BACTEEzOLOGICAL REviEws, Dec. 1976, p. 793-802Copyright 0 1976 American Society for Microbiology

Vol. 40, No. 4Printed in U.S.A.

"Who Discovered Bacteriophage?"DONNA H. DUCKWORTH

Department ofImmunology and Medical Microbiology, University ofFlorida, Gainesville, Florida

INTRODUCTION.............................................................. 793TWORrS DISCOVERY ......... ............................ 794D'HERELLE'S DISCOVERY ............... ...................... 795THE DISCOVERY OF TWORTS DISCOVERY ............................... 797THE CONTROVERSY GRIOWS .................. ................... 798D'HERELLE IS DISCREDITED...................................... 799DID D'HERELLE DISCOVER PHAGE IN 1910? .............................. 800CONCLUSIONS ........... ................................ 800LITERATURE CITED..................................... 801

INTRODUCTIONAlthough most authors state that Twort in

1915 and d'Herelle in 1917 independently dis-covered bacterial viruses (8, 9, 26), some havemade allusions which suggest that d'Herellemay not have been altogether honest in claim-ing to have had no knowledge of Twort's 1915discovery when he published his 1917 work. Inhis excellent book, The Molecular Biology ofBacterial Viruses, Gunther Stent writes: "Gra-tia [sic] drew attention to Twort's forgotten-or, rather, never noted" - discovery (33). And inhis introduction to Phage and the Origins ofMolecular Biology, a similar ambiguity ap-pears: "Bacterial viruses were discovered in1915 by the English microbiologist F. W. Twort,and two years later-perhaps independently,perhaps not-by the French-Canadian F.d'Herelle" (34). Stent had no proof thatd'Herelle had been dishonest (G. Stent, per-sonal communication), but was very much in-fluenced by Andre Lwoffwho felt that d'Herellemay have been dishonest (G. Stent, personalcommunication; A. Lwoff, personal communi-cation). Since I thought the idea that d'Herellehad had knowledge of Twort's work was com-pletely incompatible with the joyous enthusi-asm with which he described his discovery (18),I investigated the matter to see how this accu-sation might have arisen and to see whether,indeed, I could discover who discovered bacte-riophage. It would seem, on the surface, thatwho discovered what would be a relatively sim-ple matter to resolve. Actually, though, thepages of the history of science are strewn withdisputes, often sordid, over priority (31). It ap-pears that the territorial imperative is as realin the Kingdom of Science as it is in the junglesof Africa.To unravel the many possible reasons for

these "territorial" disputes is rather more diffi-cult than to prove that they exist. Their origins

undoubtedly lie in a complex of unresolvedquestions involving the sociology and philoso-phy of science. The disputes may, for instance,arise from the actual frequency of simultaneousdiscovery. But to understand why so many dis-coveries are made by different individuals in-dependently at the same time depends on an in-tricate philosophical analysis of the way inwhich science progresses. Discoveries may notbe simple events occurring at definite timepoints, but rather may result from more elabo-rate extended processes involving large num-bers of people. This has been extensively dis-cussed by Thomas Kuhn (28). In this view,claims to priority would be somewhat artificial,if not meaningless.One could also invoke egotism as the cause of

the many acrimonious controversies over prior-ity that scientists have engaged in. The sociolo-gist, Robert Merton, has rejected this idea,however, and sees the disputes as "signpostsannouncing the violation of the social norms" ofthe scientific establishment. In his view, origi-nality in science is at a premium, and rewardscome from demonstrations of originality. Buthumility and allocation of credit is also an insti-tutional norm. Under the stresses imposed bythe system, the balance between these opposingvalues may be upset, and pathogenic concernwith original discovery may result in "conten-tiousness, self-assertive claims, secretiveness." (31).Many disputes also surely arise from the lack

of a clear-cut definition of what constitutes dis-covery. Is, for instance, an observation a discov-ery, or must all the ramifications of the obser-vations be recognized to constitute a discovery?These questions, however-what is discoveryand what is the origin of the fervor to be first-are beyond the scope of this paper. As it turnsout, even the question posed in the title is notanswered. Instead, it is my intent to present

793

794 DUCKWORTH

some of the heretofore neglected circumstancessurrounding the discovery of bacteriophageand, I hope, to clear the name of Felix d'Herellefrom the aspersions that have been cast upon it.

