GENDERED MEDICAL SCIENCE: PRODUCING A DRUG FOR WOMEN

SUSAN E. BELL

Feminist Studies, 21, (3), 469-500, 1995

With the publication of The Doctor's Case against the Pill; Witches, Midwives, and Nurses: A History of Women Healers; and Complaints and Disorders: The Sexual Politics of Sickness, in the late 1960s and early 1970s, feminist scholars and activists began to examine ways in which medicine produces diagnoses and treatments that are harmful to women, depicts women in textbooks and scholarly reports in stereotypical and negative ways, and is not objective and value free. (l) Along the way, feminists uncovered ways in which medicine has also been beneficial to women, introducing further complexity into our critique. (2) More recently, feminists have explored how medicine itself is riven with tensions, contradictions, ambiguities, and uncertainties, even at the same time that it retains power in relation to women. (3) Today, feminist scholars are exploring the extent to which medicine is not a monolithic enterprise, while they continue to analyse its consequences and resist those that are negative for women. (4)

This article explores tension in one domain of medicine, It focuses on the links between transformations in medical science and cultural ideas about women using evidence drawn from medical discourse about the safety of the first synthetic oestrogen, DES (diethylstilbestrol). In the 1970s and 1980s, North American feminists undertook the research, political action, and litigation that made DES an infamous instance of medical intervention into women's reproductive lives. (5) Like the Dalkon Shield, DES initially appeared to be a benign and exciting reproductive technology but in the long run had profound and damaging consequences for women. (6)

Many of us think of DES only as a drug prescribed to pregnant women to prevent miscarriages which put the girls born of those pregnancies at increased risk of vaginal and cervical cancer, the boys at greater risk of genital abnormalities, and the mothers at a slightly greater risk of breast cancer. But DES has an earlier and quite important and complex history in which it figures in what has been described as the "hormonalisation" of women. (7) The regulation of menopause, rather than pregnancy, was its first site of intervention. (8) In addition, the usefulness of DES as a treatment for menopause was a matter of medical dispute. In this essay, I argue that DES is an example of how medical science is uncertain and ambiguous, infused with cultural assumptions about gender, and simultaneously productive of gendered meanings which can have negative consequences for women's lives. (9)

To show how cultural imagery about gender is projected into and results from medical science, I explore the metaphors used in clinical studies of DES. Like Emily Martin I am interested in the "medical propensity" to see reproduction as pathological. (10) But I am also interested in how this process contributes to what historian Diana E. Long calls "the disappearing woman." (11) Technologies such as DES have enabled medical professionals to shift their concern from menstruating, childbearing, or menopausal women to well-regulated human females. (12) What we see in the development of DES is how medicine can reduce gender, which is a socially constructed category, to sex, which medicine defines as merely biological. The ability to pathologise reproduction and to create human females derives logically from describing the body in terms of a particular set of metaphors. (13)

In this essay I focus on the production of DES before it was released for sale in 1941. I analyse two sets of questionnaires distributed to clinical investigators and then submitted collectively by pharmaceutical manufacturers to the Food and Drug Administration as evidence of the safety of DES. I explore the metaphors of cyclicity (which refers to recurrent and periodic changes in the reproductive system regulated by the ovary), production, and hierarchy in the questionnaires to show how they reflected and reinforced social conventions about gender. In addition, I show how the work embodied in the questionnaires contributed to the disappearance of women and the "diseasing" of their reproductive bodies. I focus on discussions of menopause both because the questionnaires direct attention to it and because at that time DES was most often studied in connection with menopause. I use unpublished materials because they are more likely than published materials to reveal the extent to which medical knowledge is tentative and contested and the extent to which medicine itself both embraces and resists the language and logic of gender. (14)

GENDERED MEDICINE

On a theoretical level, feminists have moved away from an old social control model that sees a male-dominated conspiracy at work within medicine. We also no longer think that paradigms from within medicine or science itself are the only controlling frames for research. (15) Rather, it is becoming clearer that metaphors from both the broad society and from within scientific and medical research communities move back and forth, reinforcing one another or creating tensions. Cultural imagery about gender is projected into medical science. Once these images are projected, they become "naturalised." Subsequently in the dynamic, the imagery itself "naturalise[s] our social conventions about gender." (16) Our task becomes trying to understand what metaphors are at work, how they connect medicine and culture, and where their contradictions lie.

Emily Martin is widely known for her demonstration of how forms of organisation and knowledge in society are reflected in medical models of anatomy, physiology, and pathology. This reflection has implications for the way medicine perceives changes in the basic organisation of the body. For example, in the early twentieth century, "models of the body [were] related in form and function to . . . Fordist mass-production systems." In the Fordist economic system, order is conferred by hierarchical pyramids; success is gained through steadily and regularly "producing large quantities of standardised products put together from standardised components." Conversely, order is lost when hierarchy breaks down. This breakdown causes failure - uneven or irregular production or simply no production at all. (l7)

To medicine a Fordist body looks like an information processing system with a hierarchical structure. When the brain is fully in control at the top of the anatomical pyramid, the body functions properly. When the authority system fails, the body's ability to produce is compromised. Changes in the body which lead to a breakdown in its authority system are pathological; interferences with its ability to produce continuously are disease states. (18)

According to Martin, if the body is viewed as a hierarchical pyramid for purposes of continuous production, then menopause is necessarily a disease. Because a menopausal body is undisciplined and unproductive, menopause is a pathological state. A menopausal body is undisciplined because its brain loses control over the ovaries, follicles, and other organs. A menopausal body is unproductive in two respects - it no longer reproduces and its ovaries no longer produce (enough) oestrogen. Not surprisingly, scientific papers and medical textbooks about menopause in the early- and mid-twentieth century are replete with images of failed functions, atrophy, disturbance, and decline. (19)

The metaphor of cyclicity, in which the body is endowed with rhythms timed by the ovary, is another example of how medical science reflects and constructs cultural imagery about gender. (20) Models of the body infused with the cyclical metaphor connect human bodies with other animal bodies. According to this metaphor, the human menstrual cycle, like cycles in other animals, fluctuates in response to periodic and recurrent fluctuations in the levels of ovarian hormones. The cycles occur "at curiously different rates in different animals," but they are nonetheless "universal in nature." For example, George W Comer's discussion of the ovary as a timepiece in human reproduction begins with a description of "milling herds of bison. . . eddying and wheeling about" every August and September as the males follow and mate with the females during the yearly period of oestrus. (21) Although the metaphor of cyclicity has been used to understand human existence for thousands of years, its link to femininity did not appear until the 1930s. (22) This connection came when sex endocrinologists began to understand the role of the ovary and its product (oestrogen) in the menstrual cycle. The name "oestrogen," given to one of the ovarian hormones, refers to cyclical "changes in the vagina characteristic of oestrus, the period of sexual activity and fertility" in females. (23) Choosing the name oestrogen firmly attached femininity to cyclicity. (24) The equation of femininity and cyclicity has suggested ways that medicine can control female bodies (by lengthening, shortening, and in general standardising women's cycles with oestrogen and other hormones). (25)

The metaphor of cyclicity introduces further complexity to the Fordist model. Whereas the Fordist model of production evokes images of women as productive or unproductive, disciplined or undisciplined labourers, the metaphor of cyclicity evokes images of bodies having productive or unproductive, regular or irregular cycles. The two metaphors simultaneously infused the language of research reports about DES.

DES. DES has had a long and controversial history. Not only has it been widely used during menopause but it has also been used to suppress lactation after childbirth, to induce bleeding the morning after unprotected intercourse, and to treat prostate cancer. It was used to prevent miscarriage until the early 1970s even though clinical studies published twenty years earlier demonstrated that it was ineffective for this purpose. (26) Physicians simply disagreed about how to interpret the evidence. Only after DES was identified as the first transplacental carcinogen in the early 1970s did Food and Drug Administration officials circulate a warning to physicians to stop prescribing it. DES is still marketed in the United States, although it is no longer approved for use during pregnancy.

