins-reared-apart design, on the other hand, is a powerful

design for estimating genetic influence (Lykken, Geisser, & Tellegen,

1978), but it does not provide an estimate of common environmental

influence. Estimation of the magnitude of common environmental

influence for the MBTI would be highly desirable. Although previous

studies suggest little common environmental influence in personality, this

outcome cannot be assumed for all traits. The significant influence of

common environmental influence for three of the six facets of

Conscientiousness in Jang et al. (1996) highlights the danger of such an

assumption. In addition, results from a limited set of kinships and unique

samples, such as the MZA and DZA twins used in this study, should

always be checked using complementary designs and new samples.

Useful information regarding the magnitude of both the additive and

nonadditive genetic influence on SN and JP could be obtained from

biological parent x adopted-away offspring correlations if both members

of the kinship were measured as adults. Given the large data bases

available on the MBTI, such information should not be impossible to

gather. Our findings predict that adult biological parent x adult

adopted-away offspring correlations for EI and TF should be very

modest (near zero), while the correlations for SN and JP will

approximate .20. Data from a large sample of adult twins reared together

would allow for an additional check on the heritability estimates and

provide an estimate of common family environmental influence.

An issue not discussed in this paper, but of considerable importance to

many users of the MBTI, is that of types. The title of the instrument

includes the terms "Type Indicator," and it purports to classify

individuals into taxonomic categories. This is a controversial claim and

virtually no mainstream personality researchers adopt this view.

Extraversion, for example, is represented in virtually every widely used

personality inventory and none of the test authors or users take a

typological stance. Part of the problem is that confirming a taxonomic

hypothesis is extremely difficult (Gangestad & Snyder, 1985). Even in

the domain of psychopathology (e.g., schizophrenia and depression).

where a taxonomic approach is assumed and genes underlying the

disorders are being sought as causal mechanisms, confirmation of the

hypothesis has been slow in coming, with most findings being

nonreplicable. New statistical methods for detecting taxonomic entities

are becoming available (Waller & Meehl, in press; Waller & Putman,

1996) and in the future we will apply them to the twins-reared-apart data

set. If the latent traits underlying the MBTI are truly categorical rather

than continuous, as our analysis assumes, it is still likely to be the case

that the influences underlying the categories are strongly genetic in

origin.

Behavior genetic studies of personality have seriously challenged the

standard socialization model that has been dominant in developmental

psychology for so many years (Harris, 1995; Scarr, 1997) and which is

also held by many personality psychologists. They are now challenging

dimensional models and theories based on simple phenotypic correlations

that are incapable of reflecting the complex and varying genetic

mechanisms that must underlie all personality traits.

1. Dominance ([V.sub.d]) referes to intralocus gene interactions and is

limited to .25 in the equation below. Interlocus interactions are an

additional form of interaction (epistactic) that could be allowed in model.

We restrict ourself to additive-dominance model because of the modest

sample sizes and limited number of kinships.

REFERENCES

Akaike, H. (1987). Factor analysis and AIC. Psychometrics, 52,

317-332.

Bouchard, T. J., Jr. (1994). Genes, environment, and personality,

Science, 264, 1700-1701.

Bouchard, T. J., Jr. (1997). The genetics of personality. In K. Blum &

E. P. Noble (Eds.), Handbook of psychiatric genetics (pp. 273-296).

Boca Raton, FL: CRC Press.

Bouchard, T. J., Jr., Lykken, D. T., McGue, M., Segal, N. L., &

Tellegen, A. (1990). Sources of human psychological differences: The

Minnesota study of twins reared apart. Science, 250, 223-228.

Bouchard, T. J., Jr., & McGue, M. (1990). Genetic and environmental

influences on adult personality: An analysis of adopted twins reared

apart. Journal of Personality, 58, 263-292.

Costa, P. T., & McCrae, R. R. (1992). NEO PI-R, professional

manual. Odessa, FL: Psychological Assessment Resources.

Eaves, L. J., Eysenck, H. J., & Martin, N. G. (1989). Genes, culture,

and personality: An empirical approach. New York: Academic Press.

Furnham, A. (1992). Personality at work: The role of individual

differences in the workplace. London: Routledge.

Furnham, A. (1996). The big five versus the big four: The relationship

between the Myers-Briggs Type Indicator (MBTI) and NEO-PI five

factor model of personality. Personality and Individual Differences, 21,

303-307.

Furnham, A., & Stringfield, P. (1993). Personality and work

performance: Myers-Briggs Type Indicator correlates of managerial

performance in two cultures. Personality and Individual Differences, 14,

145-153.

Gangestad, S., & Snyder, M. (1985). "To carve nature at its joints": On

the existence of discrete classes in personality. Psychological Review,

92, 317-349.

Goldberg, L. R. (1993). The structure of phenotypic personality traits.

American Psychologist, 48, 26-34.

Harris, J. R. (1995). Where is the child's environment? A group

socialization theory of development. Psychological Review, 102,

458-489.

Harvey, R. J., & Murry, W. D. (1994). Scoring the Myers-Briggs Type

Indicator: Empirical comparison of preference score versus latent-trait

methods. Journal of Personality assessment, 62, 116-129.

Jang, K. L., Livesley, W. J., & Vernon, P. A. (1996). Heritability of

the big five personality dimensions and their facets: A twin study.

Journal of Personality, 64, 577-591.

Loehlin, J. C. (1992). Genes and environment in personality

development. Newbury Park, CA: Sage.

Lykken, D. T. (1978). The diagnosis of zygosity in twins. Behavior

Genetics, 8, 437-473.

Lykken, D. T., Geisser, S., & Tellegen, A. (1978). Heritability

estimates from twins studies: The efficiency of the MZA design.

Unpublished manuscript, Department of Psychology, University of

Minnesota.

Lykken, D. T., McGue, M., Tellegen, A., & Bouchard, T. J., Jr.

(1992). Emergenesis: Genetic traits that may not run in families.

American Psychologist, 47, 1565-1577.

Lykken, D. T., & Tellegen, A. (1993). Is human mating adventitious or

the result of lawful choice? A twin study of mate selection. Journal of

Personality and Social Psychology, 65, 56-68.

McCrae, R. R., & Costa, P. T., Jr. (1989). Reinterpreting the

Myers-Briggs Type Indicator from the perspective of the five-factor

model of personality. Journal of Personality, 57, 17-40.

McGue, M., & Bouchard, T. J., Jr. (1984). Adjustment of twin data for

the effects of age and sex. Behavior Genetics, 14, 325-343.

Myers, I. B., & McCaulley, M. H. (1985). Manual: A guide to the

development and use of the Myers-Briggs Type Indicator. Palo Alto,

CA: Consulting Psychologists Press.

Neale, M. C. (1995). Mx: Statistical modeling. Richmond, VA:

Department of Human Genetics, Box 3 MCV.

Plomin, R., DeFries, J. C., & McClearn, G. E. (1990). Behavioral

genetics: A primer (3rd ed.). New York: W. H. Freeman.

Saunders, D. (1987). Manual: Type Differentiation Indicator: A scoring

system for Form J of the Myers-Briggs Type Indicator (Research ed.).

Palo Alto, CA: Consulting Psychologists Press.

Scarr, S. (1997). Behavior genetics and socialization theories of

intelligence: Truce and reconciliation. In R. J. Sternberg & E. L.

Grigorenko (Eds.), Intelligence: Heredity and environment (pp. 3-41).

New York: Cambridge University Press.

Vandenberg, S. G. (1967). Hereditary factors in normal personality

traits (as measured by inventories). In J. Wortis (Ed.), Recent advances

in biological psychiatry, 9, 65-104. New York: Plenum Press.

Waller, N. G., Bouchard, T. J., Jr., Lykken, D. T., Tellegen, A., &

Blacker, D. M. (1993). Creativity, heritability, familiarity: Which word

does not belong? Psychological Inquiry, 4, 235-237.

Waller, N. G., & Meehl, P. E. (in press). Multivariate taxometric

procedures: Distinguishing types from continua. Beverley Hills. CA:

Sage.

Waller, N. G., & Putman, F. W. (1996). Types of dissociation and

dissociative types: A taxometric analysis of dissociative experiences.

Psychological Methods, 1, 300-321.

This research was partially supported by grants to the Minnesota Study

of Twins Reared Apart from the University of Minnesota Graduate

School; the Pioneer Fund; the Seaver Institute; the Koch Charitable

Foundation; the Spencer Foundation; Harcourt, Brace and Jovanovich;

and the National Science Foundation (BNS-7026654). We wish to thank

the numerous graduate and undergraduate students who helped us collect

the date over the years, as well as our numerous collaborators on

MISTRA.

Named Works: Myers-Briggs Type Indicator (Test) - Usage

(Special Issue on Human Behavioral Genetics: Synthesis of Quantitative and Molecular Approaches)

Jeffrey W. Gilger; Scott L. Hershberger.

Abstract: Studies in the field of human behavioral genetics enable people to learn more about themselves and the genetic attributes that make them function as such. Sexuality, learning, intelligence as well as physical and psychological disorders are all rooted in peoples' genetic makeup. Providing an accurate description of the physiological process of these genes in development and how such genes interact with other genes and with the environment are among the problems that have to be resolved by the field.

The more we discover as to physiological activity and inheritance, the more difficult it becomes to imagine any physical or chemical description or explanation which could in any way cover the facts of persistent coordination.

• J.S. Haldane, 1931, p. 12

Science, like the individual person, matures and develops. Things that seemed impossible at one time are often realizable later on. The opening quote by Haldane represents a common belief held at that time that we would never be able to identify genes, understand their molecular complexity, or unlock the genetic code. However, it turns out that this cynicism was premature, and today we are in the midst of a revolution in genetic science where many once unimaginable things are now possible.

Early in this century the field of genetics was born. At first it was applied to plants and nonhuman animals, and gradually it developed to a level of understanding such that the genetic microscope could be directed at ourselves. Long before Haldane made his statement, Galton wrote about variation in human abilities and talents being a function of genes, and he conducted the first formal twin and adoption studies providing evidence for his hypotheses (Galton 1869). Some time after that and about the same time of Haldane's statement, the first inborn error of metabolism was identified that had an effect on human cognition (phenylketonuria) (Folling 1932). Since then a wide variety of syndromes and genes have been identified that have cognitive and personality sequelae. This is not surprising in that we now know that approximately 30% of the estimated 70,000 to 100,000 genes in the human genome are expressed primarily in the brain (Adams et al. 1991), and of course it is the brain that governs everything humans do, think, or perceive.

This special issue of Human Biology is a collection of papers that provides a bit of history about how the field of human behavioral genetics has developed in the twentieth century. These papers review the current state of genetics as applied to individual and group differences in human behavior and provide data on what is now known about the genetics of various complex psychological human phenotypes. Advances in biometric methods and molecular techniques have pushed the growth of this field in recent years.

We have organized the papers into several sections. First is a section containing papers by Segal and MacDonald, Geary, and Ganger and Stromswold. These three reviews set the stage for the later papers in that they describe how evolution may have led to genetic variation effecting key human behaviors. Thus there is reason to suspect that the selective forces thought to operate on physical characteristics might also operate on behavioral characters. Such an approach to behavior is part of a fairly new and increasingly better defined interdisciplinary approach called evolutionary psychology.

The next five sections deal with specific behaviors. Each section is designed to include a paper that approaches behavior from the classic biometric or quantitative approach long popular in behavioral genetics (e.g., twin, adoption, and family studies) and an accompanying paper that addresses behavior from a more molecular approach. All the papers describe the difficulties of studying human behavior and how some of the new molecular and statistical techniques being developed may result in the identification of specific genes. Although both papers of each section may take different perspectives, the reader will see that the current thought is that quantitative and molecular methods will best advance the field when used in conjunction.

The behaviors we could include in this issue were of course constrained by space limitations, and there are many more behaviors we could have considered [see for instance Sherman et al. (1997)]. We chose to focus on reading and learning disabilities, intelligence, language development, sexuality, and child psychopathology. Study of each of these behaviors has recently produced some intriguing results in both the quantitative and qualitative areas.

Reading disability is currently the best understood learning disorder and complex "psychological" trait in terms of its genetics. It is also the first such trait for which the location of at least one specific gene has been identified through linkage work (Smith et al.). In the accompanying paper Light et al. show how behavioral genetic methods can also be used to describe the etiological relationship between various learning problems that may co-occur in the same individual (i.e., comorbidity). For example, are math and reading problems due to the effects of the same (pleiotropy) or different genes?

Bouchard and Daniels et al. describe the current thought on the inheritance of IQ. Bouchard describes the heritability of IQ and specific cognitive abilities and concludes that most of the variance in IQ (particularly in the later adult years) is due to genetics. This is a politically hot topic [see discussion by McGue (1997)], but a vast number of studies support Bouchard's claims. As discussed by Daniels et al., there may even come a time when we will be able to locate the key loci (and eventually the genes) responsible for much of the variation observed for IQ at the population level. In their work Daniels et al. have begun to identify potential regions of the genome and candidate genes that will be explored by means of molecular and linkage methods.

In the section on language Brzustowicz and Stromswold provide an overview of how genes may be involved in deviations from normal language development. Brzustowicz describes how molecular methods, given the appropriate phenotype, may lead to the discovery of key genes for language processing, whereas Stromswold provides a review of family designs and other research on language development. Again, it is increasingly clear that genes play a role in normal and abnormal linguistic skills.

T he origin of heterosexuality, homosexuality, and variants thereof are discussed by Pattatucci and Pillard and Bailey. The genetic work in this area is new, and both papers point to data that suggest that at least some of the individual and group differences in sexuality have a genetic component.

In the last section are papers on psychopathology and personality by Nigg and Goldsmith and Alsobrook and Pauls. As with reading disability, the issue of how we can understand comorbidity of psychiatric disorders is discussed (Nigg and Goldsmith). Alsobrook and Pauls describe some of the new linkage techniques that are being applied to psychiatric disorders, including Tourette's syndrome.

In closing, this special issue presents information about where the field of behavioral genetics is heading. We are at the stage of development where the very real possibility exists that we will identify the key genes for such behaviors as sexuality, learning, and intelligence. But we must not stop there. The real problems facing the field in the future will be to accurately describe the physiological process of these genes in development (from conception on), how these genes interact with other genes, and particularly important, how these genes interact with the environment. Thus someday we will go beyond the description of heritability estimates or the localization of specific genes. And like Haldane's perspective, although this seems perhaps an insurmountable task today, it may prove to be otherwise.

Adams, M.D., J.M. Kelley, J.D. Gocayne et al. 1991. Complementary DNA sequencing: Expressed sequence tags and human genome project. Science 252:1651-1656.

Folling, A. 1932. Uber Ausscheidung von Phenylbrenztraubensaure in den Harn als Stoffwechselanomalie in Verbindung mit Imbesilliat. Hoppe-Seyler's Z. Physiol. Chem. 227:169.

Galton, F. 1869. Hereditary Genius: An Inquiry into Its Laws and Consequences. London, England: Macmillan Press.

Haldane, J.S. 1931. The Philosophical Basis of Biology. Garden City, NY: Doubleday, Doran.

McGue, M. 1997. The democracy of genes. Nature 388:417-418.

Sherman, S., J.C. DeFries, I.I. Gottesman et al. 1997. Recent developments in human behavioral genetics: Past accomplishments and future directions. Am. J. Hum. Genet. 60:1265-1275. Article A20816222

Nancy L. Segal; Kevin B. MacDonald.

Behavioral geneticists and evolutionary psychologists have generally pursued human behavioral analyses with little theoretical or methodological exchange. However, significant benefits might accrue from increased communication between these disciplines. The primary goals of this article are (1) to identify meaningful junctures between behavioral genetics and evolutionary psychology, (2) to describe behavioral genetic research designs and their applications to evolutionary analyses, and (3) to reassess current personality research in light of behavioral genetic and evolutionary concepts and techniques. The five-factor model of personality is conceptualized as subsuming variation in normative species-typical systems with adaptive functions in the human environment of evolutionary adaptation. Considered as universal evolved mechanisms, personality systems are often seen in dynamic conflict within individuals and as highly compartmentalized in their functioning between settings. However, genetically influenced individual differences in personality may also be understood within an evolutionary framework. Studies of the heritability of personality traits indicate broad-sense heritabilities in the 0.40-0.50 range with evidence of substantial nonadditive genetic variation and nonshared environmental influences. Evidence indicates that evolutionary theory (e.g., inclusive fitness theory) predicts patterns of social interaction (e.g., cooperation and bereavement) in relatives. Furthermore, variation in personality may constitute a range of viable strategies matching the opportunities available in the complex niche environment of human societies. Within this wide range of viable strategies, personality variation functions as a resource environment for individuals in the sense that personality variation is evaluated according to the interests of the evaluator (e.g., friendships, coalitions, or mate choice). KEY WORDS: EVOLUTIONARY PSYCHOLOGY, PERSONALITY TRAITS, BEHAVIORAL GENETICS, FIVE-FACTOR MODEL OF PERSONALITY

Behavioral genetics and evolutionary psychology remind us of ships passing in the night. For the most part, issues and problems have been separately pursued with little reference to the rich set of concepts and methods that each might offer the other. Formal definition of both these disciplines has been recent, yet progress has been swift. Accumulation of new data and interpretations has been impressive and reappraisal of existing material has been exciting. These developments have not gone unnoticed: A modest cadre of researchers [e.g., Freedman (1968), Buss (1987, 1990), Crawford and Anderson (1989), Belsky et al. (1991), Mealey and Segal (1993), Segal (1993), Rowe (1993), and Scarr (1995)] have persuasively argued that there is territory for fruitful exchange between these natural allies. Unfortunately, attempts at reconciliation are long overdue because of theoretical and methodological differences, or what Scarr (1995) has termed differences in intellectual geography. The primary goals of this article are (1) to identify meaningful junctures between behavioral genetics and evolutionary psychology, (2) to describe behavioral genetic research designs and their applications in evolutionary analyses of behavior, and (3) to reassess current personality research in light of behavioral genetic and evolutionary concepts and techniques.

The five-factor model (FFM) has received a great deal of support within personality psychology [e.g., Digman (1990)], and our ideas will draw heavily on the FFM literature. The FFM emerged from factor-analytical studies of English-language trait descriptors. Across a wide variety of studies of trait descriptive terms five factors consistently emerge in self-ratings or ratings by observers (Goldberg 1981, 1992). The FFM derives from a large and representative set of trait descriptors, not only in English but also in other languages, suggesting that the FFM can be generalized cross-culturally. In addition, personality traits proposed by other personality theories typically appear as subsets of the FFM traits (Costa and McCrae 1992). As a result of this converging body of evidence, the FFM has achieved preeminence within scientific personality psychology.

Factor 1 of the FFM is often labeled surgency or dominance, and it is associated with variation in sociability, sensation seeking, impulsivity, attraction to reward, and social dominance. Factor 2 of the FFM is agreeableness or love, and it is positively associated with variation in nurturance, warmth, empathy, and charitableness and negatively associated with cruelty. Factor 3 of the FFM is conscientiousness/behavioral inhibition, which is associated with variation in the ability to defer gratification, persevere in unpleasant tasks, pay close attention to detail, and behave in a responsible, dependable manner. Factor 4 is neuroticism, and it is associated with variation in tendencies toward being anxious, worrisome, moody, and emotionally reactive. Factor 5 is labeled openness to experience, and it is associated with variation in imaginativeness and creativity, intellectual and aesthetic interests, and broad-mindedness and unconventionality. These five factors and their associated attributes are summarized in Table 1.

[TABULAR DATA FOR TABLE 1 OMITTED]

Evolutionary Psychology: Focus on Universal Psychological Mechanisms

Following in the footsteps of traditional ethological research [e.g., Eibl-Eibesfeldt (1989)], evolutionary psychology is concerned with the origins, functions, and adaptive significance of human behavior (Segal 1993, 1997b; Buss 1995). In contrast to behavioral geneticists whose concerns are with within-species variability, evolutionary psychologists focus on universal psychological mechanisms underlying behavior. Applied to personality psychology, the evolutionist conceptualizes personality as a set of systems that evolved to solve adaptive problems (MacDonald 1995). Human psychological adaptations underlying personality are thus conceptualized as a set of normative universal mechanisms designed by natural selection to solve particular problems presented by the environment in our evolutionary past. Just as the vertebrate eye was designed by natural selection to respond to the properties of light and the structure of surfaces as enduring and recurrent features of the environment, it is proposed here that the behavioral approach personality systems (factor 1 of the FFM) are designed to motivate organisms to approach sources of reward (e.g., sexual gratification) that occurred as enduring and recurrent features of the environments in which humans evolved. Furthermore, just as creatures with highly sophisticated visual abilities were favored by natural selection because they were able to solve recurrent problems presented in navigating a three-dimensional world, the behavioral-approach system served to promote biological fitness by motivating organisms to acquire resources, including sexual resources and social status, related to reproductive success. Finally, just as some individuals can see better than others, the fact that there is variation in the systems underlying vision is not incompatible with a functionalist interpretation of these systems viewed as normative universal mechanisms. Barring organic visual impairment, all people have eyes with a lens that functions to regulate the amount of light entering the eye, but some individuals see better than others. Similarly, we all have behavioral approach systems conceptualized as universal human adaptations. However, some of us are more predisposed toward social dominance, reward seeking, sensation seeking, etc. than others.

The functionalist account of the systems underlying the factor space of the FFM is strengthened by findings that individual differences in personality are associated with individual differences in physiological systems common to all humans. In fact, considerable evidence links personality systems with specific brain regions and neurochemicals (Eysenck 1982; Eysenck and Eysenck 1985; Gray 1982, 1987; MacDonald 1995). Moreover, functionally and neurophysiologically similar systems are apparent in animal research [e.g., Gray (1982, 1987)]. There is evidence for personality traits in wolves (MacDonald 1983), and there is even evidence that individual differences in personality among chimpanzees can be understood within the FFM framework (Figueredo and King 1996; King and Figueredo 1994). Also adding to the plausibility of an evolutionary approach is the finding that sex differences within the five-factor space conform well to expectations resulting from the evolutionary theory of sex (MacDonald 1995).

The evolutionary theory of sex differences predicts sex differences in the FFM dimensions. For example, evolutionary theory predicts that males are more likely to be risk takers and sensation seekers because males benefit more from multiple matings and must compete with other males to do so. In contrast, females, as the high-investment sex because of limited gametes, can afford to adopt a more conservative strategy and have less to gain by adopting a risk-taking strategy. Sensation seeking is a component of factor 1 of the FFM, and males do indeed have higher average scores than females on sensation seeking in cross-cultural samples (Zuckerman 1979, 1984, 1990, 1991).

Functions of Trait Distributions. The adaptationist perspective proposes that the trait distributions of dominance and sensation seeking (factor 1), conscientiousness/behavioral inhibition (factor 3), and neuroticism (factor 4) reflect phenotypic (observable) variation in systems that serve three critical adaptive functions (MacDonald 1995): (1) the need to approach the world and accumulate resources, including sexual resources and other rewarding stimulation, and, in social species, social dominance (behavioral approach); (2) the need to monitor the environment for dangers and impending punishments and to persevere in tasks that are not intrinsically rewarding (conscientiousness/behavioral inhibition); and (3) the need to mobilize behavioral resources by moderating arousal in acutely demanding situations in the service of both approach and avoidance behaviors (affect intensity). Affect intensity may be viewed as a general behavioral engine that is used in the service of both behavioral approach and behavioral avoidance. It is a behavioral scaling system that allows the organism to scale its responses to current environmental opportunities and threats. In addition, agreeableness or love (factor 2) is proposed to underlie adaptive relationships of intimacy and other long-term relationships, especially family relationships involving reciprocity and transfer of resources to others (e.g., maternal and paternal investment in children). Finally, the openness to experience factor (factor 5) taps variation in what one might term optimal Piagetian learning, that is, intrinsically motivated curiosity and interest in intellectual and aesthetic experience combined with imagination and creativity in these areas.

Implications of the Universalistic/Systemic Perspective: Conceptualizing Interactions with Environmental Variation and Interaction among Personality Systems. Some of the traditional problems of personality theory come into clearer focus within an evolutionary framework. As described, personality systems are adaptations designed to approach environmental opportunities and avoid dangers in the environment. Personality systems are therefore responsive to environmental cues, resulting in what one might term system x situation interaction (MacDonald 1995). Thus the conscientiousness/behavioral inhibition system is triggered by perceptions of impending punishment or danger. People perceiving themselves to be in such situations experience emotions of fear and anxiety and are motivated to alter the situation by, for example, removing or escaping from the danger. Behavioral approach mechanisms may be externally cued, as when males perceive attractive, nubile, sexually receptive females, or they may be internally cued, as with a system such as sensation seeking in which individuals actively seek out exciting, stimulating, and even dangerous environments.

Besides system x situation interaction, there can also be system x system interaction, that is, interactions between systems at the physiological level. For example, there is evidence of mutual inhibitory influences between the mechanisms underlying behavioral inhibition and behavioral approach. Mesulam (1986) described reciprocal inhibition between the frontal and parietal lobes; the parietal lobe was viewed as an approach system characterized by diffuse attention and impulsive responding. Gray (1987) also noted reciprocal inhibitory influences between the behavioral inhibition system (BIS) and the reward-based approach system, and in the rat at least, the inhibitory influences from the BIS are more powerful than in the reward-based approach system. Both systems can be aroused in particular situations, as when a previously rewarded behavior is punished. Tucker and Derryberry (1992) also emphasized the inhibitory role of the frontal cortex. They noted that individuals with frontal lesions are prone to impulsive responding, to the "disinhibition syndrome" (Luria 1980), and to inadequate anticipation of aversive outcomes.

The model implied by the evolved systems perspective is therefore useful for conceptualizing conflicts between evolved systems and for situation specificity. For example, the behavioral approach associated with dominance and sensation seeking and the behavioral avoidance systems associated with conscientiousness/behavioral inhibition are psychometrically independent, implying that individuals can be more or less sensitive to rewards and more or less sensitive to punishments and deferring gratification. Nevertheless, each system has inhibitory effects on the other system, so that in a situation with both potential rewards and potential punishments both systems are activated. Individuals high on dominance and sensation seeking evaluate the risks involved and engage in behavioral approach. In contrast, introverts, being less attracted to the potential rewards, are more likely to have approach tendencies inhibited by mechanisms underlying conscientiousness, such as Gray's (1982) BIS. A situation characterized overwhelmingly by potential danger activates the BIS, even for individuals moderately high on dominance and sensation seeking, whereas a situation characterized overwhelmingly by potential reward with little risk activates the reward-approach systems, even for individuals with powerful inhibitory tendencies. The result is what one might term system x trait x situation variation, in which the system is understood as a universal mechanism responsive to particular perceived environmental contingencies, and the trait represents individual differences in proneness to activating particular systems.

Within the perspective developed here, personality is intimately bound up with affect and motivation. Among humans this conceptualization would be compatible with a hierarchical model in Emmons's (1989) sense. Specifically, behavior related to personality occurs at several levels, based ultimately on the motivating aspects of evolved personality systems. For example, the emotions of anxiety and/or relief consequent to achieving safety are central to mechanisms designed to avoid perceived danger. Within Emmons's (1989) scheme these emotions act as motive dispositions. People are able to engage in a wide range of lower-level behaviors directed at approaching or avoiding these affective motive dispositions. Humans are thus able to deduce imminent danger by using an elaborate array of open-ended general-purpose information processing mechanisms and learning (e.g., uncovering an elaborate plot as a result of sophisticated knowledge of computers). This appraisal of imminent danger would then trigger the BIS, checking behaviors, and feelings of anxiety. Humans are also able to devise a variety of flexible open-ended strategies that are not significantly constrained by evolutionary processes in an attempt to lower their anxiety (e.g., formulating an elaborate scheme to escape imminent danger).

A hierarchical model is also applicable to behavioral approach and nurturance/love. Buss (1991) emphasized the importance of considering evolutionary goals in personality psychology, such as acquiring mates, negotiating dominance hierarchies, and forming reciprocal alliances. The goals of affiliation-intimacy and power have been central to several important recent approaches to personality [e.g., McAdams (1985)]. Variation in these systems appears to influence the extent to which people are motivated to seek out discrete types of stimulation that satisfy evolutionarily derived reward systems. Thus an individual high on nurturance/love is motivated to seek out the rewards of intimate long-term relationships with family members and close friends; such a person values these relationships highly and actively maintains such relationships. Individuals high on dominance are highly motivated to control others, whereas individuals high on sensation seeking are attracted to short-term sexual relationships (presumably motivated, at least in part, by the pleasure of sexual intercourse), risk taking, and physically dangerous situations.

The result is a hierarchical model in which the highest levels often involve affective goals subsumed by the evolved systems underlying what MacDonald (1991, 1995) termed evolved motive dispositions. Such a perspective provides a robust role for general-purpose cognitive processes (schemas, tasks, and strategies) used in attempting to achieve these goals and in evaluating situations relevant to achieving affective goals. Secondarily reinforcing rewards, such as money, would then be conceptualized as lower-level goals that can be used to facilitate the attainment of evolved motive dispositions at the highest level.

