Journal of Social Issues, Vol. 53, No. 2, 1997, pp.

233258

Beyond Difference: A Biologists Perspective

Anne Fausto-Sterling*
Brown University

This piece addresses some of the challenges to social scientific thought originating from the field of evolutionary psychology and from social scientists committed to making biological arguments an important component of the sociological discussion of gender. I provide my own biological perspective on these contemporary discussions, noting especially the different kinds of explanatory value one can expect from particular kinds of argument. I also call for the application of stricter scientific standards to arguments about the evolution of sex differentiated human behaviors. I introduce some basic genetic concepts, such as the norm of reaction, which focus attention on the interaction between genes and environment during intra-generational development. Finally, I challenge the idea that there is a necessary correlation between hormone levels and reproductive behavior in vertebrates. Do biologists know things of use to social scientists who study gender? Answering this question requires a bit of context. During the past few years we have been subjected to a series of blows, both within the academy and without. The left hook hits us from the media. Lead stories in Time (August 15, 1994), Newsweek (March 27, 1995; June 3, 1996), and The New Republic (Wright, 1994a) (to name but a few) tell us that infidelity and physical desire are both in our genes, that feminists have hopelessly mislead their flock by attacking the first wave of sociobiological claims about gender and social change, and that differences in brain anatomy might explain why women are more intuitive (Newsweek, 1995, p. 54). [Figure 1] The right upper cut comes from within the academyas part of what has come to be called the science wars. Paul Gross and Norman Levitt followed up on their book entitled Higher Superstition: The Academic Left and Its Quarrels with Science (1994) with a prestigious conference

*I would like to acknowledge the help and encouragement of Jackie James in transforming this talk into the written word. Correspondence should be addressed to Anne Fausto-Sterling, Brown University, Department of Molecular and Cell Biology and Biochemistry, Box G-160J, Providence, RI 02912. 233
1997 The Society for the Psychological Study of Social Issues

234

Fausto-Sterling

Fig. 1. The covers of Newsweek and The New Republic advertising articles about the biological origins of sex differences. (Reprinted with permission).

on The Flight from Science and Reason (sponsored by the New York Academy of Sciences, May 31June 2, 1995), and the donnybrook burst into broader view with the publication of the Spring/Summer 96 issue of Social Text. Included among many serious pieces was an article on quantum gravity which turned out to be a hoax intended to show the shoddy nature of liberal critiques of science (Sokal, 1996). (As someone who takes the social studies of science very seriously, I was appalled that Sokals article slipped through the peer review process. Those cultural critics of science who know very little about the subjects they critique

Biology and Social Science

235

deserve to be criticized in serious, open debate. The vast majority of science studies scholars, however, know a great deal about the workings of science and we should all be attending to their analyses.) (See also Edge, 1996). The popular writings and the academic attacks contain related messages. In his New Republic piece (1994a), journalist Robert Wright mixes sarcasm with old-fashioned red-baiting, characterizing feminists and communists as people who fail to understand human nature because they place ideology before objective truth. At the same time the popular media presents us with the evolutionary take home message that if a behavior is in our genes, using social means to engineer behavioral change probably wont work. The Newsweek piece on The Biology of Beauty, for example, includes the following: Given the slow pace of evolutionary change, our innate preferences arent likely to fade in the foreseeable future (Cowley, 1996, p. 66). Newsweeks cover story on The New Science of the Brain (1995) embeds a message about social change into its layout design. The lower half of the cover (Figure 1) prominently displays a woman wired for brain activity measurements next to the text Why men and women think differently. In the upper right-hand corner, however, sits a banner advertising a seemingly unrelated article on affirmative action. The graphic design relates questions of social policy to brain function without directly putting the connection into words. The design of the table of contents (p. 3) reinforces the association. Brightly-colored positron emission tomography (PET) scans of a mans and a womans brain in the act of thinking differently sit prominently above a photograph of women demonstrating in favor of affirmative action. The latter image advertises a political report on the White Houses search for an affirmative action policy to call their own. One possible subtext: if men and women think differently, then perhaps affirmative action is misguided social policy. The Newsweek issue uses pictures, rather than words to suggest the possibility. Meanwhile, back at the academy, we are currently experiencing an antifeminist, anti-left backlash. Gross and Levitt, for example, lump together feminism, alternative medicine, cultural constructivism, creationism, postmodernism, science studies and other evils as belonging to an effort on the part of the academic left to replace academic reason with ideology. Sometimes a younger generation of feminists, anxious to differentiate themselves from their older forebears, join in the attack (Liesen, 1995). In general, the academic discourse has been hostile and the rhetoric so overblown that good discussion has been impossible; some cooler heads have called for a damping down of the rhetoric (Martin, 1996; Mysterud, 1996). I have been just as guilty as the next person of expressing anger in debates on the science wars. Indeed, anger and heated argument provide a good indicator that important issues are at stake. Nevertheless, I recognize that calmer, even forgiving exchanges may eventually further knowledge formation about the nature and structure of science. So let me give it a try. The popular medias publicity blitz

236

Fausto-Sterling

describes the encroachment of biological understandings on arenas of social behavior previously seen to be the bailiwick of the social sciences. Some reporters and scientists present these new biological understandings as certain beyond question. To be sure, according to this view, some gaps in our knowledge remain, but the current wave of findings has solidified the biological framework on which we hang our understandings of sex differences. Shouldnt we be happy to have these vexing questions settled once and for all? After years of believing otherwise, should we now lay down our metaphoric arms and agree with J. Richard Udry, who, in his 1994 Presidential Address to the Population Association of America, said, The existence of gendered social structure is not evidence for gendered behavior norms. Gender norms may be consequences, not causes, of sex differences. The existence of gendered social structure is not evidence of sex discrimination. Udry assures the reader that he is not against social change. Instead, he says, The future of gender can, should, and will be determined by ideology. But if our theory of gender is incorrect, we will not know how to achieve our goals. (Udry, 1994, p. 572) Writers such as Udry and Wright configure themselves as social liberals. They want change, they say, but argue that it must be reality-based. In the case of gender, reality is buried deeply within the body. And so it is there, rather than the school system or the legislature, that we must first look. What tools can I, as a biologist, offer to social scientists wishing to understand the thought currents of the past few years? What can biology tell us about human sex differences? My quarrels with many of the arguments about innate human differences and their evolution stem from understandings of the nature of biological explanation, which differ significantly from those of writers such as Wright, Udry, and the various Time and Newsweek authors. The biological work to which I am drawn contains solid theory and detailed empirical information; using such work as a starting point enables us to build complex, interesting accounts of behavior. Such chronicles are not inherently at odds with socially engineered change. Nor do they obviate the need for a highly sophisticated descriptive and experimental social science, with explanations of human behavior and social change offered in its own terms, without recourse to the language of biology. Why Frogs Jump Ill start with a seemingly simple question: Why does a frog jump? (Figure 2). The answer can be given at several different levels. Within the moment of watching the frog jump, we can suggest a holistic answerthe frog jumps as part of an ecosystem. It senses a predator nearby and gets out of the way. A more mechanistic, reductionist approach would focus on the frogs leg muscles: it jumps because its muscles twitch. Even more specifically, we can reduce the level of explanation to a discussion of how and why muscles twitch, and we can say its jumping because two proteins, actin and myosin, have molecular properties that

