To the Editors:

I was dismayed by Richard Horton’s misleading account (“Is Homosexuality Inherited?” NYR, July 13) of the paper which Dr. Shang-Ding Zhang and I published concerning the white gene’s effect on fruit fly sexual behavior (“Misexpression of the white (w) gene triggers male–male courtship in Drosophila,” Proceeding of National Academy of Sciences, USA, Vol. 92 (June 6, 1995), pp. 5525–5529). Dr. Horton’s assertions are false. We never “draw extravagant parallels with human beings,” nor did we “force [our] experimental results with fruit flies to fit [our] preconceived notions of homosexuality.” Indeed, even a cursory reading of our article would confirm these facts. Contrary to Horton’s representation of our motives for doing this work, we, in fact, accidentally stumbled onto this rather startling find while carrying out experiments designed to understand the role certain genes play in Drosophila brain development. We found that the runaway expression of a single gene, white, which we used as a “utility tool” in our experiments, induces homosexual courtship among mature males. It is remarkable that white is one of the most studied genes in man’s quest to understand genetic control and function in Drosophila (described in 1910 by Thomas Hunt Morgan as the first example of sex-linked inheritance in flies) and yet this phenomenon had not previously been reported. Using this male–male courtship as an assay, we carried out experiments to test hypotheses addressing causal mechanisms. These efforts led to the observation that wild-type (genetically unaltered) male flies will participate in homosexual courtship when exposed to a vigorous male–male courtship environment, demonstrating that both genetic factors and environmental cues influence this behavior. It should be noted that we never made or even implied the anthropomorphic interpretations of these data attributed to us by Horton.

I am particularly distressed that your reviewer either failed to grasp or deliberately misrepresented our study. I urge NYR readers to read the original version. Such shoddy scholarship hurts not only those who are trying to make sense of the complexities surrounding biological origins of sexual behavior, it also erodes public support for basic research. For this reason, I hope that your future reviews of scientific publications are both more accurate and responsible.

Ward F. Odenwald

The Neurogenetics Unit

Laboratory of Neurochemistry


Bethesda, Maryland

To the Editors:

Richard Horton (“Is Homosexuality Inherited?” NYR, July 13) says, “Historians of homosexuality will judge twentieth-century ‘science’ harshly when they come to reflect on the prejudice, myth, and downright dishonesty that litter modern academic research on sexuality.” He goes on to cite some examples.

Obviously, Horton’s review of Dean Hamer’s work was written before John Crewdson of the Chicago Tribune reported on the allegations of “scientific misconduct” by Hamer. These allegations were made by a junior researcher who performed research crucial to Hamer’s claimed discovery. These allegations are now being investigated behind closed doors by the National Institute of Health’s Office of Research Integrity.

As a man who “chooses” to love other men for rational reasons, I predict that history will soon be talking about LeVay and Hamer as “the Kato Kaelins of Science.”

Frank R. Aqueno

New Orleans, Louisiana

To the Editors:

“Is Homosexuality Inherited?” asks Richard Horton’s review (NYR, July 13), thereby colluding, however inadvertently, with a one-hundred-year-old anti-homosexual tradition. That politics keeps positing homosexuality as “the problem.” Heterosexuality, it seems, is problem free.

Why doesn’t a NYRB headline ever ask: “Is Heterosexuality Inherited?” Or: “Has Heterosexuality Always Existed?” Why not ask: “Are There Such Timeless, Unitary, Biological Things as Heterosexuality. Homosexuality, or Sexuality That Can Possibly be Inherited?” Why not an essay asking: “What The Hell Are Those ‘Male Typical’ and ‘Female Typical’ Behaviors and Feelings That LeVay talks About?” When will the NYR head a piece: “Is Desire to Read the NYR Inherited?” The phrasing of our questions implies loaded first assumptions about the shape of the world.

The reviewer praises my history of the heterosexual norm, The Invention of Heterosexuality, then charges that my stand against biological determinism is “uninformed,” a case of “extreme intellectual reductionism.” I’m accused of accepting “the naive dualism of nature vs. nurture” when those two “forces” are “not in opposition.”

