INSTITUTE FOR SCIENTIFIC (N FORMATIONe 3501 MARKET ST PHILADELPHIA PA 19104 Postmature scientific Discovery and the SexuaI Recombination of Baeteria— The Shared Perspectives of a Scientist and a sociologist

.,. ––1---- . . .-,.,. Numr3er 3 January ID, IYUY

Forty-three years ago recognition, one aspect of prematurity, as and Edward L. Tatum, then at the Osbom in the cases of Peter D. Mitchel17,S and the Botanical Laboratory, Yafe University, New Higgs boson.g, 10Currently we are prepar- Haven, Connecticut, published a landmark ing an essay that deafs with the topic of de- paper in Nature reporting the discovery of layed recognition. It will appear in the next bacterial sex. 1Its importance convinced me few months. to reprint the piece here. Recently we have Some time ago sociologist Robert K. Mer- afso reprinted reminiscences by Lederherg, ton, , honored me by now president, The Rockefeller Universi- writing the foreword to my book Citation ty, New York, on his work in genetics,z Indexing-Its Theory and Application in but as an aspiring swiologist of I Science, , and Humanities. 1I was even more attracted to the paper that There, he reflected on the possibility that the he coauthored with Harriet Zuckerman, De- Science Citation index@ was an ‘‘inevi~- partment of , Columbia Univer- ble” discovery. In a recent publication, he sity, New York, in the December 18, 1986, has also mentioned the reprint on posh-nature issue of Nature. In this paper, which we are scientific discovery that follows as an afso reprinting here, Zuckemmn and Letier- example of “sociological autobiogra- berg view this discovery from a sociologi- phy’’–combining the complementary ad- cal point of view. 3 In reading this remark- vantages of both Insider and Outsider per- able “case study” of the Lederberg-Tatum spectives, while minimizing the disadvan- discovery of bacteriaf sex, my thoughts res- tages of each. 12 Reading the Zucker- onated to their analysis of” postmature sci- man-Lederberg definition of postmature entific discovery. ” discovery made me realize the interesting Some Current Confer@ readers may not relationship between these: the inevitabili- be aware that Harriet and Josh coined this ty of discovery and postmaturity. Language term as a parallel to’ ‘premamre discovery, ” plays tricks on us—inevitability of discov- the phenomenon described by Gunther S. ery seems the right way to describe, for ex- Stent, professor of molecular biology, Uni- ample, simple techniques that are so ‘‘ob- versity of California, Berkeley, and vious” once they are invented. In patent others4,5 that we have discussed on many law, “inevitability” is often expressed by occasions. (A prototypical example is the the phrase “obvious to anyone versed in the case of botanist Gregor Mendel.G) Indeed, art. ” Many a patent application has Eeen de- we at ISI” have used citation anafysis to nied on that principle. study cases of premature discovery. These Both the inevitable “discovery” of cita- longitudinal studies have covered many tion indexing and the postmature discovery years of data and fmus especially on delayed of bacterial sex show the three attributes

16 specified by Zuckerman-LexJerberg. In ret- plication. While intuition tells us that this rospect, (1) it must be judged to have been has indeed been the case, this assertion is technically achievable at an earlier time with diftlcult to prove. Whether the ‘‘sociaf” en- methods then available; (2) it must be judged gineering of science through better under- to have been understandable, capable of be- standing of the discovery process (social sci- ing expressed in terms comprehensible to ence) can claim similar efficiencies remains working scientists at the time; and (3) its im- to be seen. But certainly the management plications must have been capable of hav- of the global science enterprise will pay a ing been appreciated. g smafl price for the continued investigation In an era of Big Science, with goals such of such processes that might accelerate both as the conquest of cancer, AIDS, and so on, research and technology transfer. Zucker- it is important to understand the limitations man and Lederberg have made an impor- of science, but also to understand how its tant contribution to that process. development may be accelerated. When I first entered the field of information science, ***** it was a tenet of our missionary zeal that in- formation technology could have a signifi- My thanks to Peter Pesavento for his help cant catalytic effect on discovery and its ap- in the preparation of this essay. ~1909IN