Biology and Social Science

237

allow them to contract, something they do in response to a nerve impulse. Each of these levels of explanation is valid for particular purposes, but one cannot necessarily substitute for the other. Suppose, for example, an ecologist wants to understand how many predators a system can sustain before the food supply crashes, and decides to estimate predator density indirectly, by looking at frog movement. He or she would want to know why the frog jumps, i.e., does it jump to avoid a predator, to approach a potential mate, or to seek food? Questions of ecosystem function would be paramount, but the molecular physiology of muscle contraction would be of little interest. If, on the other hand, a mysterious ailment afflicted the frog population, causing partial paralysis in many of the animals, using reductionist methodology to understand muscle physiology might be the analytical approach of choice. Focusing on either the contemporary ecosystem or adult muscle function leaves out two important types of biological analysis and explanationdevelopment (embryological, juvenile, and adult) and evolution. Both mechanistic and holistic analyses of current behaviors describe what is. In contrast, examining individual development or species evolution allows us to talk about how things get to be the way they are. In frogs, for example, certain cells in the retina of the eye can distinguish between the shadow cast by a marsh hawk and that of a dragon fly.

Fig. 2. Levels of biological explanation. (Inspired by Lewontin, Rose and Kamin, 1984 p. 281.)

238

Fausto-Sterling

A reductionist neuroscientist would say that a frog decides to jump rather than dart out its tongue to snag a meal because cells in the retina send electrical impulses to an appropriate part of the brain. These electrical signals, generated when sodium and potassium ions flow in and out of the nerve cells, stimulate particular muscle cells to twitch. But how did these talented retinal cells arise in the first place? There are two types of answers. A developmental biologist would respond with a descrip- tion of the genetics, anatomy, and molecular biology of eye development. He or she would ask questions such as, When during embryological development do neural and pigmented retina cells first differentiate from one another? How does the neural retina hook up with the optic nerve and how do fibers in the optic nerve connect to particular parts of the central nervous system and then from the central nervous system to the hind leg muscles? Evolutionists, however, describe events that take place over many generations. An evolutionist would speak in terms of genetic variation, natural selection, adaptation, and geographic isolation (to name the most commonly studied mechanisms of evolution). He or she might postulate that within a population, different individuals had different abilities to distinguish shadow shape. For evolution to occur, that difference would have to be inherited. This is really true by definition, since the modern synthesis defines evolutionary divergence between two populations as a change in the relative frequency of genes in the gene pool (the sum of all genetic variants in a population). Evolution, then, might occur if marsh hawks picked off frogs that were slow to differentiate between a predatory bird and a proteinaceous fly. The more precise the shadow detection, the greater the chances of survival. This would be an example of evolution by natural selection and the capabilities of the retina would be an adaptation. Evolution can occur in other ways, however. A beaver family could have interrupted the water flow in a frogs stream, creating two separate ponds. By chance, the frogs in one pond might contain individuals that are genetically invariant for degrees of shadow differentiation. Whatever trait they had at the time the new pond formed would characterize them from then on. If the frogs couldnt migrate between ponds, then evolutionthe creation of a genetically different, interbreeding populationwould have occurred, but geographic isolation rather than natural selection would have been the culprit. The retinal capacities of each population would not have been an adaptation to selection, but a chance event. Developmental answers respond to the why question for individual frogs within a particular generation, while evolutionists look at changes in gene frequency at the population level. Suppose I ask why two groups of frogs jump differently. Why, for example, do bullfrogs jump further than leopard frogs? One can stick with a strictly reductionist explanation. On average, bullfrogs jump further because they are a lot bigger. And one could reduce that answer to details about the number of muscle fibers and the comparative biophysics of fibers of different lengths and thicknesses. But one could add in a second question: why are bullfrogs larger than

Biology and Social Science

239

leopard frogs? One answer might be developmentalthey are bigger because bullfrogs spend two years as tadpoles and thus have a much longer growing period than leopard frogs, which metamorphose in a single growing season. Both explanations are valid, but offer different kinds of information. One could push matters further, asking why the bullfrog life cycle differs from that of the leopard frog. An answer here would require data about the evolution of the two species. When and under what circumstances did they diverge from a common ancestor? What kind of life cycle did the progenitor species have, i.e., did leopard frogs lose a year of development or did bullfrogs gain one? And why? Was it due to natural selection, geographical isolation or genetic drift? Or, does each species even now exhibit developmental plasticitythe ability to change its life cycle in appropriate environmental circumstances? Answers to these questions can be acquired empirically. Indeed, solving problems such as these is what many evolutionary biologists do for a living. It will not do simply to assume the answer. Many social scientists study gender inequity. Some hope to figure out how to equalize difference. Others believe that whatever differences theyve observed cannot (and perhaps should not) change. Each of these groups has an underlying and not always articulated theory about the biological basis of difference; when pressed, they will invoke it to their advantage. But when dueling social scientists use biological tools to do battle, they do not always think clearly about the level and type of biological explanation with which they argue. Thus the first thing social scientists need to think through when they use biology to ask the difference question is for which level of explanation do they seek an answer. Such a question is neither innocent nor far-removed from questions of policy. If one finds a sex difference, does it help to know its immediate physiological cause, or does one really seek a developmental or even an evolutionary explanation? I assume that in a trivial sense anything we do has an immediate physiological cause, i.e., if I run from danger, it is not only because I can recognize the cultural signs of a dangerous situation. It is also because I am physically able to see, hear, and run, in the same way that a frog can jump because of the actin and myosin in its muscles. But that knowledge is not particularly interesting if what we want to understand is why I consider a particular situation to be dangerous in the first place. The answer to such a question depends very much on why one wants to study difference in the first place. Does one want to learn about it in order to explain the way things are, or to design policies to change the status quo? I suggest that currently available types of evolutionary explanations work best in the former case, whereas developmental understandings are most useful in the latter. As the following discussion makes plain, understandings at the level of molecular physiology are usually the least useful for a researcher interested in designing effective social policy.