Horton, a medical doctor, favors a complex determinism in which fickle forces, “nature” and “nurture,” “biology” and “environment,” together seal our fates. “Many different genes, together with many different environmental factors, will interact in unpredictable ways to guide behavioral preferences.” Sounds reasonable, don’t it? That’s because it’s today’s dominant, “common sense” understanding of sexuality.

But that dualistic, bio-enviro fatalism reinstates biological determinism as half of Horton’s paradigm. And that mushy middle model (so named by anthropologist Carole S. Vance) is just as politically loaded as any other way of viewing the world. Try telling feminists that biology and environment destine women for house-work and men-tending. Try telling African Americans that nature and nurture determine their difficulty getting into college.


Horton’s nature and nurture determinism, LeVay’s biological determinism, and my social-historical understanding are conflicting ways of imagining the basic phenomena at issue. Horton charges that my anti-biological determinist view is political and ideological, implying that his bio-enviro model is free of such bias. Excuse me, I don’t think so!

Dr. Horton is critical of the loony-toons analogizing of “homosexuality” in fruit flies and human beings. I agree. Such biological comparisons completely erase the historically specific social systems in which humans interact with others, learning to feel masculine, feminine, and erotic—and to call themselves “gay,” “straight,” or whatever. (Straight-faced, straight-thinking researchers who compare fruit flies and men long derided as “fruits” should also be criticized for insensitivity to the comic ironies of language.) But the ludicrous analogizing of insects and humans is based on the same biological fatalism that Horton defends as half of his biological-environmental determinism.

I honestly don’t understand how biology can play any role in determining all the different, discontinuous forms of human relationship revealed by historians: “heterosexual,” “homosexual,” and “bisexual,” “lipstick lesbian” and “old butch,” “ancient Greek pederasty,” “Victorian true love,” “romantic friendship,” “early-colonial sodomy,” and the Native American “berdache” (so-called by the French colonizers).

I don’t see how biology can determine the social, historical, and political use of sexual preference to create two dominant and subordinate classes, “heterosexuals” and “homosexuals.”

With many others, I question the medical model that defines heterosexuality, homosexuality, and sexuality as essential biological things to be explained by scientists. With others, I’m asking basic questions about those objects we’ve all taken for granted under the rule of the medical model: “heterosexuality,” “homosexuality,” and “sexuality”; “sexual orientation”; “the body,” “biology,” and “nature”; the idea of a natural “determinism” of desire.

Dr. Horton pits my “ignorance” of nature and biology against his (and LeVay’s and Hamer’s) scientific “evidence.” Such criticism delegitimizes the view of us non-scientist gay people. Horton’s bio-enviro model makes us ignoramuses, with no authority to comment on and reject a paradigm that (supposedly) defines the origin of our feelings. His, and LeVay’s and Hamer’s, rebiologizing of homosexuality and sexual preference puts the “experts” in charge again. Their biologizing makes homosexuality and sexual desire a “scientific” problem, not essentially a social, historical, and political phenomenon, joined at the hip with power.

But it’s political power that privileges the words of medical doctors and scientists, and makes the rest of us incompetents, too “uninformed,” “naive,” and “ignorant” to speak with authority.

The way to deal with the power issues implicit in all formulations of sexual research problems is to make that politics explicit, not assume its absence in the work of doctors and scientists.

In The Invention of Heterosexuality I suggest that the recent popular boom in biological theories of homosexuality shores up slippage in the 100-year-old distinction between heterosexual and homosexual. Defenders of hetero privilege are perceiving that heteros are just like homos, except for the sex of their sex partners—and they’re scared. The hetero/homo distinction upholds heterosexual supremacy. In any case, belief in the heterosexual dictatorship and biological determinism are the problems that need explaining, not homosexuality.

