CHAPTER 11

Genetic Ratios and Medical Numbers Give Rise to Big Data Ambitions in America, 1902-1920

Take the single matter of feeble mindedness. I presume there are over a score of institutions for such people in the country and in many of them as I know by correspondence, not only records of heredity but detailed studies of families [have] been made.

—Charles B. Davenport (18 Mar. 1909)

It would seem quite as proper and important that each person of the family in which we are interested should have a numerical name as that each plant should be so designated in the plant breeding nursery. A name with eleven letters is not over long.

—Willet M. Hays (1912)

Although Mendelian genetics had its origins in experimental breeding, it was applied almost immediately to human abnormalities. While “blended” characters, as Francis Galton called them, seemed to resist Mendelian analysis, the new geneticists looked for discrete, or “segregated,” factors underlying continuous variability. Congenital and lingering conditions such as tuberculosis as well as mental deficiency and insanity had long been supposed hereditary. Even a known infectious agent left room for an inherited diathesis (a nonspecific factor) that could determine susceptibility to disease. Some conditions, however, could not be neatly classified, with insanity and feeblemindedness in the first rank of diagnostic ambiguity. Mendelian medical heredity, in contrast to the more generalized “biometric” or biostatistical variety, was scarcely thinkable without neat categories. The suppleness of biometry, in which Pearson took such pride, appeared to Bateson as its Achilles heel, encouraging its application to sloppy data to reach meaningless conclusions.1 Pearson countered that Mendelism, too, depended on numerical data, adding that Mendelians often tortured their data to provide spurious confirmation of what they supposed to be already known.

Bateson’s plant experiments, like Mendel’s, were data intensive, and chromosome mapping in Thomas Hunt Morgan’s fly lab at Columbia University was an exercise in applied probability. The most ambitious eugenic project in America was undertaken in the name of Mendelism by an American biologist who followed and then fell out with Pearson. Yet Charles B. Davenport, commander in chief of eugenic (or social) Mendelism, lived his life for data. Although his research was funded at a princely level, he recognized that his laboratory could never come close to matching the financial resources of state institutions. He soon discovered that the doctors and psychologists in these institutions had already built up impressive expertise on methods of researching heredity. He allied with them to train and send out eugenics fieldworkers, who constructed pedigrees of patients from dozens of institutions. Davenport turned up the classic Mendelian proportions, 3:1 and 1:1, for almost every condition, however ill-defined, that mattered to eugenics. Although these claims were subjected to withering criticism almost from the beginning, his project appeared for some time as a remarkable success story. Afterward it became one of the best-known tales of eugenic and genetic hubris.

Animal Models and Human Heredity

When Raymond Pearl, another biometric pioneer, arrived in London for an apprenticeship with Pearson, he could scarcely believe what cold, cramped quarters confined the professor and his assistants, or how little time remained after teaching in which to carry out such prodigious research. “The great biometric laboratory of University College is all comprised of one room with two windows, the size of the room being just that of Room 3 at Ann Arbor.”2 Pearl, like Davenport, was among those captains of science made possible by the plutocratic wealth of the gilded age. Already, in America, dollars made the visionary. Davenport, in an autobiographical reflection, stressed the strong mathematical education he had received at Brooklyn Polytechnic. He began teaching biology from a statistical perspective in the early 1890s, and his book on statistics and biological variation introduced his colleagues to Pearson’s methods.3 By 1902, when he met Galton, Weldon, and Pearson in England, he was esteemed, not least by the Carnegie Institution, as one of America’s leading scientists. From Paris, a few months later, he wrote to Willet Martin Hays of the Agricultural Experiment Station in Minnesota to propose that the newly created American Breeders’ Association (ABA) “urge upon the trustees of the Carnegie fund the establishment of a station for studying heredity.” By early 1904, the station was almost ready, with Davenport, the well-paid director, hatching schemes for chicken breeding.4

These birds were a secondary topic in his first letter to Bateson, whom he invited to serve as a correspondent of the new laboratory. Bateson, accepting, endorsed the merits of chickens for his experiments then reminded Davenport that postage to Europe was 5 cents per 1/2 ounce. While English scientists minded their pennies, the Carnegie solicited their expert advice on grants to American scientists they thought undistinguished, promising resources beyond their dreams. In a 1907 report to the Carnegie Institution on American research, Pearson complained of men who gather data indiscriminately then wonder what to do with it. “The first point in any work in this field is to have a real biological, anthropological, or sociological problem which needs solution.” Three years later, he removed Davenport from the board of editors of Biometrika. The great statistician was simply hostile to Mendelism, Davenport complained, while Pearson charged Davenport with careless and superficial work that threatened to discredit biometry.5 An American textbook by the geneticist William Castle sized up the situation in 1916, declaring Pearson’s data to be sound though his analysis was corrupted by hostility to Mendelism, while American eugenic data was unreliable on account of fieldworkers trained to assume that inheritance reduced to the presence or absence of a Mendelian factor. Davenport claimed in 1926 that “the development of Mendelism has led to the general introduction of mathematics into genetics.” Pearl, a few years earlier, called one of Davenport’s books a “bad tabulation of bad statistics.”6

Davenport never thought of limiting his genetics to farm animals; he had been keen on eugenics at least since his first discussions with Galton and Pearson. His correspondence with Alexander Graham Bell demonstrates the interpenetration of agricultural and eugenic breeding. Bell, a learned advocate for the deaf, was also a noted breeder of sheep at his estate on Cape Breton Island, Nova Scotia. Davenport initiated the correspondence in 1904 as he was setting up his laboratory, asking for four-nipple sheep to use in hybridization experiments that, he hoped, would demonstrate inheritance of this “sport,” or anomaly, “in accordance with Mendel’s laws.”7 Bell responded that his sheep had had multiple nipples for generations and so were not true sports, but he was happy to assist the research. An ensuing request for six-nipple sheep flummoxed him, since, though reluctant to donate an animal so rare and valuable, he considered himself a benefactor of research rather than a profiteer. In May 1906, he sent Davenport the breeding results of sheep born that year, and in July he arranged for Davenport to receive the recently issued census report on the blind and deaf of the United States, which had been prepared under his direction. The chapter on marriages of the deaf and their progeny was especially worthy of attention, he said. “It has generally been assumed that the laws of heredity that are known to apply to animals also apply to man but I do not know of any large collection of statistics that demonstrate the proposition with the exception of the Census returns relating to the Deaf.” He volunteered to ask the Census Office about their unprinted tables on marriages of the blind. “It has just occurred to me that this might be a matter of interest to the Carnegie Institution, Station for Experimental Evolution, and that you might perhaps like to obtain copies of these Tables.”8