Behavioral Genetics in Evolutionary Perspective: Focus on Individual Differences

Behavioral genetics aims to identify genetic and environmental influences underlying individual differences in behavior. Ever since formal definition of the field by Fuller and Thompson (1960) in their comprehensive volume, Behavior Genetics, research activity has focused on demonstrating genetic influence across a wide range of human behaviors. Genetic influence has proven to be pervasive: In addition to explaining variation in intellectual abilities and personality traits, recent analyses have demonstrated genetic influence on social attitudes (Posner et al. 1996), religious beliefs (Waller et al. 1990), job satisfaction (Arvey et al. 1989), and divorce (McGue and Lykken 1992). One of the most significant recent developments in the field has been characterization of the environment. Beginning with Plomin and Daniels's (1987) landmark review, there has been growing evidence that the critical environmental factors relevant to the development of many intellectual and personality traits are those that are unique to individuals (idiosyncratic or nonshared environmental factors) rather than those that are shared with family members (common or shared environmental factors).

Another challenging direction for behavioral genetics has been defined by advances in molecular biological techniques that may identify DNA sequences associated with phenotypic variation (Aldous 1992; Plomin et al. 1994; Martin et al. 1997). In addition, there has been considerable progress in developing sophisticated biometric models for dissecting behavior into genetic and environmental components (Neale and Cardon 1992). Behavioral genetics is now a respected discipline, yet misconceptions linger. Perhaps most prevalent among them is the belief that genetic influence on behavior implies fixity; however, there are numerous examples to the contrary (e.g., dietary practices may reverse the adverse effects of phenylketonuria; parental encouragement may reduce shyness in shy children). Currently, the vast body of data demonstrating genetic sources of variation underlying so many human behavioral phenotypes has made it virtually impossible for researchers to easily dismiss.

Behavioral Genetics Research Designs. Twin, family, and adoption designs (and combinations of these designs) are the methods of choice for behavioral geneticists. The general focus has been estimating the proportion of trait variation associated with genetic and environmental factors [although questions of developmental mechanisms are attracting considerable attention; see, for example, Goldsmith (1993) and Plomin (1988, 1994)]. According to quantitative genetic theory, the degree of resemblance between relatives for continuous traits (traits influenced by multiple genes) should vary with the degree of genetic relatedness. For example, if genetic factors influence trait variation, full siblings, who share 50% of their genes on average, should show greater resemblance than half-siblings, who share 25% of their genes on average.

Continuous traits are also influenced by environmental factors. Biological relatives living together may display similarities or differences because of genetic or environmental influences. Environmental sources of variance can be organized into shared and nonshared components. If resemblance among family members is not detected for a given trait, this indicates that neither common genetic nor common environmental factors are influencing the development of that trait. Further discussion of these issues is covered by Plomin et al. (1990). As indicated, behavioral geneticists use twin and adoption methods to disentangle the contributions of genes and environments to behavioral variation. These methods and some relevant concepts are briefly reviewed in what follows.

Classic Twin Design. In the classic twin study greater similarity between monozygotic (MZ) cotwins than between dizygotic (DZ) cotwins is consistent with a genetic contribution to individual differences in a measured trait. Trait-relevant environmental influences (i.e., factors that affect the dependent variable in predictable ways) are assumed to be similar among MZ and DZ twin pairs, a concept known as the equal environments assumption. Critics of twin research have questioned this assumption, yet there is scant evidence of meaningful associations between similarity in treatment and outcome (Hettema et al. 1995). A number of informative variants of the classic twin design, as well as applications and challenges, are described by Segal (1990) and Bouchard and Propping (1993).

Adoption Design. Adoption studies include unrelated individuals raised together or biological relatives living apart. Behavioral resemblance between unrelated individuals living together is associated with common environmental factors. Similarity between biological relatives raised apart is associated with their common genes, given an absence of correlated trait-relevant rearing environments. Problems associated with adoption designs, such as selective placement (i.e., congruence between features of the biological and nonbiological families), are described by Plomin et al. (1997) and Turkheimer (1991).

Heritability. Heritability refers to the proportion of trait variation associated with genetic differences among people. Broad heritability includes additive and nonadditive components. Additive components result from summing effects across many different genes, whereas nonadditive components result from interactions among many different genes. It is the additive component on which selection operates and which is responsible for parent-child resemblance; this is termed narrow heritability. Heritability estimates are not fixed values, nor do they imply fixity of the trait. When members of a population experience similar environments, heritability increases; when interindividual environments are less similar, heritability decreases.

Gene-Environment Interaction. Gene-environment interaction refers to differential behavioral or physical outcomes produced by particular combinations of genes and environments. Not all genotypes show similar expression in a given environment; for example, one individual may thrive in a stimulating setting, while another may do poorly.

Gene-Environment Correlation. Gene-environment correlation refers to associations between genotypes and environments. Passive gene-environment correlation describes transmission of correlated genes and environments from parents to children. Active gene-environment correlation refers to individuals seeking out people, places, and experiences compatible with their genetically based tendencies. Reactive gene-environment correlation describes elicitation of responses from others based on individuals' own genetically based characteristics.

Behavioral Genetics and Evolutionary Theory. Several investigators have applied twin and adoption methods to specifically assess evolutionary-based hypotheses. Evolutionary theory offers behavioral genetics (and virtually all subdisciplines within the psychological field) a new theoretical framework for formulating hypotheses, designing experiments, and interpreting findings (Wilson et al. 1996; Segal 1997b). This section provides a sampling of available studies and suggestions for future analyses.

Studies based on Inclusive Fitness Theory. An important link between behavioral genetics and evolutionary theory is inclusive fitness theory (Hamilton 1964). Hamilton reasoned that natural selection favors alleles predisposing individuals to behave in ways that favor preservation of those alleles in future generations. Behaviors that enhance the reproductive fitness of close relatives, even at cost to the benefactor, facilitate representation of shared genes, a concept called inclusive fitness. Behaviors that reduce the reproductive fitness of individuals may thus be selected if those behaviors increase the fitness of relatives. This theory and the body of work that flows from it have yielded novel perspectives on play (Fagen 1987), psychiatric disorder (McGuire et al. 1994), suicide (de Catanzaro 1995), sociopathy (Mealey 1995), maternal investment (Mann 1992), resource utilization (Charlesworth 1996), and other behaviors.

Behavioral genetic methods (twin, family, and adoption designs) are well suited to tests of evolutionary-based hypotheses deriving from inclusive fitness theory. Daly and Wilson (1987) showed that stepchildren were more likely targets for abuse than biological children. Crawford and Anderson (1989) noted that studying MZ twins raised apart would highlight differences in life history strategies contingent on environmental differences in the lives of genetically identical individuals. Interestingly, the exact same design was used by Bergeman et al. (1988) in an analysis of genotype-environment interaction in personality development. Collaboration between these research teams might have yielded a more comprehensive set of hypotheses and interpretations. A promising research direction using MZ twins raised apart, based on exciting new data on birth order by Sulloway (1995, 1996), is described in what follows.

Several studies have shown that interactions between MZ and DZ twins conform to evolutionary expectations based on inclusive fitness theory. Segal (1984, 1988) tested the hypothesis that cooperative behaviors occur more frequently between close genetic relatives compared with more distant relatives. MZ twins showed greater cooperation during joint puzzle completion relative to DZ twins, as demonstrated by a higher proportion of successful puzzle completions, more equidistant placement of the puzzle between partners, and other measures.

Twin studies of bereavement by Mowrer (1954) and Woodward (1988) indicated higher levels of both anticipated and genuine grief among MZ twins than among DZ twins. Given that MZ twins look and behave more alike than DZ twins on most measured traits (Plomin 1990), some environmentally oriented researchers have viewed twin group differences as reflecting differential treatment by others (Hoffman 1991). Evolutionary reasoning suggests that the level of grief experienced by survivors may vary as a function of genetic relatedness; loss of a close relative would, for example, diminish opportunities for transmission of genes into future generations. Note that evolutionary and psychodynamic explanations are not incompatible; rather, they represent different levels of analysis. Moreover, relatedness per se is not the critical variable underlying social behavior. Genetic relatedness may be a proxy for mechanisms associated with kin recognition and with the social processes that facilitate bonding between relatives.

Times of crisis may be revealing with respect to the significance of social relationships (Chagnon and Bugos 1979). With the exception of a family study of bereavement by Littlefield and Rushton (1986), genetic relatedness and its correlates have rarely been considered as influencing level of grief in an evolutionary context. Segal and colleagues (Segal and Bouchard 1993; Segal et al. 1995) showed that surviving MZ twins experienced increased levels of grief and bereavement-related symptoms relative to surviving DZ twins. There are, in addition, other twin studies of bereavement and social relations that support these findings but that were carried out in the context of other theoretical approaches; see Segal et al. (1995) for a review.

Heritable Individual Differences and Reproductive Success. Reproductive success is of considerable theoretical interest within an evolutionary framework. Mealey and Segal (1993) integrated behavioral genetic and evolutionary approaches in a life history analysis of reproduction-related behaviors and reproductive success. Responses to selected items from a life history interview and the Briggs Life History Questionnaire were examined for MZ twin pairs raised apart and were compared between males and females. Some reproduction-related behaviors were associated in a proximal sense with heritable personality and health-related factors, although this relationship did not translate into predictable differences in number of offspring. Reproduction behaviors of males showed greater variability than those of females because of both genetic and environmental factors.

Maternal Preference. Mann (1992) studied preterm twins as a "natural experiment" for understanding differences in parental response as a function of infant health. It was of interest to determine whether mothers showed preference for the relatively healthier infant, as would be anticipated by evolutionary reasoning. Specifically, investment in healthier offspring would help to ensure reproductive success. Interestingly, maternal response patterns were associated with infant health at 8 months of age, with healthier twins being favored. The twin sample, however, was quite small (N = 7 pairs) and zygosity diagnosis was neither performed nor considered in the analyses. This was unfortunate, given that differences in mothers' behaviors toward relatively healthy or unhealthy DZ cotwins may have been confounded with genetic differences between them.

Infant Social Behavior. Plomin and Rowe (1979) failed to find significant MZ-DZ twin differences in infant social behaviors directed toward the mother but found heritable variation underlying social response to a stranger. Reasoning from the research of Bowlby (1969), Plomin and Rowe suggested that the evolutionary underpinnings of attachment were not intrinsic to the mother-child bond but rather to pressure toward wariness of novelty.

New Research Study 1: Birth Order and Personality. Following an extensive review of the psychological literature by Ernst and Angst (1983), it has been generally accepted that birth order does not significantly affect personality development. Recent provocative analyses by Sulloway (1995, 1996), undertaken with an evolutionary perspective in mind, promise to force serious rethinking of this interesting area. It was anticipated that birth order differences would be real and would fall into predictable classes. Furthermore, a series of specific predictions was developed with reference to the dimensions of the FFM. For example, first-bores enjoy favored status with parents, given their greater reproductive value, relative to younger siblings. They are expected to attempt to maintain this status and so be attentive to their parents' desires. In contrast, given their lesser identification with parental authority, younger siblings may be more open to new experiences and may use this tendency to divert resources from older siblings. Examination of scientists' support for or rejection of evolutionary theory supports this view, with later-borns showing greater acceptance of a novel perspective. In this context birth order may be conceptualized as a nonshared environmental factor affecting differential parental investment in siblings.

Studies of MZ twins raised apart can importantly extend research in this area by showing how differences in life histories may affect personality development. Interestingly, twin studies show a genetic component to openness to experience, with heritability estimates of about 0.52; mean intraclass correlations for MZ and DZ twin pairs, based on 7 studies, were 0.43 and 0.17, respectively (Bouchard 1993). Informative comparisons would involve MZ twins raised apart who assumed positions as first-born and last-born in their respective adoptive families. Assessment of intrapair differences in personality associated in predictable ways with ordinal position in the family would provide an effective test of Sulloway's conclusions.

New Research Study 2: Fluctuating Asymmetry. Recent evidence suggests that greater developmental stability is reflected by reduced fluctuating asymmetry or reduced differences between left- and right-side measures of length and breadth of limbs and other bilateral traits. Associations between fluctuating asymmetry and reproductive success are of interest, and a positive association between lower fluctuating asymmetry and higher number of sexual partners has been reported (Thornhill and Gangestad 1994). In addition, a recent study found that more attractive MZ twins showed lower fluctuating asymmetry relative to their less attractive cotwins (Mealey et al. 1997). Extending this paradigm to the personality domain might reveal subtle personality differences associated with increased reproductive success. Interestingly, associations between hand preference and neuroticism in males (Mascie-Taylor 1981), between fluctuating asymmetry and intelligence (Furlow et al. 1997), and between personality traits and fitness advantages (Eaves et al. 1990) have been reported [also see Hellige (1994)].

Papers by Buss (1995) and Scarr (1995) include excellent discussions and illustrations of applying evolutionary psychological perspectives to the study of specific human behaviors and combining evolutionary reasoning with behavioral genetic analysis. They begin with general theoretical principles of evolution, proceed to middle-level evolutionary theories, and then derive specific hypotheses and predictions. Following Buss's (1995) levels of analysis scheme in the twin studies of bereavement cited earlier, mid-level theory is provided by a genetic theory of altruism. This theory yields a specific hypothesis, namely, an expected positive association between degree of genetic relatedness and frequency of bereavement-related behaviors. The specific predictions that follow correspond to the content of items and scales included in questionnaires and surveys. Comparable analyses of personality traits should be possible.

Theory. Developing linkages among evolutionary thinking, behavioral genetics, and personality psychology has proven to be a formidable task (Loehlin 1992). The demonstration that genetic variation influences individual differences is only the beginning of an evolutionary approach. To be interesting to an evolutionist, genetic variation must be important to adaptation rather than simply adaptively neutral. It is by no means obvious that this is the case. In general, narrow heritability will decrease in the event of selection pressure for an optimum of a trait [e.g., Loehlin (1992) and Tooby and Cosmides (1990)], leading to the possibility that the remaining variation is simply nonadaptive noise. The heritabilities of personality traits, provided by studies of twins, appear to be generally similar from trait to trait, approximately 0.50. This might reflect selection for intermediate trait levels and selection against extreme values, but the remaining genetic variation may be adaptively neutral and hence without evolutionary interest.

Disagreement on this issue remains [see especially Tooby and Cosmides (1990)], yet there is reason to suppose that genetic variation can be linked with adaptive processes (Bailey 1997; Bouchard et al. 1990; Loehlin 1992; MacDonald 1991, 1995; Wilson 1994). The adaptationist perspective on individual differences in personality proposes that personality distributions be conceptualized as subsuming more than one viable adaptive strategy [Gangestad and Simpson 1990; MacDonald 1991, 1995; Wilson 1994; see also a discussion by Buss (1991)]. From this perspective personality variation is a continuous distribution of phenotypes that corresponds to a continuous distribution of viable strategies.

There are theoretical reasons to suppose that genetic variation may be linked to adaptive processes. Traits under selection in fluctuating (i.e., uncertain) environments show relatively high heritabilities (Burger et al. 1989). Furthermore, Williams (1975) provided evidence that some organisms facing uncertain environments and resource scarcity resort to sexual reproduction, a mechanism that maximizes variability, but reproduce asexually during times of resource abundance and environmental predictability. Thus these results fit well with the idea that genetic variation in personality and other valued traits serves to facilitate the production of a wide range of variation (within a delimited range) that facilitates the occupation of a wide range of possible niches in the human and nonhuman environment. There is abundant evidence that intraspecific genetic variation is associated with variation in habitat preference and that genetic variation is linked with environmental heterogeneity (Futuyma and Moreno 1988; Hedrick 1986; Wilson 1994).

These findings are consistent with the proposal that at different times different segments of a personality distribution may be favored by natural selection, resulting, for example, in separate niches for risk takers and risk avoiders. However, assuming that fitness differences within the normal range of personality variation are not dramatic, the phenotypic distribution will be approximately normal, and the fitness distribution will in effect be something of a plateau because no consistent forces of selection act to eliminate variation over the great majority of the distribution. Within this perspective, although a wide range of adaptively viable personality variation is expected, extremes on these distributions are expected to be maladaptive. This approach is thus highly consistent with attempts to conceptualize psychopathology in terms of maladaptive extremes on personality dimensions [e.g., Costa and Widiger (1994) and MacDonald (1988, 1995)].

For humans a number of processes may ensure that genetic variation remains in the population for adaptive reasons: the need to confront uncertain shifting environments with different ideal trait requirements; the existence of more than one viable role within human societies that pulls for different optimal trait standings; differing ancestral environments, resulting in different trait distributions between groups; and kin selection mechanisms, such as phenotypic matching in which people favor relatives or other similar individuals in resource transactions, including marriage (Rushton 1989a). In addition, the large number of independently assorting traits implies that no one individual is likely to embody an ideal on all the dimensions, with the result that trait variation is likely to be difficult to remove from the population even if the environment is completely unchanging.

In summary, support for an adaptationist perspective on genetic variation derives from the fact that humans are greatly interested in personality variation (as well as variation in other traits, such as IQ). An evolutionary perspective proposes not only that personality variation constitutes variation in adaptive strategies but also that humans will be greatly interested in the genetic and phenotypic diversity represented by this range of viable strategies. Phenotypic variation is seen as containing cues that influence how people evaluate each other, so that different evaluations will be made depending on the putative role of the other person in their lives. Individual differences in personality and other traits are thus conceived as a resource environment for individuals. Within this perspective personality is an adaptive landscape in which "perceiving, attending to, and acting upon differences in others is crucial for solving problems of survival and reproduction" (Buss 1991, p. 471).

Personality Variation: Predictions and Mechanisms. There is indeed evidence that people are greatly interested in personality variation. Hogan (1983) proposed that personality trait terms evaluate the potential of persons as resources to others. Individual differences in personality are thus viewed as indicators of whether individuals are suited for particular roles. Buss (1991) noted that the majority of personality trait terms are evaluative, indicating a person's potential value as a resource to others. Wicklund and Braun (1987) found that individuals are more likely to ascribe traits to others if evaluation is the goal. Finally, Borkenau (1990) showed that personality traits are judged to be similar to the extent that they are similar to common extreme values on a trait dimension. This supports the idea that in making attributions of personality, individuals are not making attributions of how closely a person conforms to a central tendency for the trait but rather are describing the individual in terms of ideal-based goal categories that represent the extreme values of the trait.

The finding that an individual's personality traits are not highly predictive of behavior is not surprising in view of the role that different situations play in triggering different evolved personality systems. Nevertheless, personality traits as summary measures of individual differences are important information in our everyday lives. Individuals are expected to develop beliefs about their own relative standing on individual difference dimensions. Each individual is expected not only to appraise the phenotypic traits of others but also to evaluate these traits differentially depending on the type of relationship entered into. Thus Graziano and Ward (1992) found that teachers perceived a stronger association between conscientiousness and school adjustment than did school counselors, a finding that presumably reflects the greater interest teachers have in this trait as a component of children's adjustment. Similarly, Lusk et al. (1993) found that ideal leaders were expected to be higher than ideal friends in scales intended to tap variation in physical attractiveness, intelligence, conscientiousness, activity, and sociability but lower than the trait profile that presumably reflects individuals' criteria for being a good leader. On the other hand, ideal friends were expected to be higher than prospective leaders in athletic ability and traits that are presumably more important for a successful friendship. Moreover, subjects expected ideal friends to be more similar to the themselves than to ideal leaders, and subjects rated themselves as more similar to prospective ideal leaders and ideal friends on categories that they themselves rated highly. Finally, DeKay and Buss [cited by Buss (1996)] found that people rated different personality traits as more desirable depending on whether they were choosing persons as long-term mates, friends, or partners in coalitions.

Further supporting an adaptationist perspective on individual differences in personality is evidence that mechanisms have evolved that appraise the resource value of this variation. For example, the phenomenon of female mate choice, originally proposed by Darwin (1871) as a mechanism of sexual selection, implies mechanisms for the discrimination of phenotypic (and ultimately genetic) variants [see Andersson (1994)]. Evolutionary approaches to sexual behavior imply evolved mechanisms that assess variation in a large number of traits. Thus Buss (1989, 1994) provided cross-cultural evidence that females prefer males who control resources and are willing to invest them in the women's children. In addition, women prefer men who are intelligent, kind, healthy, physically strong, dependable, emotionally stable, ambitious, tall, and somewhat older than themselves. Individual differences have therefore been a pervasive feature of human evolution, and the appraisal of these differences has been a critical adaptive task.

Finally, the impressive level of plasticity among humans, as indicated by heritabilities of personality traits near 0.50, implies that environmental influences are an important mechanism for producing individual phenotypic variation. This phenotypic variation allowed by plasticity has evolved within the context of genetic variation for personality. Genetic variation and phenotypic plasticity are thus likely to be two aspects of the solution to the same evolutionary problem: the difficulty of dealing with uncertain complex environments. Wcislo (1989) noted the close connection between plasticity and adaptation to environmental heterogeneity: "Behavioral adaptability is important in evolution because the activity of individuals has the potential to diminish or exacerbate the influence of external environmental heterogeneity" (p. 159).

The general theory for understanding the methods by which individuals interact with the phenotypic and genotypic resource environment represented by human diversity may be termed intraspecific diversity theory, that is, the theory of the manipulation of within-species diversity as a resource environment. The methods individuals use to manipulate this diversity range from assortment and discrimination on the basis of phenotypic similarity and genes identical by common descent to methods that evaluate this diversity solely in terms of resource potential to the individual without consideration of possible genetic or even phenotypic commonality (MacDonald 1991, 1996). To the extent that these relationships are entered into voluntarily, the fundamental principle of all these methods of dealing with diversity is reciprocity. Although considerable research remains to be performed on the methods involved, the theoretical considerations and empirical support provided here are sufficient to warrant the inclusion of these methods as supporting the general importance of individual differences in an evolutionary view of human affairs.

Behavioral Genetics and Personality: Twin Research Findings. What is the evidence that personality traits are influenced in part by genetic factors? In recent years a number of studies combining twins reared together and apart have been completed [e.g., Pedersen et al. (1988), Plomin et al. (1988), Tellegen et al. (1988), and Jang et al. (1996); also see Bouchard (1994, 1996)]. Two key findings have been described. First, in general, personality characteristics show heritable influence, with broad heritabilities (which include both additive and nonadditive genetic influences) between 0.40 and 0.50. Tellegen et al. (1988) found some evidence of differential heritability among traits measured by the scales of the Multidimensional Personality Questionnaire (MPQ), and Bouchard (1996) summarized several studies indicating a lower broad heritability for agreeableness (approximately 0.35) than the other FFM traits. Tellegen et al. (1988) also noted that heritability estimates may vary with the particular genetic model fitted to the data.

Second, studies have generally failed to find a difference in the degree of personality similarity between MZ twins reared apart and those reared together. This indicates that shared environments are not responsible for personality similarity between relatives. The only exception is the finding of some evidence for shared environmental influences on agreeableness (Loehlin 1992). The pattern of familial correlations for extroversion, including twins and adoptive relatives, compiled by Rowe (1993), also supports genetic influence. Rowe noted that the correlation for MZ twins reared together ([r.sub.i] = 0.55) exceeded the correlation for MZ twins reared apart ([r.sub.i] = 0.38), a finding that might reflect special twin environments or sampling biases, a point reiterated by Bouchard (1993). Genetic influence on extroversion and neuroticism for MZ and DZ cotwins reared together has also been reported (Macaskill et al. 1994; Heath et al. 1992).

In addition, the question of who is happy and why has been the focus of considerable research interest (Diener and Diener 1996). Responses of MZ and DZ twins to items in the well-being scale of the MPQ indicated that 44-52% of the variation was associated with genetic influence (Lykken and Tellegen 1996). Retesting conducted 4.5-10 years later showed that the heritability of the stable component of subjective well-being approached 80%. Specifically, the DZ cross-twin correlation was 0.07 and the MZ cross-twin correlation was 0.40, or 80% of the retest correlation of 0.50. Transitory variations in happiness were attributable to chance. Interestingly, a new look at human happiness from an evolutionary perspective (Barkow 1997) sheds some light on the sources of happiness and on its transitional nature. Barkow pointed out that it is likely that emotions such as happiness would have enhanced the genetic fitness of our ancestors from whom these traits are derived. However, an enduring affective state would have been maladaptive, given a range of frequently changing behaviors. Happiness (or well-being) is associated with the FFM factors of neuroticism (negatively) and extroversion (positively). It is also interesting to note that Eaves et al. (1990) reported a fitness advantage among women high in extroversion and low in neuroticism.

There is evidence for nonadditive genetic variation for extroversion, neuroticism, conscientiousness, agreeableness, and a variety of other personality traits, as indicated by low correlations between first-degree biological relatives combined with a pattern in which DZ twin correlations tend to be less than half the correlations for MZ twins (Bouchard 1994, 1996; Loehlin 1992; Loehlin and Rowe 1992; Lykken et al. 1992; Plomin et al. 1997). For example, Loehlin (1992) compared broad-sense (total additive and nonadditive) and narrow-sense (additive) heritability for the FFM traits. Excluding openness to experience, the nonadditive contribution for the other four FFM factors ranged from 0.11 to 0.17, whereas the narrow-sense (additive) heritability ranged from 0.22 to 0.32. Moreover, parent-offspring correlations for personality traits tend to be approximately 0.15 (Plomin et al. 1997), indicating that despite considerable genetic influence, children's personalities are not predictable from knowledge of parents' genotypes.

This pattern (which does not occur with intelligence data) is highly compatible with the adaptationist perspective on individual differences discussed earlier: The presence of nonadditive genetic variation, low parent-offspring correlations, and the fact that personality traits are to a considerable extent open to environmental influence (approximately half the variation in personality is due to environmental variation) are all highly compatible with the proposal that environmental heterogeneity and the niche complexity of human societies are important features of human evolution. There has not been evolution for an ideal personality type. Rather, genetic and environmental influences act to make us different from even our closest relatives. Within the normal range of personality variation there is a wide range of viable strategies: Different strokes for different folks.

The presence of nonadditive genetic variation for personality is also compatible with the emphasis here on interactive relationships among personality systems. For example, the effect of genes underlying the structures related to behavioral approach may depend for their effect on the level of opposing systems related to behavioral inhibition. The functionalist approach to personality systems implies that personality systems act to pull us in different directions depending on the particular environmental situations we are in, and it emphasizes the fact that personality systems are characterized at the physiological level by mutual inhibitory influence between systems. This situation is unlike the situation for IQ. Behavioral genetics research on IQ indicates much higher parent-offspring and sibling correlations and no evidence of nonadditive genetic variation. For IQ there is no reason to suppose that there would be opponent processes that evolved to curb the tendencies for intelligent people to make sense of their surroundings in the way that nature necessarily developed behavioral inhibition mechanisms that moderate behavioral approach mechanisms. Rather, the evidence suggests directional selection for intelligence, as indicated by data on inbreeding depression for IQ [e.g., Rushton (1989b)].

For personality systems, however, it is more likely that stabilizing selection (i.e., selection against extremes) rather than directional selection occurred. Individuals who are very high or very low on particular traits would appear to be at a disadvantage, but there is a broad range of genetic variation in the middle of the distribution underlying a range of viable strategies. Indeed, as indicated, extremes on personality distributions are associated with psychopathology. Extreme sensation seeking, for example, would tend to result in dangerous risk taking and impulsivity, whereas individuals who are very low on these traits would lack motivation to pursue goals related to the accumulation of sexual and personal resources. Extremes in either direction would appear to be maladaptive. On the other hand, the wide diversity of intermediate individuals resulting both from genetic variation and developmental plasticity would be able to occupy different social roles and have differing balances between caution and impulsivity.

In this regard it is interesting that there is no evidence for nonadditive genetic variation for factor 5, openness to experience. For example, Loehlin and Rowe (1992) found that the nonadditive component was an insignificant 0.02 and that the narrow-sense heritability was 0.43. The highly interactive opponent processes characteristic of other personality systems are unlikely to be involved for a trait like openness to experience. Openness differs from the other factors also in that there are no sex differences, and there is little evidence linking extremes on the dimension to psychopathology [MacDonald 1995; but see Costa and Widiger (1994)]. Moreover, openness is correlated with intelligence (Digman 1990), and indeed a typical bipolar scale on this dimension is "intelligent-unintelligent" [see Bouchard (1996)]. As indicated, intelligence is likely to have been under directional selection, and there is no evidence for nonadditive variation for intelligence.

Evolutionary Perspective on Environmental Influences. The results of behavioral genetics research indicate that environmental variation has considerable influence on phenotypic variation, but with the exception of agreeableness, there is no evidence that environmental variation is shared within families. Evolutionists must remain dissatisfied with leaving environmental influences at the level of a statistical term resulting from a path analysis of twin data. Within an evolutionary systems perspective the conceptualization of environmental influences on personality systems is based on an understanding of the biological systems underlying behavior. Environmental influences are conceptualized as involving specific types of stimulation directed at particular evolved systems. Thus environmental influences affecting the behavioral inhibition system would be expected to be events related to fear-inducing situations, and environmental influences related to agreeableness would be expected to be warmth and affection involved in close family relationships.