240

Fausto-Sterling

From Mice to Bats and Beyond Geneticists look at evolution in a number of different ways. The following discussion comes from the work of C.H. Waddington (1975). Animal populations consist of individuals with unique genotypesthe sum of all the genes in a cell. During any particular generation, populations participate in different kinds of systems. Waddington named one of these the exploitative system. Animals, within the limits of their genetic makeup (e.g., mice cant fly from one tree top to the next; to colonize tree tops they must climb or jump to nearby branches), choose among and modify possible environments, thus creating an environmental niche. At the same time, animals in any particular generation develop under particular stresses and strains. Waddington called this development within a single generation the epigenetic system. During development, a chance environmental stress might reveal a developmental potential not ordinarily visible. Suppose, for example, our hypothetical tree mice hadnt found much food during a particular year. Some females might go through pregnancy in a condition of near starvation. As a result of that, some fetuses would die, others would develop normally, and still others might survive but exhibit some sort of limb defects, say a failure of the skin connecting the limb digits to degenerate during development as it normally does. Perhaps these abnormalities show up only when a particular genotype develops in combination with low protein availability. The result: webbed feet. (Such a scenario is not particularly far-fetched. Rats subjected to prenatal heat stress, for example, developed longer limbs.) (Siegel, Doyle & Kelley, 1977) Webbed feet might make mice slow climbers, rendering them more vulnerable to predation by a passing hawk. (This hawk is a ham. It keeps getting into the act!) But that disadvantage might be balanced out by an emerging abilitythe webbed feet enable the mouse to glide when jumping from tree to tree, thereby avoiding birds in the habit of scanning tree trunks for passing mice, and increasing their food options by enabling them to reach more distant tree tops. The mouse-hawk interactions and the ability to improve survival by niche extension form part of what Waddington called the natural selective system. If the environmental stress of low food availability remained for several generations, mice with an epigenetic system (i.e., an interaction between environment and genes to produce a new phenotype) in which development during protein deprivation produced webbed feet, might survive with higher frequency than that of same genotype developing under high food conditions. (Note how the same genotype can produce different phenotypes under different developmental conditions). Thus the low proteinwebbed feet epigenetic system would become more frequent in future generations. If the selection went on for long enough, these mice might even develop a new niche, forgoing climbing altogether. Waddington demonstrated that in some cases, new phenotypes induced by epigenetic stresses stabilize even if the original stress disappears (Waddington, 1975). Thus a new variety of webbed-footed gliding mouse might emerge and, with

Biology and Social Science

241

further selection for more efficient gliding, might even evolve into what is called in German a fledermaus, a flying mouse, or bat. I do not suggest that this is how bats actually evolved; I have merely created a plausible scenario in order to illustrate the varied biological systems involved in evolution. Using Waddingtons framework, lets turn to some of the modern-day theorizing about the evolution of the human psyche. Robert Wright (1994b) champions a new group of academics that call themselves evolutionary psychologists. In 1992, Barkow, Cosmides and Tooby published The Adapted Mind, arguably the scholarly founding volume for this emerging field. Their central premise is that there is a universal human nature, but that this universality exists primarily at the level of evolved psychological mechanisms, not of expressed cultural behaviors (p. 5). They also postulate that psychological mechanisms evolved as adaptations via natural selection and that the evolved structure of the human mind is adapted to the way of life of Pleistocene hunter-gathers. . . .(p. 5). These academics argue that we can only understand contemporary human psychology by understanding how the mind evolved. Some evolutionary psychologists have a lot to say about sex and gender. Recently, David Buss (1995), a prominent evolutionary psychologist, discussed the natural selective system faced by primitive humans:
Women face the problem of securing a reliable or replenishable supply of resources to carry them through pregnancy, and lactation. . . . especially when food resources were scarce, that is, during droughts or harsh winters. All people are descendants of a long and unbroken line of women who successively solved this adaptive challenge; for example, by preferring mates who show the ability to accrue resources and to share them. (p. 164)

Buss tells a story of much the same quality as the one I just invented about the evolution of bats. It has a certain plausibility. Proto-human females must, indeed, have had the challenge of finding enough nutrition to sustain pregnancy and lactation. But as does my bat tale, Busss lacks essential information. Without knowing when the traits of interest became a permanent part of the human lineage, we can know little about the actual environmental variations, little about the degree to which nutritional needs, via an epigenetic system, might have sharpened foraging abilities dormant within some of the genotypes in particular populations, and/or whether systems of natural selection worked to make food utilization more physiologically efficient. If Busss selective scenario played out, perhaps it fueled the development of foraging skills, including the ability to hold three-dimensional maps in ones minds eyereturning even after many years, to a spot which had previously contained a good food source. Certainly Buss can hypothesize that pregnancy and lactation led females to select males who were good providers, just as I can hypothesize that it led females to evolve well-developed spatial and memory skills. We might both be wrong, or right, but without more data and a far more specific hypothesis, we have no way of knowing. There are a lot of data about prehistoric human culture and protohominids, and it is appropriate to use them to devise hypotheses about human evolution. It is not

242

Fausto-Sterling

unreasonable to ask the hypothesis-builders of evolutionary psychology to at least postulate at what point in human or hominid history they imagine contemporary reproductive behaviors to have first appeared. Throughout the Pleistocene is pretty vague. What is the evidence that it wasnt earlier or later? What, if any, animal systems provide unnamed models? What were the food and predator stresses at that moment? Data on these points can be gleaned from the archeological and geological record. How did humans respond? Biogeographic data can be brought to bear on this point. Was there a division of labor during this early period of evolution? Or did gender-based divisions of labor evolve later (Leibowitz, 1978)? Over how long a period of time did human mating systems evolve? Or are they still evolving? For example, the earliest humans living in the heart of Africa certainly did not, as Buss suggests, experience harsh winters. How do the events of interest to evolutionary psychology relate in time to the expansion and geographical radiation of human populations? What evidence is there for a long, unbroken line of women? When and where were there genetic bottlenecks during the course of human evolution? How many of them were there? The use of molecular evidence to trace human evolution has created a great deal of ferment during recent years (Ayala, 1995; Culotta, 1995; Dorit, Hiroshi & Gilbert, 1995; Gibbons, 1994, 1995, 1996; Hammer, 1995; Piazza, 1993; Tishkoff et al., 1996). It would be nice if evolutionary psychologists were specifically to incorporate this new information into their theory building. Which evolutionary lines or kinds of adaptive behavior were lost or selected for? How much of our current gene pool do we have because of genetic drift or geographic isolation, how much because of adaptation and natural selection? Some prior work at least attempts to situate theory-making within a time line and a set of postulates about which organisms (chimps? bonobos? Australopithecus? Homo habilis?) evolved modern human mating patterns (Leibowitz, 1978; Tanner, 1981; Fedigan, 1986). Lets engage in current discussions using the best available knowledge base and the most highly detailed hypotheses available. Without addressing some of these questions, evaluating hypotheses becomes very difficult. For example, given how precarious early human existence must have been, isnt it possible that females realized that no individual male would live long enough or stay healthy enough to provide over a period of years for his offspring? Isnt it just as likely then that the females who passed on more genes to the next generation were the ones who hedged their bets and slept with more than one male? Buss and other evolutionary psychologists engage in what are, in essence, thoughtexperiments, but unless much more carefully specified hypotheses are presented, theres no way to know how the postulated starting-points relate to the actual starting-points. The development of scientifically sound theories about the evolution of human behavioral patterns and their relationship to contemporary behavior could emerge from collaborations between social scientists, evolutionists, and behavioral biologists. Specifically, those experts in the social studies of science who have been so