REFERENCES

1. Lederberg J & TaturrrE L. Gene recombination in Ercherichia co[i, Nature 158:558, 1946. 2. Garfteld E. The impact of basic research in genetic recombination-a personal account by Joshua Lederberg. Parts l&2. Current Coruerm (24):3-17, 13 June 1988; (25):3-14, 20 June 1988. 3. Zuckerrnan H & Lederherg J. Postmature scientific discovery? Nature 324:629-31, 1986. 4. Stent G S. Prematurity and uniqueness in scientific dkcovery, .Sci. Amer. 227(6):84-93, December 1972. 5. Barber B. Resistance by scientists to scientific discovery. Science 134:596-602, 1961. 6, Garfield E. Premature dkcovery or delayed recognition-why? Essays of an information scientist, Philadelphia: 1S1 Press, 19SI. Vnl, 4, p. 4S8-93. 7. ---——-. Tfre 1976 articles most cited in 1976 and 1977.2. Physical . Jbid., p. 115-26. 8. Mitchell, Peter D. (Wasson T, cd.) Nobel Prize winners. New York Wilson, 1987. p. 713-5. 9. Veltman M J G. The Higgs boson. .Sci. Amer. 255(5):76-82, 1986. 10. Garfteld E. How to use citation analysis for faculty evaluations, and when is it relevant? Part 2. OP. cit., 1984. Vol. 6. p. 363-72. 11, Merton R K. Foreword. (Garfield E.) Citation indexing-ifs theory arrd application in science, technology, orrd humanities. New York: Wiley, 1979, p. v-ix. 12. ------Some thoughts on the concept of smiological autobiography. (Riley M W, cd.) Sociological lives. Vol. 2, socicd change and the lfe course, Newbury Park, CA: Sage, 19S8. p. 17-21.

Reprmmd by permission from Nature Vol. 158, p. 558. Copyright (c) 1946 Macmillan MagazinesLtd.

Gene Recombination in Escherichia coli

Anatysis of mixed cultures of nutritional mu- ability to synthesize growth-factors. As a result tams has revealed the presence of new types which of these deficiencies they will onty grow in media strongly suggest the occurrence of a sexual pro- supplemented with their specific nutritional re- cess in the bacterium Escherichia coli. qttirementa. In these mutants single nutritional re- The mutants czmaiat of mains which differ from quirements are established at single mutational their parent wild type, strain K-12, in lacking the steps under the influence of X-ray or ultra-vio-

17 letl,2. By su-ive treatments, strains with sev- chemical requirements and phage resistance have eral requirements have beerr obtained. frequently been found. In the recombination studies herereported, two These types can most reasonably be interpreted triple mutants have been used: Y-10, requiring as instances of the assortment of genes in new threonine, Ieucirse and thiarnirr, and Y-24, requir- combinations. In order that various genes may ing biotin, phenylalanine and cystine. These have the opportunity to recombine, a cell fusion strains were grown in mixed crdtrrre in ‘Bacto’ would be required. The only apparent alternative yeast-beef broth. when filly grown, the cells were to this interpretation would be the occurrence in washed with sterile water and inoculated heavily the medium of transforming factors capable of in- into synthetic agar medium, to which various sup- ducing the mutation of genes, bilaterally, both to plements had been added to allow the growth of and from the wild condition. Attempts at the in- colonies of various nutritional types. This pro- duction of transformations in single cultures by cedure readily rdlows the detection of very small the use of sterile filtrates have been unsuccessfirl. numbers of celI types different from the parental The fusion presumably occurs only rarely, since forms in the cultures investigated only one celJ in a rr3il- The only new types found in ‘pure’ cultures of lion cars be classified as a recombination type. The the individual mutants were occasioned forms hypothetical zygote has not been detected cyto- which had reverted for a single factor, giving logically. strains which required only two of the onginrd These experiments imply the occurrence of a thrm substances. In mixed cultures, however, a sexual process in the bacterium ,flrchen’chiamfi; variety of types has been found. These include they will be reported in more detail elsewhere. wild-type strains with no growth-factor deficien- This work was supported in part by a grant from cies and single mutant types requiring ordy tbia- the Jane Coffin Childs Memorial Fund for Med- rnin or phenylalaoirw In addition, double require- icaJ Research. ment types have been obtained, including strains Joshua Lederberg* deficient in the syntheses of biotin and leucine, E. L. Tatum biotin and tfrreonine, and biotin and thiarnin re- spectively, The wild-type strains have been CJsbOm Botanical Laboratory, studied most intensively, and several independent Yale University, lines of evidence have indicated their stability and New Haven, Coon. homogeneity. Sept. 17. In other experiments, using the triple mutants mentioned, except that one was resistant to the ●FeUow of the Jane Coffin ChiJds MenmriaJ Fund for Medical Research. coli phage 11(obtained by the procedure of Luria and Delbriick3), nutritionally wild-type strains ‘Tatun, E. L , Pmc, Na. ,icad. .%i., 31, 215 (1945), lTatum, E. L., Cotd Spring Harbor, Symposia Quam. BIOI., were found both in sensitive and in resistant cat- “01 1 t (,” the press. 1946). egories. Similarly, recombination between bio- ‘Lure, S. E., and Detbriick, M.. Genetics, 2.S, 491 (1943),