Jonathan Ned Katz

New York City

To the Editors:

Richard Horton’s careful treatment of the biological aspects of homosexuality (NYR, July 13, 1995) is a step in the right direction. It’s a pity therefore, especially as the new editor of The Lancet, that he did not take the opportunity to take the further step of removing from our lexicon the phrase “genetic causality.” Horton uses a wrongheaded form of this causality by embracing an ill-defined idea of a biochemical program. With biological programs, unless we define them carefully, it is assumed we are talking about gene programs. But gene programs are linear while the origin of complex behavior, most would agree, is nonlinear. Once we begin to apply linear logic of monogenic (single gene) traits to multigenic origins of nonlinear behavior we can expect nothing but confusion. Certainly single gene diseases like some forms of muscular dystrophy or cancer can be linked to unique genes but even in these cases genetic diagnosis does not always predict the disease and other factors are involved. Monogenic diseases remain at less than 2 percent of our total disease load. Most cancer and cardiovascular diseases, responsible for roughly 70 percent of premature mortality and morbidity, are complex and show no evidence that they may be understood by linkage studies that have been so successful for monogenic diseases. Knowing this, the technological impulse has been to use such analysis to diagnose not the disease but disease tendency. But tendency diagnosis is predicted to be even more unpredictable than disease prediction when many genes are involved, especially when such multigene complexes are open to environmental signals. The idea that such complex diseases can be equated to genetic programs that are then reduced to a unit amenable to clinical application is nothing more than wild speculation. That homosexuality, or other complex human traits, may be housed in a genetic program is equally unlikely. Nevertheless, that is what the Human Genome Project is all about. Horton states (p. 40), “The question is: How do genes get you from a biochemical program that instructs cells to make proteins to an unpredictable interplay…that constitutes sexuality?” The idea that the program is genetic is reinforced in Horton’s next paragraph with statements to the effect that genes are known to be responsible for the development of lungs, mouth, speech, etc.


But we know of no such thing. We only know, as Horton points out so well in most other areas of his essay, that genes are part of an interactive matrix of enormous complexity. One strong possibility that is now making some headway in biology is that, while the organism is programmed, the program is not in the genes but is located instead in the matrix itself. A regulatory program or matrix coextensive with the cell (or organism) puts the scientific emphasis on complexity and is an idea of enormous scientific respectability (epigenesis) which has been cast aside in favor of the simplistic view of genetic causality now revitalized as genetic program. It is quite possible that Dr. Horton purposefully left the idea of a program vague. He need not have. Nonlinear dynamics provides a sound scientific approach to understanding program at the level of developmental networks that themselves regulate gene patterns rather than being controlled by genes. Epigenetic logic is inherently non-linear and chaotic but includes and contextualizes the linear determinism of a [gene—>protien] equation.

Our new generation of students…and indeed our experts like LeVay and Hamer…need to be reminded that biological programs make little sense in terms of disembodied gene complexes. More important are dynamic cellular epigenetic networks; a fusion of genes, proteins, external signals, and individual experience that instructs the genome. Examples of epigenetic instruction of the genome are everywhere in our molecular biological journals and include changing DNA chemically without changing its code so that cells express unique patterns of genes without changing their genetic heritage. These networks have a logic quite different from linear genetic determinism and understanding their operation represents a scientific approach to complex phenotypes like sexuality that are likely to be more effective than anything now contained in the Human Genome Project in eliminating “…the most oppressive of cultural forces, the prejudiced social norm.” If epigenetic biology is what Horton meant by program then he should have said so more clearly. Otherwise, we are left with a form of genetic determinism more virulent by far than any simple notion of single gene causality; we are left with a complex nonlinear causality located in a seductive linear genetic program. If you use the latter to explain the former you are bound to get to our present state of wrong answers, genetic false positives and negatives, and a host of other clinical misadventures. We also, predictably, get a perpetuation of an endless and unproductive debate centered on the separation of what, as Richard Lewontin and other population geneticists have been telling us for years, cannot be separated; the living organism and its world of experience.

Richard Strohman

Professor Emeritus,

Molecular and Cell Biology

University of California

Berkeley, California

To the Editors:

I would like to add to and expand on Richard Horton’s essay on genetics and homosexuality [NYR, July 13] to help clarify what is a complex and subtle issue.