Biology and Social Science

243

bitterly attacked in the current science wars have a great deal to offer. One model collaboration, developed in the halcyon days before science studies were taken seriously enough to be attacked, is a paper written by Bruno Latour (Mr. Science Studies!! See Latour, 1987) and Sharon Strum (Strum, 1987), a primatologist who studies baboon behavior. Latour and Strum (1986) devised a set of questions aimed at making specific hypotheses about human evolution. Using their questionnaire, they evaluated the quality of the theories constructed by both social scientists and biologists. (All failed the test pretty miserably.) I urge anyone devising theories about evolution and human behavior to use Latour and Strums nine questions [Table 1] to measure the scientific quality of their hypotheses. As Latour and Strum conclude, the difficulties of tracing human social origins goes beyond the mere speculative nature of the endeavor. Scientists have not yet come to terms with what makes an account scientific or convincing . . . when scientists are unaware of the mythic character and function of origin accounts . . . the coherence of the scientific account suffers(p. 186). When, instead of hypothesizing about past evolutionary events, biologists study evolution in contemporary populations, they collect very particular kinds of data. They monitor food supply, shelter, rainfall, predator levelssometimes for as long as a decade. At the same time, they follow individual animals as they mate, raise offspring, and die. They observe the animals, use DNA fingerprinting to see who fathered the offspring, and measure changes in animal shape, size, and behavior over several generations (Clutton-Brock, Guiness & Albon 1982; Davies, 1992). Evolutionary psychologists also obtain data about contemporary humans
Table 1. Latour and Strums Nine Questions 1. 2. 3. 4. What are the initial units of evolution (genes? individuals? the family? the species?) Which qualities do the authors think the units possess? (selfishness? self-regulation? harmony? aggressiveness?) Units with particular qualities enter into relationships with one another. Explicitly, what form do these relationships take? (exploitative, trade-offs, parasitical, competitive, cooperative?) What time delays are involved in exchanges which take place in the established relationships between fundamental units? (pre-hominid? hominid? Homo? Homo sapiens? prehistorical? last week?) What method of measurement can be used to assess answers to questions 14? (L & S write, for example, It is one thing to state that a baboon behaves as if to improve his reproductive success, but quite another to decide how he can implement this directive when he does not know who his offspring are (p. 174).) In what framework of events, is the evolutionary story embedded? L & S note that most evolutionary stories are logical, but usually not specifically historical. What agents or causes are said to play a role within the framework of events? (e.g., a shift from forest to savannah as a trigger for the evolution of socialness). What is the stated explicit methodology? What explicit political lessons do the authors of a theory draw?

5.

6. 7. 8. 9.

244

Fausto-Sterling

and try to reason backwards from what they find. But because their data come from present-day humans, they need to attend especially carefully to human epigenetic systemsto what degree specific behaviors appear in specific environments. Buss writes, Womens current mate preferences provide a window for viewing our mating past (1994, p. 23). Buss conducted surveys in both the United States and in 37 cultures worldwide, obtaining a total sample size of more than 10,000. Both men and women rated the importance of 18 different mate characteristics. In all cases women placed a higher value on men with good financial prospects than vice versa. Buss argues that the present state of affairs resulted from sexual selection. Evolution, he writes has favored women who prefer men who possess attributes that confer benefits and who dislike men who possess attributes that impose costs (p. 21). Buss further argues that the evolution of female preference for resource-rich males is ancient and not likely to change. Evidence for the latter claim comes from his observation that the feminist revolution of the 1970s and 1980s did not change this particular preference. One can look at Busss data and arguments from several points of view. Social scientists are more than competent to make judgments about data quality, sampling techniques, cross-cultural diversity, propriety of chosen statistical evaluations and such; indeed, some have made pointed criticisms of his research (numerous authors in Behavioral and Brain Sciences, 1989). But evolutionary biologists also have standards for evaluating Buss-like hypotheses. Four, in particular, have been suggested as essential to the acceptance of conjectures about the evolution of human reproductive behaviors (Wallen, 1989). First, of course, is there a good fit between the hypothesis and data? Second, is the evolutionary explanation as good as or better than some proposed alternative? Third, when using questionnaires to obtain data in support of hypotheses about reproduction, do observed or independently documented behaviors correlate with answers on the questionnaire? Fourth, do postulated characters actually relate to reproductive fitness (e.g., does marrying a wealthier man really increase a womans chance of producing more and fitter children)? In the case of Busss hypothesis that, during human evolution, natural selection favored women who prefer wealthier men, only the first of the four criteria has been reasonably met. Evolutionary arguments that meet the highest standards of evolutionary science must always hold clear the difference between obtaining data to demonstrate the workings of contemporary selective events, and using contemporary data to devise hypotheses about the past. In his popular writing, Buss often blurs this distinction. For example, he begins his discussion of human female preference for males with financial resources by reference to a field study of a bird called the gray shrike. He cites an elegant experiment in which a field biologist demonstrated that female shrikes preferred males with larger caches of food. The study shows sexual selection at work in a contemporary bird population. Buss then moves from his account of shrike behavior, to imagine a scenario that might have

Biology and Social Science

245

taken place during early human evolution. For female preference for richer mates to have evolved, he stipulates that prehistoric men would have had to have been able to accrue and control resources, that different men would have had different resource levels, and that there would have been an advantage to monogamy for the female. (As noted above, he never specifies just when during human evolution this might have been going on, so we cannot use the archeological record to evaluate his assumptions.) These conditions, he feels, are easily met among humans, and he reaches back to the contemporary world to grab as an example a Donald Trump or some Rockefeller or other. He then returns to women over evolutionary history and then back again to contemporary studies of female preference. What we have, in the end, is a mish-mash of argument in which often very beautifully done contemporary studies of mating behaviors in animals are thrown in with far less elegant surveys of contemporary human behavior. The latter are then combined with unsubstantiated but plausible postulates (like my mouse to bat story) about some unspecified earlier period of human evolution in which contemporary behavior might have had its origin. I do not argue that it is wrong to think about the evolution of human behavior. Rather, one must do it using the high standards of the best studies of behavioral evolution in animals (Fausto-Sterling, 1997; FaustoSterling, Gowaty & Zuk, 1997). And if one is going to build hypotheses about prehistoric evolution, then too, one must use the standards of the field and the rich, albeit imperfect, information already obtained from the fossil record. On Genes and Development Many biologists, sociologists and psychologists bracket the question of evolution in order to study contemporary development. Even if Buss, Wright, and others are 100% on target about the selective forces that led to our current sex/gender systems, broad sweep evolutionary arguments tell us little about specific mechanisms. In the evolutionary psychologists scenario, individual females who learned to recognize high-resource males survived and reproduced more frequently than those who didnt. But what, precisely, were the recognition mechanisms that evolved? Again, one can imagine a variety of possibilities. There might be something about the physique or physiognomy of high-resource males that females could spot (just as those frog retina cells were selected to tell the difference between birds of prey and potential food). Or, perhaps something a lot more indirect and potentially transformative of the human or hominid way of life happened. Perhaps women who talked a lot with other women could gather information through social and cultural networks; in this scenario, what would have evolved was the ability to gossip and trade information about nearby males. (Barkow, 1992, also discusses the evolution of gossip.) The result might be the evolution of elaborate cultural mechanisms, not some built-in hard-wired unchangeable brain response.