Reprinted by pmrussion from Nafure Vol. 324, pp 629-31. Copyright (c) 19S6 Macmdlan Ma@zines Ltd

Postmature scientific d~covery?

Harriet Zuckerman arrd Joshua Lederberg

New scientific discoveries do not always flow di- “ahead of their time”. ‘Ike have been examined reedy from those nradejust before. Ratf2er, several by Barberl and Stent2. Here, we suggest that varieties of discontinuity cars be identifkl in the there are also postnmture discoveries, tJsose which growth of science. Premature discoveries are those are judged retrospectively to have been ‘delayed’. that scientists do not attend to in a timely way and We arrrdyse the arguments that the discovesy of are retrospectively described as having been bacterial sex was postmature and take up the cor-

18 relative questions of how the problem was iden- cording to Butterfield, “to produce a story which tified and why Lederberg and Tattrm3.4 were is the ratification if not the glorification of the likely Candidates for snaking it when they did. present” 12. But, they are designed to serve quite This paper draws on documents, published and the contrary purpose. The ideas of premature and private, and analyses by the sociologist-observer postmature discovery provide convenient handles and the scientist-participant. Our diatectic pmce- for analysing dkcontinuities in the growth of sci- drrredeparts thm most orsd histories.% first, the entific knowledge, and support a nonfinear and procedure was iterative: as new discussion raised complex model of advancement in scientific un- further possibilities, we both searched for rele- derstanding. vant documentation; and second, we both identi- Postmature dkcoveries are not all of a piece. fied the underlying analytic questions and artic- One class results from pre-emption of scientists’ rdated tentative answers to them. We felt that per- research attention. For example, Linus Pauling sonal reminiscence had to be validated by con- observed that there was “no reasmr why” he, temporary documents and other testimony as oral himself, coufd not have dkcovered the afphrr helix history and autobiography are prone to ‘‘uncon- eleven years earlier than he acturdly did “after scious fafsitication’ ‘7. a few fmurs of work”. In fact, he was preoccupied in the intervsd by other seemingly more impor- tant and feasible inquiries 13.”. Another class of Continuities und diaeontirmitks postmature discoveries answers questions not pre- Scientific growlh, usually broadly incremental, viously recognized by scientists to be problemat- can at important times be episodic and discontin- ic. Certain assumptions, beliefs and images Is uous. Premature discoveries, one conspicuous which are also indispensable for the organization form of temporal discontinuity in science, are of scientific thought can, in specific cases, im- either passively neglected or actively resisted at pede perception of lines of inquiry. For example, the time they are made. Mendel’s discovery of Weinberg notes that physicists neglected to pur- particulate inheritance in 1865, lost to view for sue quantum field theory further in the 1930s be- thirty-five years, is the best-known historical case. cause prevailing images, conceptual schemes and Discoveries can be premature because they are attitudes toward theory and empirical evidence conce@y mismmnected with ‘canonical knowl- stood in the waylb. In our case study, both cog- edge’2, are made by an obscure discoverer, are nitive and socird processes obstructed the thirrk- published in an obscure place, or are incompati- ing of scientists about recombination in bacteria. ble with prevailing religious and political doctrine. Barriers between disciplines imposed by special- ization of inquiry afso contribute to neglect or re- Sources of neglect sistrmcc) ,glo. Although the character and sources why was recombmtion in bacteria not perceived of premature discoveries have received some as problematic before 1946? How had asexuality anrdytical attentions. 11, the pattern of postmature in bacteria come to be an unquestioned ‘truth’ and discoveries has not been identified, much less how was that view perpetuated? systematically studied. Before 1870, many believed that the different For a discovery to qualifj as pastmatore, for shapes bacteria assurrd were varieties of the same it to evoke surprise from the pertinent scientific organism, which changed under varying condi- CQmmunity that it was not made earlier, it must tions. Indeed, the doctrine of polymorphism or have three attributes. In retrospect, it must be bactcriaf plasticity became the basis for extrava- judged to have been technically achievable at an gant claims about variability through most of the earlier time with methods then available. It must nineteenth century. By 1872, Ferdinand Cohn be judged to have been understandable, capable concluded that the various shapes bacteria took of being expressed in terms comprehensible to were not different forms of the same organism; working scientists at the time, and its implications they were monomorphic and dld not change dur- must have been capable of having been appreci- ing their short lifetimes]7. Yet reports of varia- ated. tion continued until 1881 when Robert Koch in- Both prematurity and postmaturity can be rec- troduced a simple and effective means for grow- ognized onfy by retrospdon. They differ in that ing pure cultures. Kcxh’s pure-culture method, prematurity is a matter of acturd historical obser- which became a symbol of modem bacteriology vation wbife postmatority is a matter of retmsfrec- with its phobia of contamination, together with tive conjecture. Such formrdations would seem Cohn’s doctrine of monomorpbism rapidly to smack of’ Whig History’, the inclination, ac- changed bacteriologists’ views about variation.