Almost immediately after its synthesis in 1938, DES became the focus of much attention by clinical researchers, because initial reports suggested that it had great promise for the practice of gynaecology. At that time, clinicians used natural estrogens to relieve menstrual cramps, regulate menstruation, and treat menopausal symptoms (including hot flushes and sweats). These substances were difficult and costly to extract in pure form and generally were not potent orally. Usually, menopausal symptoms were treated with injections. These injections were often painful and sometimes left women with a hard nodule at the site of injection. A woman would return to her physician for injections each day or several times a week. Before the synthesis of DES, oestrogen therapy was costly, time consuming, and sometimes painful. Initial reports about DES suggested that oestrogen therapy could become less expensive, time consuming, and painful if DES were used instead of natural estrogens. First, DES could be taken orally instead of by injection, so patients could purchase it and take their pills at home instead of having to visit their physicians. This would simplify hormone therapy. Second, DES was much cheaper to produce than natural estrogens were. Third, DES appeared to be much more potent that natural estrogens, and so less of it could be used to treat more patients. Thus, DES would potentially reduce the cost of hormone therapy. These improvements could make hormone therapy more accessible to more women than ever before. (27)

Potentially, the population of menopausal women that could be treated with DES was huge. At the time DES was synthesised, oestrogen therapy was used to treat a wide range of symptoms and thus there was a large number of women who might benefit from therapy with DES. (28) In addition to hot flushes and sweats, the list of symptoms attributed to menopause was long and included emotional instability, depression, extreme irritability, insomnia, and digestive disturbances. One medical scientist observed that the list of symptoms commonly attributed to menopause "would not be complete unless nearly all the index in a textbook of medicine were included." (29)

Despite its therapeutic promise, evidence soon circulated that DES might be very toxic. (30) In the first clinical studies published in the United States, up to 80 percent of women taking DES suffered gastric intestinal distress; two-thirds of them had such severe distress that they had to stop using it. Concurrent animal studies demonstrated its carcinogenic effects. (31) However, despite these dramatic findings, medical scientists were uncertain about how to interpret them. The nausea and vomiting might be "unpleasant but otherwise innocuous concomitants" of DES, or they might indicate underlying damage to the liver or other organs. (32) They were also uncertain about how to test the possible effects of DES; liver tests, for example, were notoriously inaccurate at detecting mild toxic effects. (33) Thus, because DES was potentially so beneficial and potentially so unsafe, a great deal of medical attention was focused on it.

The resolution of uncertainty about DES was also required by law. With the passage of the Federal Food, Drug, and Cosmetic Act in 1938, Congress had for the first time given a federal agency the power to determine whether drugs were safe and on that basis decide whether they could be marketed. (Not until 1962 did drugs have to be proven effective.) According to the new legislation, before marketing a new drug, a drug company was required to file a New Drug Application (NDA) with the FDA demonstrating safety. A complete application also included a sample of the drug, labelling, and packaging, as well as a description of company controls over production and packaging. If a drug company included all of these to the satisfaction of FDA officials, an application was considered complete and permitted to become effective. If FDA officials did not act within sixty days, the application would become effective by default; FDA officials could extend this deadline up to a total of six months if they needed more time to evaluate a substance. (34) DES could not be sold until "scientific experts" gave sufficient proof to the FDA that it was safe. The medical science represented by the production of DES involved negotiations among clinical investigators, officials at the FDA, and pharmaceutical manufacturers over how to judge DES, as well as whether it was safe prior to its release for sale.

DES was not the first drug to be reviewed after the passage of the 1938 Act. However, it was the first drug that was not life saving and possibly unsafe. In addition, more drug manufacturers submitted applications to market DES than for any other drug up to that time. (35) Thus, it was used by FDA officials as an occasion to clarify procedures and to set policy for future cases. (36) Its release for sale in 1941 came "after one of the most thorough investigations ever given a pharmaceutical product." (37)

Methodology. The unpublished data about DES come from questionnaires designed and sent to fifty-four medical experts in 1941 by an ad hoc group representing pharmaceutical manufacturers interested in marketing DES. The ad hoc group, which called itself the Small Committee for Stilbestrol (Small Committee), consisted of four members representing ten pharmaceutical manufacturers. The Small Committee put together materials about DES, including evidence of safety, in consultation with officials from the FDA. (38) The completed questionnaires were jointly submitted to the FDA in 1941 by the Small Committee as evidence of the safety of DES. On the basis of the completed questionnaires, FDA officials decided that DES could be safely marketed to suppress lactation and to treat menopause, "senile" vaginitis, and juvenile gonorrhoeal vaginitis. (39)

The list of fifty-four experts was compiled by the Small Committee in consultation with FDA officials. Use of DES supplied by one of the manufacturers represented by the committee, in addition to training and occupation, underlay inclusion on the list. (40) The men (and one woman) who completed the questionnaires were members of the medical elite, positioned at the top of the medical hierarchy. Their work was valued and influential both within and beyond medicine. According to data in the sixteenth edition of the American Medical Directory forty-three of them (80 percent) were full-time specialists, compared with less than one quarter (23.5 percent) of all physicians in the United States in 1940. Thirty-seven of the full-time specialists (86 percent) were board certified, compared with 40 percent of all full-time specialists in the United States at that time. (41) They were also quite influential in medical education. Forty-one of them (76 percent) had appointments at teaching hospitals or medical schools. They were active in research and recognised by policymakers for their expertise in this capacity. Indeed, their opinions had a major impact on the regulation of new drugs, including (but not limited to) DES. When their determination of what was "good enough" evidence of safety in the case of DES was accepted by officials of the FDA, it became a standard for what was good enough in subsequent cases. (42)

The individuals who developed, tested, and approved DES were members of powerful institutions: academic medicine, the pharmaceutical industry, and an agency of the U.S. government. Furthermore, the resolution of uncertainties about DES and the standards set for future cases were in the hands of powerful male experts who acted on behalf of silent female patients. (43) The data I examine here reveal both the finely grained details of the struggle to resolve uncertainty and the extent to which this process is institutionally constrained. Neither speciality, nor age, nor location of practice can explain the difference of opinion and variation in their interpretation of the evidence.

The questionnaires were carefully crafted to include topics specifically requested by officials from the FDA in a series of meetings with the Small Committee. In the course of designing them, the committee and FDA officials determined which uncertainties were worthy of resolution and how they could be resolved. Significantly, the questionnaires asked experts about their clinical research; they did not request information about animal research, which had demonstrated a link between DES and cancer as well as its negative effects on pregnancy. (44)

The questionnaires reflect a particular context (premarketing review by the FDA) for the discussion about the meaning of medical science and the safety of DES. This context focused attention on whether DES should be released for sale; the type of data available; and what dosages, contraindications, and precautions should be included in labels and literature. It encouraged the experts to explain why their opinions should carry weight in the regulation of DES and other new drugs. In other contexts, such as in refereed journals, medical textbooks, or meetings of professional societies, one could expect these investigators to direct their attention to other questions about DES. (45)

In sum, individual researchers struggled to understand how to use DES in medical practice and how best to judge its uses. Although they did not constitute an organised subgroup, as an aggregate they belonged to the elite of academic medicine. Their struggle was witnessed and orchestrated by individuals in two organisations, the FDA and the Small Committee. The record of that struggle can be found in the questionnaires and other documents analysed here. These reveal the subtle ways that clinical researchers interpreted and used medical science to resolve uncertainty about the safety of DES and, concomitantly, to reflect and construct gender.

Clinical researchers used the metaphors of production, cyclicity, and hierarchy in their answers to questionnaires about the safety and effectiveness of DES. In the discussion of uncertainties that follows, I explore the language and logic of investigators to see how they reflected and reinforced images of factory production and cyclicity. That is, I consider how their resolution of uncertainty about DES is gendered and how this gendering contributed to the diseasing of women as well as their reduction to human females.

UNCERTAIN MEDICINE

I consider two of the uncertainties facing clinical researchers as they pondered whether and how to use DES for treating the symptoms of menopause: first, techniques for determining dosage and effects, and, second, mechanisms of nausea and vomiting. (46) Both were identified in a preliminary report about DES by the American Medical Association's Council on Pharmacy and Chemistry. (47) The council reported results from three studies about DES along with a recommendation for further investigations by institutions "which [had] facilities for proper experimentation and observation." In the meantime, .it recommended that DES should not be used by the general medical community "until further studies. . . led to a better understanding of' it. (46)

Following the council's publication, clinical and laboratory researchers continued their studies of DES, which they published and presented at professional societies. Physicians, dependent upon pharmaceutical manufacturers for supplies of DES, also summarised their results for use as evidence of the safety of DES. So many researchers were involved in the study of DES that by early 1941 the Small Committee was able to amass evidence from about 5,000 clinical cases; roughly 2,400 of these were already in the published literature. (49)

In the two examples I examine, clinical researchers self-consciously described what they believed to be a good approach to evaluating the safety of DES. Notwithstanding their disagreements over what constituted good science, they agreed that their role was to demonstrate to the FDA how best to evaluate new drugs. But embedded in their work were gender assumptions. Behind their discussions about dosage, for example, were assumptions about relationships between (male) physicians

and (female) patients, and behind their discussions about nausea and vomiting were assumptions about menopausal bodies.