246

Fausto-Sterling

Most psychologists and sociologists are interested in contemporary mechanismshow an individual develops throughout his or her life span (Waddingtons epigenetic system). What can biologists tell us about this process, the process of producing a phenotype? Lets first look, through the lens of the developmental geneticist, at the relationship between genes and environment. Figure 3B shows a scanning microscope picture of the eye of the normal fruit fly. The eye consists of light-sensing subunits called ommatidia. Figure 3C shows the eye of a mutant fruit fly that is missing a lot of the ommatidia. The cause of this Bar eye is a much-studied dominant gene on the X chromosome. There is no argument about the genetic basis of the Bar trait. Nevertheless, to predict the Bar phenotype, one needs to know some things about the environment in which the fly developed. To help with such predictions, geneticists study something that they call the norm of reaction. They look at development as it occurs in different environments. Figure 3A shows the norms of reaction for a normal, so-called wild type fly eye and for two different genetic variants of Bar. The eye of a normal fruit fly grown at 15 degrees centigrade contains about a thousand ommatidia, while the eye of a fly grown at 30 degrees has only about 800. In the infrabar allele temperature operates in the opposite direction. There are fewer ommatidia at low temperatures and more at high temperatures, while in ultrabaran allele at the same genetic locusthe

Fig. 3. The graph in 3A shows that the size of the eye in three different genotypes of the fruit fly Drosophila depends on temperature. 3B shows a fruit fly head. The large pebbled object is the eye, which consists of up to 1,000 ommatidia. 3C shows a mutation called bar eye. The number of ommatidia is drastically reduced and the overall eye shape is that of a bar or narrow oval. (Reprinted with permission from Lewontin, 1982).

Biology and Social Science

247

norm of reaction runs in the same direction (but with a different slope) as for the wild type. Furthermore, infrabar and ultrabar eyes grown at 15 degrees centigrade are the same size and thus not phenotypically distinct. The same two alleles, in flies grown at 28 degrees, however, produce easily distinguishable phenotypes. Studying norms of reaction reminds us that most of the time there is no such thing as a fixed genetic trait. Genetic traits can only be defined in a particular environment.

Fig. 4. The growth of cuttings from seven different individual plants of Achillea when planted at three different elevations. (Reprinted with permission from Lewontin, 1982).

248

Fausto-Sterling

Figure 4 illustrates the same point with a different example. Each vertical column represents cuttings from a particular plant. Each horizontal line of plants comes from a different seed. Because the verticals all were derived asexually from the same plant, they are genetically identical. The horizontals, all derived from sexual reproduction, differ genetically from one another. Some were grown at sea level, some at medium altitude, and some at high altitude. Genetically identical plants grow very differently at different altitudes. Furthermore, you cannot predict from one genetically diverse individual to the next just what the altitude response will be. In this example, the tallest plant at low elevation is the smallest at the middle height and again the tallest at high elevations, while the 4th smallest at low heights, is the tallest at middle and intermediate at high elevations. With norms of reaction, there is no way to predict an outcome. The question is entirely empirical. Knowing about norms of reaction highlights the most common fallacy encountered in discussions about the genetic basis of behaviorthe presumption that individuals have constant environments, and that, in the absence of empirical measures, one can predict for any or all environments, the phenotype produced by a particular gene. If physical phenotypes are plastic, isnt it likely that behavioral phenotypes are plastic in spades? Lets turn to an example of behavior in eastern bluebirds. While almost never physically aggressive to bluebird females, the males will defend scarce nesting sites, which they broker in exchange for sex. But what happens when there are plenty of places to nest, for example, such as when people start lining fields with nesting boxes? The female doesnt need to trade sex for lodging. Under these circumstances male bluebirds change strategies. They help females feed their babies. One approach may be particularly effective in locations or seasons when food is scarce. While at other times and places, females may be able to fend for themselves. In other words, both male and female bluebirds exhibit behavioral plasticity. Under plentiful conditions, one ornithologist whos been studying bluebirds predicts that females will mate with whichever and how many males strike their fancy (Gowaty, 1996; Gowaty & Bridges, 1991a, 1991b). In these cases, it makes more sense to think of the plasticity, rather than the specific behavior, as being under genetic control. In fact, geneticists studying animals and plants have amassed experimental data showing that plasticity is a trait under genetic control, and can evolve via natural selection. Among ecologists and quantitative geneticists, the evolution and genetics of plasticity is a very hot topic right now (for example: Newman, 1988; Gomulkiewicz & Kirkpatrick, 1992; Scheiner, 1993; Schlichting & Pigliucci, 1994; West-Eberhard, 1989). Yet neither the concepts of the norm of reaction or phenotypic plasticity have appeared as a serious part of evolutionary theories about human sex differences. How an individual looks or behaves, then, results from the full panoply of gene expression evoked in a particular environment. Having the gene is not enough. If it stays silent throughout the life cycle, its effect is never seen, and we certainly do not express all of the gene sequences found in our DNA. But even within a carefully

Biology and Social Science

249

defined environment and with (hypothetically) full knowledge of all interacting genes, a third factor contributes to the outcomerandom chemical fluctuation. Random processes occur in every developmental generation, but they are not part of a genetically controlled developmental program. Thus they are not directly selected for. Biologists sometimes refer to these one-time events as developmental noise. At the biochemical level, one can think of noise in probabilistic terms. Suppose some key biochemical has a particular atomic configuration 98% of the time, but 2% of the time, the atoms in the molecule flip into a different configuration (which may be energetically less stable, for example). A key developmental event happening during that 2% of the time, could produce a different outcome than when the more stable molecular state mediates the step in question. Figure 5 depicts the right and left halves of the thorax of an individual fruit fly. Although a number of genes control bristle development, and the right and left halves of this fly have identical sets of genes, the mirror image halves do not have precisely the same bristle number or configuration. This phenomenon of slight left-right asymmetry is very widespread (run home and count the vibrissae on either side of your pet cat or hamsters snout) and is called fluctuating asymmetry (FA). FA results from developmental noise, which is the third component contributing to phenotypic development.

Fig. 5. The black-rimmed boxes enclose identical regions of the left (5A) and right (5B) halves of the thorax of an individual fly, as seen with a scanning electron microscope. The tapered objects are bristles. The left side has 12 bristles originating within the boxed-off area. The right side has 7 bristles. In addition to differences in bristle number, bristle size and precise location also differ on the left and right halves of the body. (Reprinted with permission from Lewontin, 1982).