Davenport certainly did want those tables, but he was now focusing on “inheritance of color” in offspring of whites and blacks of the human variety and was planning a trip to Jamaica. Bell told him of the Maroons of Jamaica, descendants of slave refugees now living across the bay from Halifax, Nova Scotia. Potentially of still greater value were the census records, including manuscript returns going back to 1810, stored in the rooms of the Volta Bureau, which he had established to study deafness. Bell, who had used them to investigate the ancestry of deaf-mutes, was sure that Davenport could extract similar results for inheritance of color.9 Records like these were now Davenport’s eugenic sustenance. Bell inquired a month later if he had ever investigated “Deafness in White Cats that have blue eyes,” which Darwin had mentioned. Early in 1907, they discussed whether American eugenics should continue to be organized within the American Breeders’ Association. Davenport said yes, on the grounds that the laws of heredity were indifferent to species. Mendel’s principles could not have been worked out in full from studies of man alone.10

Hopelessly Vicious Protoplasm

In these years, the preeminent journal of genetics in America, and perhaps the world, was the American Naturalist. Davenport was an active member of its sponsoring society, the American Society of Naturalists. Its business brought him into contact with men like anthropologist Franz Boas and psychologist Edward Thorndike.11 It was, however, the ABA that, in 1906, formed a committee on eugenics as one of the “general subjects” in a miscellany that included Animal Hybridizing, Breeding for Dairy Production, Cooperative Work in Plant Breeding, Prize Competitions, and Theoretical Research in Heredity.12 Hays, now Theodore Roosevelt’s assistant secretary of agriculture, was equally devoted to eugenics. In 1910, when the society upgraded its newsletter to the American Breeders Magazine: A Journal of Genetics and Eugenics, Davenport held forth there in praise of eugenics. Man’s nature “follows the laws of the rest of the organic world,” and the affiliation with breeders would provide “dignity and safety” against quacks and popularizers. “Our greatest danger is from some impetuous temperament who, planting a banner of Eugenics, rallies a volunteer army of Utopians, freelovers, and muddy thinkers to start a holy war for the new religion.”13 Soon, he raised eugenics into a section of the society while letting the other sections wither. He envisioned a swarm of committees and subcommittees devoted to feeblemindedness, insanity, deafness, eye defects, and so on.14 The work required oceans of data and would not be cheap. Free-lovers, indeed, were not welcome, but the religious fervor he swore to lock out had already established a stronghold within. Ten millions to “redeem mankind from vice, imbecility and suffering,” Davenport proclaimed, does far more good than ten millions for charity. There loomed, after all, a satanic enemy. “Society must protect itself; as it claims the right to deprive the murderer of his life so also it may annihilate the hideous serpent of hopelessly vicious protoplasm.”15

In March 1907, when Hays inquired what tasks this new eugenics committee might pursue, Davenport was ready. The problem, he began, is so large that twenty men could work at it full time. Their work must be statistical. Thinking, no doubt, of Galton’s book English Men of Science, he proposed to study the ancestries of James McKeen Cattell’s “1000 first men of science in America.” He next mentioned the inheritance of physical characters in man and then hereditary mechanisms, including prepotency, hybridization, transmission of acquired characters, and the possible effect of maternal impressions on the fetus.16 “Supply blanks to lying-in hospitals,” he inserted here, in unwitting obeisance to asylum tradition of hereditary investigation based on filled-out forms. Also on the list were “insanity in various forms, criminal tendencies, various idiosyncrasies.”17 Early in 1909, the committee worked up a proposal to add eugenic questions to the US census. The suggestions were too late to be incorporated into the 1910 census, as they promptly learned from the House and Senate committee chairmen.18

Undaunted by this failure, the ABA formed a committee in 1912 to explore the possibilities of securing data through the Census Bureau, the Bureau of Health, “and other societies and institutions.” Davenport sketched out his hopes in a letter to Elmer Ernest Southard, a pathologist for the Massachusetts asylum system: “In brief, what is needed is a qualitative census. I suppose the original purpose of the census was military and a simpleton will stop a bullet as well as a genius but now-a-days it is the fashion to take stock of natural resources and it would seem that the State should take stock of its germ plasms and their product.”19 An undated document from about the same time outlines a project for cooperation with the U.S. Bureau of Industry and Immigration to map the geography of flawed heredity. “Are there foci of criminality and defectiveness in Europe which are supplying an exceptionally large proportion of the immigrants who become charges on the State? If so, locate centers and find their characteristic defects.” At other moments he emphasized non-European races such as American Indians alongside these immigrants from less desirable parts of Europe.20 But government statisticians did not cooperate. In 1917, he transmitted a modest resolution to the director of the census, Samuel Rogers, proposing that in 1920 and thereafter the census enumerators should record, for every male, the father’s name. It sounds easy, Rogers replied, but the United States has 100,000,000 inhabitants, “and the task of collecting, compiling, and tabulating these data is an enormous one.”21

Davenport’s most ambitious data scheme did not depend on legislation or federal agencies. He would line up experts to mine the records of institutions for persons with distinctive hereditary traits. Mainly, this meant socially undesirable characters for which states were increasingly taking responsibility. Each type of institution had its experts, who, he soon discovered, were just as interested in heredity as he was. This realization opened a new world of opportunity to him. He articulated his updated dream to David Starr Jordan, the biologist-president of Stanford University and Davenport’s choice as chairman of the ABA Eugenics Committee. They should create subcommittees to investigate “feeble-mindedness, pauperism, psychiatry, the deaf, dumb and blind, cripples, criminals.” Happily, they would not need to start from scratch. “There is a large number of institutions devoted to aberrant individuals belonging to each of these classes and some of them have departments of research with directors interested in the subject of heredity. Take the single matter of feeble mindedness.”22