Dosage. The questionnaires asked investigators what dosage of DES to use in treating menopause, how to determine this dosage, and whether their objective was control of menopausal symptoms. (50) In their answers, the investigators demonstrated uncertainty. Gender assumptions, both about gender relations and about the function of female bodies, pervaded the resolution of their uncertainty. Most of these clinicians favoured the use of clinical judgement to determine dosage. By "clinical judgement" they meant an "ability to see patients as individuals and to watch their individual responses closely." (51) This preference was based on their assumption that clinicians were best qualified to observe and treat the experiences of their patients. Clinicians could be more attentive to the needs of individual patients than pathologists or technicians, who read vaginal smears (the precursor to the Pap smear) in spaces far removed from real patients. (52) Clinical judgement also gave preference to the observations of physicians over those of patients.

The language of cyclicity and hierarchical control was embedded in the discussion about dosage, particularly when investigators focused on how long to take DES. Notably, they wrote about how to (re)create an appropriate menstrual cycle, what could happen if DES was not given in correct amounts, and how to control women's bodies, in language that was evocative of a failed and restored Fordist body.

The metaphor of cyclicity infused the discussion about how long to take DES. Of the thirteen investigators who warned against continuous use of DES, roughly one-half prescribed it cyclically. As one of them wrote, "It is to be expected those who will attempt to use this preparation will be acquainted with the cyclic variations of the estrogenic phases." (53) With the knowledge of these variations, investigators could "repeat what occurs naturally, that is, attempt to carry out a cyclic form of therapy in which the estrogens are given for a period of three weeks and then therapy instituted once more." (54)

According to this point of view, DES could be used to replace the oestrogen that a woman's body was no longer producing in sufficient quantities. Substituting DES would enable physicians to control the level of oestrogen in a woman's body: One consequence of this control would be to restore her body to a standard, four-week cycle. Under this regimen, a woman's system would once again function in a disciplined, standardised (but nonreproductive) way. A minority of investigators believed that the ability to replace natural estrogens with synthetic ones would enable them to control menopause even further; they could improve upon nature and shorten the perimenopausal period to "one third its usual length by judicious use of endocrine therapy." (55)

The metaphor of hierarchical control was also present in these discussions. A danger in prescribing DES cyclically was that (nonmenstrual) bleeding would occur. A not uncommon side effect experienced by older women who took DES was cyclical bleeding. To prevent this, investigators recommended withdrawing DES gradually. According to Robert T. Frank, clinical professor of gynaecology at the College of Physicians and Surgeons (Columbia), "patients should be treated for a month, and then the [DES] gradually withdrawn to avoid uterine bleeding. . . . Another course of [DES] is then given, the main object being to increase the intervals between courses until the patient is finally symptomless without medication." (56) The trick was to find the dosage of DES that was strong enough to control symptoms and consistent enough to standardise a woman's cycle.

Too much or inconsistent amounts of DES could have negative effects, just like the irregular and decreasing amounts of oestrogen produced by a woman's own body. (57) These negative effects included "symptoms of pelvic congestion - backache, lower abdominal pain - bleeding." (58) E. Kost Shelton, an associate clinical professor of medicine at the University of Southern California School of Medicine, listed a variety of negative effects that warranted changing dosage, including "nausea and vomiting, anorexia, abdominal pain. . . tingling of breasts, post menstrual bleeding, increased libido in women with no sexual outlet, etc." (59) The correct and consistent amount of DES would control a woman's body and ensure that it functioned smoothly; That is, DES would create a well-regulated human female (appropriately libidoed).

Even in its questions, the first questionnaire contained images of authority, hierarchy, and control: "in treating the menopause, do you consider the control of symptoms the objective desired?" (60) In effect, the question presumed that menopause is pathological rather than natural: [women and] women's bodies are out of control, signalled by the use of the term "symptoms" to describe women's experiences during menopause. Underlying the question is an assumption that with the decline in oestrogen production the hierarchical authority structure of women's bodies breaks down. (61) Medicine can intervene with DES to control women's bodies and restore order to the system. The language of this question is consistent with the medicalised view of menopause that medical experts were in the process of constructing. (62)

As for the answers to the question about control of symptoms, most of the investigators signalled their agreement with the images embedded in it by simply responding "yes." A few gave justifications for or elaborations of their answers which were also consistent with this image. M.E. Davis, associate professor of obstetrics and gynaecology at Chicago, for example, wrote that there are many "symptoms" during menopause, "which are part of the bizarre picture of' it, but only hot flushes are "pathognomonic." Therefore, "control of the hot flushes represents the objective to be obtained by estrogenic medication." (63)

At this time, however, medical scientists were ambivalent about the meaning of menopause and the extent to which the experiences women had during menopause were caused by the decline in oestrogen and/or evidence of disorder. (64) Two investigators explicitly challenged the assumptions underlying the question. One of them wrote: "I do not treat menopause. I do treat the patient with menopausal syndrome. Hormonal therapy is only adjunctive in our regime. Only about 35-40% need it!" (65) Another wrote: "The control of symptoms is the objective desired by the patient and since the menopause is not a disease the physician need not go further in providing adequate treatment." (66)

In sum, assumptions about gender surfaced in the language of cyclicity, hierarchy, and control used to discuss how long to prescribe DES, how much to prescribe, and what the goal of DES therapy was. Assumptions about gender can also be seen in the discussion about how to measure dosage correctly, where two opposing views of measurement surfaced. One argued in favour of the use of clinical judgement, whereas the other favoured laboratory tests. (67)

One point of view was that the dosage of oestrogen used in clinical practice should be decided on the basis of symptoms. That is, a woman should be given the "minimum dose which seemed to get symptomatic control." (68) According to this view, clinical judgement (the observation of symptoms) constitutes good medicine. A different approach was that the dosage of oestrogen used in clinical practice should be decided on the basis of vaginal changes, evaluated with a vaginal smear. A woman should be given enough oestrogen to produce changes in the vaginal smear ("a full follicular smear") and "symptomatic relief." (69) According to this view, laboratory tests (the visual assessment of change in the vagina by looking through a microscope at a slide which contains stained vaginal secretions) constitute better medicine.

Most of the investigators preferred to use clinical judgement instead of laboratory methods in diagnosing and treating their patients with DES. By doing this, investigators were choosing to listen to women patients instead of merely testing their bodies. In this respect, they resisted the introduction of the well-regulated human female, even at the same time that the language of their reports constructed her. Clinical judgement represented a traditional view of medical practice, whereas the use of controlled laboratory experiments with precise measurement of results represented an emergent one. To those investigators favouring clinical judgement, the use of vaginal smears to diagnose and treat the complaints of their patients entailed using "mass statistics" instead of individual observations and becoming aloof from, instead of engaged with, patients. (70) Clinical judgement, on the other hand, involved seeing patients as individuals and watching their individual responses closely. (71)

The contest was not between scientific control over dosage and patient control over it. All the investigators believed that patients' experiences should be interpreted by experts. Rather, the contest was between experts who believed that diagnosis and treatment should be based upon interpretations drawn from evidence contained in laboratory tests and those who believed that clinicians should interpret signs given by patients (and their bodies). Emil Novak, an associate professor of obstetrics and gynaecology at Johns Hopkins, and a strong proponent of clinical judgement, exemplifies this point of view in his 1941 textbook, Gynaecology and Female Endocrinology, in which he advised that the vaginal smear

“cannot be considered essential for the intelligent handling of [menopausal] cases. . . . There is wide variation in the response of individual vaginas to oestrogen dosage, and there is no exact parallelism between dosage. histological effects, and the degree of relief of the patient's subjective symptoms. If the latter are properly evaluated, with special concentration on the characteristic vasomotor phenomena. they can be most conveniently used as the therapeutic target and index.”