250

Fausto-Sterling

Recently, sociobiologists have been studying FA in animals in the hope that it will provide a measure of genetic fitness. Some animals seem to prefer more symmetrical mates. Are the more symmetrical more fit? The jury is out on this one, and so far empirical studies are giving mixed answers. This use of the concept of FA, however, differs from that of more traditional evolutionary and developmental biologists. The latter are interested in developmental stability as a trait influenced by natural selection. Of Hormones and Sociologists In his 1994 Presidential address to the Population Association of America, demographer J. Richard Udry said that all vertebrates exhibit sex dimorphism, which he describes as behavior controlled by hormones. Noting that animal studies have taught us how hormones work, he suggests that in all animals hormones control sex dimorphic reproductive behaviors which lie at the heart of gender (p. 562). He argues further that hormones that cause between-sex variance also cause within-sex variance. For example, just as testosterone might make men more assertive or aggressive than women, more assertive and aggressive women might have a higher testosterone level compared to their more passive sisters (for a discussion of this point, see Fausto-Sterling, 1992). And finally, he says, testosterone masculinizes the male brain during embryogenesis and in testosterones absence nature makes female genitalia and a female brain (Udry, 1994, p. 568). I began this paper by noting some of Udrys political conclusions. Here I want to examine the strength of his biological arguments. Do all vertebrates exhibit hormonally controlled sexually dimorphic behavior? The simple answer is no. Many vertebrates do have an associated mating pattern of the sort to which Udry presumably refers. On television nature shows we see rams in rut battering each other in competition for a female. Their testicles have grown to a huge size, their hormones surge and theyre ready to mate. High hormone levels, a high sperm count, highly developed testes and vigorous mating behavior all coincide in time (Figure 6). But not all vertebrates operate in this fashion; instead, their mating behavior and gonadal activity are dissociated. The male northern garter snake, for example, emerges from hibernation in early spring. Despite the fact that his testosterone levels and gonad weights are low, he mates immediately, transferring sperm stored in the vas deferens since the previous fall. After mating is complete, his hormone levels climb, his gonads grow and he makes another batch of sperm, which he again stores through the winter for use the following spring. In the female, mating causes a rise in estrogen level, which in turn stimulates oocyte development and maturation. Fertilization does not actually occur until more than a month has passed (Crews, 1982). Nor are garter snakes all that peculiar in this respect. Table 2 lists the variety of animal species with completely dissociated mating patterns, as well as those for which males and females have different patterns. One interesting point that emerges

Biology and Social Science

251

Fig. 6. In individuals with an associated pattern, gonads are fully developed and circulating hormones at a high level at the time of mating. In animals with a dissociated pattern, mating occurs when the gonads are small, and sex steroid hormones at low concentrations. (Reprinted with permission from Crews, 1984).

from Table 2 is that the organisms in which males and females both have associated patterns include the common laboratory and domesticated animals that we use so heavily as biological models. But in contrast to rats, human females have a dissociated pattern while human males, discounting diurnal fluctuations, have tonic hormone levels, and thus a mating pattern that is neither associated or dissociated. All of which is to say that Udrys claim that in all vertebrates sex hormones control sex dimorphic reproductive behavior (Udry, 1994, p. 562) is an inadequate account of what we know about the variety of hormone-behavior relationships in the animal world. And it inaccurately describes the relationships between hormones and reproductive behaviors in humans. Udry also relies on the widely believed-in model that prenatal exposure to male hormones permanently organizes the neural structures of the male brain. Yet here too, the story is not only much more complex for animals, it also probably applies to a far lesser extent or not at all to humans. His claim derives from studies on rodents. In adult rats, the hypothalamus stimulates the pituitary to produce socalled gonadotropic (because they move toward the gonads) hormones. These hormones stimulate the gonads to make the familiar reproductive hormones that control the production of eggs and sperm. In males, the pituitary signals the testes at a constant or tonic rate, while in females stimulation is cyclical, producing the estrus cycle. Experimenters have shown that the presence of testosterone during the early development of male rodents suppresses the cyclical activity of the

252

Fausto-Sterling Table 2. Relationship between gonadal activity and mating behavior in vertebrates (adapted from Crews). Male

Female Associated

Associated Many lab and domestic animals Most birds Temperate and tropical lizards Crocodile family Musk shrews Some birds Some lizards Some snakes Some catfish

Dissociated Most temperate turtles Some snakes Some amphibians Some teleost fish Hibernating bats Some snakes Some amphibians Some teleost fish

Dissociated

hypothalamus. Furthermore, in rats this effect seems permanent: the rat remains impervious to the most vigorous attempts to evoke, after puberty, a cyclical hormonal response. But what works for rats does not necessarily work for primates. In fact, the hypothalamus of an adult male rhesus monkey, if stimulated with an appropriate hormone cocktail, will respond with a female-typical surge of hormone production, and the same is true for humans (for a fuller discussion see Bleier, 1984, pp. 8587). In other words, the best established case for a specific physiological reorganization of brain function in rodents does not carry over to primates. In focusing on some of the general features of vertebrate reproduction, Udry has made a basic error. The fact that all vertebrates have the same hormones and that these hormones are, in each case, involved in some fashion with reproductive physiology, cannot translate into the claim that a universal theory of sex dimorphism applies to all vertebrates (p. 562). Instead, using similar building principles and construction tools (anatomy and hormones), differently shaped and managed houses have been constructed. Biologists simultaneously study similarity (deriving from common evolutionary ancestry) and difference (deriving from species diversity), a fact which social scientists tempted to use theories drawn from animal physiology and behavior need to keep uppermost in their thoughts. Even for rodents, hormones do not tell the whole behavior story. Psychologist Celia Moore has studied mother-son interactions in the development of the ejaculation response in male rats. To be sure, testosterone plays an important role, but it acts indirectly. Male pups have a distinctive odor that is produced only when their testes make testosterone around the time of birth. The pup odor stimulates the mother to lick her sons genitals more often than her daughters. Juvenile male rats also spend much more time grooming their genitalia than do females, and this additional stimulation affects the rapidity with which males ejaculate. Licking behavior is also

Biology and Social Science

253

Fig. 7. 7a shows a partial cascade of effects of perinatal testosterone. Although not shown, an effect on one element may partially mediate effects on others. 7b shows a partial causal web centered on the maternal behavior of licking pups. Both mother and pups contribute elements, not all of which are identified in the diagram. (Reprinted with permission from Moore, 1985).

responsive to salt and water balance, pup leg extension, and urine releaseeach of which differs in male and female pups (Moore, 1985). And these all interact with the mothers lactation response (Figure 7). Thus the emerging picture of the development of sexual behavior paints the brain as one among several elements affected by early hormone exposure. Some of the elements are anatomical, some physiological, some behavioral, and some social, and they all interact. Hormone treatment also affects muscle and nerve development outside of the brain. For example, male rats have a set of three muscles attached to the penis which are needed for erection and ejaculation. Nerve cells growing out of a particular region of the spinal cord connect to these muscles. The muscles and the nerves accumulate androgen, which the muscles require for sexual function. In female rats, one of these muscles degenerates shortly after birth unless the female receives androgen during a particular period of development. No one has determined whether testosterone-mediated changes for female mounting behavior, often attributed to central nervous system brain differences, might be related to the presence of this muscle, which is normally lost during development (for a fuller, more technical account, see Fausto-Sterling, 1995). Not only is the development of rodent behavior more complex than a simple hormonal-brain organizational model allows, different genetic strains of animals exhibit different patterns of sexual activity. One worker, for example, describes a strain of rats in which intact males exhibit significant levels of lordosis (Sderstein, 1976), which is often considered to be an exclusively female behavior. And even within a single rodent species, different genetic strains can exhibit different behaviors in response to testosterone. Again, we see the principles of unity and diversity at work.