A folder in the archives names six possible members of the subcommittee on heredity of the feebleminded, four for insanity, five for epi.tify, two for deaf-mutism, and four for criminality. He could only think of one candidate, W.E.B. DuBois, for the subcommittee on “heredity in Negro-white crosses,” and one also for the heredity of eye defects. Six more bodily conditions, such as cancer and physical strength, remained, for the moment, “unorganized,” as did his envisioned subcommittee on “laziness and physical basis of poverty.”23 The focus on persons “under the control of the state” was not by choice, Davenport explained some years later in a letter to Bell. Originally, his office had intended to study gifted persons as well, but it is “very much harder” to secure their cooperation. He now recognized the advantages of this unintended focus: these traits had already been “very carefully analyzed” and “minutely classified,” and the fact of state control made the individuals and their families much more accessible to research. Finally, state institutions covered half the costs and conferred a measure of public recognition on the work of eugenics.24

Data for Mendel

Davenport’s ideal subcommittee included a man of consequence as a figurehead chairman and an industrious secretary to organize the labor. He installed himself as secretary of the whole Committee on Eugenics and set to work on its first and most urgent need, data. Happily, he now realized, the data were all around.

They lie hidden in records of our numerous charity organizations, our 42 institutions for the feeble-minded, our 115 schools and homes for the deaf and blind, our 350 hospitals for the insane, our 1,200 refuge homes, our 1,300 prisons, our 1,500 hospitals and our 3,500 almshouses. Our great insurance companies and our college gymnasiums have tens of thousands of records of the characters of human blood lines. These records should be studied, their hereditary data sifted out and properly recorded on cards and the cards sent to a central bureau for study in order that data should be placed in their proper relations in the great strains of human protoplasm that are coursing through the country.

With these reservoirs of information, he could pinpoint the “lines which supply our families of great men” as well as the insane and feeble-minded, blind and deaf, prisoners, criminals, and paupers. The same data would advance understanding of genetic mechanisms, the “method of heredity of human characteristics.”25

The Eugenics Record Office (ERO), which he founded in 1910 next door to the Station for Experimental Evolution, became a storehouse of hereditary data. The records of so many schools, prisons, and asylums provided a million points from which to launch investigations. A series of donations by Mary Williamson Harriman, heiress to a railroad fortune, and others provided funds to train and to employ, over the next fifteen years, 257 eugenics fieldworkers. He always looked to the institutions where they labored to pick up as much as possible of the cost. The work drew on interpersonal skills that he regarded as distinctively feminine: a degree of tact in dealing with defective persons and the grace to coax their relatives into discussions of conditions that ran in the family. He took immense pride in the system of fieldworkers, which, he thought, raised his data to an incomparably higher level than the familiar institutional tables.

He was not shy, however, about using pedigrees from earlier researches, going back to Dahl.26 He also followed Galton’s example, encouraging eugenic-minded citizens to assemble records of as many family members as they could and to send them to the ERO. The office printed The Family-History Book, incorporating forms and schedules supplied by physicians and institution directors, to guide these volunteers, and The Trait Book, laying out standards.27 The family relationships could be reduced to a pedigree of ancestors and other relatives and annotated with markers of notable talents or, more often, psychic and bodily defects, based on descriptive histories of each individual. All these individual and family histories were recorded on cards and filed in drawers designed to preserve them indefinitely. The enthusiasm of the Honorable James Wilson, US secretary of agriculture, in his presidential address to the ninth annual meeting of the ABA, was so fierce as to sound like mockery. But these men were crippled by irony deficiency. “You have developed in your eugenics section a great experiment station and institution of research . . . containing fire-proof vaults. . . . Your one or two dozen scientific eugenists who are devoting their time to assembling the genetic data of thousands of families in those fire-proof vaults are making records of the very souls of our people, of the very life essence of our racial blood.”28 The files were designed for research and as a resource for eugenic counseling. Davenport routinely assumed such counseling would be grounded in Mendelism.29

Most of his subcommittees never got off the ground. Bell accepted the invitation to chair a subcommittee on deafness but favored a more positive eugenic approach. He had his own institutions for recording families of the deaf, and in place of Mendelian predictions of deafness, he called for the exploitation of family histories gathered by life insurance companies to anticipate longevity. Bell’s cousin, Elias J. Marsh, had studied this question in his capacity as the medical director of the Mutual Life Insurance Company of New York.30 Davenport’s man to chair the subcommittee on criminality was Professor R. G. Henderson of the University of Chicago. He found time to complain “that the name of Heredity of Criminality does not quite express what is desired,” but left the work to the secretary, Max G. Schlapp of Cornell Medical School, who agreed on this point and wondered if a “Committee on Criminality” or on “Causes of Criminality” might be “more to the point.” Davenport countered that the sociological aspects of criminality had been much studied already and that the charge of this committee was to take on “the hereditary basis which makes criminality possible . . . , the sensitive protoplasm rather than the stimulus.” And then the correspondence petered out.31

Only two subcommittees, those concerned with insanity and feeblemindedness, were able to generate any momentum. Not by accident, these were conditions already endowed with abundant family data and with allied techniques of analysis. These committees, each with Davenport’s forceful encouragement, concluded that the data on their defect confirmed its status as a Mendelian recessive trait. In each case, the institutional expert—Aaron J. Rosanoff for insanity and Henry Herbert Goddard for feeblemindedness—had professional reasons to hesitate at the leap of evidential faith required. In each case, eugenics fieldworkers had an important role—Gertrude J. Cannon and Florence I. Orr for insanity, Elizabeth Kite for feeblemindedness. The stories of these subcommittees, which interpenetrate, reveal to what extent hereditary investigations were driven by the historical trajectories of data and expertise at these medical-social institutions. Davenport, instantiating science, coordinated the work, and his Mendelian language of unitary factors was vital for the public relations.