The "vasomotor phenomena" to which Novak referred are hot flushes and sweats. To decide whether treatment of these phenomena is warranted and whether it is successful requires listening to the patient and using her account to handle her case. (72)

The debate about whether to determine dosage with clinical judgement or laboratory science took on added significance in the context of the FDA's review of applications to market DES. First, by law, the FDA was required to seek the opinions of "scientific experts" in judging the safety of drugs, and presumably these could be nonphysician scientists. If laboratory tests could replace clinical judgement in reviewing the safety of DES, then nonphysician scientists could replace physicians in negotiations with the FDA. This threatened both the cultural authority and the political power of medicine; the clause "scientific experts" had been hotly debated for precisely this reason prior to passage of the 1938 Federal Food, Drug, and Cosmetic Act. (73) Second, there appeared to be a relationship between the method used to determine dosage and the side effects of DES. Investigators who favoured laboratory science methods found a high incidence of nausea and vomiting in women to whom they had prescribed DES and suspected that they would find liver damage if they could test for it. (74) Investigators favouring clinical judgement found a much lower incidence and believed that if DES was used "cautiously, it [would be] a very helpful and essentially safe material," (75)

Two questions in the first questionnaire provided an occasion for investigators to discuss dosage as well as the relative merits of clinical judgement and laboratory tests in the practice of medicine: "What dose do you use for each condition?" and "By what means do you judge adequacy of dosage?" In their answers to these questions, many of the physicians discussed the relative merits of the vaginal smear in clinical medicine. They also implicitly clarified what they meant by clinical judgement.

There was a great deal of disagreement about the amount of DES to prescribe, especially for menopause. Twenty-eight of the investigators prescribed one milligram or less per day; seventeen others prescribed five milligrams or less per day. One physician gave up to ten milligrams per day and another up to fifteen (seven of the questionnaires could not be tabulated for this question). Over one-half of the respondents (thirty-two) judged adequacy of dosage on the basis of symptoms alone; the rest (fifteen) used both symptoms and vaginal smear. (76)

Many of the investigators described a range of dosages for treating specific conditions because, as one stated forcefully, "No two patients are alike!!" (77) For example, Emil Novak wrote that the dosage he used "varie[d] between 0.1 mg. and 1.0 mg." (78) The implication is that although a maximum and minimum standard might be set, the actual dosage for each patient for each condition depends on a person's individual response to DES and on a clinician's ability to determine the "desired response on the minimum dose or the side reactions." (79) In other words, the determination of dosage depends upon clinical judgement. Clinicians must rely on their judgement to prescribe the appropriate dosage; in turn, this depends upon having clinical experience to draw from and the willingness to attend diligently to patients.

Answers to the question about how much DES to use were congruent with answers to the question about how to determine how much, framed by disputes about the relative value of clinical judgement versus laboratory tests. Most of the investigators did not explain why they judged adequacy by either clinical response or both clinical response and vaginal smears. A few of the investigators who judged the adequacy of dosage on the basis of symptoms alone explained why they thought that the vaginal smear should not be used. Novak wrote that he felt "very strongly that it is not necessary to control the vaginal smear or to produce changes in the endometrium in order to achieve worthwhile results ...[because] the primary objective of treatment is the control of subjective symptoms." (80) Because there is no necessary connection between changes in the endometrium and changes in a woman's subjective experience, a good clinician will prescribe DES on the basis of clinical judgement.

A few of the investigators who judged adequacy by both vaginal smears and the subjective response of patients shared Novak's interpretation. For example, Conrad G. Collins, assistant professor of clinical gynaecology at Tulane, believed "that if one can control the subjective symptoms of the menopause one has accomplished the desired result even though objective signs of improvement such as changes in the vaginal smear and endometrium are absent." (81) The viewpoint represented by these investigators is that clinical judgement, listening to and observing patients, is a necessary and sufficient technique for determining the proper dosage of DES.

In all the above examples, investigators explain why clinical judgement is superior to the use of vaginal smears in prescribing DES. In each of them, an underlying assumption is that medical therapeutics should be based upon face-to-face encounters between physicians and patients. Physicians, not other experts, should "read" the evidence. Their claim for the superiority of clinical judgement served to protect their power and authority in relation to other experts.

Preference for clinical judgement also preserved their cultural authority in relation to their patients. This cultural authority enabled them to "interpret signs and symptoms, to diagnose health or illness, to name diseases. . . ," and, in short, to shape "patients' understanding of their own experience." (82) As Catherine Kohler Riessman and Alexandra Dundas Todd have shown, when the physician is male and the patient is female, the effect of cultural authority is the replication of patriarchal relations. (83)

The uncertainty about dosage was resolved in several ways. The resolution involved both how much DES to prescribe and how long to prescribe it. Regarding how much DES to prescribe, investigators agreed upon maximum and minimum dosages and also that these standards could be applied to individual patients only insofar as physicians used clinical judgement. According to this logic, if dosage was based upon relief of symptoms, then the proper dosage of DES would be prescribed, and it would not produce toxic effects. The appropriate length of time for treatment with DES was cyclic - that which mimicked a natural menstrual cycle most closely. The toxic effects resulted not from the drug but from the technique used by some investigators when they prescribed it. To put it simply, clinical judgement was preferred over the use of laboratory methods in determining how much DES to give, to whom, and for how long.

The discussion about dosage reveals medical uncertainty and ambiguity over gender. Whereas the gendered language is reductionist, the choice of clinical judgement in medical practice resists reductionism. The language of cyclicity, hierarchy, and control used to discuss dosage reinforced assumptions about how the bodies of women should function both before and after menopause. Oestrogen replacement therapy with DES re-created a well-regulated human female. The preference for clinical judgement served to preserve the women in menopausal bodies while simultaneously legitimating patriarchal social relations between these women and their physicians. Paradoxically, the preference for clinical judgement favoured well-controlled women over well-regulated human females. Although clinicians resisted the reduction of women to their bodies, they did not resist negative images of women in menopause. Careful attention to the language of dosage contained in the questionnaires reveals how the gendered language of uncertainty contributed to the diseasing of women in menopause.

Nausea and Vomiting. A second uncertainty in the clinical study of DES concerned the mechanisms of nausea and vomiting that were associated with its use. This uncertainty was noted in the preliminary report of the AMA's Council on Pharmacy and Chemistry in 1939. Investigators in the three studies reported they had observed nausea and vomiting among their patients but with varying frequency. Most alarming was a report by Ephraim Shorr and his associates at Cornell University Medical College and the Rockefeller Institute for Medical Research that 80 percent of the women in their study of DES experienced toxic symptoms, including nausea and vomiting. (84) The three research groups gave different explanations for their findings. (85)

Gendered assumptions can be seen in the metaphor of hierarchy investigators used while resolving their uncertainty about the meaning of nausea and vomiting in women treated with DES. It was commonly reported in the medical literature that women experienced gastrointestinal disturbances during menopause. (86) What clinical researchers needed to sort out in the case of DES was whether the nausea and vomiting experienced by so many menopausal women taking DES was caused by DES or by a breakdown in the system controlling their bodies. Nausea and vomiting could simply be "symptoms" of menopause. (87) With DES, investigators generally came to agree that DES could be used to treat the "symptoms" of menopause, including nausea and vomiting. In making this decision, they associated nausea and vomiting with menopause and thus reinforced the notion that menopausal bodies are undisciplined.

A question in the second questionnaire ("Have you seen cases in which nausea was a part of the menopausal syndrome?") asked investigators to consider whether the nausea they observed in older women treated with DES was a symptom of menopause instead of a toxic reaction to DES. To draw out the logic and language of the question in terms of the Fordist model of the body, the question asked whether nausea might be a symptom of a menopausal body out of control. (88) When the ovaries no longer produce oestrogen in predictable, sufficient quantities, a woman's body no longer receives appropriate orders. Hence, does the decline in oestrogen production during menopause cause nausea in menopausal women or does DES?

Some answers to the question were consistent with the notion that a decline in oestrogen production caused a breakdown in control. For example, one investigator wrote, "Nausea is often part and parcel of the menopausal syndrome. . . ." (89) The use of the word "syndrome" in reference to menopause signals that this author agrees with the view implied by the question, that menopause is a disease. Another investigator, who had a different interpretation of his observations, wrote that "Gastro-intestinal symptoms such as an irritable bowel and excessive amounts of gas are not uncommon in women who are having menopausal symptoms. Nausea may occasionally occur in association with these but I doubt if one can say that the nausea per se can be construed as a menopausal symptom." (90) Although he did not believe that nausea is a symptom of menopause, he did believe that menopause has symptoms and, as such, is abnormal.