For example, if indeed the psychobiology of the human affectional system underlying agreeableness is a reward system that makes close relationships pleasurable (MacDonald 1992), it is difficult to conceive how the primary source of environmental influences in the human affectional system under normal circumstances could be other than from adult caretakers, typically family members. If the relevant environmental stimulation is that which we label warm and affectionate, this type of stimulation is unlikely to come from other sources, at least during infancy and early childhood, and it is thus not surprising that agreeableness shows evidence of shared environmental influence. When it comes to being affectionate, parental warmth is, to a significant extent, an expression of a parental personality trait and therefore likely to be a general disposition of the parent. Although some children may be easier to love because of their own level of warmth and affection and although some children may be less favored by parents because, for example, they are physically weak, all things being equal, warm parents are thus likely to treat all their children with a fairly similar level of affection, whereas cold, indifferent parents are likely to treat their children in a similar negative manner.

On the other hand, the behavioral genetic data are compatible with supposing that the environmental influences on other systems come from a variety of sources that may well include nonfamily influences as well as family influences that are not shared by siblings. We suggest that this makes sense because, unlike the situation with agreeableness, these other systems are directed at least as much to the world beyond the family as to the world inside the family. Situations involving the behavioral approach systems would often involve social dominance situations and reward-seeking and risk-taking opportunities arising outside the family. Within the family sibling-sibling competition may well influence social dominance systems but would do so in a manner that would be unshared between siblings and perhaps lead to birth order effects, as documented by Sulloway (1995, 1996). Similar comments could be made for conscientiousness/behavioral inhibition and affect intensity: Events that strengthen or weaken these systems may well occur outside the family context, such as in schools and in neighborhoods, where children are exposed to unshared environmental influences, such as frightening events that affect one sibling and not another, different experiences with peers (e.g., bullies at school), and teachers with different expectations and effectiveness in influencing conscientiousness in their students.

Future Directions

Several promising research directions are suggested by the studies reviewed here, particularly (1) testing of evolutionary-based hypotheses by means of genetically informative subject groups, (2) applying evolutionary reasoning to understand level and changes in heritability, and (3) understanding how environmental variation influences specific evolved systems. With respect to the first direction, a new kinship group that may be of increased interest consists of unrelated children of the same age reared together since infancy. These sibling pairs uniquely replicate the rearing situations of twins. Given that these siblings are the same age and have been reared in a common environment, they represent an improvement over ordinary adoptive samples. Initial study of these pairs has indicated little IQ similarity (Segal 1997a), and it will be informative to compare the degree of personality similarity and social affiliation between these pairs with that of MZ and DZ twins with reference to both behavioral genetic and evolutionary concepts and methods.

Key advances in molecular genetics have also had a significant impact on personality research. New techniques are enabling researchers to identify specific alleles associated with personality traits. Benjamin et al. (1996) reported an association between the D4 dopamine receptor genes and measures of novelty seeking. Potential contributions from this approach to other behavioral phenotypes have been summarized by Plomin et al. (1994). This approach to understanding individual differences in personality may help to unify behavioral geneticists and evolutionary psychologists.

Literature Cited

Aldous, P. 1992. The promises and pitfalls of molecular genetics. Science 257:164-165.

Andersson, M. 1994. Sexual Selection. Princeton, NJ: Princeton University Press.

Arvey, R.D., T.J. Bouchard Jr., N.L. Segal et al. 1989. Job satisfaction: Environmental and genetic components. J. Appl. Psychol. 74:187-192.

Bailey, J.M. 1997. Are genetically-based individual differences compatible with species-wide adaptations? In Uniting Psychology and Biology: Integrative Perspectives on Human Development, N.L. Segal, G.E. Weisfeld, and C.C. Weisfeld, eds. Washington, DC: APA Press, 81-100.

Barkow, J. 1997. Happiness in evolutionary perspective. In Uniting Psychology and Biology: Integrative Perspectives on Human Development, N.L. Segal, G.E. Weisfeld, and C.C. Weisfeld, eds. Washington, DC: APA Press, 397-418.

Belsky, J., L. Steinberg, and P. Draper. 1991. Childhood experience, interpersonal development, and reproductive strategy: An evolutionary theory of socialization. Child Devel. 62:647-670.

Benjamin, J., L. Li, C. Patterson et al. 1996. Population and familial association between the D4 dopamine receptor gene and measures of novelty seeking. Natur. Genet. 12:81-84.

Bergeman, C.S., R. Plomin, G.E. McClearn et al. 1988. Genotype-environment interaction in personality development: Identical twins reared apart. Psychol. Aging 3:399-406.

Borkenau, P. 1990. Traits as ideal-based and goal-derived social categories. J. Soc. Person. Psychol. 58:381-396.

Bouchard, T.J., Jr. 1993. Genetic and environmental influences on adult personality: Evaluating the evidence. In Foundations of Personality, J. Hettema and I.J. Deary, eds. Dordrecht, Netherlands: Kluwer Academic, 15-44.

Bouchard, T.J., Jr. 1994. Genes, environment, and personality. Science 264:1700-1701.

Bouchard, T.J., Jr. 1996. The genetics of personality. In Handbook of Psychoneurogenetics, K. Blum and E.P. Noble, eds. Boca Raton, FL: CRC Press, 267-290.

Bouchard, T.J., Jr., and P. Propping, eds. 1993. Twins as a Tool of Behavioral Genetics. West Sussex, England: Wiley.

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.

Bowlby, J. 1969. Attachment. New York: Basic Books.

Burger, R., G.P. Wagner, and F. Stettinger. 1989. How much heritable variation can be maintained in finite populations by mutation-selection. Evolution 43:1748-1766.

Buss, D.M. 1987. Evolutionary hypotheses and behavioral genetic methods: Hopes for a union of two disparate disciplines. Behav. Brain Sci. 10:20.

Buss, D.M. 1989. Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behav. Brain Sci. 12:1-49.

Buss, D.M. 1990. Toward a biologically informed psychology of personality. J. Person. 58:1-16.

Buss, D.M. 1991. Evolutionary personality psychology. Annu. Rev. Psychol. 42:459-491.

Buss, D.M. 1994. The Evolution of Desire. New York: Basic Books.

Buss, D.M. 1995. Evolutionary psychology: A new paradigm for psychological science. Psychol. Inquiry 6:1-30.

Buss, D.M. 1996. Social adaptation and five major factors of personality. In The Five Factor Model of Personality: Theoretical Perspectives, J.S. Wiggins, ed. New York: Guilford, 180-207.

Chagnon, N.A., and P.E. Bugos Jr. 1979. Kin selection and conflict: An analysis of a Yanomamo ax fight. In Evolutionary Biology and Human Social Behavior, N.A. Chagnon and W. Irons, eds. North Scituate, MA: Duxbury Press, 213-238.

Charlesworth, W.R. 1996. Cooperation and competition: Contributions to an evolutionary and developmental model. Int. J. Behav. Devel. 19:25-39.

Costa, P.T., and R.R. McCrae. 1992. NEO-PI-R Professional Manual. Orlando, FL: PAR Inc.

Costa, P.T., and T.A. Widiger. 1994. Summary and unresolved issues. In Personality Disorders and the Five-Factor Model of Personality, P.T. Costa and T.A. Widiger, eds. Washington, DC: American Psychological Association.

Crawford, C.B., and J.L. Anderson. 1989. Sociobiology: An environmentalist discipline? Am. Psychol. 44:1449-1459.

Daly, M., and M. Wilson. 1987. Evolutionary psychology and family violence. In Sociobiology and Psychology: Ideas, Issues, and Applications, C. Crawford, M. Smith, and D. Krebs, eds. Hillsdale, NJ: Erlbaum Associates, 293-309.

Darwin, C. 1859. On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. London, England: John Murray.

Darwin, C. 1871. The Descent of Man and Selection in Relation to Sex. London, England: John Murray.

de Catanzaro, D. 1995. Reproductive status, family interactions, and suicidal ideation: Surveys of the general public and high-risk groups. Ethol. Sociobiol. 16:385-394.

Diener, E., and C. Diener. 1996. Most people are happy. Psychol. Sci. 7:181-185.

Digman, J.M. 1990. Personality structure: Emergence of the five-factor model. Annu. Rev. Psychol. 41:417-440.

Eaves, L.J., N.G. Martin, J.K. Hewitt et al. 1990. Personality and reproductive fitness. Behav. Genet. 20:563-568.

Eibl-Eibesfeldt, I. 1989. Human Ethology. New York: Aldine de Gruyter.

Emmons, R.A. 1989. The personal striving approach to personality. In Goal Concepts in Personality and Social Psychology, L.A. Pervin, ed. Hillsdale, NJ: Erlbaum Associates, 87-126.

Ernst, C., and J. Angst. 1983. Birth Order: Its Influence on Personality. New York: Springer-Verlag.

Eysenck, H.J. 1982. The Scientific Study of Personality. Westport, CT: Greenwood.

Eysenck, H.J., and M.W. Eysenck. 1985. Personality and Individual Differences: A Natural Science Approach. New York: Plenum.

Fagen, R. 1987. Play, games, and innovation: Sociobiological findings and unanswered questions. In Sociobiology and Psychology: Ideas, Issues, and Applications, C. Crawford, M. Smith, and D. Krebs, eds. Hillsdale, NY: Erlbaum Associates, 254-268.

Figueredo, A.J., and J.E. King. 1996. The evolution of individual differences in behavior. West. Comp. Psychol. Assoc. Observer 2(2): 1-4.

Freedman, D.G. 1968. An evolutionary framework for behavioral research. In Progress in Human Behavior Genetics, S.G. Vandenberg, ed. Baltimore, MD: Johns Hopkins University Press, 1-6.

Fuller, J.L., and W.R. Thompson. 1960. Behavior Genetics. New York: Wiley.

Furlow, F.B., T. Armijo-Prewitt, S.W. Gangestad et al. 1997. Fluctuating asymmetry and psychometric intelligence. Proc. R. Soc. Lond, ser. B, 264:823-829.

Futuyma, D.J., and G. Moreno. 1988. The evolution of ecological specialization. Annu. Rev. Ecol. Syst. 20:207-233.

Gangestad, S.W., and J.A. Simpson. 1990. Toward an evolutionary history of female sociosexual variation. J. Person. 58:69-96.

Goldberg, L.R. 1981. Language and individual differences: The search for universals in personality lexicons. In Review of Personality and Social Psychology, L. Wheeler, ed. Beverly Hills, CA: Sage Publications, v. 2, 141-165.

Goldberg, L.R. 1992. The development of markers for the Big Five factor structure. Psychol. Assess. 4:26-42.

Goldsmith, H.H. 1993. Nature-nurture issues in the behavioral genetics context: Overcoming barriers to communication. In Nature, Nurture, and Psychology, R. Plomin and G.E. McClearn, eds. Washington, DC: APA Press, 325-339.

Gray, J.A. 1982. The Neuropsychology of Anxiety. New York: Oxford University Press.

Gray, J.A. 1987. The Psychology of Fear and Stress. Cambridge, England: Cambridge University Press.

Graziano, W.G, and D. Ward. 1992. Probing the Big Five in adolescence: Personality and adjustment during a developmental transition. J. Person. 60:425-439.

Hamilton, W.D. 1964. The genetical evolution of human behaviour. J. Theor. Biol. 7:1-52.

Heath, A.C., M.C. Neale, R.C. Kessler et al. 1992. Evidence for genetic influences on personality from self-reports and informant ratings. J. Person. Soc. Psychol. 63:85-96.

Hedrick, P.W. 1986. Genetic polymorphism in heterogeneous environments. Annu. Rev. Ecol. Syst. 17:535-566.

Hellige, J.B. 1994. Handedness. In Encyclopedia of Human Behavior, V.S. Ramachandran, ed. New York: Academic Press, v. 2, 491-500.

Hettema, J.M., M.C. Neale, and K.S. Kendler. 1995. Physical similarity and the equal-environment assumption in twin studies of psychiatric disorders. Behav. Genet. 25:327-335.

Hoffman, L.W. 1991. The influence of the family environment on personality: Accounting for sibling differences. Psychol. Bull. 110:187-203.

Hogan, R. 1983. A socioanalytic theory of personality. In Nebraska Symposium on Motivation, 1982, R. Dienstbier, ed. Lincoln, NE: University of Nebraska Press, 55-89.

Horn, J.M., R. Plomin, and R. Rosenman. 1976. Heritability of personality traits in adult male twins. Behav. Genet. 6:17-30.

Jang, K.L., W.J. Livesley, and P.A. Vernon. 1996. Heritability of the big five dimensions and their facets: A twin study. J. Person. 64:577-591.

King, J.F., and A.J. Figueredo. 1994. Human personality factors in zoo chimpanzees? Paper presented at the Western Psychological Association Convention, Kona, Hawaii.

Littlefield, C.H., and J.P. Rushton. 1986. When a child dies: The sociobiology of bereavement. J. Person. Soc. Psychol. 51:797-802.

Loehlin, J.C. 1992. Genes and Environment in Personality Development, v. 2. Newbury Park, CA: Sage Publications.

Loehlin, J., and R. Nichols. 1976. Heredity, Environment, and Personality: A Study of 850 Sets of Twins. Austin, TX: University of Texas Press.

Loehlin, J., and D. Rowe. 1992. Genes, environment, and personality. In Modern Personality Psychology: Critical Reviews and New Directions, G. Capara and G.L. Van Heck, eds. London, England: Harvester Wheatsheaf, 352-370.

Luria, A.R. 1980. The Higher Cortical Functions in Man. New York: Basic Books.

Lusk, J., K. MacDonald, and J.R. Newman. 1993. Resource appraisals among self, friend, and leader: Toward an evolutionary perspective on personality and individual differences. Paper presented at the Human Behavior and Evolution Society meetings, Binghamton, New York.

Lykken, D., and A. Tellegen. 1996. Happiness is a stochastic phenomenon. Psychol. Sci. 7:186-189.

Lykken, D.T., M. McGue, A. Tellegen et al. 1992. Emergenesis: Genetic traits that may not run in families. Am. Psychol. 47:1565-1577.

Macaskill, G.T., J.L. Hopper, V. White et al. 1994. Genetic and environmental variation in Eysenck personality questionnaire scales measured on Australian adolescent twins. Behav. Genet. 24:481-491.

MacDonald, K.B. 1983. Development and stability of personality characteristics in prepubertal wolves. J. Comp. Psychol. 97:99-106.

MacDonald, K.B. 1988. Social and Personality Development: An Evolutionary Synthesis. New York: Plenum Press.

MacDonald, K.B. 1991. A perspective on Darwinian psychology: Domain-general mechanisms, plasticity, and individual differences. Ethol. Sociobiol. 12:449-480.

MacDonald, K.B. 1992. Warmth as a developmental construct: An evolutionary analysis. Child Devel. 63:753-773.

MacDonald, K. 1995. Evolution, the five-factor model, and levels of personality. J. Person. 63:525-567.

MacDonald, K.B. 1996. What do children want? An evolutionary perspective on children's motivation in the peer group. Int. J. Behav. Devel. 19:53-73.

Mann, J. 1992. Nurturance or negligence: Maternal psychology and behavioral preference among preterm twins. In The Adapted Mind: Evolutionary Psychology and the Evolution of Culture, J.H. Barkow, L. Cosmides, and J. Tooby, eds. New York: Oxford University Press, 367-390.

Martin, N., D. Boomsma, and G. Machin. 1997. A twin-pronged attack on complex traits. Nat. Genet. 17:387-392.

Mascie-Taylor, C.G.N. 1981. Hand preference and personality traits. Cortex 17:319-322.

McAdams, D.P. 1985. Power, Intimacy, and the Life Story: Personological Inquiries into Identity. Homewood, IL: Dorsey Press.

McGue, M., and D.T. Lykken. 1992. Genetic influence on risk of divorce. Psychol. Sci. 3:368-373.

McGuire, M.T., F.I. Fawzy, J.E. Spar et al. 1994. Altruism and mental disorders. Ethol. Sociobiol. 15:299-321.

Mealey, L. 1995. The sociobiology of sociopathy: An integrated evolutionary model. Behav. Brain Sci. 18:523-541.

Mealey, L., and N.L. Segal. 1993. Heritable and environmental variables affect reproduction-related behaviors but not ultimate reproductive success. Person. Indiv. Diff. 14:783-794.

Mealey, L, R. Bridgstock, and G.C. Townsend. 1997. Monozygotic co-twin comparison isolates role of symmetry in physical attractiveness. Paper presented at the Ninth Annual Meeting of the Human Behavior and Evolution Society, Tuscon, Arizona.

Mesulam, M.M. 1986. Frontal cortex and behavior. Ann. Neurol. 19:320-325.

Mowrer, E. 1954. Some factors in the affectional adjustment of twins. Am. Sociol. Rev. 19:468-471.

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

Pedersen, N.L., R. Plomin, G.E. McClearn et al. 1988. Neuroticism, extroversion, and related traits in adult twins reared apart and reared together. J. Person. Soc. Psychol. 55:950-957.

Plomin, R. 1988. Reply to Susan Oyama's review of Development, Genetics, and Psychology. Devel. Psychobiol. 21:107-112.

Plomin, R. 1990. The role of inheritance in behavior. Science 248:183-188.

Plomin, R. 1994. The Emanuel Miller Memorial Lecture 1993: Genetic research and identification of environmental influences. J. Child Psychol. Psychiatr. 35:817-834.

Plomin, R, and D. Daniels. 1987. Why are children in the same family so different from one another? Behav. Brain Sci. 10:1-16.

Plomin, R., and D.C. Rowe. 1979. Genetic and environmental etiology of social behavior in infancy. Devel. Psychol. 15:62-72.

Plomin, R., J.C. DeFries, and G.E. McClearn. 1990. Behavioral Genetics: A Primer. New York: W.H. Freeman.

Plomin, R., M.J. Owen, and P. McGuffin. 1994. The genetic basis of complex human behaviors. Science 264:1733-1739.

Plomin, R., J.C. DeFries, G.E. McClearn et al. 1997. Behavioral Genetics, 3d ed. New York: W.H. Freeman.

Plomin, R., N.L. Pedersen, G.E. McClearn et al. 1988. EAS temperaments during the last half of the life span: Twins reared apart and twins reared together. Psychol. Aging 3:43-50.

Posner, S.F., L. Baker, A. Heath et al. 1996. Social contact, social attitudes, and twin similarity. Behav. Genet. 26:123-133.

Rowe, D.C. 1993. Genetic perspectives on personality. In Nature, Nurture, and Psychology, R. Plomin and G.E. McClearn, eds. Washington, DC: APA Press, 179-195.

Rushton, J.P. 1989a. Genetic similarity, group selection, and altruism. Behav. Brain Sci. 12:503-559.

Rushton, J.P. 1989b. Japanese inbreeding depression scores: Predictors of cognitive differences between blacks and whites. Intelligence 13:43-51.

Scarr, S. 1995. Psychology will be truly evolutionary when behavior genetics is included. Psychol. Inquiry 6:68-71.

Segal, N.L. 1984. Cooperation, competition, and altruism within twin sets: A reappraisal. Ethol. Sociobiol. 5:163-177.

Segal, N.L. 1988. Cooperation, competition, and altruism in human twinships: A sociobiological approach. In Sociobiological Perspectives on Human Development, K.B. MacDonald, ed. New York: Springer-Verlag, 168-206.

Segal, N.L. 1990. The importance of twin studies for individual differences research. J. Counsel. Devel. 68:612-622.

Segal, N.L. 1993. Twin, sibling, and adoption methods: Tests of evolutionary hypotheses. Am. Psychol. 48:943-956.

Segal, N.L. 1997a. Same-age unrelated siblings: A unique test of within-family environmental influences on IQ similarity. J. Educ. Psychol. 89:381-390.

Segal, N.L. 1997b. Twin research perspective on human development. In Uniting Psychology and Biology: Integrative Perspectives on Human Development. N.L. Segal, G.E. Weisfeld, and C.C. Weisfeld, eds. Washington, DC: APA Press, 145-173.

Segal, N.L, and T.J. Bouchard Jr. 1993. Grief intensity following the loss of a twin and other close relatives: Test of kinship-genetic hypotheses. Hum. Biol. 65:87-105.

Segal, N.L., S.M. Wilson, T.J. Bouchard Jr. et al. 1995. Comparative grief experiences of bereaved twins and other bereaved relatives. Person. Indiv. Diff. 18:511-524.

Sulloway, F.J. 1995. Birth order and evolutionary psychology: A recta-analytic overview. Psychol. Inquiry 6:75-80.

Sulloway, F. 1996. Born to Rebel: Birth Order, Family Dynamics, and Creative Lives. New York: Pantheon Books.

Tellegen, A., D.T. Lykken, T.J. Bouchard Jr. et al. 1988. Personality similarity in twins reared apart and together. J. Person. Soc, Psychol, 54:1031-1039.

Thornhill, R., and S.W. Gangestad. 1994. Human fluctuating asymmetry and sexual behavior. Psychol. Sci. 5:297-302.

Tooby, J., and L. Cosmides. 1990. On the universality of human nature and the uniqueness of the individual: The role of genetics and adaptation. J. Person. 58:17-68.

Tucker, D.M., and D. Derryberry. 1992. Motivated attention: Anxiety and the frontal executive functions. Neuropsychiatr. Neuropsychol. Behav. Neurol. 5:233-252.

Turkheimer, E. 1991. Individual and group differences in adoption studies of IQ. Psychol. Bull. 110:392-405.

Waller, N.G., B.A. Kojetin, T.J., Bouchard Jr, et al. 1990. Genetic and environmental influences on religious interests, attitudes, and values: A study of twins reared apart and together. Psychol. Sci. 1:138-142.

Wcislo, W.T. 1989. Behavioral environments and evolutionary change. Annu. Rev. Ecol. Syst. 20:137-169.

Wicklund, R.A., and O.L. Braun. 1987. A psychological taxonomy of trait-descriptive terms: The interpersonal domain. J. Person. Soc. Psychol 37:395-412.

Williams, G.C. 1975. Sex and Evolution. Princeton, NJ: Princeton University Press.

Wilson, D.S. 1994. Adaptive genetic variation and human evolutionary psychology. Ethol. Sociobiol. 15:219-235.

Wilson, D.S., D. Near, and R. Miller. 1996. Machiavellianism: A synthesis of the evolutionary and psychological literatures. Psychol. Bull. 119:285-299.

Wilson, E.O. 1994. Naturalist. New York: Warren Books.

Woodward, J. 1988. The bereaved twin. Acta Genet. Med. Gemellol. 37:173-180.

Zuckerman, M. 1979. Sensation-Seeking: Beyond the Optimal Level of Arousal. Hillsdale, NJ: Erlbaum Associates.

Zuckerman, M. 1984. Sensation-seeking: A comparative approach to a human trait. Behav. Brain Sci. 7:413-471.

Zuckerman, M. 1990. The psychophysiology of sensation-seeking. J. Person. 58:313-345.

Zuckerman, M. 1991. Psychobiology of Personality, Cambridge, England: Cambridge University Press.

Karen Kuehn; George Howe Colt; Anne Hollister.

Abstract: Advances in molecular biology, genetic research and reproductive technology lead to important and controversial answers as to whether nature is a better predictor of behavior than nurture. Studies of identical twins raised in different families have produced remarkable results on the side of nature.

In the debate over the relative power of nature and nurture, there may be no more devout believers in nurture than new parents. As my wife and I, suffused with a potent mix of awe, exhaustion and ego, gazed down at our newborn daughter in the hospital, it was hard not to feel like miniature gods with a squirming lump of figurative putty in our hands. We had long believed that people could make the world a better place, and now we firmly believed that we could make this a better baby. At home our bedside tables were swaybacked by towers of well-thumbed parenting manuals. A black-and-white Stim-Mobile, designed to sharpen visual acuity, hung over the crib. The shelves were lined with books, educational puzzles and IQ-boosting rattles. Down the line we envisioned museum visits, art lessons, ballet. And if someone had tapped us on the shoulder and told us that none of this would matter--that in fact we could switch babies in the nursery and send our precious darling home with any other new parents in the hospital, and as long as those parents weren't penniless, violent or drug addicted, our daughter would turn out pretty much the same... well, we would have thwacked that someone with a Stim-Mobile.

Does the key to who we are lie in our genes or in our family, friends and experiences? In one of the most bitter scientific controversies of the 20th century--the battle over nature and nurture--a wealth of new research has tipped the scales overwhelmingly toward nature. Studies of twins and advances in molecular biology have uncovered a more significant genetic component to personality than was previously known. Far from a piece of putty, say biologists, my daughter is more like a computer's motherboard, her basic personality hardwired into infinitesimal squiggles of DNA. As parents, we would have no more influence on some aspects of her behavior than we had on the color of her hair. And yet new findings are also shedding light on how heredity and environment interact. Psychiatrists are using these findings to help patients overcome their genetic predispositions. Meanwhile, advances in genetic research and reproductive technology are leading us to the brink of some extraordinary--and terrifying--possibilities.

The moment the scales began to tip can be traced to a 1979 meeting between a steelworker named Jim Lewis and a clerical worker named Jim Springer. Identical twins separated five weeks after birth, they were raised by families 80 miles apart in Ohio. Reunited 39 years later, they would have strained the credulity of the editors of Ripley's Believe It or Not. Not only did both have dark hair, stand six feet tall and weigh 180 pounds, but they spoke with the same inflections, moved with the same gait and made the same gestures. Both loved stock car racing and hated baseball. Both married women named Linda, divorced them and married women named Betty. Both drove Chevrolets, drank Miller Lite, chain-smoked Salems and vacationed on the same half-mile stretch of Florida beach. Both had elevated blood pressure, severe migraines and had undergone vasectomies. Both bit their nails. Their heart rates, brain waves and IQs were nearly identical. Their scores on personality tests were as close as if one person had taken the same test twice.

Identical twins raised in different families are a built-in research lab for measuring the relative contributions of nature and nurture. The Jims became one of 7,000 sets of twins studied by the Minnesota Center for Twin and Adoption Research, one of half a dozen such centers in this country. Using psychological and physiological tests to compare the relative similarity of identical and fraternal twins, these centers calculate the "heritability" of behavioral traits--the degree to which a trait in a given population is attributable to genes rather than to the environment. They have found, for instance, that "assertiveness" is 60 percent heritable, while "the ability to be enthralled by an aesthetic experience" is 55 percent heritable.

Studies of twins have produced an impressive list of attributes or behaviors that appear to owe at least as much to heredity as to environment. It includes alienation, extroversion, traditionalism, leadership, career choice, risk aversion, attention deficit disorder, religious conviction and vulnerability to stress. One study even concluded that happiness is 80 percent heritable--it depends little on wealth, achievement or marital status. Another study found that while both optimism and pessimism are heavily influenced by genes, environment affects optimism but not pessimism. A third study claimed a genetic influence for the consumption of coffee but not, it seems, of tea. Critics accuse researchers of confusing correlation with causation, yet they admit the data suggest a strong genetic influence on behavior. Far less clear is how it all works. Is there a gene for becoming an astronaut? For enjoying symphonies?

Molecular biologists around the world are trying to answer such questions, searching for specific bits of DNA that may contribute to particular behaviors. In a small, windowless laboratory cluttered with bottles of chemicals, back issues of scientific journals and bags of sterile rubber gloves, there sits an aged Sears Kenmore Coldspot refrigerator. Inside are 21 plastic trays labeled with Magic Marker: College Students. Gay Men. Smokers. Shy Kids. Each tray contains 96 almond-size plastic vials. Each vial contains a smidgen of DNA. These are Dean Hamer's study subjects, his "people," as he refers to them. The refrigerator holds the blueprints for nearly 2,000 people, a database that Hamer, chief of Gene Structure and Regulation at the National Cancer Institute in Bethesda, Md., hopes will help him find the keys to why we smoke, why we get anxious, why we take risks. In what he describes as "a giant fishing expedition," Hamer is working his way through the human genome, tracking down any variation that may affect personality.

Even for someone like Hamer, who admits to genetic propensities for both optimism and risk-taking behavior, it is a daunting prospect. The human body has 100 trillion cells, each equipped with a complete set of DNA distributed among 23 pairs of chromosomes. (DNA is microscopic yet sizable: If set out in a continuous strand, the DNA from a single cell would be six feet long.) Each cell's DNA is made up of some three billion nucleic components. Most of these seem to be nonfunctioning--"junk" DNA, biologists call them--but about 3 percent are working genes. The total number of working genes is believed to be 80,000, give or take 20,000. The task: to pinpoint the one-in-three-billion bit that might contribute to a particular behavior.

Using DNA from groups of people with a high incidence of a certain trait, Hamer's lab has scanned hundreds of thousands of amplified strands of DNA, hoping to come across a variation common to those with the trait but absent in those without it. In 1993 his lab isolated a stretch of genes on the X chromosome that may be linked to male homosexuality. Three years later a gene on chromosome 11 was found to be consistently longer in people with a taste for novelty-seeking. Last year his lab linked anxiety to a gene involved in regulating levels of serotonin, a brain chemical affected by the antidepressant Prozac.