19 The two were consolidated into what was called instance to neglect of bacterial recombination, It the Cohn-Koch Dogma, which discouraged for has been argued, however, that ‘disciplinary dog- years the study of the problems of morphology, matism’ and ‘disciplinary monopoly’ have only inheritance and variation in bacteria lg. rarely impeded the development and diffusion of Cohn was convinced that bacteria were primit- scientific imovation21. ive plants which could “only reproduce by asex- Members of the Delft School of Microbiology, ual means” and in 1875 characterized all bacteria in the early part of this century, did bridge the as Schizomycefes or ‘fission fungi’. With every gap between bacteriology and genetics. Clearly use of that label, bacteriologists were reminded separating themselves from the medical bacteri- that these organisms reproduced ordy by fission ologists who maligned bacteria, they befieved that and that they were simple primitive pkrrrts, a tradi- progress in fundamental microbiology depended tion that had begun with Lecuwenhoek’s first ob- on people who ‘loved’ rnicrobeszz. Martirrus servation of bacteria in 1675. Labels, categories, Beijerinck, the main figure in the group, seems nomenclature and taxonornies usually help to or- now to have beerr the most likely candidate for ganize scientific thought but can rdso delay the investigating bacterial sex. He rejected prevaiL reexaminadon of fallacious traditions, thus becom- ing dogma on bacterial invariability, promptly ing self-fulfdling prophecies 19. In the end, the cited deVries’ finding on plant mutations and of- emergence of medical microbiology as a science fered some of the t%st coherent challenges to strict depended on the doctrinal base laid down by Cohn monomorphism~3. He also developed ‘esrrich- and the pure culture methods of Koch. Nonethe- ment culture’ methods, forerumers of the selec- less, monomorphic doctrine, when strictly con- tive techniques used later in discovering bacterisd strued, threw out the baby of bacterial variation recombination. Moreover, he was better informed with the dirty bath water of contamination. It was than most microbiologists about work on plant hy- widely assumed that observations of bacterial vari- bridization which would have been usefid in plan- ation had to result from contamination. Bacteri- ning any investigation of sex in bacteria. ologists took experiments involving variation to Beijerinck and the Delfr School were likely can- be error-prone and dlsreputablt#. Such experi- didates for investigating bacterial sex, but they ments were to be avoided as having great proce- did not. In fact, Bcijerinck strongly supported the dural difficulty and little intellectual merit. With Cohnian dogma of schizomycctes. Thtts the prob the strong incentives in science for avoiding prob lem of sexual nxombination still fell between dis- Iems notorious for leading to irreproducible re- ciplinary schools24. sults, very few scientists would elect to under- take them. Bacteria occupied an ambiguous place in the Significance of bacterial sex hierarchy of living organisms. To many, these or- By the 1930s, developments were under way that ganisms appeared w primitive that they could not led biologists to reexamine how bacteria related yet have evolved ‘differentiated genes’. This im- to other forms of life and whether bacteria redly age also reinforced the use of bacteria as exem- had genes. fnrportant among these developments plars of pre-genie levels of organization for phys- was the unification in biological thought of Men- ico-chentical anafysis. Once such complex im- deliarr genetics, quantitative population them-y and agery becomes established, special provocation Darwinian evolution, particularly the notion of is needed to splinter away one or more of its ele- species being Mendelian breedng populations or ments. isolated gene pools. The idea that sexuality was, Disciplinary emphases and the division of la- itself, an evolved genetic system proved particu- bour among the sciences also diverted attention larly provocative, with illustrations drawn from from the problem of bacterial sexuality. Bacteri- simple and complex plant life. Dobzhansky ’s ologists were principally conccmed with problems monograph, “Genetics and the Origin of Spe- in medical pathology rather than issues like the cies’ ’25, was widely read as the definitive rein- biology of bacterial reproduction. Oenedcists were terpretation of Darwinian theory of evolution and no more interested in bacterial reproduction than focused interest on the details of breeding systems bacteriologists. They were occupied with larger ss the key to understanding evolutionary develop- organisms in which the products of crossing were ment. This, in turn, sharperred interest in under- readily observed. Thus, disciplinary division of standing the evolution of organisms, like bacteria, labour and the careful choice of organisms for in- believed to be devoid of sexuaf mechanisms. quiry, both generally conducive to the develop- The bkxhemical analysis of rnicrobiaf nutrition, ment of scientific knowledge, contributed in this w=ially by Knight and Lwoffl 8, was another