Answers to this question also reflected ambivalence within medicine about the meaning of menopause: of the thirty-nine investigators whose answers were available to me, eighteen answered yes, they had seen nausea and/or vomiting; an equal number answered no. An additional three wrote that they had rarely seen these conditions as part of the menopausal syndrome.

Significantly, in their answers to another question, eight of the investigators reported that DES relieved symptoms of nausea and/or vomiting. Both a lack of clarity about the meaning of menopause and images of (loss of) control can be seen in their reports about the beneficial effects of DES. For example, one wrote: "Usually it controlled all of the nausea, just as natural estrogens do. There seems to be no correlation between the tendency to develop nausea in persons who are subject to nausea and vice-versa. Occasionally [sic] a nauseated patient became worse on stilbestrol, but usually there was no relationship between the two." (91) During menopause, it is strongly suggested, women's bodies are out of control, so much so that attempts to control them can have unexpected consequences.

Did nausea and vomiting indicate that DES was too toxic to use in medical practice, that physicians were overdosing their patients, or were they just symptoms of menopause? If DES caused nausea and vomiting regardless of dosage, then it could not be safely prescribed to patients. However, if nausea and vomiting were indications that physicians were overdosing their patients, then the problem was not DES but improper use of it. This uncertainty took on more weight because the FDA specifically asked the pharmaceutical manufacturers to explain why DES so frequently caused nausea and vomiting before it would approve DES for marketing. Clinical researchers asked questions about nausea and vomiting in the context of the FDA's drug regulation timetable and industry pressure to market DES.

Forty-five of the first set of questionnaires could be tabulated; of these thirty-nine reported instances of nausea and vomiting among patients, although the incidence varied. Three investigators (including Shorr) reported that more than 50 percent of patients had had to discontinue using DES because of these effects, two reported (25) percent and the rest reported 20 percent or less had had to discontinue its use. Most frequently (twenty-four), investigators attributed nausea and vomiting to overdosage. As one of them wrote, "In the majority of cases toxic symptoms are due to overdosage." (92) Seven investigators supported the view that nausea and vomiting were often transitory.

How did they resolve uncertainty about the causes of nausea and vomiting? They agreed that the cause of nausea and vomiting was most likely not DES, but inexperience or improper medical practice or problem patients. Over time, their answers implied, manufacturers became more proficient at synthesising DES (93) and investigators became more proficient in dosing their patients. (94) According to this view, if patients continued to experience nausea and vomiting after physicians had been prescribing DES for some time, this could be attributed to poor medical practice. Either physicians persisted in overdosing their patients or warned their patients about these effects and thereby induced nausea and vomiting by suggestion. (95) In their view, irresponsible physicians lay behind patients' suffering, thus affirming once more their power over patients either to help or to hurt them.

Two investigators resolved the uncertainty about nausea and vomiting by labelling the patients neurotic. A third advanced the theory that a woman's social class determined her response to treatment: his "private patients [had] a greater incidence of unpleasant reactions than those treated in free clinics. In the latter group the patients often tolerat[ed] the unpleasant reactions for a longer period of time than did private patients." (96)

Regarding the cause of nausea and vomiting, consensus emerged. However, uncertainty about the mechanism of these reactions persisted. Simply put, were they caused when DES was taken orally? Was the irritation confined to the gastro-intestinal tract? If so, the problem was probably local in origin and the solution lay in finding a way to reduce the irritation (taking DES with meals, for example, or giving it by injection). Alternatively, were the reactions unrelated to whether DES was taken by mouth or by injection? If so, the problem was probably central in origin and the solution was more complicated.

The question of the mechanism of nausea and vomiting was cast aside by clinical investigators. In the second questionnaire, experts were asked whether they had any "comments or theories" about the "possible mechanism by which Stilbestrol produce[d] nausea." In addition, they were asked the following: "Do you think that the mechanism by which Stilbestrol produces nausea and vomiting must be understood before Stilbestrol can be considered sufficiently safe to be made available to the medical profession generally?"

Investigators could not reach agreement about the mechanism of DES-related nausea and vomiting. Nonetheless, almost all of them argued that the mechanism did not need to be understood before DES was released for sale. First, it would be impractical to withhold DES from the market until the mechanism was known, because the mechanism might never be known and so DES might never be sold. Second, it would be impractical to hold DES to such a standard, because the mechanism producing nausea in a number of drugs then marketed was unknown and would remain a "mystery." (97) Finally, it would be irresponsible to withhold it. (98)

Ultimately, uncertainty regarding nausea and vomiting was resolved by deciding that either they could be expected to occur during menopause or they were related to the inexperience or irresponsibility of physicians or their patients. Alternatively, the "mystery" of nausea and vomiting was one that DES had in common with other drugs. For all practical purposes, the evidence demonstrated that nausea and vomiting were not evidence of the toxicity of DES; hence, it could be safely marketed.

Ironically, the effect of this interpretation of the evidence was to release DES for sale to treat menopausal symptoms, including nausea and vomiting. The resolution of this uncertainty reflected and constructed gender. Medical scientists described the bodies of their patients as uncontrolled, in that they are nauseous and vomiting. According to this logic, menopausal bodies produce insufficient amounts of oestrogen to regulate themselves. But medicine has a solution to this chaos. It can control these bodies with DES. In turn, DES legitimated one interpretation of the evidence: that nausea and vomiting are symptoms of menopause and, as such, evidence that menopause is a disease.

DISCUSSION

In the case of DES, medical uncertainties were resolved in the context of new drug regulation. Regarding dosage, investigators concluded that if practitioners prescribed enough DES to relieve the symptoms of individual patients and to make their menstrual cycles regular they would prescribe correct amounts of DES for an appropriate length of time. With regard to nausea and vomiting, investigators constructed the evidence as demonstrating that DES could be safely marketed before the mechanism of DES or menopause-related nausea and vomiting was entirely understood.

An audience of federal regulators and pharmaceutical manufacturers shaped the logic used by medical scientists in their answers to the two sets of questionnaires about DES. FDA officials shaped experts' logic by asking them to decide what was "good enough" evidence of safety to warrant approval of DES. Pharmaceutical manufacturers shaped the logic by carefully constructing specific questions to ask in the questionnaires they circulated to clinical investigators. The regulators and manufacturers designated some problems as important and others as unimportant. The investigators answered that DES is a safe drug and that what was good enough medical science to determine this fact was clinical judgement.

The audience also shaped the pace of uncertainty resolution. The FDA had sixty days in which to approve or disapprove an application to market a new drug; it was empowered by the 1938 act to extend its deadline for up to six months. If no action was taken the drug would automatically be approved. In the early years following passage of the 1938 Federal Food, Drug, and Cosmetic Act, FDA officials took the deadline very seriously. (99) In addition, drug manufacturers stood to gain by marketing new drugs as quickly as possible. There was already a market for estrogens and the monetary incentive to market DES was high. At the time, DES cost $2.00 per gram compared with the natural oestrogen estradiol, which cost $300.00 per gram. (100) Because DES was so much cheaper than the natural estrogens, it could expand the market widely. Against this background, it became "impractical" to delay marketing DES, even though medical experts were unable to resolve all their uncertainties about DES safety. As a physician from the Lahey Clinic wrote: "I personally believe that it is safe to prescribe [DES] using the same precautions to avoid toxic effects as in the use of digitalis. The mechanism by which digitalis produces nausea and vomiting was not understood when Withering wrote about it more than 150 years ago and it is fortunate that overanxiety in regard to the toxic symptoms did not lead him to withhold its use then." (101)

Others in the social studies of science have shown that the resolution of scientific uncertainty and the emergence of one theory over another does not necessarily indicate that the "winner" is a better representation of reality. (102) The DES case shows quite dramatically how the process and the "truth" itself are shaped by forces beyond the world of research. understanding how medical science comes into being hinges on the ability to see ways in which the two worlds are intimately bound together. (103)

The connection between the two worlds can be seen in the language of medical science. This language embodies and reinforces the norms and values of society, including those referring to gender. (104) I have shown how questions and answers about the effects of DES contained images of production, hierarchical control, and cyclicity. These metaphors reflected and reinforced a Fordist view of women's bodies. (105) By and large the language did not consist of stereotypical images about women. Even though some of the answers included these images, the more powerful effects of the answers came from other, apparently more neutral, language. However, the effect was the same. Once DES was approved for sale, it was used widely. Indeed, it became a major means of medicalising menopause and pregnancy: (106) The metaphors were projected on to and embedded in DES, where they naturalised the medicalisation of women's experiences.