The hoopla with which these discoveries have been greeted--"GAY GENE!" the headlines blared--has obscured the fact that other institutions have had mixed results when trying to replicate the findings. It has also made it seem as if single genes dictate specific behaviors. The reality is more complicated. Genes don't make men gay or children timid. They make proteins, which kindle complex neurological events. Biologists now believe that any given trait is shaped by a constellation of different genes. "From twin studies, we know that anxiety is 40 to 50 percent genetic," explains Hamer. "And from our data we know that the gene we isolated accounts for about 5 percent of the effect. We think there may be ten genes altogether that influence anxiety. But there may be a hundred or a thousand." In any case, he says, different people can have different combinations of those genes. People with just a few of those anxiety genes might feel nervous when they have to give a speech. Those with a few more might cringe when the phone rings. And those with a full complement might be so timid they rarely leave the house.

If, as twin studies suggest, the heritability of most personality traits is about 50 percent, that still leaves 50 percent to the environment--an environment, say behavioral geneticists, whose influence works far differently from what we once thought. Until recently the family was assumed to be the crucible in which personality was formed. In fact, children may shape parents' behavior as much as parents shape theirs. "If you are genetically a responsive, happy infant, you are going to get different mothering than if you are an irritable or rejecting child," says University of Minnesota psychologist David Lykken. The older a child gets, the more power he has to mold his own environment. "People seek out experiences and environments," Lykken says, "that are compatible with their genetic nature."

Studying adolescents adopted in infancy, University of Virginia psychologist Sandra Scarr was surprised to find that children adopted by well-educated, professional parents performed no better in school or on intelligence tests than children who had been adopted into working-class homes. "Providing children with super environments--private schooling, museum visits, lessons and so on--made no difference in their intelligence, adjustment or personality development," says Scarr. She concludes that if a child has "good enough parenting"--parents who aren't abusive or neglectful and provide a basic level of support--one set of parents is as good as another. The child will develop along paths set out by his genes. "It doesn't matter whether you take the kids fishing or to a Mozart concert," says Scarr. "As long as you do it with love, almost anything you do is going to be fine and functionally equivalent."

But don't throw out those Spock and Brazelton manuals. Even the most zealous behavioral geneticists admit that genes are not--quite--destiny. "Depending on the other genes you inherit, and on your biology and on your in utero experience, the genes will have full force or less force," explains Harvard psychologist Jerome Kagan. Upbringing and circumstance may steer someone born with a predisposition for shyness to grow up into an agoraphobe--or a great poet. Someone with a propensity for aggression might become an Adolf Hitler, but he might become a General Patton.

In any case, if genes are not commands but nudges, we can nudge back. We are the only animals on earth that can overrule our genes. And we do so constantly--whenever an alcoholic chooses not to drink or an obese person diets.

How important is it to understand our genetic makeup? Does it matter that our anxiety can be traced to a snippet of nucleic matter and not to the time Mommy spanked us for spilling our juice? Psychologist Thomas Bouchard, director of the Minnesota twin study, believes it does: "A lot of books say you can do anything you want, but we have real doubts about that. It's not that you can't, but we suspect it's done at a cost." He suggests that we not push kids in directions they're not inclined toward. "The job of a parent," says Bouchard, "is to look for a kid's natural talents and then provide the best possible environment for them."

Bethesda psychiatrist Stanley Greenspan is one of a growing number of therapists who have incorporated the findings of behavioral genetics into their practice. "When a trait appears to be influenced by genes, people assume it's not changeable," he says. "Well, we can't change the genes, but we can change the way genes express themselves. We can change behavior." Greenspan works with children and their parents to rechannel a child's genetic propensities. For a sensitive girl, so fearful of new sights and sounds that at age three she still clung to her mother for dear life, Greenspan prescribed rhythmic rocking, as well as extra doses of imaginative play. The rocking soothed the child; the play helped her to gradually become more assertive. For an aggressive girl who pushed, punched and bit her classmates--"she so craved sensory input that she literally attacked her world," says Greenspan--he designed games of dancing, shouting and beating on drums, but part of the exercise was to gradually go from fast and loud to slow and soft. "We gave her socially appropriate ways to satisfy her needs, but we taught her how to learn control."

Greenspan's work illustrates an idea at the heart of behavioral genetics today--that heredity and environment are entwined, always reacting to and building on each other. "It's not a horse race between nurture and nature," he says. "It's a dance."

By the year 2005, scientists are expected to have mapped the entire sequence of the human genome. It will be many years before they know the functions of those 80,000 genes, but ways to take advantage of this information are already being developed. Within a few decades, people who feel ill will go to physician-geneticists who will run DNA scans to check the relevant genes, make pinpoint diagnoses and prescribe drugs targeted to precise genetic needs. This will be true for depression, phobias and life-threatening obesity, as well as for less crippling traits. Just as Mary Poppins had a magic bottle from which she dispensed spoonfuls of strawberry-flavored liquid to cure Michael's fussiness, parents may supply a pill to embolden their shy child before the school dance.

Before my wife and I had our daughter, genetic counselors were able to tell us whether she had the genes for Down syndrome or Tay-Sachs disease. By the time she is ready to be a mother, genetic counselors will be able to tell her whether her fetus is genetically inclined toward depression or addiction. Such knowledge will surely lead to an ethical morass. "Where does it stop?" asks a character in The Twilight of the Golds, a recent play in which a couple decide to abort a fetus whose genes suggest it will be gay.

"What if you found out the kid was going to be ugly, or smell bad, or have an annoying laugh, or need really thick glasses?" (Not such a far-fetched question, given that three quarters of young couples in a recent survey said they would choose abortion if told their fetus had a 50 percent chance of growing up to be obese.) The morass will become still stickier when we have the technology to tinker with the genes themselves. Clinical trials are already under way using gene therapy--the introduction of healthy new genes to counteract a mutated or missing gene--to repair disorders such as cystic fibrosis, cancer and AIDS. Most of us would welcome treatment that might eliminate these afflictions. But what about depression? Aggression? Timidity?

By the time my daughter's grandchild is ready to give birth, prospective parents may design their children at the computer, scrolling through genetic menus to pick and choose, from their own DNA pools, specific gene clusters for height, weight and eye color, as well as for assertiveness, extroversion, happiness and so on. "The question is not whether the science will happen--it will," says Princeton molecular geneticist Lee Silver. "The question is, will people use it?"

Have we ever been able to restrain ourselves? The first person to study twins, 19th century anthropologist Francis Galton, finding that "nature prevails enormously over nurture," recommended breeding quotas to weed out the "unfit." The eugenics movement gathered force--from 1907 to 1965, some 60,000 people were sterilized in the U.S. for such conditions as pauperism and "feeblemindedness"--and led to the extermination programs of the Third Reich, a horror that shadows the nature-nurture debate today. Critics of behavioral genetics say the risk of misuse should preclude further research. But that, journalist William Wright argues, "makes as much sense as rejecting electricity because of daytime television."

Weighed against the potential benefits--might we end war by getting rid of aggressive genes?--is a Pandora's box of misuse. After the discovery of the so-called gay gene, for example, religious fundamentalists called for techniques to "correct that genetic defect." Caution is needed. "Do we know enough to know what we are changing?" asks Ronald Green, director of the Ethics Institute at Dartmouth. "Are we going to be wise enough to do it well, in such a way that we don't impoverish the future? In trying to avoid a Ted Kaczynski, might we destroy an Einstein?"

A few nights ago, watching my daughter arrange her 37 Beanie Babies by color and species, I felt a shock of recognition--and glanced over at my wife, who wears the same expression when she arranges Shakespeare's plays in chronological order. My lump of putty is eight now, and I don't need a DNA scan to tell me she has inherited her mother's intelligence, her father's stubbornness, her grandfather's wit. The genes may be familiar, but the mix--thank heavens--is unique. Warts and all, she is exactly the child I want.

When I look at her, I see a part of me. When I look at myself, it seems there's less of me than there once was. At a recent party, schmoozing with one last guest on my way out the door, I suddenly thought, I'm acting exactly like my father! Having spent my youth fighting to forge my own identity, I find, increasingly, that I resemble the very parent against whom I worked so hard to rebel: his social ease, his sense of humor--and, now that I am in my forties, his thinning hair and slight potbelly. Indeed, as I get older, I feel that instead of adding layers, I am shedding skins. In becoming more like my parents, I am becoming more myself. I am surprised but delighted that it all feels so comfortable--not an imprisoning but a homecoming.

Related article:

GENES AND VIOLENCE

No genetic link to criminality--other than being born male--has been proved. But that hasn't stopped people from making a connection. "The place to fight crime is in the cradle," says psychologist David Lykken, who has a controversial proposal: that biological parents be licensed. Lykken believes that a lot of crime is due to genetic predispositions for aggression and impulsiveness combined with incompetent parenting and the breakdown of the nuclear family. "We wouldn't let a crack addict, a teenager or a criminal adopt a child," he says. "Why not make the same minimal requirements for people having children biologically?"

One Minnesota state representative is trying to write a version of Lykken's views into law, while detractors have called Lykken a fascist. Indeed, when it comes to the subject of violence, behavioral genetics is particularly prickly. In 1992 a lawyer tried to stage a conference on genetics and crime, but civil rights groups forced its postponement. When it was finally held three years later, the symposium was disrupted by protesters, and a handful of attendees signed a statement labeling the proceedings "racist pseudoscience."

Critics say that linking genes and violence is blaming the victims and shifting the focus away from the real culprits: poverty, racism and unemployment. Brain research has shown that violent males tend to have low levels of the chemical serotonin, levels associated with depression, aggression and impulsivity--all traits with high heritability. But adoption studies show that children whose biological parents had trouble with the law have a far greater likelihood of having similar problems if their adoptive family had those problems too. Biology may contribute to antisocial behavior, the studies suggest, but environment helps tip the balance. In the same way, crime may be more pervasive in inner cities, not because of the genes of the people who live there but because inner cities tend to be fragmented, impoverished and racially polarized environments.

Neurobiologist Evan Balaban sees Lykken's proposal as a throwback to the early 1900s, when 15 states had laws permitting the sterilization of criminals. "The predominantly academic people making these suggestions seem to be ignorant of what attempts have been made to solve these problems by people on the firing line. It might behoove them to put some effort into learning what the real issues are."

FOR FURTHER READING

Born That Way by William Wright. Knopf, 1998.

Galen's Prophecy by Jerome Kagan. Basic Books, 1994.

The Growth of the Mind by Stanley Greenspan. Addison-Wesley, 1997.

Living With Our Genes by Dean Hamer and Peter Copeland. Doubleday, 1998.

Remaking Eden by Lee Silver. Avon, 1997.

Twins by Lawrence Wright. John Wiley & Sons, 1997. Article A20384293

Pierre Darlu; Jennifer Curtiss Gage.

Abstract: Gene mechanisms alone are insufficient to express or explain all the complexities of human behavior. Although they can account for a small portion of certain behaviors in specific situations, many other variables have impacts that can alter gene expression in areas ranging from personality to cognitive aptitudes. Genes therefore are not rigid determinisms.

From its very beginnings, this century has been under the sign of genetics. Indeed, it was in 1900 that the laws established by Mendel in the mid-nineteenth century were rediscovered. In that same year, Landsteiner identified the first human blood typing, the ABO system. At that time, agronomists, eugenicists, and physicians were the principal agents of the development of genetics. The chromosome theory of heredity was asserted beginning in 1911; it was followed in the 1940s by the understanding of the role of the gene in cellular metabolism, and in 1954 by the explanation of the structure of the DNA double helix. The pre-eminence of genetics was to increase through the second half of the century, particularly in the 1970s with the advent of genetic engineering.

In our day, the science of genetics reigns triumphant. It has extended its domain over vast sectors of economic and social activity, and it monopolizes the attention of the medical community. It stirs up political debates and it transforms and fuels many a philosophical viewpoint. If it has made possible advances in food production and the eradication or containment of several rare diseases, it has also guided several genocides and contributed to imposing a determinist and biologically oriented vision of intellectual activities and behaviors, to the point of modifying the nature of human relations at times.(1) Like many human enterprises, genetics thus exhibits the double face of Doctor Jekyll and Mr. Hyde.

Scientists in the second half of the 1970s were already issuing warnings against the "rush to treat social and political questions biologically" and against "the ascendancy of biological representation in the practice and apparatus of social control."(2) Recent progress in genetics only increases this need to reflect on the place of science in our society, in particular by fundamentally questioning its neutrality and objectivity. In our day, these reflections most often come down to ethical questions, including quite radical ones such as the following: is "progress" inexorable?(3) Are some areas of research off limits?

In order to put the current triumphalism of genetics (particularly human genetics) in its place, it is useful to cast a look backwards on some lessons that history has taught us. One of these examples is supplied by Provine, who analyzed the way in which supposedly "objective and neutral" scientific knowledge on the subject of "race crossing" or "interracial" mixing, has in reality varied with the historical evolution of various ideological and cultural positions on the matter.(4) Beginning with Sir Francis Galton at the end of the nineteenth century, and through the beginning of the twentieth century, numerous observations and experiments, performed principally but not exclusively on animals, led geneticists and anthropologists to the more or less categorical conclusion that an interracial mix could only produce undesirable results, coyly labeled "disharmonious." Certain agnostic scientists, such as Pearson around 1930, and later Huxley and Haldane, criticized such positions, but they confined themselves to pointing out the methodological or statistical insufficiency of the studies that led to such conclusions. At the dawn of World War II, there was no record of decisive progress that might effectively cast doubt upon the previous conclusions by consensus.

However, between 1938 and 1946, the scientific community underwent a veritable ideological awakening. This electroshock was clearly provoked by the excesses of the Nazi regime, which legitimized its repression on the basis of this old scientific consensus on the "disharmony" of interracial mixing. This belated awareness of a political reality that posed a challenge to their responsibilities as scientists radically altered their views of the earlier observations and experiments, leading in turn to a global and fundamental revision of the scientific interpretations of facts that until then had been considered well established. This change found concrete expression in Dobzhansky's 1946 declaration that "the widespread belief that hybrids of human races are inferior ... must be classified as a superstition."(5) Similar pronouncements were issued by institutions such as UNESCO.

Thus, once the mirage of neutrality was dispelled, research that had been considered "objective and neutral" turned out to be deceptive. Provine, who analyzed this process in detail, concluded that "The real danger is not that biology changes with society, but that the public expects biology to provide objective truth, apart from any social influence. Geneticists and the public should realize that the science of genetics is often intertwined with social attitudes and political considerations."(6)

This message, voiced twenty-five years ago, remains entirely a propos today, and it explains the caution with which questions on the heredity of cognitive aptitudes and behaviors must be approached.

These last themes are obviously far from unexplored; they were already the particular subject of intellectual disputes in the nineteenth century. Let us recall the debates on the role of social or biological heredity in the expression of criminality, of mental retardation, or of "genius." Since the work of Francis Galton, this matter of the heredity of intelligence and of behavior has continually resurfaced in the scientific literature and in the media, demonstrating that, after more than a century of research, the case is neither definitively nor clearly closed. As hackneyed as it is, this old theme continues nevertheless to provoke numerous polemics, both scientific and ideological alike.

However, research methods have been perfected, and models have been developed with greater formality and complexity, giving rise to the hope that simplistic arguments that oppose genes to environment, innate to acquired, or, to use Galton's terms, nature to nurture, might become utterly obsolete.(7) Similarly, these developments gave reason to hope that reductive or mechanistic approaches would be abandoned. Yet such is not the case. Many errors of interpretation stem from a poor understanding of genetic models and from a certain blind propensity to wilfully misunderstand the hypotheses they underlie.

In what follows, I shall try to understand and analyze how some of the deviations of human genetics have come into being and persisted, when genetics enters into territory that is no longer confined to the strictly biological.

The rediscovery of Mendel's laws was a major event in the development of the biological sciences. These laws made it possible to understand why some of the biological traits that are observed in parents can also be found in their children, and why brothers and sisters may share these traits. Since Weismann (1892), it has been known that these resemblances are due to entities known as "genes," which are present in every individual in pairs, one of which is transmitted by the mother, chosen at random from the two genes she possesses, and the other transmitted by the father, also chosen at random from the two that he possesses. A child and his father or mother therefore share half of their genes, and two brothers or two sisters have, on average, a quarter of their genes in common.

Let us recall that these laws of transmission apply to discrete biological traits such as the color of flowers or the shape of seeds. They apply straightforwardly to blood type (ABO, Rhesus factor), to tissue and immunity groups (HLA, Gm), and many other traits. Mendel's laws are so effective in resolving questions that had baffled many a great mind before him that there was a powerful temptation to use these same laws to explain any resemblance whatsoever between related persons, including those resemblances that were not considered a priori to have any obvious biological foundation. Thus, in 1911, Davenport proposed to explain criminality or nomadism by genetic heredity. Other authors, as we have seen, used similar arguments to explain that racial mixing was genetically harmful because it caused a genetic "disharmony" in offspring, thus fostering criminality, mental retardation, and laziness.

These attempts to apply Mendel's laws in such varied domains as anthropometry, physiology, psychology, or sociology received generous support and political backing at the highest level. Thus, it was thanks to Rockefeller and Harriman that Davenport founded the Eugenics Record Office, whose mission was to carry out such studies by gathering genealogical records for thousands of families. The influence of such studies was felt well beyond the middle of this century. Scholarly books in biology still bear witness to it. The genealogies of great families of mathematicians, musicians, and painters served to illustrate the heredity of aptitudes, in keeping with the views developed by Galton in his book Hereditary Genius.(8)

But the most salient offshoot of pan-Mendelism came to the fore in the 1970s, in the excesses of sociobiology Indeed, many of the theoretical developments of sociobiology were based on the concept of the gene as applied to behaviors such as altruism, selfishness, aggressivity, sociability, the capacity for innovation, and so forth.

Such was the case with Lumsden and Wilson's notion of "culturgens" (1981)(9) or Dawkins' "memes" (1976)(10). While the notion of the gene is often expanded in such approaches, two characteristics of the "true" genes of genetics are retained: the notion of the simple, discrete entity, to which sociobiologists attempt to reduce the complexity of a behavior, and that of transmission, which is sometimes generalized in order to escape Mendel's overly rigid rules. In fact, no observation can directly support |he hypothesis of a gene for such behaviors, since no one has been able to prove their universal transmission according to Mendel's laws from generation to generation. But sociobiological models exempt themselves from the requirement of proof, maintaining that the validity of the genetic hypothesis is a given as long as it furnishes an accurate explanation for certain sociological, cultural, or ethnological observations. In this reductive universe, the fitting of a model to the data is sufficient to validate the model's hypotheses.

Another illustration of the pan-Mendelian trend is manifested today in medicine and in the behavioral sciences. Research on genes is becoming an all-consuming activity, as we shall see below.

The Breakdown of Genetics

A second important advance in the field of genetics was achieved in 1981 by the statistician and geneticist Fisher. Indeed, Mendel's laws provided no explanation for resemblances among related individuals for continuous traits such as height, head dimensions, or ... mental faculties. Galton had devoted significant time to this problem, without being able to propose any satisfying explanations. Fisher developed a genetic model capable of accounting for familial correlations. He conceived of continuous traits as being determined by the existence of numerous genes that are mutually independent, each one making a small contribution to their expression. Other non- genetic causes could also contribute to the ultimate expression. Ibis dissection of a trait into different components is clearly artificial. As an example, let us consider the intelligence quotient (IQ). How can we say that with an IQ of 110, the first 90 points could be explained by genes and the remaining 20 points by environment? This breakdown also presents the major disadvantage of ignoring a term of interaction that would express the fact that the influence of genes on the trait being studied is not necessarily independent of the environment. It is particularly reductive to disregard this interaction, as in the case of IQ, since doing so is tantamount to seeing the genetic differences between the IQ values of two individuals as being of the same importance, whether these individuals are raised in a favorable environment or a disadvantaged one.

This non-interactive or "additive" model is the essential foundation of what is called quantitative genetics as applied to IQ. Its objective is to identify the causes of IQ disparities in a population or in a group of people. From a statistical point of view, this disparity or difference is expressed by a statistical parameter called "variance," which grows smaller as a function of smaller IQ differences among the individuals. The method thus proposes to break down this variance and to quantify the part that is thought to be due to genes and the part thought to be due to other causes, termed environmental as a simplification. The relation between the genetic variance and the total variance of the trait is referred to as heritability.

To estimate this heritability, it must generally be supposed that there is no correlation between the effects of the genes and those of the environment. This hypothesis amounts to the assumption that all the particular forms of a gene (called alleles) can be found with the same probabilities in any environment. There is therefore an assumption that the same alleles are found as frequently in a favorable environment as in a disadvantaged one. The estimation of heritability then relies on calculations and comparison of correlations between different types of family members, such as monozygotic or dizygotic twins, brothers and sisters, half-siblings sharing the same mother or the same father, parents and their children. Added to this are studies of adopted children and of family members living together or separated at various stages of childhood.

All of these successive hypotheses -- the additive model, the absence of correlation between genes and environment -- have the advantage of setting up conditions that promote the testing of simplified (some would say simplistic) models. The number or parameters that define this model, as for example the various components of variance, then become reasonable. They can therefore be estimated with great precision -- even if the model is false. These simplifications have the major drawback of removing any degree of universality from the conclusions. As a result, this breakdown of variance has only a very limited significance: it is valid only for a given group of persons, in a certain narrow range of environmental variation, and for a precise time period. In other words, if a different population, environment, or generation is examined, this breakdown cannot produce the same results, and comparisons of "heritability" lose all usefulness.

The objection that is often raised to such reservations about quantitative genetic models comes from the fact that these models give good results in experiments with animals and plants, as for example in efforts to improve feed grain or dairy production. But what relation is there between an experimental science on the one hand, in which crossings of animals or plants can be controlled for uncontestably biological criteria (such as the quantity of lipids contained in milk or the ratio of lean matter to fatty matter), and on the other hand, what can be cobbled together from observations of the effects of imperfectly controlled situations (such as adoption situations or socio-economic criteria) on a trait, a cognitive aptitude, or a behavior, which are particularly difficult to measure and whose biological nature is by no means evident?

The utter inanity of research on the heritability of IQ or other behavioral traits was largely demonstrated in the latter 1970s,(11) but not without spirited resistance. In our day, even the most ardent defenders of the genetics of behavior have come to admit this vacuousness,(12) without however fully accepting the implications of this admission.

Indeed, works using this model continue to flourish in the scientific literature, as witnessed by an article recently published under the title "The Heritability of IQ."(13) The originality, so to speak, of this work lies in the way it combines together in one model the results of some 212 studies already published on the subject since the beginning of the century. It consists then of a compilation of necessarily heterogenous studies, each of which takes a different approach to evaluating IQ patterns among family members (monozygotic and dizygotic twins, siblings raised together or separately, parents and children, adoption studies, and so on). The perfectly classic model used is that of the breakdown of variance whose limits we have just seen; the existence of interactions, of correlations between genes and environment, of genetic and environmental differences among the samples, are all completely disregarded. We may wonder about the innovation of such a "meta-analysis." In fact, the authors attempt to integrate into their model an additional effect, the maternal influence: the IQ's of children of the same mother tend to show greater similarity because these children shared the same prenatal conditions. The conclusions of this work therefore moderate the role of genes in determining IQ differences, according greater importance to maternal influences. Perhaps it would have been less deceptive if the article were entitled "The Implications of Maternal Effects on Similarities in IQ." The deliberate choice of another title seems to prove that the goal of this work is not so much to assert these maternal influences as it is to attempt to take them into account with the sole aim of "refining" the measure of heritability in order to give it greater weight; the trade-off requires only a minor concession as to the value of the model. In another article in the same issue of this journal, "The Democracy of the Genes," McGue points out this contradiction, but only in order to reassert the classical biology-driven refrain: "Research on the nature and nurture of IQ is converging on the view that human intellectual ability has a strong, but malleable, biological basis."(14) Everything thus boils down to biology, but in a magnanimously democratic gesture -- democratic oblige -- biology allows the little people a voice in numerous causes. According to this interpretation, genes remain the all-powerful deus ex machina, even in their very moments of abdication. We can only wonder when the democracy gene will be discovered.

A virtually irrefutable argument in favor of a genetic foundation for cognitive aptitudes, behaviors, and personality traits would be the ability to localize and materialize on chromosomes one or more genes that act directly upon them. Three methods are generally used in an attempt to achieve this end.

The first consists in looking to see whether people who show the trait possess a particular form of a gene, an allele, that is not present, or that is possessed with less frequency, by persons who do not show this trait. This association between the trait and the genetic marker would thus be the first argument in favor of a relation between the two. However, it does not constitute a proof, for such an association can have other causes, which are well documented by population genetics (selection, admixture of populations, random variations, and so forth). Nevertheless, it is possible to get closer to a proof by invoking a second method, which endeavors to investigate whether the trait and the marker are jointly transmitted from parents to their offspring, according to Mendel's laws. If the trait and the marker are linked in this way, it means they obey the same genetic cause, or that the causes are topologically close together on the same chromosome. In order to determine which of these possibilities is operative, recourse must be had to a third method, which will investigate directly the relation between the product of the gene (a protein, an enzyme, etc.) and the expression of the trait. This last approach takes us outside the realm of genetics to the field of molecular and cellular biology or to that of pharmacology.

The study of associations came into existence with the discovery of the first genetic markers in human beings. The ABO system of blood types was a natural candidate for testing this sort of association. Carried to an extreme, some such studies have reached the point of caricature. Thus Leone Bourdel devoted an entire book the relations between blood types and "temperaments."(15) According to this author, blood type A governs "intimacy, the domain of affective harmony, reaching beyond the self," along with the capacity for contemplation and passion; group B, in contrast, characterizes rational, deliberate and authoritarian tendencies ... These conclusions, drawn from observations at the level of individuals, were extended to entire peoples. Those peoples with predominantly type-A blood, for example, are said to wage war only in "self-defense of their affective intimacy," whereas the peoples with predominantly type-B blood are considered to be innately "the most spontaneously belligerent" peoples, for whom war is a "natural function."

These wild imaginings obviously have no statistical foundation, but they are reported in a highly "scientific" language in the series "Bilan de la science," directed by a scientific guru of the period (Leprince-Ringuet). The resulting mirage has played an active role in attempts to explain human nature in biological terms.

This theme was still alive and well as of 1973. For example, Gibson et al(16) reported that IQ is higher in individuals of blood types 0 and A2 (a variant of type A) than in individuals of other types. The difference is significant, although particularly small (3 percent).

The insistence on finding an association between a genetic marker and a behavior always ends up exacting a toll. Dumont-Damien and Duyme 17 report that between 1956 and 1991, the more than 140 studies of association that were carried out in connection with alcoholism explored close to 50 different markers (ABO blood types, rhesus factor, HLA, various enzymes playing a role in the metabolism of alcohol, and others), some of which were revealed to be significantly associated with certain forms of alcoholism.

Recently, an association has also been reported between a marker located in the gene for the dopamine (a neurotransmitter) receptor and the personality trait of novelty seeking. In the same vein, a number of works demonstrate associations with homosexuality, criminal tendencies, and so on. Finally, IQ itself has not escaped association studies. Plomin's team(18) thus selected three groups of persons with IQs averaging 130, 105, and 82. Then 100 genetic markers were tested for possible associations. Three of these turned out to be significant, the strongest association being with mitochondrial DNA. This is an important point, since this DNA is transmitted only by the mother. Therefore, if the relation between mitochondrial DNA and IQ is grounded in fact, we would expect IQ to be transmitted by the mother and not by the father, something for which we have no evidence as of yet.

In all of these investigations, the essential difficulties -- apart from the methodological questions already discussed above(19) -- he in the definition of the trait and the classificatory distinction between those who carry it and others. This problem is particularly flagrant for alcoholism, which is the expression of multiple risk factors: neurophysiological, psychological, familial, professional, and perhaps genetic. The same holds true for a personality trait or for IQ, which are only statistical composites of performance on multiple tests.

Methods of investigating links between a genetic marker (whose precise location on a chromosome may be known) and IQ a cognitive aptitude, or other behavioral traits are clearly more conclusive than studies of association. Research on links attempts to determine whether the "candidate" gene for a trait is transmitted at the same time as a genetic marker in family genealogies. Once again, the definition of the trait and the to some degree arbitrary decision to assign its presence or absence in an individual are rife with ambiguities. Another handicap is our ignorance with regard to specifying a priori the mode of transmission and expression of the candidate gene, which remains hypothetical and which we are attempting to concretize. Is it recessive or dominant -- that is, is it expressed only when it is present in two copies, or is one sufficient? Is the expressivity or penetrance of the hypothetical gene total or partial -- in other words, is it always expressed whatever the circumstances, or only under certain conditions, controlled or not controlled, such as those of age or environment? What is the gene's frequency in the population? What is the probability of its appearing in sporadic non-genetic cases, or the frequency of mutation de novo? The choice of a model is therefore particularly delicate, for upon it will depend the ultimate conclusions. It might therefore be said that the systematic search for a gene always has a strong probability of succeeding without undue strain. For even if one does not find the gene for the trait, there is still a chance of proving the existence of a gene with variable penetrance. And even when the latter fails, there is still the possibility of finding one or several genes for "susceptibility," that is, genes whose presence indicates fertile ground for developing the trait, always under certain conditions. Moreover, if the physical localization of the gene on the chromosomes is not possible, it can always have a "statistical" existence." And as a last resort, there is always the possibility of saying that the gene is so rare that it has been found in only one family, even in a single individual; or even that it has heretofore eluded detection but that nothing is lost by waiting ... As we can see, the notion of the gene is so elastic on this level, between physical reality and virtual existence, that it is virtually impossible to escape it.