20 major impetus to reexamining the relationship of chemistry, cytology and physiological embryol- bacteria to other forms of life. In particular, the ogy. But he wss not ignorant of classicsJ genet- discoveries that the biochemistry of microbes par- ics and the Columbia biologists were well con- alleled in many details that of higher organisms nected with the New York network of scientific inspired Beadle and Tatmn’s work on Neurospora communication about genetics. Dobzhatts& was in 1941~. They showed Neurospora’s usefuhress a central figure. Arthur Pollister was in close as a research nrrsterird for studying the genetic con- touch with researchers at the Rockefeller f.rrstitnte. trol of an organism’s development through the en- Alfred Mirsky worked at both institutions. L.e- coding of specific enzymes, known as the ‘one- derberg not only learned quickly about the gene-one-enzyme’ hypothesis. This marriage of neo-Drrrwirrian developments described earfier but biochemistry and genetics had particular signiti- he afao heard about the work of Avery et af. from cancc for the I.ederberg-Tatum workZT,ZB. Mirsky and promptly read their paper. If the There was rdso renewed speculative interest in Avery et af. work shropened Lederberg’s interest a bbchemicxd theory of the origin of life. ‘‘7?re in bacterird reproduction, Dubos’ review of the Origin of Life on Earth’ ’29 by the Russian inconclusiveness of evidence on sexual reproduc- biochemist, oparin, became available in English tion sharpened MS skepticism; the cognitive and in 1944, as did ‘‘W7rat is fife ?”~ by the physi- structural elements were coalescing. cist, Scbr6dinger. Both focused attention on ques- In Lederberg’s second year at Columbia he met tions that demanded the integration of the biology Francis Ryan, an assistant professor, who had just of viruses and microbes with the more traditional completed a postdoctorsd fellowship at Stanford biology of plants and animafs. with Beadle and Tatum. It was Ryars who first The comections between tlmsc independent de- told him about the work on biochemical genetics velopments were not always apparent at the time. and who persuaded him that chemistry and ge- But one event did call attention to their common netics were not as far apart as he had thought27. message: the dkovery by Avery, MacLemI and It was also Ryan who generously provided Le- McCarty31 in 1944 which identified DNA as the derberg with laboratory facilities, catrdyaed his transforming principle that changed rough association with Tatum, and, most importantly, non-pathogenic pnemnococc i into wnodr vindent encouraged, educated and socialized him as a sci- ones. entist. Columbia provided Lederberg with a mul- The scientific signifi~ of that discovery has tifacetcd and advantageous structural context for been examined in detai132-3s. For our purposes, his scientific development and for the initiation it higldighted two important questions: what was of a high-risk, high-stakes research progrsnmne. the structure of bacterial genes and how were they Lederberg’s plan for research was well worked transferred? Tfnrs the work by Avery et aL made out by Jrrly 1945, when he was a second-year stu- the question of bacterial sex newly consequential. dent at Columbia Medical School but continued DubosW makes it clear that had sexual reproduc- to work in Ryan’s labomtory. The research might tion been observed, it woutd have been tmderstcod have been pursued at Columbia, but Ryan and Le- and appreciated. But bacteria were so widely as- derberg both recognized that an association with sumed not to reproduce sexually that no one con- Tatum would be vahsable. In particular, his ex- sidered this problem to be important. Dogma pre- perience in microbial biochemistry could help vailed over focused curiosity. broaden L.ederberg’s education beyond the oppor- tunities available on Morrringside Heights. Fur- thermore, Tatum, then in the process of moving Structural contexts to Yale, was rapidly becoming recognized as a In retrospect, Lederberg’s position in the com- scientific leader. He cmrfd provide Lederberg with munication network and MS not yet having a ca- better access not only to information, research ma- reer in science seem consequential for his identi- terials and fellowship support, but also to the in- fying the problem of bacterial sex, for his develop visible college of the emerging Scientic discipline ing a method for its investigation and for his be- of biochemical genetics. ‘Rte impact of such infor- ing in a position to do the research. Tatum was mal ties between investigators on the dwections led to the problem independently for somewhat and pace of scientific research has yet to be prop- different reas0ns27. Lederberg was unenthusiastic erly investigated. about classicaf genetics when he arrived at Co- Lederberg’s status as a medical student was less lumbia College in 1941. His interest in “under- an obstacle to hk investigating bacterial sex than