The Subcommittee on Insanity

As in Europe, distress at the unrelenting expansion of mental illness was stimulating bold scientific initiatives in the United States, led by Massachusetts and New York. These ambitions derived partly from sources within mental health systems while reflecting a general spirit of Progressive reform. In 1896, after six years of planning, the New York Commission in Lunacy opened its Pathological Institute in New York City. The founding director, Ira Van Gieson, included in his first report a grand historical survey of treatment of the insane, the dark prelude to a luminous future. Commencing with a “Period of Revenge,” it had advanced through periods of “Indifference” and “Humanitarian and Empirical treatment,” and now stood at the threshold of “Scientific Study, Rational Treatment and Preventive Medicine.” Already, the prognosis was favorable for patients without hereditary brain defects. Such defects must have been numerous, since the report for the state system showed a recovery rate under 5% (951 recoveries out of 20,843).32 The New York asylums, including the first in Utica, had begun their advance into the last, glorious epoch by acquiring a freezing microtome for slicing brain samples along with a card catalogue “so devised that each specimen receives its special number with a history of the case and a record of each successive step in the examination.” It is characteristic of this new scientific age that laboratory specimens rather than patients were the basic units of these information technologies. Since the causes of mental disease are material ones, wrote the Utica superintendent, the filing system must naturally adapt.33

Van Gieson was partial to neurological explanations. A dislocation of the arms of a nerve cell leads to discordance of the “spheres of higher consciousness.” Over a few generations, the defect would worsen and become irremediable. Psychiatry, the reigning science of mental illness, was now “flapping about in the doldrums. . . . It has shut itself up within the asylum walls, discouraged original work and thought and met deservedly the fate of China and ancient Egypt. As a science psychiatry, at present, is dead, and a mummy may be its symbol.” The renewal so desperately needed would arise from the integration of other sciences, which were making brilliant progress. There was a new psychology of mental disease, which examined the nervous systems of cockroaches and conducted experiments on sane humans. The Institute’s psychologist could be found “at the laboratory table, gathering facts, using instruments of precision, conducting experiments.” Cellular biology was laying a new foundation for heredity based on the examination of male and female components, “intimately wrought together and distributed in equal amounts in the process of cell division.” Insanity, arising from defects of this germ plasm, was also a problem for anthropology, whose assigned role was to identify “the initial and intermediate stages in the course of degeneracy.” The anthropologist depended on precision instruments to measure skulls and test sensory acuity and on statistics to compare the number of cases of insanity without hereditary predisposition to those caused or complicated by heredity. So much paper data required an archivist, and Van Gieson even named her: Miss Marie Onuf. In a reflexive move, he announced the founding of a new journal, the Archives of Neurology and Psychopathology, and republished his long essay in its first volume as “The Correlation of Sciences in the Investigation of Nervous and Mental Diseases.”34

Van Gieson’s radical attempt to fashion an institute for basic research within the state asylum system rang alarm bells in the legislature, forcing his resignation. He was replaced by Adolf Meyer, whose ambitions favored clinical medicine over nerve cells. “A great part of psychiatry is administrative knowledge,” he wrote in 1904 in his first full report. “Psychiatry must be met with the same clear determination as a business proposition, not as a field of vague aspirations and ambitions.” As in Massachusetts, his goal in New York was to supply “full statistical particulars” so as to make the “statistical evidence” reliable. He complained that the data on the 72,228 admissions in 1901–1902 was marred by various little contradictions. The figures should be checked rigorously and inscribed onto cards. “In this way you obtain a cross-index of your material according to causes.”35

Meyer was cautious about hereditary explanations, especially after he encountered a Swiss study by Jenny Koller that seemed to challenge it (see chapter 12).36 Davenport nevertheless invited him in November 1909 to become president of the subcommittee on insanity. Possibly he was unaware of this skepticism, or perhaps it mattered less to him than Meyer’s devotion to data. It helped that the men agreed on appointing Southard, no less a data enthusiast, as secretary. Southard was not in the least daunted by Davenport’s ambitious paper forms. Quite the contrary; he already had acquired his own supply and was trying to stimulate interest among Harvard medical students. He had begun preparation of an inventory of families with two or more patients in Massachusetts institutions. Southard endorsed Davenport’s claim that research could be a profitable investment for the state by helping “to dry up the springs which feed the flood of protoplasm too weak to withstand the strain of untoward conditions.”37

In March 1910, Davenport asked his advice on several appointments, including Rosanoff for the subcommittee on inheritance of epi.tify. This suggestion met with silence, and in late October he elaborated that Rosanoff, though a Jew, was “a very nice sort of Jew.” Southard replied immediately that his hesitation had nothing to do with religion. The problem was Rosanoff’s tendency to make strong claims on insufficient data. “I think the writers on insanity are altogether premature with their statements that such and such data ‘indicate Mendelian inheritance,’ and so the actual status of matters is that many writers have not the slightest idea of the differences which exist in types of insanity.” Without rigorous diagnostic categories, Rosanoff could only make blind leaps to Mendelian conclusions.38 It is a strange comment to make to the high priest of presumptive Mendelism. Perhaps he thought Davenport, a biologist, was being led astray by flawed psychiatry.

Rosanoff’s early career exemplifies the bold ambitions of asylum medicine in the new century. About 1904 he was appointed junior assistant physician at the King Park State Hospital on Long Island, just twelve miles east of Cold Spring Harbor. The new medical superintendent, William Austin Macy, argued in the annual report that a good recovery rate was scarcely possible given the character of the city and its insane. The hospital was huge, approaching 4,000 patients when, in 1908, Rosanoff and Macy prepared a paper on institutional scale for the annual report. Even a thousand patients is too many for the superintendent to hope to know them individually, they declared, and once this threshold is passed, the state could just as well seize the advantages of a division of labor. A large institution permits administrative efficiencies and high-quality research. It can afford a well-paid pathologist and a well-equipped lab with an attendant for purely technical work. Not least among the advantages for research is the greater variety of “clinical and pathological material,” that is, living patients and dead ones. To support this point, they supplied a seventy-page case list with all kinds of patient data extending to paternal, sibling, and maternal heredity, the raw material for statistical analysis of the causes and conditions of illness. The next year’s report explained cold-bloodedly that they were proceeding with a “special study of certain organic cases. . . . Fortunately, many of these cases reach autopsy . . .”39