In the process of translating findings about the puzzling effects of DES into certainty about its safety; clinical researchers simultaneously translated uncertainty about the puzzling effects of menopause into certainty about its pathology. They demonstrated that although menopause represented a breakdown in the body's ability to reproduce a regular menstrual cycle, DES would create a new one. DES could replace oestrogen in a controlled, standardised way. Indeed, whereas women's bodies typically do not always have twenty-eight-day cycles, DES-controlled bodies could. In their resolution of uncertainty about DES, medical scientists created a new menstrual cycle and the possibility of a well-regulated human female. (107) By linking gender and the resolution of uncertainty; I have shown how medical science reflects and creates gender.

What the DES case adds to our understanding of the reflection and creation of gender in medical science is that the process of gender construction involves both ambiguity and resistance. There was not unanimity about the safety of DES, its effects on the bodies of menopausal women, or its use to create and then regulate new "menstrual" cycles. Indeed, there was a culture of resistance in medicine, both to diseasing menopause and to replacing women with well-regulated human females. Most of the medical scientists reviewing DES believed that only by judging the signs and symptoms of women could they practice good medicine. To them, the use of a laboratory test in medical practice signalled the disappearance of the physician (and, implicitly, the woman patient) from the picture. But most of them also agreed that DES could be a useful way of controlling the bodies of these women, thereby embracing the notion of the well-regulated human female. They wanted both - a menopausal woman, whom they could treat safely, and an aging human female, whose nonreproductive body they could regulate exactly.

The contradictions, ambiguities, and resistances revealed in the analysis of the development of DES for menopause remain with us. The terms of the contest have shifted, from whether oestrogen therapy should be offered to women during menopause for relief of symptoms to whether hormone therapy should be considered by postmenopausal women for prevention of future disease. Instead of regulating nonreproductive bodies for a relatively short length of time (one to five years), the potential for regulating them has extended to twenty years or more. In addition, Premarin, an oestrogen extracted from the urine of pregnant mares, is now the most popular ingredient in hormone therapy during and after menopause. (108) However, even today, medical scientists acknowledge that "there is no consensus within the medical community about even the definition of menopause, let alone the risks and benefits associated with hormone therapy." (109) As feminist scholars approaching, passing through, and living after menopause, we confront these uncertainties and contradictions both in our research and our personal lives.

NOTES

This paper was presented at Vassar College, the Ninth Berkshire Conference on the History of Women, "Transformations: Women, Gender, Power," Poughkeepsie, New York, 11 June 1993. Thanks are due to Diana Long for continuing discussions about the issues addressed in this essay, to Roberta Apfel, Adele Clarke, Evelynn Hammonds, Philip Hart, David Kertzer, Craig McEwen, and Irving Kenneth Zola for thoughtful comments on a draft of it; to Susan Reverby for suggestions at the last stages of its completion; and to Suzanne White, Scott Lauze, and Katy Biron for research assistance. My research was funded by a Grant-in-Aid from the American Council of Learned Societies and a grant from the Faculty Development Fund, Bowdoin College.

1. Barbara Seaman, The Doctor's Case against the Pill (New York: Doubleday, 1969); Barbara Ehrenreich and Deirdre English, Witches, Midwives, and Nurses: A History of Women Healers, Glass Mountain Pamphlet No.1 (New York: Feminist Press, n.d.); Barbara Ehrenreich and Deirdre English, Complaints and Disorders: The Sexual Politics of Sickness, Glass Mountain Pamphlet No. 2 (New York: Feminist Press, 1973).

2. See, for example, Catherine Kohler Riessman, "Women and Medicalisation: A New Perspective," Social Policy 14 (summer 1983): 3-18; Judith Walzer Leavitt, Brought to Bed: Childbearing in America, 1750-1850 (New York: Oxford University Press, 1986); Rima D. Apple, ed., Women, Health, and Medicine in America: A Historical Handbook (New York: Garland, 1990).

3. Diana Long Hall, "The Social Implications of the Scientific Study of Sex" (Paper delivered at the "Feminist IV: Connecting Theory, Practice, and Values Conference," Barnard College, New York City, New York, 23 Apr. 1977); Susan E. Bell, "Changing Ideas: The Medicalisation of Menopause," Social Science and Medicine 24, no. 6 (1987): 535-42; Adele E. Clarke, "Research Materials and Reproductive Science in the United States, 1910-1940," in Physiology in the American Context, 1850-1940, ed. Gerald L. Geison (Bethesda, MD.: American Physiological Society, 1987); Nelly Oudshoorn, "On the Making of Sex Hormones: Research Materials and the Production of Knowledge," Social Studies of Science 20 (1990): 5-33, and "On Measuring Sex Hormones: The Role of Biological Assays in Sexualising Chemical Substances," Bulletin of the History of Medicine 64 (summer 1990): 243-61.

4. Faye Ginsburg and Rayna Rapp, "The Politics of Reproduction," Annual Review of Anthropology 20 (1991): 311-43.

5. Joyce Bichler, DES Daughter: The Joyce Bichler Story (New York: Avon, 1981); Roberta J. Apfel and Susan M. Fisher, To Do No Harm: DES and the Dilemmas of Modern Medicine (New Haven: Yale University Press, 1984); Diana B. Dutton, Worse Than the Disease: Pitfalls of Medical Progress (Cambridge: Cambridge University Press, 1988).

6. Boston Women's Health Book Collective, The New Our Bodies, Ourselves (New York: Simon & Schuster, 1992).

7. On the early history of DES, see Susan E. Bell, "A New Model of Medical Technology Development: A Case Study of DES," Research in the Sociology of Health Care 4 (1986): 1-32. On the hormonalisation of women, see Nelly Oudshoorn, Beyond the Natural Body: An Archaeology of Sex Hormones (New York: Routledge, 1994). On recent research about DES, see National Institutes of Health, Office of Women's Health, Long-Term Effects of Exposure to Diethylstilbestrol (DES) (Washington, D.C.: GPO, 1992).

8. Bell, "Changing Ideas."

9. Here I focus on medical discourse. For a review of images of menopause in the popular press (1900-1976) see Linda S. Mitteness "Historical Changes in Public Information about Menopause," Urban Anthropology 12 (summer 1983): 161-79. She found that prior to 1950, "menopause was given both positive and negative values" (171). For a contrasting view, see Marilyn Maxwell, "Portraits of Menopausal Women in Selected Works of English and American Literature," in The Meaning of Menopause: Historical Medical and Clinical Perspectives, ed. Ruth Formanek (Hillsdale, N.J.: Analytic Press, 1990). Maxwell analysed a small number of literary texts written during this same time frame and found them filled with negative, stereotypic images of middle-aged women. See also Emily Martin, The Woman in the Body: A Cultural Analysis of Reproduction (Boston: Beacon Press, 1987) for a comparison of women's experiences and understandings of menopause with medical literature in the 1970s. Women's experiences of menopause during the 1930s and 1940s - expressed in their letters and diaries or through research interviews - have yet to be explored.

10. Emily Martin, "The Egg and the Sperm: How Science Has Constructed a Romance Based on Stereotypical Male-Female Roles," Signs 16 (spring 1991): 501.

11. Diana E. Long, "Not Just Words: 'Women' and 'Gynaecology' in the Index Catalogue of the Library of the Surgeon General's Office, 1880-1950" (Paper delivered at the conference, "Women and Science, Feminism and Science," University of Minnesota, Minneapolis, 14 May 1995).

12. Donna Haraway, "A Manifesto for Cyborgs: Science, Technology, and Socialist Feminism in the 1980s," Socialist Review 15 (March-April 1985): 65-107.

13. See Martin, The Woman in the Body.

14. Descriptions of recalcitrant instruments, materials, and techniques are usually deleted from published work. On this point see Bernard Barber and Renee C. Fox, "The Case of the Floppy-Eared Rabbits: An Instance of Serendipity Gained and Serendipity Lost," American Journal of Sociology 64 (September 1958): 128-36; Susan Leigh Star, "Scientific Work and Uncertainty," Social Studies of Science 15 (1985): 391-427.

15. Bruno Latour and Steve Woolgar, Laboratory Life: The Social Construction of Scientific Facts (Beverly Hills, California: Sage Publications, 1979); Haraway; Evelyn Fox Keller, Reflections on Gender and Science (New Haven: Yale University Press, 1985); Sandra Harding, The Science Question in Feminism (Ithaca: Cornell University Press, 1986); Martin, "Egg and the Sperm."