standing] the chemical nature of life” led him might be suppd Though much of his time was to spend much of the next four Years studying spent on course work, he was not subject to the

21 constraints that apply in the early years of study scientist as the exclusive contributor ad foeuases for the PhD. He did not, like ordinary graduate attention on the operation of organized skepticism students, have to choose a research problem that and its effect on shaping the meaning and assess- would be suitable for a thesis and publication, Be- ment of those contributions. ing marginal~ to the biological research enter- prise, he could afford to take on a high-risk prob- lem. The search for bacterial sex was definitely Conclusions high-risk; it was not one likely to produce usefrd and publishable findings. After aff, not observ- Was the investigation of sexuaf recombination in ing bacterial recombination would scarcely dem- bacteria posrmature, thatis, wnducted significant- onstrate that it did not exist. The risk of a negative ly later than it coufd have been? The problem was finding using E coli is now known precisely; bac- obscured for decades by the Cohn-Koch dogma teriaf recombination being observed in only five of monomorphism and the conviction that bacte- percent of afl strains with the techniques used in rial variation resulted ordy from contamination. 1946. This was so even for Beijerinck and members of For a different set of reasons, TatUrn coufd also the DeM School who did not subscribe to strict afford research on a high-risk problem at the time. monomorphkm, knew how to mark microbial He had a variety of projects in process in his lab- strains by their fermentative and nutritional char- oratory and could manage to take a long-shot ex- acteristics (the basis of Lederberg’s design), knew periment that required littfe time and little money, about Mendelian segregation in plants, and might For both men, bacterial recombination was a gwct have appreciated the significance of sexual recom- gamble; failure would have low marginal costs bination in bacteria were it observed. But, they for each but promised large if prospectively im- were committed to the view that bacteria repro- probable returns. High-risk investigations are not duced only by fission and did not consider the phe- equally feasible for all scientists. They fall to the nomenon problematic. In principle, the investi- comparative] y well-established or to those who gation was technically feasible by 1908, as dem- are marginal as Lederberg was in 1946. Those onstrated by Browning’s4T use of drug resistance who solve high-risk problems, having chosen as a selective marker, an early anticipation of the them in the first place, may more often come from Lederberg-Tatum work. But Browning dealt with the ranks of the well established than from neo- a different organism, reported a negative result, phytes thus contributing to the accumulation of and used terminology not readily transferable to advantage 1. Risk-taking in science is a matter the case of bacterial recombntion. In the 1930s, not only of psychological daring but also of posi- bacterial sex was still viewed as unlikely, even tion in the sociaf stmcture 14.42. as a disputable idea. Yet bad it km demonshated Atler a brief correspmdence, Tatmn invited h- experimentally, it would have been understood derberg to work at Yafe. He arrived in March and appreciated by geneticists and possibly even 1946; genetic recombination in E. coli was ex- by bacteriologists. perimentally observed early in May. The results This case study suggests that problem identifica- were so arresting that Tatrun arranged for Leder- tion and selection in science have features deserv- berg to present them at the Cold Spring Harbor ing further analysis. First, the solutions to two Symposium to be held in July. The publications classes of problems are apt to be p.stmature: those which followeds.d.’$s did not merely describe the which do not survive competition for scientists’ results of the initial labomtory investigation. They attention when they tirst appear because they ~m are the product of critical discussion of those re- insignificant, unfeasible or both and those which sults at that meeting and of follow-up experiments are obscured by prevailing cognitive mmrnhments done immediately afterwards17, The dynamics of w have no socially and cognitively defined dis- organized skepticism in science~ can be ob- ciplinary home. Second, in calculating the prob- served in the records of that meeting and later in able returns on selecting problems for investiga- responses to the papers announcing the discovery. tion, scientists assess the likelihcmd of error and Even as first pubfiahed, discoveries are not simply this contributes to the continuing neglect of cer- reports of events initially observed in the labora- ~in problems that have a history of being error- to&5,w but often are also the outcomes of ex- prone. Third, the feasibility of addressing bigh- change between contributors and their critics. risk problems in science and so of making major Treating scientific contributions as the results of ~dvances in this way is not equaf for aff investi- inquiry, criticism and subsequent work makes gators; they are left largely to the well-established problematic the custom of designating this or that who can afford them and to others who have a

22 smaller stake, for structural reasons, in their im- Uenter for Advanced Study in the Behavioral mediate record of publication. What scientists de- Sciences. fine as problematic and worthy of investigation are the products of interactions between cognitive and social processes. Harriet Zxckemum is at Ihe Deparrmenl of .%ciobgy, CoIumbia University, New York, New York 10327, USA and Joshua Research supported by NSF Foundation (SES Lc-derberg is a~ 77ie Rockefeller University, 1230 York Avenue, 80-08609), the Russell Sage Foundation and the New York, New York I(XJ21, USA.