On 14 October 1910, Rosanoff joined a group of charity professionals at the Skillman (New Jersey) Village for Epileptics in a discussion of methods of standardization. He presented there a paper, coauthored with Gertrude Cannon, who had just resigned from her post as eugenics fieldworker. According to a report, they showed “that (excluding certain types) two insane parents will have only insane offspring and that normal parents, both of whom belong to insane strains, will, in the long run, have one quarter of their offspring defective.” This assertion of Mendelian ratios must have provoked Southard’s acid remark, just two weeks later, about unsupported Mendelian assertions. Davenport, who had taught Rosanoff his genetics, defended the study to Southard, mainly on the basis of the excellent fieldwork. Cannon’s research, he said, was comparable in quality to the expert pedigrees gathered by Elizabeth Kite for Henry Goddard, institutional psychologist at Vineland, New Jersey, who also attended the Skillman meeting. The excellent charts of the fieldworkers, based on research in homes and communities, were in sharp contrast to slipshod institutional pedigrees, Davenport said. Southard, however, was no more impressed by Goddard’s data than by Rosanoff’s. Kite’s pedigrees, he objected, showed even more feeblemindedness in offspring of one feeble-minded and one normal parent than the 50% allowed by Mendel’s rules.40

We cannot know what exactly Rosanoff said at the meeting on 14 October. However, ten days earlier he had presented the same work to the New York Neurological Society, and that paper, coauthored by Cannon, came out the following May in the Journal of Nervous and Mental Disease, edited by the New York neurologist Smith Ely Jelliffe. The printed version, billed as a “preliminary study,” spoke of Mendelian inheritance—not of insanity, but of “neuropathic make-up,” scarcely a recognized term in the medical lexicon. Davenport used a similar term a month later when he described the paper to Southard as revealing “that, contrary to all our anticipations, a neuropathic condition is inherited as tho it were due to the absence of a single unit character and that this neuropathic taint may show itself indifferently in any one of a number of kinds of insanity: now in maniac depressive insanity, now in dementia praecox; now in senile dementia, now in paranoia and so on.” He added that if both parents are insane, all the children will show some type of insanity.41

Nobody, then or now, could imagine that Mendelian proportions were contrary to Davenport’s anticipations. It sounds disingenuous at first. He was surprised not by the Mendelian ratio, but by its grab-bag object, “neuropathic make-up.” The key to the mystery is preserved in Eugenics Record Office Bulletin No. 3, billed as a reprint of the Cannon-Rosanoff paper. It is dated 3 May 1911, implying simultaneous publication with the “preliminary study” in Jelliffe’s neurological journal, and the texts are identical. The ERO reprint, however, includes at the end something extra: the proceedings of the session of New York neurologists, consisting primarily of a comment by Davenport himself, who had been present as Rosanoff’s guest. He was unreservedly positive, yet on the crucial question of disease specificity, his recorded spoken words contradict the printed paper. The rules of inheritance, he had said, are always the same, whether it be dementia praecox, manic-depressive insanity, senile dementia, epi.tify, or feeblemindedness. It followed that “two feeble-minded parents could have only feebleminded offspring, and that two epileptic parents could have only feeble-minded or epileptic children.”42

Davenport’s intervention makes no sense unless Rosanoff’s original presentation was about inheritance of specific, recognized disease conditions rather than an all-embracing neuropathic make-up. Sometime within about two months of the neurological meeting on 4 October, and probably before the charity workshop at Skillman on 14 October, the authors abandoned their claim for inheritance of specific mental disease conditions. The revision was not one that Davenport could have welcomed. He had until then been proud to distinguish himself from the London biometricians by his reliance on precise diagnoses, made possible by the trained eugenics fieldworkers. Jelliffe, a neurologist, firmly rejected this catch-all category. Although he had been absent from the October neurological meeting, by Christmas he had read Rosanoff’s paper. It figured prominently in a long letter to Davenport, his old high-school classmate, dated 26 December, nominally a response to Davenport’s just-published essay on eugenics in Popular Science Monthly, yet specifically relevant to Rosanoff’s paper. Davenport’s reliance on the concept of “insanity,” Jelliffe wrote, was a “great mistake,” one that “goes back to the teaching of Morel in the early 50’s.” Insanity is so heterogeneous, with such diverse causes, that it is more properly understood as a legal category rather than a medical one. Having read Jenny Koller, Jelliffe was now skeptical of claims for the overbearing power of heredity. Mainly he insisted on the absolute need for hereditary researchers to employ a proper disease classification, such as Kraepelin’s, rather than imagining that a mere legal category could be biologically heritable. The Cannon-Rosanoff paper, a conscientious effort with deep flaws, must have misled him.43

Davenport wanted to agree on disease specificity. His comments to the neurologists referred specifically to inheritance of dementia praecox and manic-depressive illness, whose absence from the Cannon-Rosanoff paper he had explained as a reluctant capitulation to unanswerable evidence. He replied immediately that their reliance on so vague a category was “in opposition to all my prepossessions. . . . This conclusion I must say has astounded me in my attempt to interpret it.” He obviously was no mere dupe of Rosanoff’s erroneous reasoning. Davenport now proposed a distinction between eugenic realities and medical ones. For other purposes it might be appropriate to insist on distinct disease forms, but practical eugenics was something else, “and if it appears empirically that for the latter purposes the whole classification reared with so much care is useless and can be lumped, then I for one do not see why it should not be lumped in formulating eugenical advice.”44 Jelliffe, unconvinced, inserted a sharp editorial critique of all-encompassing disease categories in the January 1911 issue of his journal. The “great bane of psychiatry” in regard to factors of heredity, he declared, “has been the hopeless confusion of statistical studies for lack of fundamental nosological conceptions.”45

It remains unclear just what led Davenport and Rosanoff to give up disease specificity in the inheritance of mental illness. Clearly, Davenport was the driving force for universal Mendelism. He never asked if Mendelian ratios applied, always how. All the evidence suggests that his role in the formulation of “neuropathic make-up” was fundamental. The authors thanked him in the paper for “guidance, advice, and assistance.” Rosanoff cannot have been any happier than Davenport with this result. As a physician, he looked on accurate diagnosis as a professional responsibility, and many American psychiatrists shared Jelliffe’s conviction that Kraepelin’s nosology was at last making it possible. In 1912, Rosanoff tried out a modified genetic ontology, a “hypothetical germ-plasmic determiner for complete mental development” made up of discrete units, such as for epi.tify or manic-depressive insanity, with a hierarchy of dominance relations.46 Flummoxed, however, by the problem of distinguishing dominant from recessive characters, he and Orr finally took refuge in the supposition that degree of dominance could vary. None of this mattered for their published conclusions. The final version of their study was decked out with many more tables and an intricate coding of defects, yet they scarcely altered a word of their conclusions, which seem to have owed more to Davenport’s Mendelian theories than to painstaking pedigrees. Rosanoff, putting no faith in cures, described prevention as the task of eugenics, and treatment as “palliative measures to combat antisocial manifestations.”47