16. Martin, "Egg and the Sperm," 501.

17. Emily Martin, "The End of the Body?" American Ethnologist 19 (February 1992): 122. See also The Woman in the Body.

18. Ibid.

19. On imagery in scientific papers, see Bell, "Changing Ideas." On imagery in medical textbooks see Martin, The Woman in the Body.

20. George Corner, The Hormones in Human Reproduction (Princeton: Princeton University Press, 1947); Diana Long Hall, "Variations on a Theme: Research on the Oestrus and Menstrual Cycles, 1920-1940" (Paper delivered at the History of Science Society, Atlanta, December 1975); Oudshoorn, "On Measuring Sex Hormones."

21. Corner, 173, 64, 63.

22. Adele Clarke, "Women's Health over the Life Cycle," in The History of Women, Health, and Medicine in America: An Encyclopaedic Handbook, ed. Rima D. Apple (New York: Garland, 1990); Oudshoorn, Beyond the Natural Body.

23. Oudshoorn, "On Measuring Sex Hormones," 258.

24. Oudshoorn, Beyond the Natural Body, 60. Although the term oestrus was introduced in 1901 by physiologist Walter Heape, he used it to refer to recurrent periods of sexual excitement in both females and males, see Corner, 63.

25. Oudshoorn, "On Measuring Sex Hormones."

26. Kenneth L. Noller and Charles R. Fish, "Diethylstilbestrol Usage: Its Interesting Past, Important Present, and Questionable Future," Medical Clinics of North America 58 (July 1974): 793-810; Apfel and Fisher; Dutton.

27. See Susan E. Bell, "The Synthetic Compound Diethylstilbestrol (DES), 1938- 1941: The Social Construction of a Medical Treatment" (Ph.D. diss., Brandeis University, 1980), chap. 3; Sir Charles Dodds, "Stilbestrol and After," Scientific Basis of Medicine: Annual Review (1965): 1-16; "Chemotherapy in Gynaecology," The Lancet (14 Jan. 1939): 97-98.

28. See Bell, "Changing Ideas," for a discussion of medical uncertainty and the role of DES in the medicalisation of menopause. For more about continuing medical uncertainty about the meaning of menopause, see U.S. Congress, Office of Technology Assessment, The Menopause, Hormone Therapy, and Women's Health, OTA-BP-BA- 88 (Washington, D.C.: GPO, 1992).

29. Jean Paul Pratt, "Treatment of the Menopause," Southern Medical Journal 31 (May 1938): 564; Bell, "Changing Ideas," 53, 58.

30. Paul N. Leech, "Preliminary Report of the Council on Pharmacy and Chemistry", Journal of the American Medical Association 13 (23 Dec. 1939): 2312.

31, Ephraim Shorr, Frank H. Robinson, and George N. Papanicolaou, "Clinical Study of the Synthetic Oestrogen Stilbestrol," Journal of the American Medical Association 12 (23 Dec. 1939): 2312-18; Leech.

32. Shorr, Robinson, and Papanicolaou, 2317.

33. Susan E. Bell, "From Local to Global: Resolving Uncertainty about the Safety of DES in Menopause," Research in the Sociology of Health Care 11, (1994): 41-56.

34. For more on this topic, see Bell, "The Synthetic Compound Diethylstilbestrol," chaps. 1 and 5. 35. Ibid., 250.

36. Memorandum of a Meeting of the Small Committee of Pharmaceutical Manufacturers Interested in Stilbestrol, 24, 25 Mar. 1941, Pay ton v. Abbott; Memorandum of an Interview, 25 Mar. 1941, fiche 0152, DES NDA file. The historical materials analysed here can be found at the offices of the Food and Drug Administration, Rockville, Maryland, through a Freedom of Information Act request. They are drawn from the FDA files of New Drug Applications (NDAs) submitted by pharmaceutical manufacturers interested in marketing DES. For each firm, the FDA kept a separate file. Copies of the questionnaires are available only on microfiche, in separate files: Fiche 0163, Small Committee Report, DES NDAs; and Fiche 0051, Upjohn NDA for DES. Of the questionnaires, I could locate fifty-three of the fifty- four originally completed. The names of some experts had been deleted before reproducing the materials on microfiche; sometimes I could identify the writer and sometimes I could not. I have indicated in notes the name of the writer, if I have been able to identify it.

37. E.C. Hamblen, "Diethylstilbestrol Commercially Available," editorial, Journal of Clinical Endocrinology 1 (1941): 928-29.

38. The first questionnaire was sent in February and March of 1941. The second was sent after officials from the FDA met with drug company representatives in March and outlined some fundamental questions about the safety of DES that remained unanswered. The second questionnaire was completed by thirty-five to forty of the experts; some of them appended it to the first. Most of the names of experts have been deleted from the second questionnaire. When I could identify the writer, I have indicated this in notes.

39. Bell, "A New Model of Medical Technology Development."

40. Bell, "The Synthetic Compound Diethylstilbestrol."

41. American Medical Association, American Medical Directory (Chicago: AMA, 1940). For estimates of the proportion of specialists in 1940, see Rosemary Stevens, American Medicine and the Public Interest (New Haven: Yale University Press, 1971), 218.

42. Bell, "The Synthetic Compound Diethylstilbestrol."

43. Silent but not invisible in the process of constructing the medical science of DES were the women whose bodies metabolised DES and whose lives were affected by it. Today, women cannot be silenced so thoroughly. In the late 1930s, men took it for granted that women were excluded from participation, even as clinical researchers. Today, even if problematically, women must be included in the review of new drugs; women are included in the profession of medicine; and women patients do speak and at times resist the voice of medicine. For more on resistance to the voice of medicine, see Elliot G. Mishler, The Discourse of Medicine: Dialectics of Medical Interviews (Norwood, N.J.: Ablex Publishing Corp., 1984); Alexandra Dundas Todd, Intimate Adversaries: Cultural Conflict between Doctors and Women Patients (Philadelphia: University of Pennsylvania Press, 1989); Susan E. Bell and Roberta J. Apfel, "Looking at Bodies: Insights and Inquiries about DES-Related Cancer," Qualitative Sociology 18 (Spring 1995): 3-19.

44. See Dutton, 37. According to her review of the literature, by 1939 more than forty articles had been published that documented the carcinogenic effects of natural and synthetic estrogens (including DES) in animals. For example, in January 1939, an article linking DES exposure to breast cancer in rats had appeared in Science.

45. For more on how contexts shape the kinds of questions asked in medical science, see Adele E. Clarke, "Social Worlds/Arenas Theory as Organisational Theory," in Social Organization and Social Process: Essays in Honour of Anselm Strauss, ed. David R. Maines (New York: Aldine de Gruyter, 1990).

46. The strategies and dilemmas in my essay are the same as those in the DES case. See Susan Leigh Star, Regions of the Mind: Brain Research and the Quest for Scientific Certainty (Stanford: Stanford University Press, 1989). Here, I have chosen two uncertainties I think are especially crucial and illuminating for understanding the relationship between gender and the resolution of uncertainty. In a related essay, I focus on how medical scientists attempted to resolve local uncertainties about whether DES caused liver damage and what techniques to use for determining dosage and effects. For this discussion of uncertainty, see Bell, "From Local to Global."

47. Leech. The council was empowered by the AMA to screen new drugs and devices and to make recommendations about them to the medical profession (Bell, "The Synthetic Compound Diethylstilbestrol," 105-7).

48. Leech, 2312.

49. Bell, "The Synthetic Compound Diethylstilbestrol," 262.

50. Portions of the discussion that follows appear in Bell, "From Local to Global." 51. Bell, "Changing Ideas," 537.

52. It is beyond the scope of this article to look more deeply at the long history of political and intellectual struggle between laboratory and clinical scientists. Nelly Oudshoorn carefully documents how different types (natural or synthetic), sources (ovaries, urine), and test methods for sex hormones shaped power relationships among laboratory scientists, clinicians, and pharmaceutical companies. See her Beyond the Natural Body. By the 1930s, laboratory scientists, along with pharmaceutical companies, increasingly dominated the field of hormone research. For more detail about these interrelationships in the DES case, Bee Bell, "The Synthetic Compound Diethylstilbestrol" and "A New Model of Medicine Technology Development." 53. Francis C. O'Connor to D.C. Hines, Lily Research Laboratories, 4 Feb. 1941, Fiche 0051, Upjohn NDA for DES.