1. Barber, B. Science 134, 5%-602 (1%1). 26, Beadle, G.W & Tatum, EL. Proc. mm. Acad. USA. 27, 2. Stent. G.S. Sci. Am. 227, 12, 8’-93 ([972). 49’-506 (1941). 3. Lederbers, J. & Tatum, E.L. Nature 1S8, 558 (1946). 27. lderberg, J, Nomre 324,627-631 (1986). 4 Lederbsrg, J. & Talum, E, L. Coki Sp, Harb. Synp. 9, 28. Lderb-sx8, J.& Tatum, E. L, A. ffev. Genef. 13, l-5(l979). 113-114 (1946), 29. Oparin, A. I.7he Origin of Lifeon~t?h Trans. Am Synge 5. Nevins, A. Wif$on Lib. Bull. 60, 647-615 (1966). (Academic, New York. 1957). 6. Bension, S. in Modem Metbds in rhe HisIory of$fedicine 30. 8chr6din8er, E. ~ti&@?~e PhysimIApetioftie tiw (cd. Clarke, E.) 2S6-365 (AOdone, London, 1971) ing Cell (Cambridge University Press, Cambridge, 1%2). 7 Sanon, G. 71w Study of the History of Mathenumcs 35 (Har- 31. Avery, O. T., MaLeod.C.M.& McCarty,M.J .exp.Med. vard University press, Cambridge, 1936), 79, 5%-602 (1944). 8. Menon, R.K, 7he Socio@y of Science 371-382 (Universi- 32. Judxm, H. F.lhe Eighth &yofCreadon: Makers ofrhe Rew ty of Chicago Press, 1973). olurion in Biology (S~mon & Schuster, New York, 1979). 9. Gar6e1d, E. Essays of an Infonnarion Sciemist (1979-80) 33. McCarty, M.37w Trarufonning Princip&: Discoveries 7har 488-493 f3SI, Philadelphia, 1981). Genes Are Made of DNA(Norton, New York, 1986). 10. Cole, S. Am. J. Socio(. 76, 28t-306 (1970). 34. Olby, R.C. 7hc Parhro rhe Double Heli.r(Univemny of II. Gar6eld, E. Currenr Comments 1717, 3-10 (19S5). Washington Press, .%altle, 1974), 12. Bunerfield, H. 7?w Whig Imetpreration of J-JtsIor3 (W. W. 35. Dubos, R.J. The Professor, 17w In.WiIuleandDNA(Rocke- Norton, New York, 1%5). feller Umvemily Press, Ncw York, 1976). 13. Paulin8, L. Nafure 248, 769-771 (1974). 36. Fmmn, J.S. Molemkx dti2:Htirotim/ fisa~onrk In- 14. Zuckermsn, H, Sociol. 6tq. 48, 65-95 (1978). te~hyof&tit~ ati Bwlogywtley, New York, 1972). 