Anticipations of a match between Mendelian genetics and specific disease categories recurred like a refrain in debates about psychiatric heredity but went nowhere.48 Most eugenic pedigree charts displayed myriad defects of mind and behavior whose tangled webs could never be unraveled. Rosanoff and Davenport were unable in the end to go beyond diathesis to diagnostic precision. Jelliffe, unconvinced by his old classmate’s Mendelian discoveries, declared his preference for the hereditary researches of English biometricians. He also praised Jenny Koller, Otto Diem (who built on her research), and Wilhelm Tigges. Even so, he did not disdain to publish, a few months later, Davenport’s paper (with David Weeks of the Skillman Institution) on the Mendelian inheritance of epi.tify.49

The Feebleminded Subcommittee

Davenport’s choice as chairman of the subcommittee on feeblemindedness, Dr. Arthur C. Rogers of the School for the Feeble Minded in Faribault, Minnesota, was no figurehead. He remarked in a report to the ABA on the wealth of data being gathered by American institutions for the feebleminded and cited researches going back to 1880. Three of these institutions—in Faribault; Vineland, New Jersey; and Lincoln, Illinois—deserved special attention for having established departments of research. Vineland was where Henry Goddard, secretary of this subcommittee, worked as institutional psychologist. He attributed its abundant patient histories and pedigrees to the “full, free, and hearty” cooperation of the parents of these children as well as “adroit questioning and cross-reference” on the part of the professional staff. The trust of the families was crucial to the great goal of data accumulation. “The field worker comes to them as the superintendent’s personal representative with a letter from him recommending her and urging the parents, for the sake of the child, to tell all they possibly can, and to send her on to other relatives or to any one who may be able to give the information, which may be used to help their child, or some one’s child.” Here and elsewhere, interviewers relied heavily on the pretext that eugenic information would benefit individual patients. The subcommittee commenced its investigations at these exemplary institutions, aided by eugenics fieldworkers. Davenport and his associates paddled forward in an already-flowing stream.50

In these years, the science of feeblemindedness was brimming with new or resuscitated ideas. Davenport even thought of employing a eugenics fieldworker trained in psychoanalysis to work with Vineland patients. By examining the children with their siblings, parents, and aunts and uncles, a psychologist might be able to determine “how far the defects of the morons are specific and are inherited as units.” Superintendent E. R. Johnstone, like Davenport, had hopes of identifying particular families, often of immigrants, who had brought feeblemindedness into a region.51 Goddard, similarly, required no persuasion to take up research on patient heredity. On the contrary, he was already a master, and his first letter came as a revelation. Davenport initially wrote to Johnstone on 9 March 1909 to ask about “data concerning feeblemindedness.” Goddard, delegated to reply, explained that while the Vineland admissions forms had always requested information on heredity, the responses were meager and unreliable. He was just then engaged in drawing up new blanks to be sent out after admission, not just to the family but to other informed parties, such as their physician. “This is yielding fruit slowly,” he continued, and as evidence he passed on an early result for a newly admitted child, born in an almshouse to an unmarried, feebleminded mother with four feebleminded children. She had since been married three times and had given birth to six more feebleminded children. The Department of Research at Vineland was seeking funds to hire “a man and probably also a woman” and put them on the road to seek out information on families of its nearly four hundred inmates.52

Davenport, who was then designing his own blank forms, was stunned. He replied immediately. “I can hardly express my enthusiasm over these blanks,” he declared, “and my enthusiasm that you are planning, I trust, extensive work in the pedigree of feeble minded children.” He was especially taken by the idea of putting a man and woman on the road. Would Goddard join the ABA so that Davenport could appoint him to the subcommittee on feebleminded children? Within weeks, Davenport was making arrangements to tour the Vineland facility. He owed his proudest research innovation to professionals at a home for feebleminded children.53

Meanwhile, he had written exultantly to David Starr Jordan of his delight to have found this unexpected ally. He marveled at the “detailed family studies” as well as data collection going on in a score of institutions for the feebleminded. Goddard was suitably deferential, and Davenport’s own plan for form-filling was not working out. It also appeared that his great hopes for the Census Bureau to circulate forms in its limited enumeration areas would be frustrated. He was still awaiting the “printed blanks from our dilatory printer.” Needing information on “perhaps 1,000 large families,” he inquired if Jordan might persuade some enterprising students at Stanford University to gather up such data. Jordan said he could, and even promised to take this investigation to other universities west of the Rocky Mountains.54 At the end of March, Davenport bundled up a thousand printed forms for dispatch to Palo Alto. He would gladly send more, he said, and then he explained how these industrious students could extend the entries to a still “longer heredity” than the form called for. But fate favored fewer shorter family forms. “I have distributed most of the blanks which you sent me,” Jordan wrote in November, “but the students who take them are staggered by the large amount of writing and by their not having thought much about matters of heredity.” Jordan concluded cheerily that the forms would at least stimulate thought.55

Davenport preferred data. He complained to Meyer in May 1909 that “I have met with practically no response with the full family blanks because they deal with so many matters difficult of access and not sufficiently pointed for the physician.” He did not scale down his ambitions. A few months later he was discussing with economist Irving Fisher, another leader of American eugenics, the design of Hollerith cards for eugenic data.56 Harry H. Laughlin, at this time a school superintendent in Kirksville, Missouri, earned his position as eugenics deputy at the ERO in part by diligently registering complexion, form of hair, and eye color as Mendelian family traits. Laughlin also contributed to the solution of “the mulatto problem” by acquiring an expensive microscope. “Can we consider a blend as the Mendelian inheritance of minute units?” he asked in March 1908, referring now to shorthorn cattle. The answer was yes.57