54. Willard M. Allen to J. Murray Scott, Upjohn, 28 Feb. 1941, Fiche 0163, Small Committee Report, DES NDAs.

55. Harold D. Palmer to D.C. Hines, Lily, 14 Oct. 1940, Fiche 0051, Upjohn NDA for DES.

56. Robert T. Frank, Fiche 0163, Small Committee Report, DES NDAs.

57. In this essay, I focus on how images of continuous production and information processing lead to negative views of women. Elsewhere I focus on how negative stereotypes about aging women led to the medicalisation of menopause. See Bell, "Changing Ideas."

58. Donato A. D'Esopo, Fiche 0051, Upjohn NDA for DES.

59. E. Kost Shelton, Fiche 0163, Small Committee Report, DES NDAs.

60. All the experts who discussed this question responded yes, they did consider the control of symptoms the object desired.

61. Martin, The Woman in the Body.

62. Bell, "Changing Ideas."

63. M.E. Davis, Fiche 0163, Small Committee Report, DES NDAs.

64. Bell, "Changing Ideas."

65. Abraham R. Abarbanel, Fiche 0051, Upjohn NDA for DES.

66. Frank N. Allan, Fiche 0163, Small Committee Report, DES NDAs.

67. Questions about DES were asked in the framework of a more general debate, as clinicians considered the worth of clinical judgement versus laboratory tests in medicine. At the time, there was ambivalence among clinicians about how best to practice medicine: how closely medical science could/should approximate laboratory science. See Bell, "Changing ideas." It is beyond the scope of this essay to consider the larger debate. See Oudshoorn, Beyond the Natural Body, for a detailed analysis of this debate.

68. E.L. Sevringhaus to J.J. Durrett (FDA), 4 Jan. 1941, Fiche 0163, Small Committee Report, DES NDAs.

69. Ephraim Shorr , "The Menopause," Bulletin of the History of Medicine 16 (July 1940): 453-74.

70. Bell, "Changing Ideas."

71. New diagnostic techniques had uncertain implications for medicine. Although they could serve to strengthen medicine's cultural authority by reducing physicians' dependence on patients, they could erode this authority by increasing medicine's dependence on capital equipment, formal organisations, and nonphysician laboratory workers. See ibid.; and Paul Starr, Social Transformation of American Medicine (New York: Basic Books, 1982).

72. Emil Novak, Gynaecology and Female Endocrinology (Boston: Little Brown, 1941), 470. According to Novak, clinical judgement could be no less rigorous than laboratory science.

73. Bell, "The Synthetic Compound Diethylstilbestrol."

74. Shorr, Robinson, and Papanicolaou.

75. E.L. Sevringhaus to J.J. Durrett (FDA), 4 Jan. 1941, Fiche 0163, Small Committee Report, DES NDAs.

76. Seven experts did not answer this question or gave responses that were illegible. 77. Abarbanel, Fiche 0051, Upjohn NDA for DES.

78. Emil Novak, Fiche 0163, Small Committee Report, DES NDAs.

79. John B. Plastino, ibid.

80. Novak, ibid.

81. Conrad G. Collins, ibid.

82. Starr, 14.

83. Riessman; Todd. Patriarchal relationships between physicians and women were also supported by the investigators' assumption that if DES were sold, it should not be sold directly to the public. They did not believe that women could determine oestrogen therapy was warranted, nor could they use DES safely without medical guidance. In the first questionnaire, they were asked whether DES should "be made generally available to the medical profession." None of them questioned the assumption that lay behind this question, that physicians should be the gate-keepers to treatment with DES, in the event that it was safe enough to market. Some explicitly supported it, as in C.J. Ehrenberg's recommendation that DES "is a drug which under no circumstances, should be available for self-medication. . . " (Fiche 0163, Small Committee Report, DES NDAs).

84. Shorr, Robinson, and Papanicolaou. See also C.L. Buxton and Earl T. Engle, "Effects of the Therapeutic Use of Diethylstilbestrol" (2318-20), and Cyril MacBryde, Harold Freedman, and Ellen Loeffel, "Studies on Stilbestrol" (2320), both in Journal of the American Medical Association 13 (23 Dec. 1939). Buxton and Engle, from the Department of Anatomy at the Sloane Hospital for Women and the Department of Obstetrics and Gynaecology at Columbia, found nausea and occasional vomiting in 24 percent of their patients. The incidence of nausea was approximately 21 percent in the study conducted by MacBryde, Freedman, and Loeffel from the Department of Medicine, Washington University School of Medicine. Vomiting occurred less frequently than nausea among women in this study.

85. Shorr and his associates believed that the toxic effects were central in origin (because they occurred following injection or oral administration), that there was no relation between the size of the dose and the occurrence of toxic effects, and that women did not acquire tolerance to DES over time. Buxton and Engle found that nausea and vomiting were unrelated to dosage, and in all but one of the cases transitory. They believed that the proportion in their study was smaller than that reported in other series "because with a few exceptions the medication was placed in gelatine capsules." See Buxton and Engle, 2318. According to MacBryde, Freedman, and Loeffel, the toxic effects were caused by administering large doses of DES; they ceased when the dose was reduced.

86. Jean Paul Pratt, "Sex Functions in Man," in Sex and Internal Secretions: A Survey of Recent Research, ed. Edgar Allen, Charles H. Danforth, and Edward A. Doisy (Baltimore: Williams & Wilkins Co., 1939); Shorr.

87. See note 28.

88. See Martin, The Woman in the Body.

89. Author unknown [#100], second questionnaire, Fiche 0052, Upjohn NDA.

90. Author unknown, from the Mayo Clinic to G.R. Hazel, 21 May 1941, Fiche 0053, Upjohn NDA.

91. Author unknown [#154], second questionnaire, Fiche 0053, Upjohn NDA.

92. Norman R. Kretzschmar, Fiche 0163, Small Committee Report, DES NDAs.

93. Emmerich von Haam to J.B. Rice, 26 Feb. 1941, Fiche 0163, Small Committee Report, DES NDAs.

94. Plastino, ibid.

95. Jean P. Pratt to J.A. Morrell, 12 May 1941; von Haam to J.B. Rice, 26 Feb. 1941, both references from ibid.

96. James Albert Corscaden and Jacob P. Greenhill, ibid.; Maximilian A. Goldzieher, source unknown.

97. Cyril MacBryde, second questionnaire, Fiche 0052, Upjohn NDA. 98. Author unknown [#108], ibid.

99. Dutton, 39.

100. Bell, "The Synthetic Compound Diethylstilbestrol," 209.

101. Author unknown [#108), second questionnaire, Fiche 0052, Upjohn NDA. 102. Keller, Gender and Science; Latour and Woolgar.

103. Bruno Latour, "Give Me a Laboratory and I Will Raise the World," in Science Observed: Perspectives on the Social Study of Science, ed. Karin Knorr-Cetina and Michael Mulkay (Beverly Hills, California: Sage Publications, 1983).

104. See Evelyn Fox Keller, Secrets of Life, Secrets of Death (New York: Routledge, 1992), esp. chap. I.

105. Martin, The Woman in the Body.

106. On menopause, see Bell, "Changing Ideas." On pregnancy, see Susan E. Bell, "Technology in Medicine: Development, Diffusion, and Health Policy," in Handbook of Medical Sociology, 4th ed., ed. Howard E. Freeman and Sol Levine (Englewood Cliffs, N.J.: Prentice-Hall, 1989).

107. Long, "Not Just Words."

108. In 1992, the American College of Physicians published a set of guidelines for clinical practice, recommending that all postmenopausal women should consider hormone therapy to prevent osteoporosis and coronary heart disease. The effect of the guidelines on hormone use is not yet known. In another essay, I explore some of the uncertainties and resistances within medicine concerning hormone therapy during and after menopause. For more on this topic, see American College of Physicians, "Guidelines for Counselling Postmenopausal Women about Preventive Hormone Therapy, Annals of Internal Medicine 117 (15 December 1992): 1038-41. For critical reviews of the recommendation, see Lynn Rosenberg, "Hormone Replacement Therapy: The Need for Reconsideration," American Journal of Public Health 83 (December 1993): 1670-73; Susan E. Bell, "Technology Assessment, Outcome Data, and Social Context: The Case of Hormone Therapy," in Technology Assessment: Uses, Context, and Interpretation, ed. Philip Boyle and Daniel Callahan (Washington, D.C.: Georgetown University Press, 1996).

109. U.S. Congress, Office of Technology Assessment, The Menopause, 106.