As Dr. Horton points out, the psychic and physical characteristics of human beings, and the differences between individuals, are the consequence of an interaction between the genes that are present in the fertilized egg and the sequence of environmental circumstances that the developing organism experiences during its entire life history. With a few exceptions, like cystic fibrosis where possession of the defective genotype leads ineluctably to the disease, or language acquisition where the language spoken depends only on experience and not at all on genotype, human characteristics are all subject to this interaction of forces. There are, moreover, random events in cell growth and differentiation that are neither genetic nor environmental in the usual sense, and which play an extremely important part in development, especially in behavioral traits. (For an extensive discussion of these issues see my essay in the recent book published by The New York Review, Hidden Histories of Science.)

An important consequence of the unique interaction between internal and external forces, not emphasized by Horton, is that knowledge of genetic differences contains no information at all about whether a characteristic can be changed by environmental and social arrangements. The most elementary error about genetics and development is to suppose that “genetic” is the opposite of “changeable” and that an answer to the question “how much can a trait be changed by social, historical, and individual circumstances” is given by an answer to the question “how important are genes.” The defect in my vision that makes a blur of the computer screen as I write this essay is entirely a consequence of anatomical changes that have occurred in the shape of my eyeballs as I have grown older. Even though that anatomical defect might be written indelibly in my genes, I am able to see the screen perfectly by virtue of the invention of eyeglasses by Roger Bacon (or perhaps by the Chinese), an invention made available to me by my economic and historical circumstances. That is why the search for improved estimates of heritability is the “search for better methods of estimating useless quantities.” The only way to know whether a human behavioral characteristic is malleable is either theoretically, from a complete understanding of psychic development, not a serious proposition in our current or foreseeable state of knowledge, or practically, by trying to change it.

If, nevertheless, one is interested, for whatever reason, in knowing whether genetic differences are relevant to some characteristic, one must face a serious and deep problem of human genetics. Genetics is the study of similarity between relatives, and the problem of human genetics is that, in a species with a family and social structure and a taboo against manipulating individual life histories experimentally, there is a confounding between the similarity of relatives that arises from biological causes and the similarity that arises from social causes. Thus, the standard methods of estimating the influence of genes on complex behavioral traits depend on a number of assumptions about how children are brought up that are manifestly unrealistic. By a simple but clever application of genetic principles Dr. Hamer succeeded in designing an observational protocol, described in Dr. Horton’s essay, that makes these assumptions irrelevant and which can be applied to the study of the genetics of any human difference. That is why Professor Balaban and I asked him to lecture to our class on behavioral genetics. I cannot speak to the quality of Hamer’s data, which may contain unsuspected observational biases or be atypical in some respect, but the method is an important methodological advance in human genetics. Taking the data at face value, Hamer certainly demonstrated that genetic differences are somehow relevant to differences in sexual preference.

To say that genetic differences are relevant to hetero- and homosexuality is not, however, to say that there are “genes for homosexuality” or even that there is a “genetic tendency to homosexuality.” This critical point can be illustrated by an example I owe to the philosopher of science, Elliott Sober. If we look at the chromosomes of people who knit and those who do not, we will find that with few exceptions, knitters have two X chromosomes, while people with one X and one Y chromosome almost never knit. Yet it would be absurd to say that we had discovered genes for knitting. The possession of two X chromosomes causes an embryo, with rare exceptions, to develop into an anatomical and physiological female, while the possession of a Y chromosome leads almost always to male development, and in our culture women are taught to knit while men are not. The beauty of this example is its historical (and geographical) contingency. Had we made the observations before the end of the eighteenth century (or even now in a few Irish, Scottish and Newfoundland communities), the result would have been reversed. Hand knitting was men’s work before the introduction of knitting machines around 1790, and was turned into a female domestic occupation only when mechanization made it economically marginal. The pathway of mediation between genes and human sexual behavior may be anything and Hamer, LeVay and their colleagues have left us no wiser.