Davenport also advocated Mendelian inheritance of mental ability to Goddard, whose endorsement of these doctrines is a little surprising. As Leila Zenderland points out, Goddard made his reputation as a pioneer of Binet testing, which treated the distribution of intelligence as a continuum. Children who would never surpass a mental age of four he classified as idiots, while an age of 5 to 8 meant imbecility. Goddard’s new category, the “moron,” was for those between 9 and 12, while (in the United States) feeblemindedness took in all three classes. On what basis could he draw a sharp hereditary line between mental ages of 12 and 13? His defense was strikingly hesitant. “The writer confesses to being one of those psychologists who find it hard to accept the idea that the intelligence even acts like a unit character. But there seems to be no way to escape the conclusions from these figures.”58 Rogers, in a coauthored book of 1919 on the feebleminded “in the Vale of Siddem,” dissented from Goddard’s conclusions on just these grounds. Intelligence was too complicated to be reduced to a Mendelian law. Rogers and the fieldworker Maud Merrill quoted Lewis Terman, the new guru of IQ testing, for the doctrine that low intelligence could be defined only relative to the demands placed on the individual.59

The pose of reluctant submission to facts is one we encounter often. “Any theories or hypotheses that have been presented have been merely those that were suggested by the data themselves,” Goddard intoned. Yet his hesitation was real, even if his data were very far from raw. He began his work with a list of causes assigned by parents and physicians, a long-familiar genre. In just 173 of his 327 cases had they assigned any cause, and only 30 of these involved heredity. He reasoned that other causes, such as neuropathy, pointed to heredity, while certain diseases, such as measles, could occasion feeblemindedness only in connection with hereditary weakness. With adjustments like these, he reached 164 hereditary cases plus 34 “probable.” His field workers then sought out feebleminded relatives of these children. On the hypothesis of a recessive Mendelian unit, he had to infer from the traits of relatives the presence or absence of a single factor for feeblemindedness in each normal sibling. This often amounted to arguing in a circle. When he had no information at all, he assumed arbitrarily that exactly half bore a defective factor. He knew from Wilhelm Weinberg’s critique of one of Davenport’s papers that a recessive factor will remain invisible unless at least one offspring inherits it from both parents, and he made the appropriate correction. There were six feebleminded children born in families with two feebleminded parents, clearly violating his Mendelian law unless he supposed that the husband was not the father. And so he did. Goddard did not conceal these inferences, revisions, and corrections to the data on which his result depended. At the end he added up all the numbers and compared them with his prediction. The agreement was excellent: 704 normals to 352 feebleminded expected, 708 to 348 observed. “Such results are difficult to account for on any other basis.”60

And so, despite painful doubts, he stuck by his own makeshift numbers. Davenport’s unflagging Mendelism counted for much here. Goddard, who seems unaware of the biometric critics of monogenic causation, was sincere in his credulity. This is especially evident from the Kallikak episode, his most treasured “natural experiment,” which made him famous, and then infamous. That story involved a soldier from the era of the American Revolution, Martin Kallikak (Greek for “goodbad”), who initiated one line of defective descendants by a feebleminded girl and another of excellent ones with a sound moral wife. Zenderland shows how eugenics fieldworker Elizabeth Kite, an obliging investigator and gifted writer, convinced Goddard, Davenport, and legions of readers, most of them primed to be deceived, that she could reliably detect feeblemindedness from a glance or an anecdote.61

The Kallikak Family presented the Mendelian unit character for feeblemindedness as an assumption rather than a result. Goddard anticipated its demonstration in a big scientific book that was soon to follow. His faith was sustained specifically by the success of Mendelian psychiatry, especially by Rosanoff and Orr’s final report in 1913. They, too, acknowledged the strangeness of their Mendelian result given the motley defects they had merged into neuropathic make-up: “imbecility, epi.tify, deteriorating psychoses, periodic psychoses, paranoic conditions, involutional psychoses, the slighter psychopathic states, and certain eccentricities.” There must also be specific “unit determiners,” such as for feeblemindedness, as well as this encompassing one. In the end, Rosanoff and Goddard alike looked out to the match of their numerical results with theoretical expectation and concluded that it was too good to be denied. Both overcame grave professional doubts to preserve a faith in unit Mendelian factors. It was a relief to reach conclusions in accord with the latest discoveries of science.62

Mendelism Was Biometry for Other Ends

In March 1910, still feeling the sting of Pearson’s rejection, Davenport wrote to Frederick Adams Woods, a prospective member of the Eugenics Committee, criticizing the biometricians for a lack of “biological, let alone medical training.” “Does it not seem to you as a medical man a bad procedure to lump a lot of insane individuals of different generations to get at the hereditary index of insanity? Does it not seem more hopeful to study some well defined form of insanity such as melancholia and work it out carefully by first hand studies on a few families? Do you think it would be a fair statement that mass statistics are good when nothing better is available?”63

Before the year was over, he went over to the other side, inviting Jelliffe’s scorn. It was difficult to maintain disease specificity of inheritance when almost every pedigree chart, starting with Dahl’s, was crowded with the heterogeneous symbols of a multiplicity of defects. His reliance now on lumping to get his Mendelian proportions, however, was also, in a way, a concession to statistical values, the privileging of numerical data. Just as Pearson was satisfied with parent-child or sibling data on mental traits when he achieved correlation measures of about 0.55, Davenport scoured his data for the all-holy three to one, or sometimes one to one. It was all the evidence he needed. He criticized Goddard’s masterwork for faltering at this conclusion.64

Even before 1900, agricultural breeders had used hybridization techniques like Mendel’s to develop new varieties. By 1900, they were working on an industrial scale.65 In the 1910s, experimental animals with short life cycles, most notably the fruit flies cultivated in T. H. Morgan’s fly lab, yielded Mendelian variants in comparable abundance. Somatic medicine, meanwhile, was beginning to identify rare genetic conditions such as chorea. Even before George Huntington began recording the inheritance of this disease, and one year after Dahl’s book, the Norwegian Johan Christian Lund had described this “inherited disease” and compiled a family table.66 If a human condition so common and so costly as insanity or mental deficiency had yielded readily to rules of taxonomy and statistics, we might now have Thurnam’s, Dahl’s, or Jung’s laws of heredity instead of Mendel’s.