Finally, it is possible that there is not just one "candidate" gene to explain a trait, but rather several that can come into play alternatively, synergistically, or independently, depending on the individual.

This mad hunt for the slightest effect of a gene is one of the faces of contemporary pan-Mendelism. Obviously, it has only a minor impact in the area of public health, though it absorbs a major expenditure of energy.

The Gene That Does the Most Does the Least?

These studies of linkage have thus far never shown the existence of links between cognitive aptitude, personality traits or behaviors, with the exception of specific pathological situations. Indeed, positive results have been obtained in some cases of mental retardation, such as those provoked by phenylketonurea or the weak X chromosome. In the case of altered intellectual functions, such as in Alzheimer's disease, studies of linkage show that the genetic forms with dominant mode of inheritance involve only about 4 out of every 1000 cases, but they represent from 10 to 20 per cent of early-onset cases (before the age of sixty). Several different mutations of the genome appear to be responsible for this neurodegenerative disease.(20) Likewise, a certain Dutch family appears to demonstrate a mutation of monoamine oxydase which might be linked to a type of impulsive violent behavior.

The encouraging results in the search for genes implicated in different forms of mental retardation, mild or severe, are often used as arguments in favor of the determinant role of genes on behaviors or cognitive aptitudes. The reasoning is as follows: since there exist genes that account for mental retardation, therefore there must also, necessarily, be genes for intelligence. This deduction amounts to viewing mental deficiencies, which are incontestably of genetic origin, as being of the same order as "intelligence." But there is no reason that genes implicated in mental deficiency, if they do exist, should be the same as those that dictate normal or exceptional intelligence. The causes that he behind dysfunction are not necessarily the same as the causes of normal functioning.

This reasoning, however, has been put forth recently, not without some rather amusing -- if unwitting -- implications.(21) We know that the "candidate" genes implicated in these mental deficiencies are quite numerous, numbering somewhere between 300 and 400, and for the most part they are located on the X chromosome. This situation is not without consequences. For this chromosome is found in two copies (XX) in women, whereas in men one X chromosome is associated with one Y chromosome (XY). Since many of the genes that are responsible for mental retardation are located on the X chromosome, that, according to the author, is where the intelligence genes are also to be found. This generalization, a rather hasty one, leads to various far-fetched deductions, however logical they may appear here, as well as to some marital advice for those who wish to guarantee "superior" offspring.

Indeed, if "intelligence" is transmitted with the X chromosome, then the father no longer has any influence on the intelligence of his son, since in that case he transmits only his Y chromosome. If he wants to maximize his son's intelligence, he must therefore select his parents-in-law with great care. Likewise, whomever a woman may marry, the intelligence of her son will depend only on that of her own parents (since she gives her son the X chromosome inherited from her father or her mother). On the other hand, if she wished to insure her daughter's superior intelligence, there is no point in choosing her father-in-law carefully, since he plays no role in her daughter's brilliance (he gives her no X chromosome). But she will have to pay attention to the "intellectual quality" of her mother-in-law ...

The genealogies of famous families are given by the authors as confirmation of this simplistic theory. For example, that of Sir Charles Darwin and his cousin Sir Francis Galton, undeniably two great scholars of the nineteenth century. There again, we are sorely deceived, for the genealogy proposed by these authors is carefully purged of all the weaker minds, the unstable or suicidal individuals, those suffering from depression, dyslexics, and the like.(22) As for the women, they are never counted as "brilliant" -- probably because no one ever dreamed they might be ... But do they not possess an X chromosome, just as men do?

This example shows how the obsession with the gene at any cost, and the extension of a possible pathological reality to the realm of normality, can result in aberrations, even leading some authors to advocate certain types of behavioral strategies.

All of this pseudo-science, which appears in prestigious reviews indiscriminately mixed together with true science, furnishes solid proof -- by counter-example and by hyperbole -- that the social and moral perception of intelligence is more determinant than a genetic pseudo-definition based on multiple series of biases, errors, and falsifications, which persist despite having been disproven already a thousand times over.

From this rapid survey of the relations between genetics, cognitive aptitudes, behaviors, and personality, one is tempted to admit the relative failure, temporary powerlessness, or absence of universality of the conclusions. The methods of genetics, both classical and modern approaches alike, could no doubt explain a tiny fraction of certain behaviors linked to certain forms of deviance in a limited number of families, as the above examples have shown. It is not impossible that this knowledge might shed some light on certain molecular or neurophysiological mechanisms, and this information would enrich the current pharmaceutical repertoire. But that no longer pertains to the field of genetics.

The limits of genetics quickly become clear in the domain of cognitive sciences and social sciences. For example, a form of gene therapy that would make it possible to correct some deviations with respect to "standards" at the particularly fluctuating boundaries of these disciplines is not within the realm of the possible. Would it even fall within the reality of the desirable? Likewise, a reproductive selection based on genetic criteria (if their reality were indeed demonstrated) would raise the issue of justifying the distinction between negative selection (interrupting the transmission of unfavorable genes) and positive selection (promoting the transmission of favorable genes). This issue opens up the Pandora's box of a related problem -- that is, eugenics.

Must we consign the radiant future of the gene to the rubric of science fiction? Up until now, as we have tried to show, the biologically correct discourse has placed DNA at the center of a rather totalitarian machinery. Thus, Plomin remarks that "the current enthusiasm for genetics should not obscure the important contribution of non-heritable factors, though these are difficult to investigate."(23) But this profession of faith regarding the importance of non-genetic factors is nullified a few lines later, with a reversion to positions that are directly in line with exclusively genetic thought: "The ostensible measures of environment appear to assess genetically influenced characteristics of individuals. In a certain way, individuals create their own experiences for genetic reasons." This statement coincides with McGue's interpretation of the pre-eminence of biological, despite its high degree of malleability.

Now, it is not inconceivable that all of these discourses are merely an ideological falsification of biological reality. Let us recall the question of racial mixing. For the gene must no longer be seen as a program whose inexorable execution governs all the details of our lives -- cellular, hormonal, or social. At the level of populations, at the level of the individual, and at the molecular level, the mechanisms of the expression of genes do not obey rigid determinisms.(24) Rather, they are subject to the play of probabilities, which are affected by all the structures and molecules that surround genes, all the stimulations of the physical, chemical, and human environment, and all the interactions with other "gene carriers."

Upon this complex tapestry, time also plays its role as the creator of unique histories, those that each of us is free to live out.

(1.) Reference is made here to the growing importance accorded to biological kin-ship and to the desire to preserve one's genes for posterity through one's descendants, whereas kinship was first defined by affective bonds within an ethno-sociological context, without particular reference to biology.

(2.) Levy-Leblond et al., in Le Monde, 15 June 1977.

(3.) The lack of interest on the part of certain biologists and the ignorance of many technocrats with respect to these questions often stem from an unshakable faith in the inevitability of "progress." "One of the inevitabilities of scientific progress is that if something can be done then it will be done," according to Daniels et al., in Journal of Biosocial Sciences 28 (1996): 491-507. Moreover, some ways of accentuating and encouraging the hope that may arise out of any scientific advance is in fact only a clever tactic for leaving one's hands free to cover up research whose consequences will ultimately escape control (on this subject, see the analysis by Axel Kahn, Futuribles 223 (1997): 5-27.

(4.) Provine, "Geneticists and the Biology of Race Crossing," Science 4414 (1973): 790-796. The word "race" is to be understood in its historical context.

(5.) Dunn and Dobzhansky, Heredity, Race and Society (New York, 1946), p. 114.

(6.) Provine, "Geneticists and the Biology of Race Crossing," p. 796.

(7.) Galton himself participated actively in this effort by developing the statistical concept of correlation.

(8.) Sir Francis Galton, Hereditary Genius: An Inquiry into Its Laws and Consequences (London, 1892; 1st ed. 1865).

(9.) Charles J. Lumsden and Edward O. Wilson, Genes, Mind and Culture: The Coevolutionary Process (Cambridge, Massachusetts, 1981).

(10.) Richard Dawkins, The Selfish Gene (Oxford, New York, 1976).

(11.) In particular by Feldman and Lewontin, and in France by Jacquard in Eloge de la difference (Paris, 1978).

(12.) "Estimates of heritability apply only to the population studied at that particular time, and under environmental conditions that prevail at that point," in the words of M. Rutter and R. Plomin, "Opportunities for Psychiatry from Genetic Findings," British Journal of Psychiatry (1997), pp. 209-219.

(13.) Devlin et al., Nature 388 (1977), pp. 468-471.

(14.) McGue, "The Democracy of the Genes," Nature 388 (1977), pp. 417-418.

(15.) L. Bourdel, Sangs et temperament (Paris, 1962).

(16.) Gibson ect al., "Ig and ABO Blood Groups," Nature 246 (1973), pp. 496-499.

(17.) Dumont-Damien and Duyrne, Genetics and Alcoholism (Les Editions INSERM, 1993).

(18.) Plomin et al., Behavior Genetics 21 (1995): pp. 31-48.

(19.) And a few other questions whose technical aspects exceed the scope of the present article.

(20.) Campion et al., "Les facteurs genetiques clans l'etiologie de la maladie d'Alzheimer," Medecine/Sciences 12 (1996): pp. 723-731.

(21.) G. Turner, "Intelligence and the X Chromosome," The Lancet 347 (1996): pp. 1814-1815.

(22.) Resta, "Whispered Hints?", American Journal of Medical Genetics 59 (1995): pp. 131-133.

(23.) Plomin, Owen, and McGuffin, "The Genetic Basis of Complex Human Behaviors," Science 264 (1994): pp. 1733-1739.

(24.) Kupiec and Sonigo, "Du genotype au phenotype: instruction ou selection?," in Pour Darwin, ed. Patrick Tort (Paris, 1997), pp. 1025-1034. Article A20844670

Helene Gjone; Jim Stevenson.

Objectives: To assess the longitudinal covariance between emotionality,

activity, and sociability (EAS) temperamental traits and anxious/depressed

behavior, attention problems, delinquent behavior, and aggressive behavior

and to assess the significance of genetic and common environmental

influences on the temperament and behavior relations. Method: Parental

responses to the Child Behavior Checklist and the EAS Temperament Survey

were collected from five national cohorts of Norwegian same-sex twins. The

final sample consisted of 759 twin pairs aged 7 through 17 at 2-year

follow-up. Results: High emotionality predicted anxious/depressed behavior,

attention problems, delinquent behavior, and aggressive behavior. The

influence on delinquent and aggressive behavior was stronger in boys.

Aggressive behavior was further predicted by high activity scores, especially

in younger children. Significant genetic influence was found for the

covariance between emotionality and attention problems and emotionality and aggressive behavior. Conclusion: Emotionality was the strongest

temperamental predictor of behavior problems. The mechanisms involved in

the associations between temperament and behavior problems appeared to

differ with kind of behavior problems. Key Words: emotionality, activity,

sociability, temperament, Child Behavior Checklist, behavior problems,

follow-up, twins.

Several studies have reported associations between temperament and behavior problems or psychiatric disturbance in children. Different temperament concepts are applied in these studies and complicate the comparison between studies. Difficult temperament, mainly consisting of nonadaptability and negative emotionality, is found to predict behavior problems in the study by Thomas et al. (1968) and in later studies applying derivatives of the original Thomas and Chess temperamental dimensions (Bates and Bayles, 1988; Caspi et al., 1995; Oberklaid et al., 1993; Prior et al., 1992; Thomas and Chess, 1977; Wolfson et al., 1987). The difficult temperament concept is closely related to negative emotionality (Bates, 1989). Studies applying the EAS temperamental dimensions - emotionality, activity, and sociability (Buss and Plomin, 1975, 1984) - report associations between emotionality and later behavior problems or psychiatric disturbance (Goodyer et al., 1993; Rende, 1993; Rende and Plomin, 1992).

Various mechanisms are hypothesized for the association between

temperament and behavior problems (Rutter, 1989). Difficult child

temperament could influence parents' response directly by eliciting negative

responses and thus create difficulties in parent - child interaction. Difficult

temperament could also act indirectly in social interaction through a poor fit

between the child and the social environment (Thomas et al., 1968).

Temperament may moderate the possible impact of stressful experiences, such as family conflicts, on child behavior (Rutter and Quinton, 1984). Alternatively or additionally, constitutional factors could account for the continuity between temperament and psychiatric disturbance, with

temperament being the genetic contribution to behavior in childhood (Graham

and Stevenson, 1987; Stevenson and Graham, 1982). Graham and Stevenson

(1987) recognized a three-factor structure of temperament and behavior

problems. They discussed whether the Buss and Plomin typology of

temperament (emotionality, activity, sociability) are the ordinary, not

pathological, behaviors that in extreme form become three types of abnormal

or maladaptive behavior, namely emotional disorder, hyperactivity, and

antisocial disorder. Thus they proposed a continuity between particular

temperament traits and the type of childhood psychiatric disorder occurring

later. With some modification, they regarded Buss and Plomin's EAS

temperaments to fit into such a typology. Lack of sociability could be a part of

later antisocial behavior. Emotionality includes both fear and anger

components, and fearfulness could be a part of later neurotic disorder. A

hyperactivity component of behavior problems could have high activity as a

precursor.

Few studies have addressed the specificity of temperament - behavior

associations. Associations with difficult temperament are found both with

internalizing and externalizing behavior (Bates and Bayles, 1988; Caspi et al.,

1995), though they are stronger for externalizing than internalizing problems

(Caspi et al., 1995; Maziade, 1989). Rende (1993) investigated the hypothesis

of continuity and specificity as proposed by Graham and Stevenson (1987)

with a sample of young children. High emotionality predicted Child Behavior

Checklist (CBCL) Anxious/Depressed behavior and Attention Problems in

boys. Anxious/Depressed behavior was predicted by high emotionality and

low sociability in girls. Thus the study did to some extent support a

hypothesis of specific associations between EAS temperaments and particular

behavior problems in this age group.

In this study, associations between EAS temperaments and the CBCL

Anxious/Depressed, Attention Problems, Delinquent, and Aggressive

Behavior scales were investigated in a population-based sample of child and

adolescent Norwegian twins. The first purpose was to investigate the extent to

which parentally perceived EAS temperament traits predict parent-reported

behavior problems in children and adolescents. The second purpose was to

investigate whether there are specific associations between EAS temperament

and behavior problems as proposed by Graham and Stevenson (1987). A

third issue to be addressed in this article is whether a predicted covariance

between temperament and behavior problems consists of genetic components

indicative of a direct constitutional relationship as suggested by Graham and

Stevenson (1987).

METHOD

Sample

Time 1([T.sub.1]). Five national birth cohorts of same-sex twins born in

1977, 1978, 1979, 1983, and 1986 were drawn from the Norwegian Medical

Birth Register, and all pairs with both twins alive were included (1,529 twin

pairs). The families were contacted by mail in 1992 and one parent, preferably the mother, was asked to answer questionnaires about the twin's temperament (EAS Temperament Survey [EAS]) and behavior problems (CBCL). Parents of 916 pairs (59.9% of the total) responded in the first wave of data collection. The procedure is described in further detail elsewhere (Gjone and Novik, 1995).

Time 2([T.sub.2]). A follow-up with mailed questionnaires to the [T.sub.1]

respondents was conducted 2 years after [T.sub.1]. One or both parents of

782 pairs (85.5% of the [T.sub.1] respondents) responded. Complete CBCL

and EAS ratings were obtained from 772 pairs. For the subsequent analyses

only the cases with the same respondent to the EAS-1992 and to the

CBCL-1994, or with one respondent at one time point and a joint response at

the other time point, were retained. The sample then consisted of 758 pairs.

Representativeness of Sample

Time 1 Respondents. There were no significant differences in sex

distribution, place of living, birth weight, rates of cesarean section, or Apgar

scores between [T.sub.1] respondents and nonrespondents. There was a

slightly lower mean duration of pregnancy in the nonrespondent group (Gjone

and Novik, 1995). No information on socioeconomic status (SES) was

available for the nonrespondents. Fathers in respondent families showed a

significantly higher SES when compared with the general Norwegian male

population by the Norwegian standard classification of SES ([[Chi].sup.2] =

61.9, df = 4, p [less than] .001).

Time 2 Respondents. [T.sub.1] respondents who did not reply at 2-year

follow-up or who provided incomplete responses ([T.sub.2] nonresp: n = 144

pairs) were compared with respondents at both time points ([T.sub.2] resp: n

= 772 pairs). There were significantly more female twins among the [T.sub.2]

nonrespondents (59% versus 51.4% in [T.sub.2] resp; [[Chi].sup.2] = 5.66,

1 dr, p = .017). Duration of pregnancy was slightly though significantly

lower among the [T.sub.2] nonrespondents (262.0 versus 266.0 days; t =

-3.27, p = .001) and so was birth weight (2,514.9 versus 2,680.3 g; t =

-4.78, p [less than] .001). More [T.sub.2] nonrespondents were found in the

most urban category when compared on place of residence ([[Chi].sup.2] =

18.73, 6 df, p = .005). There were significantly more manual workers among

the fathers in the [T.sub.2] nonrespondent families than in the [T.sub.2]

respondent families ([[Chi].sup.2] = 26.03, 4 df, p [less than] .001). No

significant difference appeared in number of pairs living with both parents

(79% of [T.sub.2] nonresp versus 83.6% of [T.sub.2] resp), living with a

single parent (4.2% of [T.sub.2] nonresp versus 2.3% of [T.sub.2] resp),

living with one parent and with a separate address for the other parent (16.1%

of [T.sub.2] nonresp versus 13.5% of [T.sub.2] resp), or twins living

separately (0.7% of [T.sub.2] nonresp versus 0.5 % of [T.sub.2] resp). The

corresponding numbers for nonrespondents at both time points were 79.7%,

4.3%, 15.6%, and 0.5%. When this group was added in a [[Chi].sup.2]

comparison, the slight overweight of single parents among nonrespondents at

one or both time points was significant at the .05 level ([[Chi].sup.2] =

13.02, 6 df, p = .043).

Mean behavior problem scores and temperament scores for [T.sub.2]

respondents and [T.sub.2] nonrespondents are shown in Table 1, with a

significantly higher sociability score among the [T.sub.2] nonrespondents;

otherwise there were no differences.

Behavior Problems. Parental CBCL (Achenbach, 1991; Achenbach and

Edelbrock, 1983) ratings were obtained for both twins. The 113 CBCL

behavioral problem items are scored as follows: 0 if the item is not true of the

child, 1 if the item is somewhat or sometimes true, and 2 if it is very true or

often true. The 1991 version of the CBCL generates eight syndrome scale

scores across sex and age (Achenbach, 1991). Four syndromes -

Anxious/Depressed, Attention Problems, Delinquent Behavior, and

Aggressive Behavior - were selected for the subsequent analyses. Studies

have shown no one-to-one relationship between separate CBCL subscales and

DSM-III diagnoses (Edelbrock and Costello, 1988; Gould et al., 1993). The

above subscales do show correspondence with diagnoses related to emotional

problems, hyperactivity, and antisocial disorder, respectively. The selection of

these subscales also allows comparison with the study by Rende (1993). The

[T.sub.2] correlations between the behavior syndromes in the present study

were all significant. The Anxious/Depressed scale correlated .49 with the

Attention Problems scale, .39 with the Delinquent scale, and .51 with the

Aggressive scale. The Attention Problems scale correlated .52 with the

Delinquent scale and .63 with the Aggressive Scale. The Delinquent scale

correlated .67 with the Aggressive scale.

Temperament. Emotionality, activity, and sociability were assessed by the

EAS Temperament Survey (Buss and Plomin, 1984). The EAS temperament

survey is based on a theory of temperament as appearing in infancy, being

relatively stable and predominantly genetically influenced personality traits

(Buss and Plomin, 1975, 1984). Each EAS dimension was computed as the

average score from five items on a 5-point scale from 1 = "not characteristic or typical for your child" to 5 = "very characteristic or typical for your child."

The items included describe child temperament as follows: Highly emotional

children are characterized by a tendency to cry easily; they tend to be

somewhat emotional, often fuss and cry, get upset easily, and react

intensively when upset. Active children are characterized as always on the go;

they do not move slowly about; they are off and running as soon as they are

up in the morning, are very energetic, and do not prefer quiet, inactive games

to more active ones. Sociable children characteristically like to be with people,

prefer playing with others rather than playing alone, find people more

stimulating than anything else, are not loners, and feel isolated when alone.

These last five items constitute the sociability part of the Sociability/Shyness

scale. Buss and Plomin (1984) report an average test-retest reliability for the

adult version of the EAS of 0.82. Lower test-retest reliability was found for a

sample of younger children: 0.72 for emotionality, 0.80 for activity, and 0.58

for sociability/shyness. No test-retest data were obtained for the present

sample. Buss and Plomin (1984) report a significant correlation of .16

between activity and sociability and other correlations being insignificant.

Higher and significant correlations were obtained in the present work: .36

between activity and sociability, .25 between emotionality and sociability, and .07 between activity and emotionality. An unforced factor solution yielded

five factors with eigenvalues greater than one in the present sample,

explaining 54.6% of the total variance. With varimax rotation there were no

dimensions corresponding totally with the generated EAS factors. A structure

with one emotionality, one sociability, and one shyness scale was obtained. In

addition, two sociability/activity dimensions emerged. Similar findings of

adequate but not total correspondence have been reported for the EASI scales

(Buss and Plomin, 1975; Gibbs et al., 1987). The EASI scale, being the

precursor of the EAS scale used in the present study, has shown more

sufficient factor study than scales based on the Thomas and Chess dimensions

(Gibbs et al., 1987).

Mean 1992 CBCL and 1992 EAS Scores in 1994 Responders and

Nonresponders

1994 Responders 1994 Nonresponders (n = 1,544) (n = 288)

Variable Mean SD Mean SD t

CBCL

Anxious/Depressed 2.28 2.75 2.41 3.21 0.74

Attention Problems 1.78 2.17 2.00 2.55 1.54

Delinquent 1.11 1.57 1.29 1.76 0.84

Aggressive 4.86 5.24 5.12 5.38 0.78

EAS

Emotionality 2.49 0.84 2.53 0.87 0.59

Sociability 3.63 0.60 3.73 0.56 2.76(1)

Activity 3.77 0.68 3.69 0.68 -1.72

Note: CBCL = Child Behavior Checklist; EAS = Emotionality,

Sociability, Activity.

* p [less than] .01 (t test).

Zygosity. The twins were identified as identical (MZ) or fraternal (DZ) by a

twin similarity questionnaire (Cohen et al., 1973). By a procedure described

in detail elsewhere (Gjone et al., 1996), 57.5% of the 1992 twin sample were

assigned to the MZ group and 42.5% to the DZ group. There were 56.7% MZ

and 43.3% DZ pairs among the 1994 respondents.

Age and sex differences in level of [T.sub.2] behavior problems and

[T.sub.1] temperamental traits were analyzed in a two-way analysis of

variance. Associations between [T.sub.1] EAS traits and [T.sub.2] CBCL

traits were first assessed by computing correlations separately for boys and

girls, in four age groups (aged 7 through 8,10 through 11, 14 through 15,

and 16 through 17 at follow-up). Second, each pair of significant

temperament and behavior syndrome correlations was selected for further

multiple regression analyses with the temperament score, sex, and age entered into the regression on a first step and the interaction terms on a second step.

Interactions were computed as product terms between EAS traits and sex

(dummy-coded 0 = male, 1 = female) or between EAS traits and [T.sub.1]

chronological age, to investigate the influence of sex and age on the

temperament and behavioral syndrome associations. The explanatory

temperament variables that were significantly associated with the dependent

(behavioral) variable were selected for the next set of analyses together with

sex, age, and interaction terms if significant. Third, when these explanatory

variables had been selected separately for all CBCL behavior syndromes, they

were entered simultaneously for each behavior syndrome to assess significant

and independent contributions to the explained variance of each behavior

problem syndrome.

For further investigation of genetic and environmental contributions to the significant associations between

[T.sub.1] temperament and behavior problems, the data were analyzed by

multiple regression with one twin in a pair as proband and the other as cotwin

(DeFries and Fulker, 1985, 1988; LaBuda et al., 1986). The regression model

tests the significance of genetic and common environmental influences. This is

based on the extent to which cotwins of identical probands have scores more

similar to their deviant proband and more different from the total sample mean

than cotwins of fraternal probands. With this unselected twin sample, either

member of the pair could be designated as the proband. Consequently, a

double entry that allows both twins to be selected as probands/cotwins was

used. The standard errors and t values were corrected accordingly. Obtained

standard error x [-square root of] 2 gives an appropriate estimate of corrected

standard error, with this large number of twins. The regression model (1) is

identical with the one proposed by LaBuda et al. (1986) for estimating

(univariate) heritability ([h.sup.2]) and common environment ([c.sup.2]) as a

function of proband score. Heritability denotes the relative impact of genes,

while common environment includes the relative impact of all nongenetic

factors that give rise to similarities between family members (here twins) for

the trait or disorder being studied. To assess the heritability and the common

environmental influence on the four selected behavior problem syndromes,

proband [T.sub.2] behavior problem score was entered as the explanatory

variable (proband score) and the corresponding cotwin score as the dependent

variable. To assess bivariate relations between temperament and behavior

problems, significant [T.sub.1] EAS traits were entered as explanatory

variable (proband score) and the [T.sub.2] behavior problems score as

dependent variable (cotwin score). The regression equation is:

C = [b.sub.3]P + [b.sub.4]R + [b.sub.5]PR + A (1)

C is the cotwin score, P is the proband score, and R is the coefficient of

relationship (1 for MZ and 0.5 for DZ twins). From this equation [b.sub.3]

tests the significance of common environmental influences and [b.sub.5] tests

the significance of genetic influences on the temperament-behavior covariance;

[b.sub.4] is a function of twice the difference between MZ and DZ cotwins.

In the univariate case [b.sub.5] equals heritability and [b.sub.3] relative

common environmental influence. Because of differences in distribution and

scaling for the behavioral and temperamental measures, the bivariate results

are not interpreted as quantitative estimates of heritability and common

environmental influences on the temperament and behavior problems

associations. The regression equation 1 was extended to test for significant

sex or age influences on the temperament-behavior covariance if significant

effects of sex, age, or interactions were found in the previous

temperament-behavior regression analyses:

C = [b.sub.6]P + [b.sub.7]R + [b.sub.8]PR + [b.sub.9]K + [b.sub.10]PK +

[b.sub.11]RK + [b.sub.12]PRK + A (2)

K denotes sex or age. From equation 2, [b.sub.9] tests the independent

influence of sex or age, [b.sub.10] tests the significance of age or sex effects

on the common environmental influences, and [b.sub.12] tests the

significance of age or sex effects on the genetic influences on the

temperament-behavior covariance. The interpretation of [b.sub.6] corresponds

to [b.sub.3] in equation 1, [b.sub.7] to [b.sub.4] in equation 1, and [b.sub.8]

to [b.sub.5] in equation 1, with the influence of sex or age removed.

The observed distributions of the CBCL scales were considerably skewed.

Analyses of the distributions of residuals for the different models showed

statistically significant but small deviations from normal distributions. In

earlier work with the CBCL Internalizing and Externalizing scales, we have

shown that the skewed distribution did not violate the assumption of normal

error distribution in a way that affected the significance testing (Gjone et al.,

1996). Consequently the data are not transformed for the following analyses.

Results are considered statistically significant when p [less than] .01.

Two-way analysis of variance showed significantly lower scores with

increasing age for [T.sub.2] Delinquent Behavior ([F.sub.age] = 6.107, p

[less than] .01) and Aggressive Behavior ([F.sub.age] = 29.065, p [less than]

.001). There were sex differences with higher Anxious/Depressed scores

([F.sub.sex] = 19.734, p [less than] .001) and lower Delinquent ([F.sub.sex]

= 11.671, p [less than] .001) and Aggressive Behavior ([F.sub.sex] = 9.501,

p [less than] .01) in girls. The [T.sub.1] temperament scores were lower with

increasing age for emotionality ([F.sub.age] = 34.023, p [less than] .001),

sociability ([F.sub.age] = 19.301, p [less than] .001), and activity

([F.sub.age] = 14.222, p [less than] .001). Boys were scored lower than girls

on sociability ([F.sub.sex] = 14.967, p [less than] .001).