We come finally to Hamer’s motivation in studying the heritability of sexuality, given the errors of simple genetic determinism. In his lecture at Harvard, and in other communications, he offers an explicitly political motive, the defense of homosexuality against prejudice and assault. If, according to Hamer, it could be shown that homo- and hetero-sexuality are simply another example of natural human physical variation, just like brown eyes and blue eyes, rather than a matter of willful choice, then the social stigma of homosexuality would be abated or even disappear. But this argument is contradicted by the entire history of the West during the last two centuries. The claim of biological and genetic differences between blacks and whites did not protect Africans from the slave block and the overseer’s lash nor did the assertion of their ineradicable genetic degeneracy keep Jews and Gypsies out of the gas chambers. On the contrary, once depravity is defined, the claim that it is “natural depravity” leads only to extermination. What Hamer has misunderstood is that the critical element in his program is not that variation in sexual orientation can be shown to be “natural,” but that it is “just like brown eyes and blue eyes,” that is, without any normative content. How the choice of sex partner is to be made as socially unfreighted as a preference for red or white wine is the problem. To begin with, it must be moved from the sphere of morality to the sphere of taste. But even that is not enough. We must not forget that “disgusting” means literally bad tasting. Unfortunately, de gustibus disputandum est, and sometimes to the death.

R. C. Lewontin

Cambridge, Massachusetts

Richard Horton replies:

I reply to your correspondents in descending order of their censure. Ward Odenwald chastises me for misrepresenting his work on the sexual behavior of fruit flies, “misleading” New York Review readers with “false” statements, and indulging in “shoddy scholarship.” In addition, he strongly denies drawing parallels with behavior patterns in other species, including humans. I am afraid that the unexpected surge of media interest in his wilder early speculations has led him to make an embarrassing about turn. I can understand why.

In his original June 1995 paper, Odenwald and his colleague, Shang-Ding Zhang, conclude by drawing explicit parallels between their data on w gene misexpression in mature male fruit flies and homosexual behavior in rats, rabbits, and cats.1 Going well beyond their preliminary evidence, they write that “elements of the basic machinery controlling male sexual behavior may be highly conserved between taxonomically distinct organisms.”

That is, the behavior of mammals may be understood by studying fruit flies. One might regard such rhetoric within the confines of the scientific literature as excusable. To repeat it in the pages of a popular news magazine is sheer recklessness. Here is Larry Thompson’s account of Odenwald’s views in the June 12 issue of Time.2

Without a wink or a chuckle, Odenwald claims that these male fruit flies are gay—and that he and Zhang made them that way. The scientists say they transplanted a single gene into the flies that caused them to display homosexual behavior. And that’s very interesting, they assert, because a related gene exists in human beings, although there is no evidence yet that the human gene has an effect on sexual preference.

The mistaken extrapolation from a “gay gene” in fruit flies to its potential role in humans is made abundantly clear and was evidently highlighted by Odenwald in his talks with Time.

Frank Aqueno raises the serious matter of an investigation currently being undertaken by the Office of Research Integrity into Dean Hamer’s work. The Chicago Tribune first drew attention to this inquiry in its June 25 issue and an article in Science3 added further details to the initial report. The investigation concerns Hamer’s 1993 paper in which he and his colleagues reported a highly significant statistical association between a region on the X chromosome—designated Xq28—and homosexual male behavior. One of Hamer’s coworkers has subsequently raised questions about methods used to collect data from forty pairs of gay brothers who were the subjects of this study.

It should be emphasized, however, that the specific charges against Hamer have not been made public under the rules of the ORI’s secret inquiry. Nor has Hamer been found guilty of scientific misconduct. But he has not been allowed to publicly rebut the claims of his accuser. If he breaks his silence, he could face disciplinary action by his employer, the National Institutes of Health. This gagging has left Hamer without the right of public reply. Under the procedures of the ORI, there is no right to due process and ORI officials act as judge, jury, and executioner. Once an accusation has been made against the scientist, the burden of proof lies with that person to privately demonstrate his or her innocence. By any standard of justice, the system by which Hamer will be tried is hardly one in which we can put much faith; but until both the charges and his defense are open to public examination, it would be unfair to give credence to any accusations against him.