254

Fausto-Sterling

If it is hard to generalize about the hormone-brain-behavior relationships within genetically diverse rat populations (let alone knowing much about the relationship of highly structured laboratory settings to behaviors in the wild), then it seems we are on very shaky grounds in assuming that the rat model will work in humans. One final rat example emphasizes experience as an important component of rat development (for additional examples see Fausto-Sterling, 1995). Dutch psychologist Francien de Jonge and her co-workers (de Jonge, 1986) removed the ovaries of adult female rats, some which had prior sexual experience and some of which had not. They then induced sexual behaviors by injecting testosterone (or for controls, plain oil). Test animals could choose between mounting a male or a female. Inexperienced females preferred the company of males when they got testosterone but showed no preference without it (score one for hormone specificity). In contrast, females with prior mount experience with other females continued to prefer females regardless of whether they received oil or testosterone (score one for prior experience). If, instead, their prior experience had been with males, they subsequently showed no particular sexual preference. In other words, if given hormones that made them want to mount and thrust, female rats preferred to do it with other females if they had had previous female-female experience. But if they had no prior experience or only experiences with males, then they didnt care which sex they coupled with! Udry also repeats the widely held idea that while testosterone masculinizes the brain during embryogenesis, in its absence nature makes a female brain. It has been a standard viewpoint that sex determination can be considered the equivalent of being determined as a male (Wolf, 1995, p. 325). Yet many researchers in this field understand that such a formulation is problematic at best because female sexual development is undoubtedly not spontaneous but [involves] a process of cellular differentiation and morphogenesis (Wolf, p. 325). For many years the idea of the female as some natural ground state was unquestioned. Nobody asked whatin terms of mechanismssuch an account could possibly mean. I used to offer this viewpoint when I lectured to my embryology students. Then one day, postfeminism, I stopped mid-sentence and thought, That cant be right. Females cant just happen. Something must control and organize their development also. And of course, there are genes that control female development (Fausto-Sterling, 1987; 1989). Indeed, the presence-absence model has been so pervasive that only a few have thought that there was anything to look at with regard to female development. Conclusion The discussions of sex differences that one reads in various settings, from scholarly texts to the media, frequently slip from one category of biological explanation to another. This slippage makes it difficult to assess the strengths and limitations of particular knowledge claims. One finds oneself simultaneously coping with

Biology and Social Science

255

reductionist explanations for behavior (frogs jump because their muscles twitch) and evolutionary explanations, which address rather different biological questions and demand different types of proof. To have intelligent discussions and arguments about the role of biological difference in the genesis of gender difference, we must attend to what level of explanation is being offered, and to hold those with whom we debate these questions to a higher standard of explanatory clarity than has hitherto been offered. Evolutionary explanations of difference often entail elegant theories based on very partial knowledge of contemporary cultures and on analogies from animals, but without any foundation in the specific history of human evolution. There are no studies of human evolution comparable to those on red deer (Clutton-Brock, Guiness & Albon, 1982) or chimpanzees (Goodall, 1986). And the two species of chimps, for example, have strikingly different mating systems. Who shall we choose as our model female? Females of the better-known chimp species have an associated pattern of hormones and copulation, but the bonobo female has sex constantly with both males and females, she apparently uses sex not just for reproduction, but as a medium of social mediation (Parish, 1994; de Waal, 1995). Evolutionary explanations of human sex differences usually ignore an entire literature on norms of reaction and phenotypic plasticity. Using this strong and interesting literature in basic genetics and ecology could lead to a very different kind of story-telling. It is primarily the feminist Darwinists such as Jane Lancaster, Barbara Smuts, and Patricia Gowaty who have incorporated contextual diversity and polymorphism into their evolutionary accounts of human behavior (Fausto-Sterling, Gowaty & Zuk, 1997). So, do biologists know things of use to social scientists? The answer is certainly yes. But which aspects of biological knowledge are most useful for social scientists? The answer depends on what the knowledge project is in the first place. Suppose one is interested in educational reform. If ones unstated starting assumption is that there are likely to be irreducible cognitive differences between boys and girls, one is more likely to incorporate biological knowledge which suggests fixed brain differences between males and females and which goes on to offer evolutionary explanations for the origins of such difference. Belief in hormonallyinduced, hard-wired brain differences of very ancient evolutionary origin is easiest to reconcile with difference-based reformencouraging boys and girls to develop their special but rather different skills. If, on the other hand, one assumes that on average anyone can learn just about anything (and ought to do so if they want to), then the views of geneticists who focus on norms of reaction, adaptive plasticity and context-dependent gene action will appeal. These latter understandings of biology are more compatible with equity-based reformthe belief that given the right circumstances almost all students can excel. We each gravitate to the accounts of biology that most suit our social belief systems. Rather than becoming mired in often inaccurate or partial renditions of

256

Fausto-Sterling

biological knowledge, social scientists can offer something to the world that biologists cannotthick, complex, multivariate descriptions of human behavior. It would be a shame if social scientists gave up their skills and knowledge base in favor of strictly biological accounts of complex human behaviors. Were biological approaches to stand alone as our sole method of understanding, however interesting they may sometimes be, they would provide impoverished versions, indeed, of the knowledge systems offered us by anthropologists, sociologists and psychologists. References
Ayala, F. (1995). The myth of Eve: Molecular biology and human origins. Science, 270, 19301936. Barkow, J. H. (1992). Beneath new culture is old psychology: Gossip and social stratification. In J. H. Barkow, L. Cosmides, & J. Tooby, (Eds.), The adapted mind: Evolutionary psychology and the generation of culture (pp. 627637). New York: Oxford University Press. Barkow, J. H., Cosmides, L., & Tooby, J., Eds. (1992). The adapted mind: Evolutionary psychology and the generation of culture. New York: Oxford University Press. Behavioral and Brain Sciences, 12 (1989) pp. 149. (Target article by David Buss and responses) Bleier, R. (1984). Science and gender: A critique of biology and its theories on women. New York: Pergamon. Buss, D. M. (1995). Psychological sex differences: Origins through sexual selection. American Psychologist, 50, 164168. Buss, D. M. (1994). The evolution of desire. New York: Basic. Clutton-Brock, T. H., Guinness, F. E., & Albon, S. E. (1982). Red deer: The behavior and ecology of two sexes. Chicago: University of Chicago Press. Crews, D. (1984). Gamete production, sex hormone secretion and mating behavior uncoupled. Hormones and Behavior, 18, 2228 Crews, D. (1982, November). The ecological physiology of a garter snake. Scientific American, 159168. Cowley, G. (1996, June 3). The biology of beauty. Newsweek, 6066. Culotta, E. (1995). Asian hominids grow older. Science, 270, 11161117. Davies, N. B. (1992). Dunnock behavior and social evolution. New York: Oxford University Press. de Jonge, F. (1986). Sexual and aggressive behavior in female rats: Psychological and endocrine factors. Amsterdam: Nederlands Instituut voor Hersenonderzoek. de Waal, F. B. M. (1995, March). Bonobo sex and society. Scientific American, 8288. Dorit, R. L., Akashi, H., & Gilbert, W. (1995). Absence of polymorphism at the ZFY locus on the human Y chromosome. Science, 268, 11831185. Fausto-Sterling, A. (1997) Feminism and behavioral evolution: A taxonomy. In P. A. Gowaty, (Ed.), Evolution and Feminism (pp. 4260). New York: Chapman Hall. Fausto-Sterling, A. (1995). Animal models for the development of human sexuality: A critical evaluation. Journal of Homosexuality, 28, 217236. Fausto-Sterling, A. (1992). Myths of gender: Biological theories about women and men (2nd ed.). New York: Basic. Fausto-Sterling, A. (1989). Life in the Xy corral. Womens Studies Int. Forum, 12, 319331. Fausto-Sterling, A. (1987). Society writes biology, biology constructs gender. Daedalus, 116, 6176. Fausto-Sterling, A., Gowaty. P., & Zuk, M. (1997). Evolutionary psychology and Darwinian feminism. Feminist Studies. (Forthcoming) Fedigan, L. (1986). The changing role of women in models of human evolution. Annual Review of Anthropology, 15, 2566. Gibbons, A. (1996). The peopling of the Americas. Science, 274, 3133. Gibbons, A. (1995). The mystery of humanitys missing mutations. Science, 267, 3536.