Correlations between [T.sub.1] EAS traits and [T.sub.2] Anxious/Depressed,

Attention Problems, Delinquent, and Aggressive scores are shown in Table 2,

separately for boys and girls, and by age group at follow-up. All behavior

problem scales showed significant positive correlations with emotionality

across sex and age groups, and few significant correlations with activity or

sociability.

For each pair of significant temperament-behavior problem associations,

explanatory variables (temperament, sex, age, and interaction terms)

[TABULAR DATA FOR TABLE 2 OMITTED] that contributed significantly

to predict the behavior problem syndrome are shown in Table 3. To assess the

independent influence of these explanatory variables, they were then entered

simultaneously to regression models for each behavioral syndrome. A high

emotionality score was the strongest predictor of Anxious/Depressed

behavior, in addition to a significant sex influence. Emotionality was the only

significant predictor of Attention Problems. [TABULAR DATA FOR TABLE

3 OMITTED] The significant influence of emotionality on Delinquent

Behavior was greater in boys. Several temperament predictors were found for

Aggressive Behavior, with emotionality as the strongest predictor, particularly

in boys. In addition, high activity scores, especially in younger children,

predicted Aggressive Behavior (Table 3).

A significant relative genetic influence (heritability) was found for all behavior

syndromes when analyzed by univariate regressions across the full sample.

The [T.sub.2] heritabilities with standard errors for the estimates in

parentheses were .77 (.12) for Anxious/Depressed, .79 (.13) for Attention

Problems, .59 (.10) for Delinquent Behavior, and .85 (.11) for Aggressive

Behavior. The corresponding estimates for the [T.sub.1] EAS temperament

traits were 1.24 (.11) for activity, 1.29 (.11) for sociability, and .71 (.09) for

emotionality.

Genetic and common environmental contributions to the EAS traits and

behavioral syndrome covariances were then investigated by using the

regression procedure from LaBuda et al. (1986) with bivariate longitudinal

data (Table 4). For Anxious/Depressed and Delinquent Behavior no

significant genetic or common environmental contribution of the EAS traits

was found. There was a significant influence of genetic factors on the

emotionality and Attention Problems association and on emotionality and

Aggressive Behavior. No significant common environmental influence was

found for any [TABULAR DATA FOR TABLE 4 OMITTED]

temperament-behavior association. There were no significant sex or age

influences.

A strong and specific association between temperament and behavior

problems would be expected according to the hypothesis of continuity by

Graham and Stevenson (1987). Emotionality was the strongest EAS predictor

of the examined behavior syndromes. Rende (1993) found emotionality to be

the strongest predictor of later Anxious/Depressed behavior and Attention

Problems in early childhood up to 7 years of age. The mean behavior problem

levels in the present sample are lower than the Achenbach norms (Achenbach,

1991). They are also lower than in Rende's sample except for Delinquent

Behavior. The present sample includes a greater age span and older children.

According to the Achenbach norms (Achenbach, 1991), the level of

Delinquent Behavior increases with increasing age. The present sample

showed decreased scores on the Delinquency scale with increasing age,

indicating cultural differences between the samples in terms of age-dependent

changes in level of delinquent behavior. It may be that a relatively larger and

thus significant part of the variance in Delinquency scores is associated with

emotionality in the present sample; no predictive influence of emotionality was

reported by Rende (1993).

Following Graham and Stevenson (1987), one would expect low sociability scores to be linked to antisocial behavior. The

pattern of correlations in the present study did not confirm this. The present

study agreed with Rende (1993) in finding no specific associations between

activity and Attention Problems. The 1991 Attention Problems scale of the

CBCL (Achenbach, 1991) is closely similar to the former Hyperactive scale

across sex and age (Achenbach and Edelbrock, 1983). As the CBCL 1983

Hyperactive scale has been shown to correspond closely to diagnoses of

ADHD (Shekim et al., 1986; Steingard et al., 1992), this indicates that activity

as measured by the EAS is not a correlate to attention problems and

hyperactivity.

To our knowledge, there are no other studies reporting on the specificity of EAS

temperaments as predictors of behavioral disturbance. Goodyer et al. (1993)

investigated associations between EAS temperaments and major depression on

a cross-sectional level in 11- to 16-year-olds and found associations between

major depression and emotionality especially, but not exclusively, in girls. No

sex difference was found in the present study.

So far it is not reasonable to claim specificity between EAS temperaments and

certain behavioral syndromes in the direction proposed by Graham and

Stevenson (1987).

Significant genetic influences were found for all EAS temperaments. The univariate heritability estimates for activity and sociability exceeding 1.0 indicate that more specific model-fitting procedures,

including influences other than additive genetic factors, common environment,

and specific environment are needed to characterize these temperament factors

further. The present findings of significant genetic factors in the association

between emotionality and later Attention Problems and Aggressive Behavior

support the hypothesis of emotionality as at least in part a genetic contribution

to these behavior problems. It could be directly shared genetic influences, or

the influence of emotionality could be mediated through gene-environment

correlation by influencing how children are exposed to or select different

environments that in turn contribute to different psychiatric disturbance or

behavior problems. Aggressive Behavior shows a significant genetic

covariance with temperamental traits that is not apparent for Delinquent

Behavior, and it also has a higher heritability, indicating that Aggressive

Behavior is more constitutionally based than Delinquent Behavior.

The present work has several limitations and must be regarded as preliminary.

The unreliability of the instruments, the test-retest unreliability, and the rater

bias for the proband and cotwin contribute to measurement error in these

analyses. Specific environmental factors include measurement error in

addition to environmental influences that makes the individual child (or twin)

different from other family members. Such influences are likely to be of

importance for the temperament-behavior associations. Studies applying more

specific models that allow for further testing of the importance of these

factors, and studies including more specific measures of environmental

adversities, are needed.

The results represent only parentally perceived temperament. The results may

be confounded by genetic and environmental factors involved in parental bias

and attitudes toward ratings of child temperament and behavior. By using the

same rater at two time points, the temperamental prediction of child and

adolescent behavior problems could to some extent reflect shared rater bias.

However, Bates and Bayles (1988) found independent temperamental

prediction of preschool behavior problems after controlling for potential

confounding variables such as maternal personality and observed mother-child

interaction. Using the EASI scale (Buss and Plomin, 1975), Lyon and Plomin

(1981) found no evidence of parental projection while Stevenson and Fielding

(1985) reported that maternal ratings of the emotionality scale were subject to

maternal projections. This was supported by later analyses with more specific

models (Neale and Stevenson, 1989). Thus results from studies investigating

the influence of parental projection on child temperament differ.

The [T.sub.1] sample comprised only about 60% of the total twin population

in these cohorts. The representativeness of the [T.sub.1] sample has been

discussed elsewhere (Gjone and Novik, 1995; Gjone et al., 1996). A possible

loss of disturbed children through [T.sub.1] and [T.sub.2] nonresponse may

have left genetic and environmental associations for severely disturbed

children undetected.

The present findings support other studies in emphasizing temperament,

particularly negative emotionality, as an important factor in the development of

behavior problems. As discussed by Rutter (1989), developmental pathways

from temperament to behavioral disturbance may involve different

mechanisms. For Anxious/Depressed behavior no significant genetic influence

was found, indicating temperamental poor fit between child and caretaker or

temperament as moderator of stressful experiences as possible mechanisms.

For Aggressive Behavior and Attention Problems, the present findings

indicate genetic influences on the association with negative emotionality,

suggesting that constitutional factors or gene-environment covariance to some

extent account for these temperament-behavior continuities.

In clinical work with behaviorally disturbed children, the temperament

perspective is beneficial for caretakers in developing coping strategies. The

possible variability in mechanisms involved in the development of behavior

problems, based on interaction between temperamental vulnerability and

external stressful experiences, must be acknowledged. Several authors

suggest two components of difficult temperament or negative emotionality

related more specifically to fear or anger/frustration (Bates, 1989; Buss, 1989;

Derryberry and Reed, 1994). If the temperament-behavior associations are

characterized by different mechanisms as indicated in the present study,

children with vulnerability toward aggressive problems may benefit from

increased family awareness of their tendency to elicit or be involved in

conflictual situations. Children with temperamental vulnerability toward

anxious/depressed behavior may benefit more from an increased focus on

cognitive and emotional coping in stressful situations. To elaborate further the

mechanisms involved in temperament-behavior associations, more specific

concepts than negative emotionality or difficult temperament are needed.

This work is supported by grants from the Norwegian Research Council, the

Norwegian Council for Mental Health, Anders Jahres Foundation, Solveig

and Johan P. Sommers Foundation, and the Norwegian Centre for Child

Research. The Medical Birth Register of Norway is acknowledged for

providing the twin sample.

Achenbach TM (1991), Manual for the Child Behavior Checklist/4-18 and

1991 Profile. Burlington: University of Vermont Department of Psychiatry

Achenbach TM, Edelbrock C (1983), Manual for the Child Behavior

Checklist and Revised Child Behavior Profile. Burlington: University of

Vermont Department of Psychiatry

Bates JE (1989), Concepts and measures of temperament. In: Temperament in

Childhood, Kohnstamm GA, Bates JE, Rothbart MK, eds. Chichester,

England: Wiley, pp 3-26

Bates JE, Bayles K (1988), Attachment and the development of behavior

problems. In: Clinical Implications of Attachment, Belsky J, Nezworski T,

eds. Hillsdale, NJ: Erlbaum, pp 253-299

Buss AH (1989), Temperament as personality traits. In: Temperament in

Childhood, Kohnstamm GA, Bates JE, Rothbart MK, eds. Chichester,

England: Wiley, pp 321-355

Buss AH, Plomin R (1975), A Temperament Theory of Personality

Development. New York: Wiley Series in Behavior

Buss AH, Plomin R (1984), Temperament: Early Developing Personality

Traits. Hillsdale, NJ: Erlbaum

Caspi A, Henry B, McGee RO, Moffitt TE, Silva PA (1995), Temperamental

origins of child and adolescent behavior problems: from age three to age

fifteen. Child Dev 66:55-68

Cohen DJ, Dibble E, Grawe JM, Pollin W (1973), Separating identical from

fraternal twins. Arch Gen Psychiatry 29:465-469

DeFries JC, Fulker DW (1985), Multiple regression analysis of twin data.

Behav Genet 15:467-473

DeFries JC, Fulker DW (1988), Multiple regression analysis of twin data:

etiology of deviant scores versus individual differences. Acta Genet Med

Gemellol (Roma) 37:205-216

Derryberry D, Reed MA (1994), Temperament and the self-organization of

personality. Der Psychopathol 6:653-676

Edelbrock C, Costello A J (1988), Convergence between statistically derived

behavior problem syndromes and child psychiatric diagnoses. J Abnorm

Child Psychol 16:219-231

Gibbs MV, Reeves D, Cunningham CC (1987), The application of

temperament questionnaires to a British sample: issues of reliability and

validity. J Child Psychol Psychiatry 28:61-77

Gjone H, Novik TS (1995), Parental ratings of behaviour problems: a twin

and general population comparison. J Child Psychol Psychiatry 36:1213-1224

Gjone H, Stevenson J, Sunder JM, Eilertsen DE (1996), Changes in

heritability across increasing levels of behavior problems in young twins.

Behav Genet 26:419-426

Goodyer IM, Ashby L, Altham PME, Vize C, Cooper PJ (1993),

Temperament and major depression in 11 to 16 year olds. J Child Psychol

Psychiatry 34:1409-1423

Gould MS, Bird H, Jaramillo BS (1993), Correspondence between

statistically derived behavior problem syndromes and child psychiatric

diagnoses in a community sample. J Abnorm Child Psychol 21:287-313

Graham P, Stevenson J (1987), Temperament and psychiatric disorder: the

genetic contribution to behaviour in childhood. Aust N Z J Psychiatry

21:267-274

LaBuda MC, DeFries JC, Fulker DW (1986), Multiple regression analysis of

twin data obtained from selected samples. Genet Epidemiol 3:425-433

Lyon ME, Plomin R (1981), The measurement of temperament using parental

ratings. J Child Psychol Psychiatry 22:47-53

Maziade M (1989), Should adverse temperament matter to the clinician? An

empirically based answer. In: Temperament in Childhood, Kohnstamm GA,

Bates JE, Rothbart MK, eds. Chichester, England: Wiley, pp 421-435

Neale MC, Stevenson J (1989), Rater bias in the EASI temperament scales: a

twin study. J Pers Soc Psychol 56:446-455

Oberklaid F, Sanson A, Pedlow R, Prior M (1993), Predicting preschool

behavior problems from temperament and other variables in infancy.

Pediatrics 91:113-120

Prior M, Smart D, Sanson A, Pedlow R, Oberklaid F (1992), Transient

versus stable behavior problems in a normative sample: infancy to school age.

J Pediatr Psychol 17 : 423-443

Rende RD (1993), Longitudinal relations between temperament traits and

behavioral syndromes in middle childhood. J Am Acad Child Adolesc

Psychiatry 32:287-290

Rende RD, Plomin R (1992), Relations between first grade stress,

temperament, and behavior problems. J Appl Der Psychol 13:435-446

Rutter M (1989), Temperament: conceptual issues and clinical implications.

In: Temperament in Childhood, Kohnstamm GA, Bates JE, Rothbart MK,

eds. Chichester, England: Wiley, pp 463-379

Rutter M, Quinton D (1984), Parental psychiatric disorder: effects on

children. Psychol Med 14:853-880

Shekim WO, Cantwell DP, Kashani J, Beck N, Martin J, Rosenberry J

(1986), Dimensional and categorical approaches to the diagnosis of attention

deficit disorder in children. J Am Acad Child Psychiatry 25:653-658

Steingard R, Biederman J, Doyle A, Sprich-Buckminster S (1992),

Psychiatric comorbidity in attention deficit disorder: impact on the

interpretation of Child Behavior Checklist results. J Am Acad Child Adolesc

Psychiatry 31:449-454

Stevenson J, Fielding J (1985), Ratings of temperament in families of young

twins. Br J Dev Psychol 3:143-152

Stevenson J, Graham P (1982), Temperament: a consideration of concepts

and methods. In: Temperamental Differences in Infants and Young Children

(Ciba Foundation Symposium 89). London: Pitman Books Ltd, pp 36-50

Thomas A, Chess S (1977), Temperament and Development. New York:

Brunner/Mazel

Thomas A, Chess S, Birch HG (1968), Temperament and Behavior Disorders

in Children. New York: New York University Press

Wolfson J, Fields JH, Rose SA (1987), Symptoms, temperament, resiliency,

and control in anxiety-disordered preschool children. J Am Acad Child

Adolesc Psychiatry 26:16-22

Dr. Helene Gjone is with the National Centre for Child and Adolescent

Psychiatry, University of Oslo, Norway. Dr. Jim Stevenson is Professor of

Psychology, University of Southampton, U.K.

Wray Herbert.

Abstract: Mental illness, criminal behavior, and other problems traditionally associated with environmental factors are being ascribed biological or genetic origins. This shift in thinking has become very politicized and is influencing policymaking.

Laurie Flynn uses the technology of neuroscience to light up the brains of Washington lawmakers. As executive director of the National Alliance for the Mentally Ill, she marshals everything from cost analysis to moral pleading to make the case for laws banning discrimination against people with mental illness. But her most powerful advocacy tool by far is the PET scan. She takes a collection of these colorful brain images up to Capitol Hill to put on a show, giving lawmakers a window on a "broken" brain in action. "When they see that it's not some imaginary, fuzzy problem, but a real physical condition, then they get it: 'Oh, it's in the brain.' "

The view of mental illness as a brain disease has been crucial to the effort to destigmatize illnesses such as schizophrenia and depression. But it's just one example of a much broader biologizing of American culture that's been going on for more than a decade. For both political and scientific reasons--and it's often impossible to disentangle the two--everything from criminality to addictive disorders to sexual orientation is seen today less as a matter of choice than of genetic destiny. Even basic personality is looking more and more like a genetic legacy. Nearly every week there is a report of a new gene for one trait or another. Novelty seeking, religiosity, shyness, the tendency to divorce, and even happiness (or the lack of it) are among the traits that may result in part from a gene, according to new research.

This cultural shift has political and personal implications. On the personal level, a belief in the power of genes necessarily diminishes the potency of such personal qualities as will, capacity to choose, and sense of responsibility for those choices--if it's in your genes, you're not accountable. It allows the alcoholic, for example, to treat himself as a helpless victim of his biology rather than as a willful agent with control of his own behavior. Genetic determinism can free victims and their families of guilt--or lock them in their suffering.

On the political level, biological determinism now colors all sorts of public-policy debates on issues such as gay rights, health care, juvenile justice, and welfare reform. The effort to dismantle social programs is fueled by the belief that government interventions (the nurturing side in the nature-nurture debate) don't work very well--and the corollary idea that society can't make up for every unfortunate citizen's bad luck. It's probably no coincidence that the biologizing of culture has accompanied the country's shift to the political right, since conservatives traditionally are more dubious about human perfectability than are liberals. As Northeastern University psychologist Leon Kamin notes, the simplest way to discover someone's political leanings is to ask his or her view on genetics.

Even so, genetic determinism can have paradoxical consequences at times, leading to disdain rather than sympathy for the disadvantaged, and marginalization rather than inclusion. Cultural critics are beginning to sort out the unpredictable politics of biology, focusing on four traits: violence, mental illness, alcoholism, and sexual orientation.

The nature of violence. To get a sense of just how thorough--and how politicized--the biologizing of culture has been, just look at the issue of urban gang violence as it is framed today. A few years ago, Frederick Goodwin, then director of the government's top mental health agency, was orchestrating the so-called Federal Violence Initiative to identify inner-city kids at biological risk for criminal violence, with the goal of intervening with drug treatments for what are presumed to be nervous-system aberrations. Goodwin got himself fired for comparing aggressive young males with primates in the jungle, and the violence initiative died in the resulting furor. But even to be proposing such a biomedical approach to criminal justice shows how far the intellectual pendulum has swung toward biology.

The eugenics movement of the 1930s was fueled at least in part by a desire to get rid of habitual criminals, and many attempts have been made over the years to identify genetic roots for aggression, violence, and criminality. A 1965 study, for instance, found that imprisoned criminals were more likely than other people to have an extra Y chromosome (and therefore more male genes). The evidence linking this chromosomal aberration to crime was skimpy and tenuous, but politics often runs ahead of the evidence: Soon after, a Boston hospital actually started screening babies for the defect, the idea being to intervene early with counseling should personality problems become apparent. The screening was halted when further study showed that XYY men, while slightly less intelligent, were not unusually aggressive.

As with many psychopathologies, criminal aggression is difficult to define precisely for research. Indeed, crime and alcohol abuse are so entangled that it's often difficult to know whether genetic markers are associated with drinking, criminality--or something else entirely, like a personality trait. A 1993 National Research Council study, for example, reported strong evidence of genetic influence on antisocial personality disorder, but it also noted that many genes are probably involved. Getting from those unknown genes to an actual act of vandalism or assault--or a life of barbaric violence--requires at this point a monstrous leap of faith.

Yet it's a leap that many are willing to make. When geneticist Xandra Breakefield reported a possible genetic link to violent crime a few years ago, she immediately started receiving phone inquiries from attorneys representing clients in prison; they were hoping that such genetic findings might absolve their clients of culpability for their acts.

Mutations and emotions. Just two decades ago, the National Institute of Mental Health was funding studies of economic recession, unemployment, and urban ills as possible contributors to serious emotional disturbance. A whole branch of psychiatry known as "social psychiatry" was dedicated to helping the mentally ill by rooting out such pathogens as poverty and racism. There is no longer much evidence of these sensibilities at work today. NIMH now focuses its studies almost exclusively on brain research and on the genetic underpinnings of emotional illnesses.

The decision to reorder the federal research portfolio was both scientific and political. Major advances in neuroscience methods opened up research that wasn't possible a generation ago, and that research has paid off in drugs that very effectively treat some disorders. But there was also a concerted political campaign to reinterpret mental illness. A generation ago, the leading theory about schizophrenia was that this devastating emotional and mental disorder was caused by cold and distant mothering, itself the result of the mother's unconscious wish that her child had never been born. A nationwide lobbying effort was launched to combat such unfounded mother blaming, and 20 years later that artifact of the Freudian era is entirely discredited. It's widely accepted today that psychotic disorders are brain disorders, probably with genetic roots.

But this neurogenetic victory may be double edged. For example, family and consumer groups have argued convincingly that schizophrenia is a brain disease like epilepsy, one piece of evidence being that it is treatable with powerful antipsychotic drugs. Managed-care companies, however, have seized upon the disease model, and now will rarely authorize anything but drug treatment: it's efficient, and justified by the arguments of biological psychiatry. The American Psychiatric Association just this month issued elaborate guidelines for treating schizophrenia, including not only drugs but an array of psychosocial services--services the insurance industry is highly unlikely to pay for.

The search for genes for severe mental disorders has been inconclusive. Years of studies of families, adoptees, and twins separated at birth suggest that both schizophrenia and manic-depressive illness run in families. But if that family pattern is the result of genes, it's clearly very complicated, because most of the siblings of schizophrenics (including half of identical twins, who have the same genes) don't develop the disorder. Behavioral geneticists suspect that several genes may underlie the illness, and that some environmental stress--perhaps a virus or birth complications--also might be required to trigger the disorder.

On several occasions in the past, researchers have reported "linkages" between serious mental illness and a particular stretch of DNA. A well-known study of the Amish, for example, claimed a link between manic-depression and an aberration on chromosome 11. But none of these findings has held up when other researchers attempted to replicate them.

Even if one accepts that there are genetic roots for serious delusional illnesses, critics are concerned about the biologizing of the rest of psychiatric illness. Therapists report that patients come in asking for drugs, claiming to be victims of unfortunate biology. In one case, a patient claimed he could "feel his neurons misfiring"; it's an impossibility, but the anecdote speaks to the thorough saturation of the culture with biology.

Some psychiatrists are pulling back from the strict biological model of mental illness. Psychiatrist Keith Russell Ablow has reintroduced the idea of "character" into his practice, telling depressed patients that they have the responsibility and capacity to pull themselves out of their illness. Weakness of character, as Ablow sees it, allows mental illness to grow. Such sentiment is highly controversial within psychiatry, where to suggest that patients might be responsible for some of their own suffering is taboo.

Besotted genes. The best that can be said about research on the genetics of alcoholism is that it's inconclusive, but that hasn't stopped people from using genetic arguments for political purposes. The disease model for alcoholism is practically a secular religion in this country, embraced by psychiatry, most treatment clinics, and (perhaps most important) by Alcoholics Anonymous. What this means is that those seeking help for excessive drinking are told they have a disease (though the exact nature of the disease is unknown), that it's probably a genetic condition, and that the only treatment is abstinence.

But the evidence is not strong enough to support these claims. There are several theories of how genes might lead to excessive drinking. A genetic insensitivity to alcohol, for example, might cause certain people to drink more; or alcoholics might metabolize alcohol differently; or they may have inherited a certain personality type that's prone to risk-taking or stimulus-seeking. While studies of family pedigrees and adoptees have on occasion indicated a familial pattern for a particular form of alcoholism (early-onset disorder in men, for example), just as often they reveal no pattern. This shouldn't be all that surprising, given the difficulty of defining alcoholism. Some researchers identify alcoholics by their drunk-driving record, while others focus on withdrawal symptoms or daily consumption. This is what geneticists call a "dirty phenotype"; people drink too much in so many different ways that the trait itself is hard to define, so family patterns are all over the place, and often contradictory.

Given these methodological problems, researchers have been trying to locate an actual gene (or genes) that might be involved in alcoholism. A 1990 study reported that a severe form of the disorder (most of the subjects in the study had cirrhosis of the liver) was linked to a gene that codes for a chemical receptor for the neurotransmitter dopamine. The researchers even developed and patented a test for the genetic mutation, but subsequent attempts to confirm the dopamine connection have failed.

The issues of choice and responsibility come up again and again in discussions of alcoholism and other addictive disorders. Even if scientists were to identify a gene (or genes) that create a susceptibility to alcoholism, it's hard to know what this genetic "loading" would mean. It certainly wouldn't lead to alcoholism in a culture that didn't condone drinking--among the Amish, for example--so it's not deterministic in a strict sense. Even in a culture where drinking is common, there are clearly a lot of complicated choices involved in living an alcoholic life; it's difficult to make the leap from DNA to those choices. While few would want to return to the time when heavy drinking was condemned as strictly a moral failing or character flaw, many are concerned that the widely accepted disease model of alcoholism actually provides people with an excuse for their destructive behavior. As psychologist Stanton Peele argues: "Indoctrinating young people with the view that they are likely to become alcoholics may take them there more quickly than any inherited reaction to alcohol would have."

Synapses of desire. It would be a mistake to focus only on biological explanations of psychopathology; the cultural shift is much broader than that. A generation ago, the gay community was at war with organized psychiatry, arguing (successfully) that sexual orientation was a lifestyle choice and ought to be deleted from the manual of disorders. Recently the same community was celebrating new evidence that homosexuality is a biological (and perhaps genetic) trait, not a choice at all.

Three lines of evidence support the idea of a genetic basis for homosexuality, none of them conclusive. A study of twins and adopted siblings found that about half of identical twins of homosexual men were themselves gay, compared with 22 percent of fraternal twins and 11 percent of adoptees; a similar pattern was found among women. While such a pattern is consistent with some kind of genetic loading for sexual orientation, critics contend it also could be explained by the very similar experiences many twins share. And, of course, half the identical twins did not become gay--which by definition means something other than genes must be involved.

A well-publicized 1991 study reported a distinctive anatomical feature in gay men. Simon LeVay autopsied the brains of homosexual men and heterosexual men and women and found that a certain nucleus in the hypothalamus was more than twice as large in heterosexual men as in gay men or heterosexual women. Although LeVay couldn't explain how this neurological difference might translate into homosexuality, he speculates that the nucleus is somehow related to sexual orientation. The hypothalamus is known to be involved in sexual response.

The only study so far to report an actual genetic connection to homosexuality is a 1993 study by Dean Hamer, a National Institutes of Health biologist who identified a genetic marker on the X chromosome in 75 percent of gay brothers. The functional significance of this piece of DNA is unknown, and subsequent research has not succeeded in duplicating Hamer's results.

Homosexuality represents a bit of a paradox when it comes to the intertwined issues of choice and determinism. When Hamer reported his genetic findings, many in the gay community celebrated, believing that society would be more tolerant of behavior rooted in biology and DNA rather than choice. LeVay, himself openly gay, says he undertook his research with the explicit agenda of furthering the gay cause. And Hamer testified as an expert witness in an important gay-rights case in Colorado where, in a strange twist, liberals found themselves arguing the deterministic position, while conservatives insisted that homosexuality is a choice. The argument of gay-rights advocates was that biological status conveyed legal status--and protection under the law.

History's warning. But history suggests otherwise, according to biologist and historian Garland Allen. During the eugenics movement of the 1920s and 1930s, both in the United States and Europe, society became less, not more, tolerant of human variation and misfortune. Based on racial theories that held Eastern Europeans to be genetically inferior to Anglo-Saxon stock, Congress passed (and Calvin Coolidge signed) a 1924 law to restrict immigration, and by 1940 more than 30 states had laws permitting forced sterilization of people suffering from such conditions as "feeblemindedness," pauperism, and mental illness. The ultimate outcome of the eugenics craze in Europe is well known; homosexuals were not given extra sympathy or protection in the Third Reich's passion to purify genetic stock.

Allen is concerned about the possibility of a "new eugenics" movement, though he notes that it wouldn't be called that or take the same form. It would more likely take the form of rationing health care for the unfortunate. The economic and social conditions today resemble conditions that provided fertile ground for eugenics between the wars, he argues; moreover, in Allen's view, California's Proposition 187 recalls the keen competition for limited resources (and the resulting animosity toward immigrants) of the '20s. Further, Allen is quick to remind us that eugenics was not a marginal, bigoted movement in either Europe or the United States; it was a Progressive program, designed to harness science in the service of reducing suffering and misfortune and to help make society more efficient.

These concerns are probably justified, but there are also some signs that we may be on the crest of another important cultural shift. More and more experts, including dedicated biologists, sense that the power of genetics has been oversold and that a correction is needed. What's more, there's a glimmer of evidence that the typical American may not be buying it entirely. According to a recent U.S. News/Bozell poll, less than 1 American in 5 believes that genes play a major role in controlling behavior; three quarters cite environment and society as the more powerful shapers of our lives. Whether the behavior under question is a disorder like addiction, mental illness, or violence, or a trait like homosexuality, most believe that heredity plays some role, but not a primary one. Indeed, 40 percent think genes play no role whatsoever in homosexuality, and a similar percentage think heredity is irrelevant to drug addiction and criminality. Across the board, most believe that people's lives are shaped by the choices they make.

These numbers can be interpreted in different ways. It may be that neurogenetic determinism has become the "religion of the intellectual class," as one critic argues, but that it never really caught the imagination of the typical American. Or we may be witnessing a kind of cultural self-correction, in which after a period of infatuation with neuroscience and genetics the public is becoming disenchanted, or perhaps even anxious about the kinds of social control that critics describe.