(This paper is part of an address entitled"The Filterable History of Discovery," pre-sented at the University of Florida on 2 March1976 as one of the President's Scholars LectureSeries.)

TWORT'S DISCOVERYF. W. Twort, the son of a country doctor and a

hard-working mother, was trained in medicineat St. Thomas Hospital in London, but as thefortunes of his family were small he was unableto pursue his keen interest in pathology. "Itwas imperative," relates Twort in his article,"The Discovery of the Bacteriophage" (37),"that I should earn at least sufficient to be ableto pay for lodgings and food. Accordingly, Iaccepted the first paid post available, whichwas that of Assistant Superintendent of theclinical laboratory at the hospital." The postenabled him, nonetheless, to gain experience inpathology, and in 1909 Twort was appointedSuperintendent of the Brown Institution, Lon-don University. The Brown Institution hadbeen founded in 1871 to provide a hospital "forthe care and treatment of Quadrupeds orBirds useful to man." In his capacity as direc-tor, Twort was allowed to carry out research inany branch of pathology or bacteriology, "pro-vided expenditure could be kept within the lim-its of the small income available." He chose tostudy the growth requirements of primitiveforms of life. He had one success, obtaining anessential substance that would allow Johne'sbacillus, the causative agent of a serious dis-ease of cattle, to grow in artificial medium. Hisessential substance has since been identified asvitamin K (20, 37).He then went on to perform experiments in-

tended to indicate how viruses could be grownin artificial media. This seems an unfortunateavenue of research to us as we now know thatviruses can only grow, ifindeed we can say theygrow at all, inside living cells. But Twort wasproceeding upon a not unreasonable hypothesiswhich was that, since viruses were the small-est, hence simplest, forms of life, they musthave, at one time in evolutionary history, beenable to grow in a medium devoid of any livingmatter.As Twort explains in his famous note to the

Lancet on 4 December 1915 (35), "attempts tocultivate these (viruses) from such materials assoil, dung, grass, hay, and water from pondswere made on specially prepared media. It is

impossible to describe all these in detail butgenerally agar, egg, or serum was used as abasis, and to these varying quantities ofcertainchemicals or extracts of fungi, seeds, and etc.,were added." He would then inoculate thesevarious media with extracts of his various soils,dung, or sera, which had been filtered throughcandle filters to remove all bacteria. Every oneof his hundreds of experiments was negative.He never succeeded in growing a filter-passingagent on artificial media. He did, however, ob-serve something else. He had inoculated anagar medium with some fluid (unfiltered) com-monly used for smallpox vaccinations. He no-ticed that although the vaccinia virus did notgrow, a bacteria-a micrococcus- did grow.The bacteria, however, appeared to be afflictedwith some disease -"inoculated agar tubes of-ten showed watery-looking areas, and in cul-tures that grew micrococci it was found thatsome of these colonies could not be subcultured,but if kept they became glassy and transpar-ent." In retrospect, this phenomenon, whichcame to be known as the glassy transformationof Twort, seems exceedingly strange. It is sorare that, in the more than 14 years that I haveworked with bacteria and their viruses, I havenever observed it- at least that I know of. Fur-thermore, I have seen only one reference in theliterature to a similar phenomenon, and thatwas Sir Alexander Fleming's observation thatwhat was later found to be penicillin causedcolonies of staphylococcus to appear like "dropsof dew" (2). Twort, however, made some veryinteresting observations about the glassy trans-formation. (i) The affected colonies would notgrow on any medium. (ii) Examination of theglassy areas revealed only minute granulesand no bacteria. (iii) If a pure culture of mi-crococcus was touched with a small portion ofone of the glassy colonies, the growth at thepoint touched started to become transparentand gradually made the whole colony trans-parent. (This is the startling part of his obser-vation as regards bacterial viruses. Usually bythe time bacteria have grown long enough toform a visible colony, they have reached aphase such that they cannot be attacked by avirus.) (iv) After filtration of the glassy mate-rial through a Chamberland candle, it retainedits ability to cause the "glassy transformation."(v) The "transformation" could be conveyed tofresh cultures for an indefinite number of gen-erations.Twort concluded that the cause of the glassy

transformation was an infectious, filterableagent that killed bacteria and in the processmultiplied itself. He could have very logically

BACTERIOL. REV.