15. Holton, G.J. J7wnatic Origins of Sm”entific l’bought: Kepler 37, Lederberg, J. Narure 239,234-236(1972). to .Eim6?in. (Harvard University Press, Cambrid8e, 1973). 38. Cohen, J. S.& Portugal, F. H. Persp. Biol. Med 1S,204-207 16. Weinberg, S. Daeddus 106, 17-35 (1977). (1975). 17. Geison, G.L. in Dicl. Sci. Biog. 3, (at. Gillispie, CC.) 39. Dubos, R.J. 7he BacWialCell 181, (Harvard University 335S41 (Scribners, New York, 1971). Press, Cambridge, 1945). 18. Dubm, R.J. J3ie Bacteria/ Cell 135 (Harvard University 40. Gieryn, T.F. &Hirsch, R, F.Suc S/. of Sci. 13,87-106 Press, Cambrid8e, 194S). (1983) 19. Merton, R.K. Social 7heoIY and Social Smcrure 475-490 41. Merton, R.K. 7he Sociology of Science 439-590Jniversily (Free Press, New York, 1968). of Chicx3g0 Press, Chicago, 1973). 20. Zuckennan, H. in Deviance and .%cial Change (cd. SaSrarin, 42 Mulkay, M.J. Jle Sucial Procw of Scienrifc /movalion! E. A.) 87-138 (Sage F’ub[ications, Beverly Hills, California, A St@ in rlu .%ciofogyof.?cience, 49-51 (Macmitlsn, lJJn- 1977). don. 1972). 21. Ben-David, J. in Culture ad /(s Creatiom (cds Ben-David, 43. Tatum, E.L. &L@4erberg J.J, JkacI. S3,673-684 (1947). J. & Clark, T.) 255-65 (University of Chica80 Press. CM- 44 Merton, R.K. 7he Sociology of Science 267-27g (Univer. CJ380, 1977), sityof Chicago Press, ChiC480, 1973). 22. Kluyve,, Albert Jan: His Life and Work. (eds Kamp, A. R., 45. Latour, B,& Woolgar, S. LdoralOry Life: 7he&xial Con- La Rivi.&e, J. W.M. & Verhceven, W.) 186 (Amsterdam, smxtionoftit-@cFczw (Sage Publications, Beverly HIJs, NordI Holland, 1959). 1979). 23. Beijerinck, M. Versl. A&f. Wemsch. (Amsterdam) 9,310 46. Knom-Cetina, K. 7he Manufacture of Knowledge: An Er- (1901), sayondw Co@~”tit& Contemual NaiureofScience fPer- 24. Aronson, N. Why Weren ‘r Vitamins Discovered fidiet’? (un- 8am0n,0.ford, 1981). published, 1984). 47. Browning, C.H. J. Path, EacI. 12, 166-193,(1908). 25. Dobzhanskv. T.G. Genetics and the Ork.ifl of Suecies (Co-

Joshua kderbert

23