Whatever's going on, it's clear that this new mistrust of genetic power is consonant with what science is now beginning to show. Indeed, the very expression "gene for" is misleading, according to philosopher Philip Kitcher, author of The Lives to Come. Kitcher critiques what he calls "gene talk," a simplistic shorthand for talking about genetic advances that has led to the widespread misunderstanding of DNA's real powers. He suggests that public discourse may need to include more scientific jargon--not a lot, but some--so as not to oversimplify the complexity of the gene-environment interaction. For example, when geneticists say they've found a gene for a particular trait, what they mean is that people carrying a certain "allele"--a variation in a stretch of DNA that normally codes for a certain protein--will develop the given trait in a standard environment. The last few words--"in a standard environment"--are very important, because what scientists are not saying is that a given allele will necessarily lead to that trait in every environment. Indeed, there is mounting evidence that a particular allele will not produce the same result if the environment changes significantly; that is to say, the environment has a strong influence on whether and how a gene gets "expressed."

It's hard to emphasize too much what a radical rethinking of the nature-nurture debate this represents. When most people think about heredity, they still think in terms of classical Mendelian genetics: one gene, one trait. But for most complex human behaviors, this is far from the reality that recent research is revealing. A more accurate view very likely involves many different genes, some of which control other genes, and many of which are controlled by signals from the environment. To complicate matters further, the environment is very complicated in itself, ranging from the things we typically lump under nurture (parenting, family dynamics, schooling, safe housing) to biological encounters like viruses and birth complications, even biochemical events within cells.

The relative contributions of genes and the environment are not additive, as in such-and-such a percentage of nature, such-and-such a percentage of experience; that's the old view, no longer credited. Nor is it true that full genetic expression happens once, around birth, after which we take our genetic legacy into the world to see how far it gets us. Genes produce proteins throughout the lifespan, in many different environments, or they don't produce those proteins, depending on how rich or harsh or impoverished those environments are. The interaction is so thoroughly dynamic and enduring that, as psychologist William Greenough says, "To ask what's more important, nature or nurture, is like asking what's more important to a rectangle, its length or its width."

The emerging view of nature--nurture is that many complicated behaviors probably have some measure of genetic loading that gives some people a susceptibility--for schizophrenia, for instance, or for aggression. But the development of the behavior or pathology requires more, what National Institute of Mental Health Director Stephen Hyman calls an environmental "second hit." This second hit operates, counterintuitively, through the genes themselves to "sculpt" the brain. So with depression, for example, it appears as though a bad experience in the world--for example, a devastating loss--can actually create chemical changes in the body that affect certain genes, which in turn affect certain brain proteins that make a person more susceptible to depression in the future. Nature or nurture? Similarly, Hyman's own work has shown that exposure to addictive substances can lead to biochemical changes at the genetic and molecular levels that commandeer brain circuits involving volition--and thus undermine the very motivation needed to take charge of one's destructive behavior. So the choice to experiment with drugs or alcohol may, in certain people, create the biological substrate of the addictive disorder. The distinction between biology and experience begins to lose its edge.

Nurturing potentials. Just as bad experiences can turn on certain vulnerability genes, rich and challenging experiences have the power to enhance life, again acting through the genes. Greenough has shown in rat studies that by providing cages full of toys and complex structures that are continually rearranged--"the animal equivalent of Head Start"--he can increase the number of synapses in the rats' brains by 25 percent and blood flow by 85 percent. Talent and intelligence appear extraordinarily malleable.

Child-development experts refer to the life circumstances that enhance (or undermine) gene expression as "proximal processes," a term coined by psychologist Urie Bronfenbrenner. Everything from lively conversation to games to the reading of stories can potentially get a gene to turn on and create a protein that may become a neuronal receptor or messenger chemical involved in thinking or mood. "No genetic potential can become reality," says Bronfenbrenner, "unless the relationship between the organism and its environment is such that it is permitted to be expressed." Unfortunately, as he details in his new book, The State of Americans, the circumstances in which many American children are living are becoming more impoverished year by year.

If there's a refrain among geneticists working today, it's this: The harder we work to demonstrate the power of heredity, the harder it is to escape the potency of experience. It's a bit paradoxical, because in a sense we end up once again with the old pre-1950s paradigm, but arrived at with infinitely more-sophisticated tools: Yes, the way to intervene in human lives and improve them, to ameliorate mental illness, addictions, and criminal behavior, is to enrich impoverished environments, to improve conditions in the family and society. What's changed is that the argument is coming not from left-leaning sociologists, but from those most intimate with the workings of the human genome. The goal of psychosocial interventions is optimal gene expression.

So assume for a minute that there is a cluster of genes somehow associated with youthful violence. The kid who carries those genes might inhabit a world of loving parents, regular nutritious meals, lots of books, safe schools. Or his world might be a world of peeling paint and gunshots around the corner. In which environment would those genes be likely to manufacture the biochemical underpinnings of criminality? Or for that matter, the proteins and synapses of happiness?

Mag.Coll.: 88F0689

Bus.Coll.: 100Y2077

(child personality development)

Marc Peyser; Anne Underwood.

Abstract: Researchers have found that genes play a large role in shaping a child's emotional makeup, but a child's personality traits are also profoundly affected by his or her environment. Genetic and environmental factors combine in complex ways to shape a child's psychological development.

The wizards of genetics keep closing in on the biological roots of personality. It's not your imagination that one baby seems born cheerful and another morose. But that's not the complete picture. DNA is not destiny;, experience plays a powerful role, too.

IF ANY CHILD SEEMED DESTINED TO GROW UP afraid of her shadow and just about anything else that moved, it was 2-year-old Marjorie. She was so painfully shy that she wouldn't talk to or look at a stranger. She was even afraid of friendly cats and dogs. When Jerome Kagan, a Harvard professor who discovered that shyness has a strong genetic component, sent a clown to play with Marjorie, she ran to her mother. "It was as if a cobra entered that room;' Kagan says. His diagnosis: Marjorie showed every sign of inherited shyness, a condition in which the brain somehow sends out messages to avoid new experiences. But as Kagan continued to examine her over the years, Marjorie's temperament changed. When she started school, she gained confidence from ballet classes and her good grades, and she began to make friends. Her parents even coaxed her into taking horseback-riding lessons. Marjorie may have been born shy, but she has grown into a bubbly second grader.

For Marjorie, then, biology--more specifically, her genetic inheritance-- was not her destiny. And therein lies our tale. In the last few years scientists have identified genes that appear to predict all sorts of emotional behavior, from happiness to aggressiveness to risk-taking. The age-old question of whether nature or nurture determines temperament seems finally to have been decided in favor of Mother Nature and her ever-deepening gene pool. But the answer may not be so simple after all. Scientists are beginning to discover that genetics and environment work together to determine personality as intricately as Astaire and Rogers danced. "If either Fred or Ginger moves too fast, they both stumble," says Stanley Greenspan, a pediatric psychiatrist at George Washington University and the author of "The Growth of the Mind." "Nature affects nurture affects nature and back and forth. Each step influences the next." Many scientists now believe that some experiences ,:an actually alter the structure of the brain. An aggressive toddler, under the tight circumstances, can essentially be rewired to channel his energy more constructively. Marjorie can overcome her shyness--forever. No child need be held captive to her genetic blueprint. The implications for child rearing--and social policy- are profound.

While Gregor Mendel's pea plants did wonders to explain how humans inherit blue eyes or a bald spot, they turn out to be an inferior model for analyzing something as complex as the brain. The human body contains about 100,000 genes, of which 50,000 to 70,000 are involved in brain function. Genes control the brain's neurotransmitters and receptors, which deliver and accept mental messages like so many cars headed for their assigned parking spaces. But there are billions of roads to each parking lot, and those paths are highly susceptible to environmental factors. In his book "The New View of Self," Dr. Larry Siever, a psychiatry professor at Mount Sinai Medical Center, writes about how the trauma of the Holocaust caused such intense gene fir: scrambling in some survivors that their children inherited the same stress-related abnormalities. "Perhaps the sense of danger and uncertainty associated with living through such a time is passed on in the family milieu and primes the biological systems of the children as well," says Siever. He added that that might explain why pianist David Helfgott, the subject of the movie "Shine," had his mental breakdown.

A gene is only a probability for a given trait, not a guarantee. For that trait to be expressed, a gene often must be "turned on" by an outside force before it does its job. High levels of stress apparently activate a variety of genes, including those suspected of being involved in fear, shyness and some mental illnesses. Children conceived during a three-month famine in the Netherlands during a Nazi blockade in 1945 were later found to have twice the rate of schizophrenia as did Dutch children born to parents who were spared the trauma of famine. "Twenty years ago, you couldn't get your research funded if you were looking for a genetic basis for schizophrenia, because everyone knew it was what your mother did to you in the first few years of life, as Freud said," says Robert Plomin, a geneticist at London's Institute of Psychiatry. "Now you can't get funded unless you're looking for a genetic basis. Neither extreme is right, and the data show why. There's only a 50 percent concordance between genetics and the development of schizophrenia."

SCIENTISTS HAVE BEEN DEVOTING enormous energy to determining what part of a given character trait is "heritable" and what part is the result of socialization. Frank Sulloway's book "Born to Rebel," which analyzes the influence of birth order on personality, opened a huge window on a universal--and largely overlooked-environmental factor. But that's a broad brushstroke. Most studies focus on remarkably precise slivers of human emotions. One study at Allegheny University in Pennsylvania found that the tendency for a person to throw dishes or slam doors when he's angry is 40 percent heritable, while the likelihood a person will yell in anger is only 28 percent heritable. The most common method for determining these statistics is studying twins. If identical twins are more alike in some way than are fraternal twins, that trait is believed to have a higher likelihood of being inherited. But the nature-nurture knot is far from being untied.

The trick, then, is to isolate a given gene and study the different ways environment interacts with it. For instance, scientists believe that people with the longer variety of a dopamine-4 receptor gene are biologically predisposed to be thrill seekers. Because the gene appears to make them less sensitive to pain and physical sensation, the children are more likely to, say, crash their tricycles into a wall, just to see what it feels like. "These are the daredevils," says Greenspan. But they need not be. Given strict boundaries, Greenspan says, thrill-seeking kids can be taught to modulate and channel their overactive curiosity. A risk-taking child who likes to pound his fist into hard objects can be taught games that involve hitting softly as well. "If you give them constructive ways to meet their needs," says Greenspan, "they can become charismatic, action-oriented leaders."

Shyness has been studied perhaps more than any other personality trait. Kagan, who has monitored 500 children for more than 17 years at Harvard, can detect telltale signs of shyness in babies even before they're born. He's found that the hearts of shy children in the womb consistently beat faster than 140 times a minute, which is much faster than the heartbeats of other babies. The shy fetus is already highly recan become charismatic, action-oriented leaders."

Shyness has been studied perhaps more than any other personality trait. Kagan, who has monitored 500 children for more than 17 years at Harvard, can detect telltale signs of shyness in babies even before they're born. He's found that the hearts of shy children in the womb consistently beat faster than 140 times a minute, which is much faster than the heartbeats of other babies. The shy fetus is already highly reactive, wired to overmonitor his environment. But he can also outgrow this predisposition if his parents gently but firmly desensitize him to the situations that cause anxiety, such as encouraging him to play with other children or, as in Marjorie's fear of animals, taking her to the stables and teaching her to ride a horse. Kagan has found that by the age of 4, no more than 20 percent of the previously shy children remain that way.

Will the reprogramming last into adulthood? Bemuse evidence of the role of genes has been discovered only recently, it's still too early to tell. But studies of animals give some indication. Stephen Suomi at the National Institute of Child Health and Human Development works with rhesus monkeys that possess the same genetic predisposition to shyness that affects humans. He's shown that by giving a shy monkey to a foster mother who is an expert caregiver, the baby will outgrow the shyness. Even more surprising, the once shy monkey will become a leader among her peers and an unusually competent parent, just like the foster mom. Though she will likely pass along her shyness genes to her own child, she will teach it how to overcome her predisposition, just as she was taught. And the cycle continues--generations of genetically shy monkeys become not just normal, but superior, adults and parents. The lesson, says Suomi: "You can't prejudge anyone at birth. No matter what your genetic background, a negative characteristic you're born with may even turn out to be an advantage."

But parents aren't scientists, and it's not always easy to see how experience can influence a child's character. A baby who smiles a lot and makes eye contact is, in part, determining her own environment, which in turn affects her temperament. As her parents coo and smile and wrinkle their noses in delighted response, they are reinforcing their baby's sunny disposition. But what about children who are born with low muscle tone, who at 4 months can barely hold up their own heads, let alone smile? Greenspan has discovered that mothers of these kids smile at the baby for a while, but when the affection isn't returned, they give up. And so does the baby, who over time fails to develop the ability to socialize normally. "If you move in the wrong patterns, the problem is exacerbated," Greenspan says. He has found that if parents respond to nonsmiling babies by being superanimated-like Bob Barker hosting a game show--they can engage their child's interest in the world.

The ramifications of these findings clearly have the potential to revolutionize child-rearing theory and practice. But to an uncertain end. "Our society has a strong belief that what happens in childhood determines your fate. If you have a happy childhood, everything will be all right. That's silly," says Michael Lewis, director of the Institute for the Study of Child Development in New Jersey and the author of "Altering Fate." Lewis estimates that experience ultimately rewrites 90 percent of a child's personality traits, leaving an adult with only one tenth of his inborn temperament. "The idea that early childhood is such a powerful moment to see individual differences in biology or environment is not valid," he says. "We are too open to and modifiable by experience." Some scientists warn that attempting to reprogram even a narrow sliver of childhood emotions can prove to be a daunting task, despite research's fascinating new insights. "Children are not a 24-hour controlled experiment," says C. Robert Cloninger, a professor of psychiatry and genetics at the Washington University School of Medicine in St. Louis. "If you put a child in a Skinner box, then maybe you could have substantial influence." So, mindful of the blinding insights of geneticists and grateful for the lingering influences of environment, parents must get on with the business of raising their child, an inexact science if ever there was one. Article A19324386

Linda Gottfredson.

Abstract: Highly regarded scholars from various social science disciplines have voiced their support for a collective statement called "Mainstream Science on Intelligence". The experts coming from diverse fields such as anthropology, behavior genetics, neuropsychology, sociology, and psychology agree that all 25 conclusions that comprise the statement are mainstream issues. The statement discusses the nature, origin, and practical consequences of individual and group differences in intelligence. Fifty two out of 100 experts signed the statement indicating support for its validity.

The following statement was first published in the Wall Street Journal, December 13, 1994.

Since the publication of "The Bell Curve," many commentators have offered opinions about human intelligence that misstate current scientific evidence. Some conclusions dismissed in the media as discredited are actually firmly supported.

This statement outlines conclusions regarded as mainstream among researchers on intelligence, in particular, on the nature, origins, and practical consequences of individual and group differences in intelligence. Its aim is to promote more reasoned discussion of the vexing phenomenon that the research has revealed in recent decades. The following conclusions are fully described in the major textbooks, professional journals and encyclopedias in intelligence.

The Meaning and Measurement of Intelligence

1. Intelligence is a very general mental capability that, among other things, involves the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly and learn from experience. It is not merely book learning, a narrow academic skill, or test-taking smarts. Rather, it reflects a broader and deeper capability for comprehending our surroundings--"catching on," "making sense" of things, or "figuring out" what to do.

2. Intelligence, so defined, can be measured, and intelligence tests measure it well. They are among the most accurate (in technical terms, reliable and valid) of all psychological tests and assessments. They do not measure creativity, character, personality, or other important differences among individuals, nor are they intended to.

3. While there are different types of intelligence tests, they all measure the same intelligence. Some use words or numbers and require specific cultural knowledge (like vocabulary). Other do not, and instead use shapes or designs and require knowledge of only simple, universal concepts (many/few, open/closed, up/down).

4. The spread of people along the IQ continuum, from low to high, can be represented well by the bell curve (in statistical jargon, the "normal curve"). Most people cluster around the average (IQ 100). Few are either very bright or very dull: About 3% of Americans score above IQ 130 (often considered the threshold for "giftedness"), with about the same percentage below IQ 70 (IQ 70-75 often being considered the threshold for mental retardation).

5. Intelligence tests are not culturally biased against American blacks or other native-born, English-speaking peoples in the U.S. Rather, IQ scores predict equally accurately for all such Americans, regardless of race and social class. Individuals who do not understand English well can be given either a nonverbal test or one in their native language.

6. The brain processes underlying intelligence are still little understood. Current research looks, for example, at speed of neural transmission, glucose (energy) uptake, and electrical activity of the brain.

7. Members of all racial-ethnic groups can be found at every IQ level. The bell curves of different groups overlap considerably, but groups often differ in where their members tend to cluster along the IQ line. The bell curves for some groups (Jews and East Asians) are centered somewhat higher than for whites in general. Other groups (blacks and Hispanics) are centered somewhat lower than non-Hispanic whites.

8. The bell curve for whites is centered roughly around IQ 100; the bell curve for American blacks roughly around 85; and those for different subgroups of Hispanics roughly midway between those for whites and blacks. The evidence is less definitive for exactly where above IQ 100 the bell curves for Jews and Asians are centered.

9. IQ is strongly related, probably more so than any other single measurable human trait, to many important educational, occupational, economic, and social outcomes. Its relation to the welfare and performance of individuals is very strong in some arenas in life (education, military training), moderate but robust in others (social competence), and modest but consistent in others (law-abidingness). Whatever IQ tests measure, it is of great practical and social importance.

10. A high IQ is an advantage in life because virtually all activities require some reasoning and decision-making. Conversely, a low IQ is often a disadvantage, especially in disorganized environments. Of course, a high IQ no more guarantees success than a low IQ guarantees failure in life. There are many exceptions, but the odds for success in our society greatly favor individuals with higher IQs.

11. The practical advantages of having a higher IQ increase as life settings become more complex (novel, ambiguous, changing, unpredictable, or multifaceted). For example, a high IQ is generally necessary to perform well in highly complex or fluid jobs (the professions, management); it is a considerable advantage in moderately complex jobs (crafts, clerical and police work); but it provides less advantage in settings that require only routine decision making or simple problem solving (unskilled work).

12. Differences in intelligence certainly are not the only factor affecting performance in education, training, and highly complex jobs (no one claims they are), but intelligence is often the most important. When individuals have already been selected for high (or low) intelligence and so do not differ as much in IQ, as in graduate school (or special education), other influences on performance loom larger in comparison.

13. Certain personality traits, special talents, aptitudes, physical capabilities, experience, and the like are important (sometimes essential) for successful performance in many jobs, but they have narrower (or unknown) applicability or "transferability" across tasks and settings compared with general intelligence. Some scholars choose to refer to these other human traits as other "intelligences."

Source and Stability of Within-Group Differences

14. Individuals differ in intelligence due to differences in both their environments and genetic heritage. Heritability estimates range from 0. 4 to 0. 8 (on a scale from 0 to 1), most thereby indicating that genetics plays a bigger role than does environment in creating IQ differences among individuals. (Heritability is the squared correlation of phenotype with genotype.) If all environments were to become equal for everyone, heritability would rise to 100% because all remaining differences in IQ would necessarily be genetic in origin.

15. Members of the same family also tend to differ substantially in intelligence (by an average of about 12 IQ points) for both genetic and environmental reasons. They differ genetically because biological brothers and sisters share exactly half their genes with each parent and, on the average, only half with each other. They also differ in IQ because they experience different environments within the same family.

16. That IQ may be highly heritable does not mean that it is not affected by the environment. Individuals are not born with fixed, unchangeable levels of intelligence (no one claims they are). IQs do gradually stabilize during childhood, however, and generally change little thereafter.

17. Although the environment is important in creating IQ differences, we do not know yet how to manipulate it to raise low IQs permanently. Whether recent attempts show promise is still a matter of considerable scientific debate.

18. Genetically caused differences are not necessarily irremediable (consider diabetes, poor vision, and phenylketonuria), nor are environmentally caused ones necessarily remediable (consider injuries, poisons, severe neglect, and some diseases). Both may be preventable to some extent.

19. There is no persuasive evidence that the IQ bell curves for different racial-ethnic groups are converging. Surveys in some years show that gaps in academic achievement have narrowed a bit for some races, ages, school subjects and skill levels, but this picture seems too mixed to reflect a general shift in IQ levels themselves.

20. Racial-ethnic differences in IQ bell curves are essentially the same when youngsters leave high school as when they enter first grade. However, because bright youngsters learn faster than slow learners, these same IQ differences lead to growing disparities in amount learned as youngsters progress from grades one to 12. As large nation al surveys continue to show, black 17-year-olds perform, on the average, more like white 13-year-olds in reading, math, and science, with Hispanics in between.

21. The reasons that blacks differ among themselves in intelligence appear to be basically the same as those for why whites (or Asians or Hispanics) differ among themselves. Both environment and genetic heredity are involved.

22. There is no definitive answer to why IQ bell curves differ across racial-ethnic groups. The reasons for these IQ differences between groups may be markedly different from the reasons for why individuals differ among themselves within any particular group (whites or blacks or Asians). In fact, it is wrong to assume, as many do, that the reason why some individuals in a population have high IQs but others have low IQs must be the same reason why some populations contain more such high (or low) IQ individuals than others. Most experts believe that environment is important in pushing the bell curves apart, but that genetics could be involved too.

23. Racial-ethnic differences are somewhat smaller but still substantial for individuals from the same socioeconomic backgrounds. To illustrate, black students from prosperous families tend to score higher in IQ than blacks from poor families, but they score no higher, on average, than whites from poor families.

24. Almost all Americans who identify themselves as black have white ancestors--the white admixture is about 20%, on average--and many self-designated whites, Hispanics, and others likewise have mixed ancestry. Because research on intelligence relies on self-classification into distinct racial categories, as does most other social-science research, its findings likewise relate to some unclear mixture of social and biological distinctions among groups (no one claims otherwise).

25. The research findings neither dictate nor preclude any particular social policy, because they can never determine our goals. They can, however, help us estimate the likely success and side-effects of pursuing those goals via different means.

Rarely do scientists join in making statements to the public about the state of their discipline. As a rule, they do not readily agree among themselves or speak in the public arena.

There is, of course, no dearth of public pronouncements from scientific associations and committees. It is unusual, however, for a broad spectrum of unaffiliated (and often unacquainted) scientists to issue a public statement (see Page, 1972, for an example concerning human heredity). It is unprecedented that one should coalesce as quickly as did the "Mainstream" statement. A fuller understanding of this event is provided by recounting its origins.

The controversy over The Bell Curve (Herrnstein & Murray, 1994) was at its height in the fall of 1994. Many critics attacked the book for supposedly relying on outdated, pseudoscientific notions of intelligence. In criticizing the book, many critics promoted false and highly misleading views about the scientific study of intelligence. Public miseducation on the topic is hardly new (Snyderman & Rothman, 1987, 1988), but never before had it been so angry and extreme.

I therefore approached the editorial features editor, David Brooks, at the Wall Street Journal to see if he would be interested in my writing an essay on the rising crescendo of misinformation on intelligence. He was not. He said he would, however, consider a short statement signed by 10 to 15 experts on what knowledge they do, in fact, consider to be mainstream in the study of intelligence. Timeliness required that any statement be submitted within 2 weeks.

Invitations

In the next few days, I drafted a statement that addressed the most common claims and misconceptions in the public media, whether in book reviews, opinion pieces, letters to the editor, or in TV and radio commentary. I wanted to fashion a primer of sorts by outlining briefly the most basic, well-accepted conclusions in the field. The draft was faxed to half a dozen leaders in the field (including the editor of the journal Intelligence), with a request that they review its accuracy and suggest revisions. I also solicited comments on the draft's comprehensibility from several nonexperts.

In the meantime, I compiled a list of experts who could be invited to sign the statement. The aim was to gather a large group of highly knowledgeable researchers who represented a wide spectrum of disciplines and perspectives in the scientific study of intelligence. Names were obtained from four sources: (1) lists of individuals elected as fellows (for their distinguished contributions to psychology) by relevant divisions of the American Psychological Association such as educational psychology; school psychology; industrial and organizational psychology; and evaluation, measurement, and statistics; (2) lists of editorial board members of Intelligence; (3) tables of contents of books and journals devoted to the science of intelligence; and (4) suggestions from other people more knowledgeable than I am about some of the subdisciplines in the study of intelligence. The final list ranged from individuals I was sure would sign to those I was sure would not (I was sometimes wrong on both counts). I invited only academics, because nonacademic researchers are often constrained in the public statements their employers allow them to make. The experts represented a variety of disciplines, including anthropology, behavior genetics, mental retardation, neuropsychology, sociology, and various specialties in psychology such as psychometrics, child development, educational psychology, and personnel selection.

Early the next week, my assistant and I began faxing the statement to individuals for whom we could obtain fax numbers. My one-page letter recounted the Wall Street Journal editor's suggestion for such a statement and invited their signatures. Recipients were advised that the deadline for my submitting the signed statement to the Journal was that Friday at 5:00 p.m. and that the statement would also be published as a signed editorial in Intelligence. Invitees were given no opportunity to revise the statement. Nor was anyone told (and only one person asked) who else had been invited or who had already signed.

The letter of invitation asked recipients to return an accompanying signature form, regardless of whether they chose to sign it, so that I could confirm that the invitation had been received. We attempted to telephone all individuals from whom I did not receive a response within 24 to 48 hr.

No inferences can be drawn about who declined to sign the statement, because many worthy scholars were either inadvertently omitted from the list or were unavailable the week I attempted to contact them.

A total of 131 invitations was issued, and 100 responses were obtained by the deadline. The signature form asked respondents to check either "yes" or "no," and if "no," to check one of three options explaining why they declined to sign: "I don't agree that the statement represents the mainstream," "I don't know enough to say for sure," and "other reason." Many nonsigners wrote comments or letters explaining their decision. Those comments will be discussed here. No comments were solicited from signers, but about two thirds either telephoned or wrote brief comments; these were usually praise, appreciation, or rewordings they would have preferred.

Table 1 shows that, of the 100 individuals who responded, 48 declined to sign--7 because they thought the statement did not represent the mainstream, 11 because they did not know whether it did, and 30 for other reasons. The bottom panel of Table 1 categorizes the nonsigners (excluding the 11 individuals who "do not know enough") according to the major reason each gave for not signing the statement. It is clear that declining to sign the statement did not necessarily mean disagreement with it.

TABLE 1 Responses to Invitation to Sign "Mainstream" Statement

Responses From the Experts Successfully Contacted (N = 100)

Signed the statement 52 Decided not to sign the statement Statement does not represent the mainstream 7 Do not know enough to say 11 Other reasons 30 (Not located before deadline) (31)

Reason for Not Signing the Statement (N = 37)(a)

Disagreed with 1 or 2 specific items 3 Disagreed with 3-5 specific items 2(b) Disagreed with statement's conception of intelligence 4(c) Disagreed in general or vague way 2 Did not dispute content of statement, but disagreed 6 with its mode of presentation Agreed with statement, but feared that signing it 4 would jeopardize their position or project Mostly agreed with statement, but uncomfortable 4(b) being associated with it or potential signers Did not want to sign "at this time" 2 Gave no explanation 10

(a) Excludes the 11 individuals who "do not know enough."

(b) Two individuals marked "does not represent mainstream."

(c) Three individuals marked "does not represent mainstream."

Of the 27 who gave a reason, 11 explicitly disagreed with the content of the statement (or that its claims are "mainstream"). In three cases the individuals disagreed with only 1 or 2 of the 25 items. Two disagreed with 3 to 5 items, another 4 disputed the concept of general intelligence itself ("it is not a useful concept"), and 2 expressed nonspecific disagreement ("I agree with part but not all," "much ... is oversimplified, does not adequately represent what is known, and incorrect").

Fourteen individuals declined to sign the statement despite seeming to agree, sometimes strongly, that its content is "mainstream." Six of them disagreed with the way the statement was written (submitting that it did not mention enough complexities and qualifications) or how it was published (as a group statement) or where (a newspaper, nonscientific, or "conservative" outlet). Four nonsigners were specific about the possible political repercussions to them of signing it (such as loss of funding or other support). Another four expressed discomfort with the possibility of being caught up in controversy ("getting in a no-win fight") or seeming to associate with certain unnamed individuals ("about whom I have serious reservations"). Two other individuals, by stating that they "did not want to sign at this time," also seemed to signal that they agreed with the statement but thought it prudent not to endorse it.

"Mainstream Science on Intelligence" is a collective statement that was first issued in order to inject some scientific rigor into an increasingly vitriolic and wrongheaded controversy concerning intelligence. That it garnered such immediate support from so many highly regarded scholars testifies to their confidence both that it represents the mainstream and that their joint testimony to that effect was needed in the public realm.