The Science article also reported claims that completed work by George Ebers’s group at the University of Western Ontario does not confirm Hamer’s finding of X chromosome linkage. Yet the work of Hamer and Ebers is unlikely to be directly comparable. Each team followed a different research protocol and Ebers’s technique may have lacked sufficient sensitivity to detect the Xq28 association. Moreover, Ebers’s work has not yet been subjected to the peer review process.

Hamer has presented additional evidence to support his original discovery from a second study, which remains in the final stages of review at Nature Genetics; his unreviewed findings were reported at a recent meeting of the Behavioral Genetics Association, held in Richmond, Virginia.4 In a new group of thirty-three homosexual brothers, the Xq28 marker was significantly associated, but less so than before, with homosexual behavior. Heterosexual brothers were also studied—a crucial missing element of the earlier study—and some were found not to have the Xq28 marker, exactly what one would expect if Xq28 was having an influence on a homosexual behavior preference. Still, I should stress, as I did in my review, that Hamer’s conclusions remain tentative. Until they are replicated by other researchers using sufficiently rigorous study designs, the Xq28 data must be regarded as preliminary. However, the debate about the authenticity of these results rests on the scientific evidence, not on vague ad hominem attacks.

Jonathan Katz rightly points out that the title of my essay can be read as making unfortunate assumptions. However, he goes too far by claiming that my emphasis on complexity—I referred to a “quantum view” of behavior—is “today’s dominant” model. It is not. The dominant model is in fact a stark polarization between genetic advocates (Odenwald) and genetic skeptics (Katz), where each side claims that the position of its opponent reflects a fatal bias.

Katz seems to deny any place for biology in understanding human behavior. Yet, in the very broadest sense, all human (sexual) behavior has a biological basis. Environmental influences must always exert their effects through the body. And, of course, gene expression never occurs in an environmental vacuum. There will always be extraneous factors, whether in the cell, the organ, the body, or the environment, which will modify gene expression. Katz calls this perspective a “mushy middle” view; but his own reductionism holds that sexual preferences are easily accounted for by a single model. Again, they are not.

Katz seems to believe that those who think that biology can make a contribution, however marginal, to understanding homosexuality aim to confuse the public; scientists, he writes, seek to exclude non-scientists from debate through the obscurity of their language and method. His error is to believe that the epistemology of science is inherently different from other disciplines, that its processes, in some mystical, scientistic way, exert “authority.” This view of science has long been held by some scientists, but it is a myth and I am surprised that Katz accepts it so uncritically. Scientific observations must be replicated if they are to be accepted as valid. Hypotheses are often proven wrong; theories are frequently undermined. Hamer’s evidence will stand or fall as a result of attempts to corroborate it, not because of venomous attacks such as Katz’s on science itself.

A much more convincing criticism, which I recommend to Jonathan Katz, comes from Richard Strohman. The notion of “genetic causality” implies a linearity I certainly do not subscribe to, and I should not have used the phrase. My quantum explanation of behavior should have gone some way to suggest my view that influences involved in producing a particular pattern of behavior act in non-linear ways. But such non-linearity of effects does not rule out the influence of one or more genes, as both Katz and Aqueno seem to believe. Whether the sum of genetic influence is a necessary component in the balance of variables determining sexual preference remains to be investigated.

Finally, in Richard Lewontin’s amplification of my arguments, he accepts that Hamer’s method is “an important…advance.” Yet, remarkably, even Lewontin’s position betrays some features of genetic determinism. The clinical manifestations of cystic fibrosis do not “ineluctably” follow from “a defective genotype.” Many different mutations have been identified in the gene responsible for cystic fibrosis, and some of these are associated with varying degrees of disease severity. The expression of a mutated gene is not linearly associated with disease. Also, equating “genetic” with “the unchangeable” is widely recognized as a fallacy. The whole basis of gene therapy is that one can modify the effect of gene misexpression. It is this possibility that understandably causes anxiety and alarm for many of those now observing the gradual isolation of genetic contributions to sexual orientation. Attempts to identify these genetic influences undoubtedly could pose dangers, and scientists should be vigilant about any tendency to oversimplify the relation between genes and sexual preference.

This Issue

November 2, 1995