WHO DISCOVERED BACTERIOPHAGE? 795

concluded that he was dealing with a bacterialvirus. But he does not do so. He says:

from these results it is difficult to draw definiteconclusions.... [This] may be living proto-plasm that forms no definite individuals or anenzyme with the power of growth.... In anycase, whatever explanation is accepted, the pos-sibility of its being an ultra-microscopic virushas not been definitely disaproved because wedo not know for certain the nature of such avirus.

Then a little later, "On the whole it seemsprobable, though by no means certain, that theactive transparent material is produced by themicrococcus, and since it leads to its own de-struction and can be transmitted to freshhealthy cultures it might almost be consideredas an acute infectious disease of micrococci." Atthe end of the paper he laments, "I regret thatfinancial considerations have prevented mycarrying these researches to a definite conclu-sion" (35).

Here, as happened also with the discovery ofplant viruses and animal viruses, we have aremarkable discovery faced by its discovererwith uncertainty. Although it was by this timeknown that there were filterable viruses thatcould grow in either plant cells or animal cells,Twort was reluctant to believe he had discov-ered a bacterial virus-if indeed he had-al-though he did consider this a possibility. Hencehe considered Beijerinck's old idea of a fluidform of life, introduced to explain the filterabil-ity of the agent of tobacco mosaic disease (3),and also the possibility that the micrococcuswas producing an enzyme that could grow andmultiply. This latter idea, which seems ridicu-lous now, was the focal point of a controversythat raged in the scientific literature for manyyears. The controversy did not start immedi-ately, however. In fact, Twort's paper went un-recognized until 26 March 1921-5 years, 3months, and 22 days.

D'HERELLE'S DISCOVERYFelix d'Herelle, the man who was heralded

for 4 years as the sole discoverer of bacterialviruses-or bacteriophage, as he called them-lived a life that was more befitting of a Magel-lan, a true explorer, than a mere bacteriologist.Born in 1873 in Montreal of a French-Canadianfather and a Dutch mother, he was educated inFrance and Holland, and during his life trav-eled and worked in Guatemala, Mexico, SouthAmerica, Egypt, Algeria, Tunisia, Indochina,and Russia. He was a professor at Yale for 5years, and an associate ofhis there has recently

said of him that everywhere d'Herelle was,there were fireworks (Florence Mack, personalcommunication). The fireworks materializedrather unfortunately in Russia, where d'Herellewent several times during the 1930s to foundinstitutes for the study of bacteriophage. Dur-ing one of his trips his trusted associate,Eliava, was arrested and shot (1, 7, 29, 32). Heleft Yale for reasons that are not entirely clear,and "several confidential letters exchanged be-tween Dean Winternitz and Professor d'Herellecannot be released" (A. Ebbert, personal com-munication).

It was in Mexico that d'Herelle first observedwhat later came to be called bacteriophage. Hisaccount ofthis discovery written in 1949- about40 years after he had made the discovery - israrely equalled in the scientific literature oftoday for its ingenuous enthusiasm (18). Hesaid:

In 1910, I was in Mexico, in the state of Yuca-tan, when an invasion of locusts occurred; theIndians reported to me that in a certain placethe ground was strewn with the corpses of theseinsects. I went there and collected sick locusts,easily picked out since their principal symptomwas an abundant blackish diarrhoea. Thismalady had not as yet been described, so I stud-ied it. It was caused by bacteria, the locustcoccobacilli, which were present almost in thepure state in the diarrhoeal liquid. I could startepidemics in columns ofhealthy insects by dust-ing cultures of the coccobacillus on plants infront of the advancing columns: the insects in-fected themselves as they devoured the soiledplants.

During the years which followed, I went fromthe Argentine to North Africa to spread thisillness. In the course of these researches, atvarious times I noticed an anomaly shown bysome cultures of the coccobacillus which in-trigued me greatly, although in fact the obser-vation was ordinary enough, so banal indeedthat many bacteriologists had certainly made itbefore on a variety of cultures.The anomaly consisted of clear spots, quite

circular, two or three millimetres in diameter,speckling the cultures grown on agar. Iscratched the surface of the agar in these trans-parent patches, and made slides for the micro-scope; there was nothing to be seen. I concludedfrom this and other experiments that the some-thing which caused the formation of the clearspots must be so small as to be filtrable, that isto say, able to pass a porcelain filter of theChamberland type, which will hold back allbacteria.