No individual or group has systematically rebutted the statement. Some people might construe the 24-page "Intelligence: Knowns and Unknowns" (Neisser et al., 1996) to be an alternative. However, that report was the result of 6 months' work by an 11-member task force created by the American Psychological Association's Board of Scientific Affairs. (Three of the task force members were also signers of the "Mainstream" statement.) That report differs in purpose, emphasis, and degree of equivocation, but its conclusions only reinforce the claim that the contents of the "Mainstream" statement are squarely within the mainstream. It too concludes, for example, that differences in intelligence exist, can be measured fairly, are partly genetic (within races), and influence life outcomes.

It is obviously not the case that there is no disagreement about these important issues or that scientific truth is a matter of majority rule. A significant minority of the experts who were contacted disagreed in part or in whole with the statement, and many of the signers would have written the statement somewhat differently. Rather, the lesson here is that what have often been caricatured in the public press as discredited, fringe ideas actually represent the solid scientific center in the serious study of intelligence. As Snyderman and Rothman's (1988) survey of IQ experts and journalists revealed, the media, among others, have been turning the truth on its head.

Many of the conclusions outlined in "Mainstream" are ones that many scholars have reached only recently and reluctantly (Gottfredson, 1996). The mainstream shifted slowly but steadily in recent decades as accumulating research evidence changed our understanding of the nature, measurement, origins, and consequence of differences of intelligence. The press and public have yet to catch up to the new mainstream.

Social and political pressure, both internal and external to the field of intelligence, continues to make scholars reluctant to share their conclusions freely. Over one third of the individuals who declined to sign the "Mainstream" statement expressed reasons that signal such reluctance.

It is also understandable that some respondents wanted the statement's 25 items to be stated with a fuller account of their complexity. It is difficult for knowledgeable and precise scientists to make simple summary statements that do not do full justice to the topics they know so well, especially ones subject to controversy. Indeed, many books have been written about most of the individual items in the "Mainstream" statement. As a practical matter, people are more likely to reach consensus on general principles than highly particular ones. More importantly, it is sometimes wiser to focus on the forest than the trees--certainly when public perceptions are 180 degrees in the wrong direction.

Furthermore, only a strong collective voice is likely to be heard when popular opinion has been aroused against particular ideas, as had been the case with intelligence for some years. For many of us who signed the "Mainstream" statement, this joint effort was the only corrective letter of the many we individually wrote to the media that was ever published.

Scientists should not have to issue public statements about what is most basic in their fields. However, responsibility to science and society sometimes demands that they do so. What effects such statements have is uncertain--except that pundits can no longer assert their falsehoods without fear of contradiction.

REFERENCES

Gottfredson, L.S. (1996). What do we know about intelligence? American Scholar, Winter, 15-30.

Herrnstein, R.J., & Murray, C. (1994). The bell curve: Intelligence and class structure in American life. New York: Free Press.

Neisser, U., Boodoo, G., Bouchard, T.J., Boykin, A.W., Brody, N., Ceci, S.J., Halpern, D.F., Loehlin, J.C., Perloff, R., Sternberg, R.J., & Urbina, S. (1996). Intelligence: Knowns and unknowns. American Psychologist, 51, 77-101.

Page, E.B. (1972). Behavior and heredity. American Psychologist, 27, 660-661.

Snyderman, M., & Rothman, S. (1987). Survey of expert opinion of intelligence and aptitude testing. American Psychologist, 42, 137-144.

Snyderman, M., & Rothman, S. (1988). The IQ controversy, the media and public policy. New Brunswick, NJ: Transaction.

BIBLIOGRAPHY

The following bibliography is provided as an entry point into the vast literature on intelligence. It samples the major books since 1980 (with the addition of one 1975 classic). Consulting any subset of entries will quickly reveal many other important works.

Some of the books listed here examine issues that are now considered settled for the most part (e.g., test bias), and others represent newer, quickly evolving fields of inquiry (e.g., biological bases of intelligence). Date of publication is therefore a fallible guide to currency.

Some of the volumes synthesize work on a single major question (e.g., Jensen, 1980; Spitz, 1986); others survey the variety of expert opinion on an issue (e.g., Detterman and Sternberg, 1982); yet others represent separate threads of research on a fast-breaking topic (e.g., Vernon, 1993). All, however, give a sense of the ways in which researchers have tried to puzzle out the meaning and measurement of intelligence. By illustrating the kind and amount of evidence on particular questions, as well as debates over how compelling we should consider that evidence, these volumes help to illustrate not only what we know but also how we know it.

The bibliography provides general documentation for the "Mainstream" statement. It was culled from documentation for each of the statement's 25 specific items, which, in turn, had been obtained by asking signers of the "Mainstream" statement to provide the best one or two citations for each item. That list of more than 150 book and journal citations ("Selected Documentation for 25 Items in `Mainstream Science on Intelligence'") is available from the author.

Short Books for General Audience

Dunn, J., & Plomin, R. (1990). Separate lives: Why siblings are so different. New York: Basic Books.

Jensen, A.R. (1981). Straight talk about mental tests. New York: Free Press.

Seligman, D. (1992). A question of intelligence: The IQ debate in America. New York: Citadel Press.

Textbooks

Anastasi, A. (1996). Psychological testing (7th ed.). New York: Macmillan.

Brody, N. (1992). Intelligence (2nd ed.). San Diego: Academic Press.

Cronbach, L.J. (1990). Essentials of psychological testing (5th ed.). New York: HarperCollins.

Kaufman, A.S. (1990). Assessing adolescent and adult intelligence. Boston: Allyn & Bacon.

Plomin, R., DeFries, J.C., McClearn, G. E., & Rutter, M. (1997). Behavioral genetics (3rd ed.). New York: W.H. Freeman.

More Technical or Specialized Volumes

Braden, J. P. (1994). Deafness, deprivation, and IQ. New York: Plenum.

Carroll, J.B. (1993). Human cognitive abilities: A survey of factor-analytic studies. New York: Cambridge University Press.

Detterman, D.K. (Ed.). (1994). Current topics in human intelligence: Vol. 4. Theories of intelligence. Norwood, NJ: Ablex.

Detterman, D.K. (Ed.). (1996). Current topics in human intelligence: Vol. 5. The environment. Norwood, NJ: Ablex.

Detterman, D.K., & Sternberg, R.J. (Eds.). (1982). How and how much can intelligence be increased. Norwood, NJ: Ablex.

Eysenck, H. J. (1995). Genius: The natural history of creativity. Cambridge: Cambridge University Press.

Gottfredson, L.S. (Ed.). (1986). The g factor in employment [Special issue]. Journal of Vocational Behavior, 29(3).

Hetherington, E.M., Reiss, D., & Plomin, R. (Eds.). (1994). Separate social worlds of siblings: The impact of nonshared environment on development. Hillsdale, HJ: Erlbaum.

Jensen, A.R. (1980). Bias in mental testing. New York: Free Press.

Locurto, C. (1991). Sense and nonsense about IQ: The case for uniqueness. New York: Praeger.

Loehlin, J.C., Lindzey, G., & Spuhler, J.N. (1975). Race differences in intelligence. San Francisco: Freeman.

Modgil, S., & Modgil, C. (Eds.). (1987). Arthur Jensen: Consensus and controversy. New York: Falmer Press.

Plomin, R. (Ed.). (1994). Genetics and experience: The interplay between nature and nurture. Beverly Hills: Sage.

Plomin, R., & McClearn, G.E. (Eds.). (1993). Nature, nurture, and psychology. Washington, DC: American Psychological Association.

Reynolds, C.R., & Brown, R.T. (Eds.). (1984). Perspectives on bias in mental testing. New York: Plenum.

Rowe, D.C. (1994). The limits of family influence: Genes, experience, and behavior. New York: Guilford Press.

Salkofske, D.H., & Zeidner, M. (Eds.). (1995). International handbook of personality and intelligence. New York: Plenum.

Snyderman, M., & Rothman, S. (1988). The IQ controversy, the media and public policy. New Brunswick, NJ: Transaction.

Spitz, H.H. (1986). The raising of intelligence: A selected history of attempts to raise retarded intelligence. Hillsdale, NJ: Erlbaum.

Sternberg, R.J. (Ed.). (1988). Advances in the psychology of human intelligence. Hillsdale, NJ: Erlbaum.

Sternberg, R.J., & Grigorenko, E. (Eds.). (1996). Intelligence: Heredity and environment. New York: Cambridge University Press.

Vernon, P.A. (Ed.). (1993). Biological approaches to the study of human intelligence. Norwood, NJ: Ablex.

Wigdor, A. K., & Garner, W.R. (Eds.). (1982). Ability testing: Uses, consequences, and controversies. Part I: Report of the Committee. Part II: Documentation section. Washington, DC: National Academy Press.

Wolman, B. B. (Ed.). (1985). Handbook of intelligence: Theories, measurements, and applications. New York: Wiley.

RELATED ARTICLE:

The following professors-all experts in intelligence and allied fields-have signed this statement:

Richard D. Arvey, University of Minnesota

Thomas J. Bouchard, Jr., University of Minnesota

John B. Carroll, Un. of North Carolina at Chapel Hill

Raymond B. Cattell, University of Hawaii

David B. Cohen, University of Texas at Austin

Rene V. Dawis, University of Minnesota

Douglas K. Detterman, Case Western Reserve Un.

Marvin Dunnette, University of Minnesota

Hans Eysenck, University of London

Jack Feldman, Georgia Institute of Technology

Edwin A. Fleishman, George Mason University

Grover C. Gilmore, Case Western Reserve University

Robert A. Gordon, Johns Hopkins University

Linda S. Gottfredson, University of Delaware

Robert L. Greene, Case Western Reserve University

Richard J. Haier, University of California at Irvine

Garrett Hardin, University of California at Santa Barbara

Robert Hogan, University of Tulsa

Joseph M. Horn, University of Texas at Austin

Lloyd G. Humphreys, University of Illinois at Urbana-Champaign

John E. Hunter, Michigan State University

Seymour W. Itzkoff, Smith College

Douglas N. Jackson, Un. of Western Ontario

James J. Jenkins, University of South Florida

Arthur R. Jensen, University of California at Berkeley

Alan S. Kaufman, University of Alabama

Nadeen L. Kaufman, California School of Professional Psychology at San Diego

Timothy Z. Keith, Alfred University

Nadine Lambert, University of California at Berkeley

John C. Loehlin, University of Texas at Austin

David Lubinski, Iowa State University

David T. Lykken, University of Minnesota

Richard Lynn, University of Ulster at Coleraine

Paul E. Meehl, University of Minnesota

R. Travis Osborne, University of Georgia

Robert Perloff, University of Pittsburgh

Robert Plomin, Institute of Psychiatry, London

Cecil R. Reynolds, Texas A & M University

David C. Rowe, University of Arizona

J. Philippe Rushton, Un. of Western Ontario

Vincent Sarich, University of California at Berkeley

Sandra Scarr, University of Virginia

Frank L. Schmidt, University of Iowa

Lyle F. Schoenfeldt, Texas A & M University

James C. Sharf, George Washington University

Herman Spitz, former director of research E.R. Johnstone Training and Research Center, Bordentown, N.J.

Julian C. Stanley, Johns Hopkins University

Del Thiessen, University of Texas at Austin

Lee A. Thompson, Case Western Reserve University

Robert M. Thorndike, Western Washington Un.

Philip Anthony Vernon, Un. of Western Ontario

Lee Willerman, University of Texas at Austin

(sex difference research)

Sharon Begley.

Abstract: Women tend to view emotional infidelity as more threatening than sexual infidelity, while men are more upset about sexual infidelity than emotional infidelity. Sex differences in jealousy are increasingly being explained as a result of genetic differences.

Think of a committed romantic relationship that you have now, or that you had in the past. Now imagine that your spouse, or significant other, becomes interested in someone else. What would distress you more:

• Discovering that he or she has formed a deep emotional attachment to the other, confiding in that person and seeking comfort there rather than from you?

• Discovering that your partner is enjoying daily passionate sex with the other person, trying positions rarely seen outside the Kamasutra?

While this makes for an interesting party game--though we don't advise trying it around the family Christmas table--the question has a more serious purpose. Researchers have been using such "forced choice" experiments to probe one of the more controversial questions in psychology: why do more men than women say sexual betrayal is more upsetting, while more women than men find emotional infidelity more disturbing.'? Psychologist David Buss of the University of Texas, Austin, first reported this gender gap in 1992. Since then other researchers have repeatedly found the same pattern. But when it comes to explaining why men and women differ, the battle rages.

The year now ending brought claims that genes inherited from our parents make us risk takers or neurotic, happy or sad. In the new year, watch out for ever more studies on how genes passed down from Neanderthal days make us what we are. "There is tremendous interest in evolutionary perspectives in psychology," says John Kihlstrom of Yale University, editor of the journal Psychological Science. And not just among scientists. In 1996, magazine articles waxed scholarly on how evolution explains, for instance, Dick Morris's extramarital escapades. Basically, his DNA made him do it.

The debate shapes up like this. Evolutionary psychologists argue that sex differences in jealousy are a legacy of humankind's past, a biological imperative that no amount of reason, no veneer of civilization, can entirely quash. In other words, genes for traits that characterized the earliest humans shape how we think, feel and act, even if we are doing that thinking, feeling and acting in cities rather than in caves. In particular, men fly into a rage over adultery because to do so is hard-wired into their genes (not to mention their jeans). The reason is that a man can never be altogether sure of paternity. If, at the dawn of humanity, a man's partner slept around, he could have wound up inadvertently supporting the child of a rival; he would also have had fewer chances of impregnating her himself. That would have given him a poor chance of transmitting his genes to the next generation. Or, put another way, only men who carried the gene that made them livid over a spouse's roaming managed to leave descendants. Says UT's Buss, "Any man who didn't [do all he could to keep his wife from straying sexually] is not our ancestor."

FOR A WOMAN, THE STAKES WERE DIFFERENT. IF her partner sired another's child, his infidelity could have been over in minutes. (OK, seconds.) But if he became emotionally involved with another woman, he might have abandoned wife No. 1. That would have made it harder for her to raise children. So women are evolutionarily programmed to become more distressed at emotional infidelity than sexual infidelity.

The journal Psychological Science recently devoted a special section to the controversy. Leading off: a study by Buss, working with colleagues from Germany and the Netherlands, in which 200 German and 207 Dutch adults were asked the standard "which is more upsetting" question. As usual, more men than women in both cultures said that sexual infidelity bothered them more than emotional infidelity. "This sex difference is quite solid," says Buss. "It's been replicated by our critics and in cross-cultural studies, giving exactly the results that the evolutionary theory predicts."

Critics of the evolutionary paradigm say it is dangerous to call the jealousy gender gap a product of our genes. "This theory holds profound implications for legal and social policy," says psychologist David DeSteno of Ohio State University. "Men could get away with murder [of a sexually unfaithful spouse] by attributing it to their biology and saying they had no control over themselves." What's more, he argues, the theory is wrong. First, if there are genes for jealousy, they can apparently be influenced by culture. Although in every country more men than women were indeed more upset by sexual infidelity than the emotional variety, the differences between the sexes varied widely. Three times as many American men than women said that sexual treachery upset them more; only 50 percent more German men than women said that. The Dutch fell in between. So the society in which one lives can change beliefs, and thus make the gender gap larger or smaller.

More problematic for evolutionary psychology is another repeated finding. Yes, more men than women find sexual infidelity more disturbing. Something like 45 percent of men and 10 percent of women, or 80 percent of men and 8 percent of women (the numbers depend, says Buss, on how the question is worded), were more upset by the idea of sexual betrayal. But look more closely at the numbers for men. If 45, or $0, percent say that sexual betrayal disturbs them more, that means that most (55 percent, 70 percent) are not disturbed more by it. Yet evolutionary theory prediets that, even though men should not be indifferent to emotional infidelity, they should care more about the sexual kind.

Scientists who have been skeptical about the "my genes made me think it" theory have a different explanation for the jealousy gender gap. What triggers jealousy depends not on ancient genes, they argue, but on how you think the opposite gender connects love to sex and sex to love. Or, as psychologists Christine Harris and Nicholas Christenfeld of the University of California, San Diego, propose, "reasonable differences between the sexes in how they interpret evidence of infidelity" explain the gender gap. In other words, a man thinks that women have sex only when they are in love; if he learns that a woman has had sex with another man, he assumes that she loves him, too. Thus sexual infidelity means emotional infidelity as well. But men believe also that women can be emotionally intimate with another man without leaping into bed with him. A woman's emotional infidelity, then, implies nothing beyond that. By this reasoning, men see sexual betrayal as what Peter Salovey of Yale University and OSU's DeSteno call a "double shot" of infidelity. Sexual infidelity is therefore more threatening than mere emotional infidelity.

A woman, on the other hand, notices that men can have sex without love. Thus a man's sexual betrayal does not necessarily mean that he has fallen in love with someone else. So adultery bothers her less than it does men. But a woman also notices that men do not form emotional attachments easily. When they do, it's a real threat to the relationship. Says DeSteno, "Whichever type of infidelity represents a double shot would bother someone more."

Now scientists are designing experiments to show whether the mind's ability to reason, rather than genes, can explain the jealousy gender gap. The UCSD team asked 137 undergraduates the "which distresses you more" question. As expected, more men than women picked sexual infidelity as more upsetting. But the researchers also found differences in men's and women's beliefs. Women thought that, for men, love implies sex more often than sex implies love. And men said that, for women, sex implies love about as strongly as love implies sex. This difference in assessments of the opposite sex, argue the UCSD psychologists, explains all the gender gap in jealousy. Of course a woman is more bothered by a man's emotional infidelity than by sexual betrayal: a man in love is a man having sex, they figure, but a man having sex is not necessarily a man in love. Now, there's a shock.

OTHER EXPERIMENTS UNDERMINE as well the "my genes made me think it" argument. DeSteno and Salovey asked 114 undergraduates, and then 141 adults ages 17 to 70, how likely it is that someone of the opposite sex who is in love will soon be having sex, and how likely that someone of the opposite sex who is having sex is or will be in love. Anyone, man or woman, who believed that love is more likely to mean sex than sex is to mean love was more upset by emotional infidelity than by sexual infidelity. And anyone, man or woman, who believed that someone having sex is someone in love found sexual infidelity more upsetting. These data, says DeSteno, "argue against the evolutionary interpretation. Which infidelity upsets you more seems related to [gender] only because [gender] is correlated with beliefs about whether sex implies love and love implies sex."

Evolutionary psychologists don't buy it. Buss points to studies showing that a woman is at greatest risk of being battered, and even murdered, by her partner when he suspects her of sexual infidelity. "Men's sexual jealousy is an extremely powerful emotion. It makes them go berserk," says Buss. "The 'rational' arguments don't square with [the fact that] jealousy [eels 'beyond rationality.' This vague implication that culture and socialization [cause sex differences] is very old-social-science stuff that sophisticated people don't argue anymore ... Sometimes I feel that I am amidst members of the Flat Earth Society."

For all the brickbats being hurled, there is some common ground between the opposing camps. Buss and colleagues believe that jealousy, like other emotions sculpted by evolution, is "sensitive to sociocultural conditions." And those who scoff at evolutionary psychology agree that, as DeSteno says, "of course evolution plays a role in human behavior." The real fight centers on whether that role is paramount and direct, or whether biology is so dwarfed by culture and human reason that it adds little to our understanding of behavior. Spinning stories of how Neanderthal genes make us think and act the way we do undeniably makes for a lively parlor game. (Example: men prefer women in short skirts because they learned, millennia ago on the savanna, that women in long skirts tended to trip a lot and squash their babies.) And it is one that will be played often in 1997. If there is a lesson here, it may be this: be wary of single-bullet theories advanced so brilliantly that their dazzle gets in the way of their content.

Mag.Coll.: 86K0491

Abstract: Research indicates that the personality trait neuroticism is linked to two alleles of a gene that encode a transporter for the neurotransmitter serotonin. The genetic strategies of candidate allele testing and reverse genetic linkage are compared.

Why do humans behave so differently from one another? There is perhaps no question more fascinating, important, or controversial. Heritability studies tell us that if we look hard enough, variants of genes that influence behavior will be identified. And in fact, much of the variation in personality traits and in diseases such as schizophrenia, bipolar affective illness, and alcoholism is genetic in origin, at least when examined in one population at a single time point. On page 1527 of this issue, Lesch et al. provide a very clear example of this principle: The amount of neuroticism, a personality trait that can be quantified by testing (1), is influenced by two alleles of a gene encoding a transporter for the neurotransmitter serotonin. One allele results in more protein - and more neuroticism - and the other, less protein and less neuroticism.

One of the alleles of the transporter (called l for long allele) contains a 44-amino acid insertion in a regulatory region of the gene. As a result, this l allele is transcribed more efficiently than the short (s) allele and more protein is made, leading to twice as much serotonin uptake. There was a significant association of the presence of the s allele with higher scores for neuroticism. This is the second recent analysis of quantitatively analyzed personality traits that is taking behavioral genetics back to its Galtonian roots. Two reports related the presence of an allele of the dopamine DRD4 receptor that contains a 16-amino acid repeat to novelty seeking (2).

Why examine neuroticism, which is seemingly much harder to get a handle on than the psychiatric diseases whose genetic bases have been a greater focus of attention? Arguably replicated linkages - but not yet genes - have been identified for both bipolar affective illness (3) and schizophrenia (4). For alcoholism, both aldehyde dehydrogenase and alcohol dehydrogenase alleles are protective. In fact, neuroticism too is a reliably measured trait, although like the behavioral diseases, it has a complex architecture and highly heterogeneous origins. In addition, neuroticism predicts other behavioral phenotypes such as anxiety and depression and can also tell us about psychiatric illnesses, which may be extremes of behavioral continua.

The identification of the functional serotonin promoter variant is illustrative of an old, but recently neglected, strategy that is nevertheless decisively successful and deserves a new look by psychiatric geneticists, This strategy - testing candidate alleles - is a forward genetics approach, which looks for a relation between the phenotype and a variant that alters gene structure and the function or expression of the gene product. Thus, what is being searched for is not identity of the genetic marker by descent but identity by state. It is the actual presence of the allele that is important, not, as is true for reverse genetics, linkage to a nearby structural difference in the DNA. Reverse genetic linkage, with markers chosen for informativeness and genomic location, has advantages and disadvantages in power and genome coverage. A whole genome can be analyzed by linkage with panels of highly informative markers; today, the candidate gene approach can only examine a few thousand DNA bases at a time. However, for common phenotypes that are determined by multiple genes, as are most behaviors, the power of linkage to detect loci responsible for a small portion of the variance (for example, <10%) is very low (5). Thus, under realistic models of frequency and transmission, no region of the genome can be excluded. Even less likely may be the detection of most or all of the loci that contribute a larger portion of the variance (for example, 50%). Furthermore, the low prior probability to detect valid linkages by whole-genome scanning predicts that many detected linkages will prove to be artifacts, exactly as has occurred. Finally, to detect association in populations, disequilibrium of the marker with a functional variant is required. Until the marker density is much higher, perhaps on the order of hundreds of thousands of markers (5), this will dilute, perhaps fatally, the power to detect the effect of the nearby allele.

Relating a functional variant of a gene to a phenotype - the candidate allele approach - has its own pitfalls. One difficulty is evaluating the significance of an association. It is true that the prior probability for legitimate association is higher with an allele that alters function. Also, if the mode of action of the allele is known, the grouping of genotypes by the "presence of allele-absence of allele" for statistical analysis can be justified. The reader will have to judge whether new results from the Lesch et al. data from 10 individuals of three genotypes establish a dominant mode of action of the s allele or whether their grouping of homozygous s/s and heterozygous l/s genotypes is ad hoc. This "presence of allele-absence of allele" grouping method has frequently been misapplied in marker association studies where there were no functional data or where the marker alleles were nonfunctional [as with the marker for the DRD2 dopamine receptor (6)]. These arbitrary groupings of data remove degrees of freedom and inflate the significance levels, leading to nonreplications later. Indeed, the DRD24/novelty seeking association (2), which was based on arbitrary genotype groupings, did not replicate in a large sample of Finns (7).

The serotonin transporter polymorphism is one of an important but still very small list of candidate alleles for behavior that alter synthesis or function of the gene product in humans. Others include the serotonin receptors 5HT2A [His.sup.452] Tyr (which alters transduction) (8) and 5HT2C [Ser.sup.23] Cys (ligand affinity) (9), and the dopamine DRD2 [Ser.sup.311] Cys variant (transduction) (10) and the DRD4 16-amino acid repeat (11). Identification of candidate alleles is one important approach for elucidating the origin of the measured heritabilities of behavioral traits and psychiatric diseases.

Fortunately, the brevity of the list of candidate alleles is a situation that cannot last. The number of potential candidate genes - genes that can in any way alter brain function - is formidable: the cloned neurotransmitter biosynthetic metabolic, receptor, and transductional genes alone number more than 200. Many of these will possess functional variants that contribute differently to behavior - after all, alleles are why behaviors are heritable. As Barton Childs wrote, "...such is the nature of mutation that whenever a student meets the gene product - an enzyme, a receptor, a cell-adhesion molecule - he should think of potential variation and potential disease" (12).

References

[1.] K.-P. Lesch et al., Science 274, 1527 (1996). [2.] R. B. Ebstein et al., Nature Genet. 12, 78 (1996); J. Benjamin et al., ibid., p. 81. [3.] W. Berrettini et al., Proc. Natl. Acad, U.S.A. 91, 5918 (1994); O. C. Stine et al., Am. J. Hum. Genet. 57, 1384 (1995). [4.] R. E. Straub et al., Nat. Genet. 11, 287 (1995). [5.] N. Risch and K. Merikangas, Science 273, 1516 (1996). [6.] K. Blum et al., J. Am. Med. Assoc. 263, 2055 (1990). [7.] A. K. Malhotra et al., Mol Psychol. 1, 388 (1996). [8.] N. Ozaki et al., ibid., in press. [9.] M. Okada et al., unpublished data. [10.] A. Cravchik et al., J. Biol. Chem. 271, 26013 (1996). [11.] H. H. M. Van Tol et al., Nature 358, 149 (1992). [12.] B. Childs, Am. J. Hum. Genet. 52, 224 (1993). Article A18965607

an experimental synthesis of Hirschian and Benzerian perspectives. (Papers from a National Academy of Sciences Colloquium on Memory: Recording Experience in Cells and Circuits) Tim Tully. Proceedings of the National Academy of Sciences of the United States Nov 26, 1996 v93 n24 p13460(8) View abstract and retrieval choices

Erik Parens.

Author's Abstract:

Discussions of information produced by genetics research are often guided by two mistaken theoretical moves. Enthusiasts tell us not to worry because genetic tinkering can alter only our body, never our soul; worriers suggest that discovering links between our behavior and our different bodies threatens important democratic ideas like moral equality. If we understand the body and soul to be inseparable and equality to be undiminished by difference, we can begin to take seriously the information produced by genetics research.

As information about the genetic component of human behavior increases, so, of course, does the number of opportunities for its abuse. It is troubling to imagine simple-minded criminologists who aspire to use genetic information about behavior to identify and control "antisocial" individuals. It is perhaps more troubling--because far more practicable--to imagine insurers and employers who use such information to deny insurance or employment to people deemed "predisposed to" costly behaviors. Given the sordid history of attempts to use pseudobiological explanations to justify the stratification of our society,[1] perhaps most troubling of all is to imagine that our society will use such information to reinforce the view that current forms of stratification are "natural." Peter Breggin, director of the Center for the Study of Psychiatry in Bethesda Maryland, suggests this last fear in vivid terms: "Behavioral genetics is the same old stuff in new clothes.... It's another way for a violent, racist society to say people's problems are their own fault, because they carry `bad' genes."[2]

Such pronouncements may tempt some to ignore research that explores the genetic component of human behavior. Merely taking such research seriously could be construed as legitimating an inherently evil enterprise. But ignoring such research won't make the abuses go away. Indeed, the crafting of careful critiques will be needed to effectively fight the abuses. In this essay I attempt to make a small contribution to that project by calling attention to two errors that are sometimes made in discussions of the implications of genetics research in general, and behavioral genetics research in particular. Actually, the first error is not usually made by critics of behavioral genetics, but by those who tend to be its friends. They make their error in the course of trying to fend off critics, who in their view exaggerate the power of genetics to alter important human behaviors. Their error is to conceive of the body and soul (or mind) as separable--and to assume, for example, that no matter how much altering of the body we might do with genetics, altering the soul (and the complex behaviors associated with it) is forever beyond our reach. The second error tends to be made by those who are largely unfriendly toward genetics research, and who want to expose its dangers. That error is to think that facts and policies are related in an obvious and inevitable manner--to think, for example, that facts about the genetics of behavior will inevitably be used to undermine valuable political ideas or to undergird hateful social programs.

Before turning to those errors, I should say a word about the modest contribution that molecular genetics has already made, and is expected to make, toward understanding some complex behaviors.[3] It has already contributed toward understanding some behaviors associated with some forms of, for example, Alzheimer disease, Tourette syndrome, and Lesch-Nyhan syndrome. Further, much research suggests that genetics may help to explain a partial but significant component of some forms of, for example, schizophrenia, bipolar disorder, and depression.[4]