With its paper forms and files, the ERO put forth a legion of Mendelian unit traits. Davenport, who contributed a botanical classification and then a eugenic one for Melvil Dewey’s decimalized library system, used a card catalogue for triple-filing of individual records by name, disease form, and place.67 In letter-sized cabinets, he arranged and filed individuals according to distinctive behaviors, talents, and disease forms. He gathered records of insanity and feeblemindedness from institutions using established research techniques that began with a patient or inmate, the propositus, and moved outward, weaving a web of parents, grandparents, siblings, and other relatives (see chapter 12). Since many social problems seemed to arise from lack of self-restraint, he worked up a series of studies on the “feebly-inhibited.” One, Hereditary Nomadism, for example, was about persons who gave in to the temptation to drift, run away, or emigrate. His cards revealed that such behaviors ran in families and could be grouped to generate the ratios he wanted. The wonderful miscellany of behaviors represented in these files was made possible by eugenic-minded volunteers who sent in printed forms such as the “Single Trait Sheet” that he distributed to one and all. It began with a blank line on which to enter a trait, then a line for the propositus, and then a set of ordered spaces on which to inscribe names and recount behaviors in affected relatives.68

The proof of the Mendelian factor was a numerical ratio. There was no other. Pearson’s group joked that Mendelians were simply incompetent biometricians. It has been shown over and over that Pearson’s categorical rejection of Mendelism is a myth. He even gave mathematical arguments for its theoretical compatibility with the biometricians’ law of ancestral inheritance. To be sure, he opposed the assumption of complete dominance, and he mobilized data to challenge various particular claims for Mendelian inheritance.69 There was plenty of low-lying fruit. About 1911, for example, Radcliffe Nathan Salman extended his expertise on potato breeding to the “racial” heredity of Jewishness. Jewish distinctiveness had remained mostly constant since the fifth century, he explained, and consisted mainly of facial traits that were recessive to gentile ones. Pearson’s student Harold J. Laski responded in Biometrika, comparing the work to Davenport’s claims about Mendelian heredity of the mulatto and Mudge’s rat pedigrees, concluding: “It is not, indeed, too much to say that the endeavor to make man a complex of sharply-defined unit characters has failed, and failed completely.”70 Aleš Hrdlička, a member of the ABA Committee on Eugenics and expert on interbreeding between whites and Indians, was similarly dismayed by Davenport’s explanation of skin color. He wrote: “You blame the anthropologist for making so little use of ‘the new points of view’; is the cause of this not to be found in the fact, perhaps, that . . . he finds so little that agrees with his more general observations and helps him to understand the vital phenomena with which he is confronted?”71

The biometricians were especially harsh on Mendelian explanations of moral and mental characteristics. Provoked by praise for Davenport, Rosanoff, and Goddard (“the American researchers”) during the International Eugenics Congress in London in the summer of 1912, David Heron of the Eugenics Laboratory initiated a series of attacks. He focused on bad data practices such as inconsistent categorization of a single individual in different pedigree charts. He found much circular reasoning, when, for example, the parents were symptom free but had one or more mentally defective children, and the theory was made to work by assuming the presence of a recessive factor. These were not mere errors of execution. They were codified in the manuals that instructed fieldworkers to search for defects in ever more remote relatives when the condition of a descendant required a recessive factor, and otherwise to stop. Heron made clear that the scientists, not the workers, were to blame.72

The second installment of this critique was by Pearson and Gustav Jaederholm, a Swedish school expert who was working in Pearson’s laboratory. Jaederholm described the factors considered by nurses, medical officers, and teachers in deciding to place a student in a special school. They included “capacity and willingness to learn, power of self-control, habits of cleanliness, moral order, power or desire of attention to instruction, fits, possibly epi.tify, and a vast variety of semi-physical deficiencies.” Heron called attention to descriptions used by Rosanoff to justify categorization as neuropathic, including “high strung, cries easily,” and “very queer, lives alone, boards out cats.”73 Were these suitable criteria for diagnosing a biological anomaly? Pearson’s justification for rejecting the Mendelian factor for feeblemindedness did not depend on mathematics. Segregation between types, what he had called Mendel’s most valuable contribution, was nowhere apparent for mental weakness. School performance showed no sharp breaks, only continuous curves. Jaederholm came over from Sweden bearing data, the most welcome of gifts. Pearson calculated that almost half of the Swedish children who were three or more years behind normal could be explained as the tail of a Gaussian bell curve.74

Heron’s diatribe, picked up by the New York Times in November 1913, became a terrible embarrassment to Davenport and his allies. Rosanoff, claiming his right to an immediate reply in the AJI, invoked medical “judgment and experience” against this mere layman, who “assumes an attitude, in relation to a psychiatric issue, which is in opposition to a view universally held by psychiatrists.” He had, he said, merely extended well-known genetic principles to a new case.75

The story is not of isolated failings by a few bad scientists. David Barker has shown that in Britain as well as America, the very biologists given credit by historians for discrediting eugenic dogmas endorsed “naïve” one-factor explanations of epi.tify, insanity, and feeblemindedness. Bateson, one of the first, celebrated the achievements of the “American students” in 1913. In 1921, he anticipated progress “of a magnitude that statesmen perhaps have never conceived” once jurisprudence learned to reason with the Mendelian factor for criminality.76 A paper by Hamish Spencer and Diane Paul shows that the critiques by Heron and Pearson include all the main points made by the new human geneticists of the 1930s.77 R. A. Fisher was, like Pearson, a far more active eugenicist than Bateson and almost incomparably expert on the mathematics of heredity. Yet he was an early convert to Davenport’s single-gene explanations. As late as in 1924, he referred to “strong evidence of a mendelian factor” and still stronger evidence of clear differentiation between the normal and the feebleminded. On this basis he anticipated that strict eugenic policies could reduce mental defect far more quickly than many scientists had supposed.78 Barker, exaggerating, comments that the flaws in his reasoning should have been apparent to an undergraduate geneticist. Fisher had by then replaced Pearson as the world’s leading statistician.

The biometric critique of Mendelian psychiatry was well known to human geneticists in the 1920s and 1930s. Lionel Penrose, a forceful opponent of eugenic policies with no reservations about Mendelism, cited Heron and Pearson and set out from their lines of argument when he challenged single-gene explanations in his 1933 book on mental defect. His own conclusions on its inheritance relied primarily on statistical analysis of phenotypic (rather than genetic) data from medical-social institutions. Insanity and mental deficiency remained the most urgent topics of human genetics through the 1930s.79 The reshaping of psychiatric heredity by Mendelian genetics was dubious and fleeting. It is absurd to describe biometricians as standing in the way of a promising Mendelian program.