However, the appearance of these clearpatches was inconstant. I sometimes wentweeks without seeing a single one, and I couldnot reproduce the phenomenon at will; I there-fore could not study it.

VOL. 40, 1976

796 DUCKWORTH

In March, 1915, during the first World War, alarge invasion of locusts appeared in Tunisia,threatening to destroy the harvests which werethen so vital; I was given the job of starting anepidemic amongst them. As the result of theinfection there was a considerable mortality,and, even more interesting, when the followingyear all the rest of North Africa was againinvaded, Tunisia remained free.

In the course of this campaign, I again ob-served my clear spots, and before returning toFrance I stayed for a time at the Institut Pas-teur in Tunis, to investigate their significance.

I showed them to Charles Nicolle, then direc-tor ofthe Institute, and he said to me: 'That maybe the sign of a filtrable virus carried by yourcoccobacilli, a filtrable virus which is the truepathogenic agent, while the coccobacillus isonly a contaminant.' So I filtered emulsions ofcultures grown on agar and showing the clearspots, and tried to infect healthy locusts withthe filtrate, but without result.On my return to Paris in August 1915, I was

asked by Dr. Roux to investigate an epidemic ofdysentery which was raging in a cavalarysquadron, then resting at Maisons-Laffitte. Ithought the hypothesis, put forward for the lo-custs' illness might be helpful in understandinghuman dysentery. I therefore filtered emulsionsof the faeces of the sick men, let the filtrates acton cultures ofdysentery bacilli and spread themafter incubation on nutritive agar in petridishes: on various occasions I again found myclear spots, but the feeding of these cultures toguinea pigs and rabbits produced no disease.At this time we often got cases of bacillary

dysentery in the hospital of the Institut Pasteurin Paris. I resolved to follow one of these pa-tients through from the time of admission to theend of convalescence, to see at what time theprinciple causing the appearance of the clearpatches first appeared. This is what I did withthe first case which was available.The first day I isolated from the bloody stools

a Shiga dysentery bacillus, but the spreading onagar ofa broth culture, to which had been addeda filtrate from the faeces of the same sick man,gave a normal growth.The same experiment, repeated on the second

and third days, was equally negative. Thefourth day, as on the preceding days, I made anemulsion with a few drops of the still bloodystools, and filtered it through a Chamberlandcandle; to a broth culture of the dysentery bacil-lus isolated the first day, I added a drop of thefiltrate; then I spread a drop of this mixture onagar. I placed the tube of broth culture and theagar plate in an incubator at 37°. It was the endof the afternoon, in what was then the mortu-ary, where I had my laboratory.The next morning, on opening the incubator,

I experienced one of those rare moments of in-tense emotion which reward the researchworker for all his pains: at the first glance I sawthat the broth culture, which the night before

had been very turbid, was perfectly clear: allthe bacteria had vanished, they had dissolvedaway like sugar in water. As for the agarspread, it was devoid of all growth and whatcaused my emotion was that in a flash I hadunderstood: what caused my clear spots was, infact, an invisible microbe, a filterable virus, buta virus parasitic on bacteria.Another thought came to me also: 'If this is

true, the same thing has probably occurred dur-ing the night in the sick man, who yesterdaywas in a serious condition. In his intestine, as inmy test-tube, the dysentery bacilli will havedissolved away under the action of their para-site. He should now be cured.

'I dashed to the hospital. In fact, during thenight, his general condition had greatly im-proved and convalescence was beginning.'(From F. d'Herelle, The bacteriophage, Sci.News 14:44-59, 1949. Copyright X PenguinBooks Ltd 1949. Reprinted by permission ofPenguin Books Ltd.)