Biology and Social Science

257

Gibbons, A. (1994). Rewritingand redatingprehistory. Science, 263, 10871088. Gomulkiewicz, R., & Kirkpatrick, M. (1992). Quantitative genetics and the evolution of reaction norms. Evolution, 46, 390411. Goodall, J. (1986). The chimpanzees of Gombe: Patterns of behavior. Cambridge: Harvard University Press. Gowaty, P. A. (1996). Field studies of parental care in birds: New data focus on variation among females. In C. T. Snowdon & J. S. Rosenblatt (Eds.), Advances in the study of behavior (pp. 476531). New York: Academic Press. Gowaty, P. A., & Bridges, W. C. (1991a). Nestbox availability affects extra-pair fertilizations and conspecific nest parasitism in eastern bluebirds, Sialia sialis. Animal Behavior, 41, 661675. Gowaty, P. A., & Bridges, W. C. (1991b). Behavioral, demographic and environmental correlates of extrapair fertilizations in eastern bluebirds, Sialia sialis. Behavioral Ecology 2, 339350. Gross, P., & Levitt, N. (1994). Higher superstition: The academic left and its quarrels with science. Baltimore: Johns Hopkins. Hammer, M. F. (1995). A recent common ancestry for human Y chromosomes. Nature, 378, 376378. Harris, G. W., & Levine, S. (1965). Sexual differentiation of the brain and its experimental control. Journal of Physiology, 181, 379400. Latour, B. (1987). Science in Action. Cambridge: Harvard University Press. Latour, B., & Strum, S. C. (1986). Human social origins: Oh please, tell us another story. Journal of Social Biological Structure, 9, 169187. Leibowitz, L. (1978). Females, males, families: A biosocial approach. North Scituate, MA: Duxbury Press. Lewontin, R. C. (1982). Human Diversity. New York: W. H. Freeman. Lewontin, R. C., Rose, S., & Kamin, L. J. (1984). Not in our genes. New York: Pantheon. Martin, E. (1996). Meeting polemics with irenics in the science wars. Social Text, 46/47, 4360. Moore, C. L. (1985). Another psychobiological view of sexual differentiation. Developmental Review, 5, 1855. Moore, C. L., Dou, H., & Juraska, J. M. (1992). Maternal stimulation affects the number of motor neurons in a sexually dimorphic nucleus of the lumbar spinal cord. Brain Research, 572, 5256. Moore, C. L., & Rogers, S. (1984). Contributions of self-grooming to onset of puberty in male rats. Developmental Psychobiology, 17, 243253. Mysterud, I. (1996). Communicating ideas about humans and evolutionary theory. Trends in Ecology and Evolution, 11:7, 310. Newman, R. A. (1988). Adaptive plasticity in development of Scaphiopus couchii tadpoles in desert ponds. Evolution, 42, 774783. Newsweek. (1995, March 27). The new science of the brain. Newsweek, 4854. Newsweek. (1996, June 3). The biology of beauty: what science has discovered about sex appeal. Newsweek, 6169. Parish, A. R. (1994). Sex and food control in the uncommon chimpanzee: How bonobo females overcome a phylogenetic legacy of male dominance. Ethology and Sociobiology, 15, 157179. Piazza, A. (1993). Who are the Europeans? Science, 260, 17671769. Scheiner, S. M. (1993). Genetics and the evolution of phenotypic plasticity. Annual Review of Ecology and Systematics, 24, 3568. Schlichting, C. D., & Pigliucci, M. (1994). Gene regulation, quantitative genetics and the evolution of reaction norms. Trends in Ecology and Evolution, 9, 154168. Siegel, M. I., Doyle, W. J., & Kelley, C. (1977). Heat stress, fluctuating asymmetry and prenatal selection in the laboratory rat. American Journal of Physical Anthropology, 46, 121126. Sderstein, P. (1976). Lordosis behavior in male, female and androgenized female rats. Journal of Endocrinology, 70, 409420. Sokal, A. D. (1996). Transgressing the Boundaries: Toward a transformative hermeneutics of quantum gravity. Social Text, 4647, 217252. Strum, S. (1987) Almost Human. New York: Random House. Tanner, N. M. (1981). On Becoming Human. Cambridge: Cambridge University Press. (1994, August 15). Infidelity: It may be in our genes. Time, 4452. Tishkoff, S. A., et al. (1996). Global patterns of linkage disequilibrium at the CD4 locus and modern human origins. Science, 271, 13801385.

258

Fausto-Sterling

Udry, J. R. (1994). The nature of gender. Demography, 31:4, 561573. Waddington, C. H. (1975). The evolution of an evolutionist. Ithaca, NY: Cornell University Press. Wallen, K. (1989). Mate selection, economics and selection. Behavioral and Brain Sciences, 12, 3738. West-Eberhard, M. J. (1989). Phenotypic plasticity and the origins of diversity. Annual Review of Ecology and Systematics, 20, 249278. Wolf, U. (1995). The molecular genetics of human sex determination. Journal of Molecular Medicine, 73, 325331. Wright, R. (1994a, November 28). The Female Mind. The New Republic, 3446. Wright, R. (1994b). The moral animal: Why we are the way we are: The new science of evolutionary psychology. New York: Pantheon.

ANNE FAUSTO-STERLING is Professor of Medical Science and Womens Studies in the Department of Molecular and Cell Biology and Biochemistry at Brown University. Her early research focused on the developmental genetics of fruit fly (Drosophila) embryos, but her interests have broadened to include the field of gender and science specifically, and science studies more generally. She is the author of Myths of Gender: Biological Theories about Women and Men (New York: Basic Books, 1992, 2nd ed.) and is currently working on a new book, tentatively entitled Body Building: How Biologists Construct Sexuality. In the laboratory she has turned her attention to the developmental ecology and life cycle evolution of freshwater flatworms.