In d'Herelle's first communication on bacte-riophage, published in 1917 and entitled "On anInvisible Microbe Antagonistic to DysenteryBacteria" (13), he does not mention his observa-tion of the clear spots in the cultures of thelocust bacteria, and I think he later came toregret this fact. He tells only of his work inParis with dysentery. The main observationsreported were that an agent capable of killingthe dysentery bacterium appeared in the intes-tines of persons recovering from dysentery, butnot from individuals suffering from the acutestage of the disease or from normal individuals;that this agent could be filtered through aChamberland filter and still be active; that itcould be serially transmitted and hence mustbe a "living germ"; that it would not grow onany artificial medium, and therefore had to bean obligatory parasite, a "bacteriophage"; and(mirabile dictu) that this antagonist could "im-munize" rabbits against a lethal dose of thedysentery bacteria. Although, for a historicalrecord, d'Herelle's conclusion that he had founda living organism that would grow only in bac-teria (a bacterial virus) is the most noteworthy,for d'Herelle and many others it was this latterobservation, that this "antagonist" might bethe agent of immunity to bacterial disease, thatwas the most thrilling. He ends his paper byexpressing his belief that this "immunity" maybe a general phenomenon. He does not mentionthe work of Twort that had preceded his byabout a year and a half; in fact, there are noreferences at all in the paper, although this wasnot uncommon for papers presented to theAcademy of Science at that time. Beijerinck,who is often credited exclusively with the origi-nal discovery of viruses, had also not cited the

BACTERIOL. REV.

WHO DISCOVERED BACTERIOPHAGE? 797

earlier work of Ivanowski (3)-work that hadpreceded his by 6 years.

THE DISCOVERY OF TWORTSDISCOVERY

For the next 4 years, d'Herelle lived a lifethat every scientist must dream of. Duringthose years, people in all parts of the worldwere isolating and studying agents oftransmis-sible bacterial lysis-bacteriophage- and mostpeople referred to what they were studying asthe "Phenomenon of d'Herelle." It was hoped bymany that these bacterial killers could be usedtherapeutically in the treatment of bacterialdisease, and it seemed as if Pasteur's idea ofmicrobes fighting microbes or Erlich's "magicbullet" had at last materialized. Hundreds ofpeople cited d'Herelle's work, and although hemay not have been universally regarded, hewas certainly universally acknowledged. Then,in 1921, a paper appeared that was to end hisglory. The paper, by J. Bordet and M. Ciucaand entitled "Remarks on the History of Re-search Concerning Transmissible BacterialLysis," was presented at the 26 March 1921meeting of the Belgian Society of Biology(5). Trouble had actually started in 1920. Adispute had arisen over the true nature of the"d'Herelle phenomenon." Kabeshima (27),working on immunizing animals using phage,had come to the conclusion that the lytic agentcould not be a living being but was a chemicalcatalyst produced from a "pro-ferment" found inall microbes. The pro-ferment could, suppos-edly, be activated by an outside influence or bythe catalyst itself. Bordet and Ciuca (4) hadthen shown that a leukocytic exudate couldcause normal bacteria to produce an agent oftransmissible lysis. They hypothesized that thebacteria were first stimulated by leukocytes toproduce a lytic enzyme and that this enzymecould then, in the absence of the leukocytes,stimulate other bacteria to produce more of theenzyme. This idea was similar to Kabeshima'sbut also to Twort's idea that the transmissiblelysis was caused by an enzyme with the powerof growth. Hence, in their March 1921 paper(5), Bordet and Ciuca say:

it is currently admitted that d'Herelle has beenthe first to observe the lysis which he attributesto a bacteriophage, but which we believe tohave shown represents an autolytic phenome-non that one can start in perfectly normal mi-crobes by a leucocytic exudate. The burden ofanexact history makes it necessary for us to cite aprevious work which d'Herelle has not known[italics mine] and that we ourselves have beenignorant of until now that contains the observa-

tions that d'Herelle has made. This remarkablework by F. W. Twort appeared in Lancet in1915, that is to say, two years before the re-search of d'Herelle.

They go on to describe Twort's observations andhis various interpretations. Concerning thepossibility raised by Twort that what he ob-served was due to an autolytic enzyme pro-duced by the microbe itself, they say: "this au-thor has expressed the idea for which we haveproved the solid basis." They conclude: "with-out wanting to diminish the interest of theobservations of d'Herelle we believe that it is aduty to recognize the incontestable priority ofTwort in the study of this question" (5). It wasobviously a duty they took delight in perform-ing!A very similar thing had happened to Beijer-

inck when he published his paper on filterabil-ity of tobacco mosaic virus in 1898 (3). Iva-nowski pointed out, in a short communicationin 1899, that he had observed the same thing 6years earlier, but in this case Beijerinck hadapologized. d'Herelle, however, did not react asBeijerinck had.On 11 May 1921, d'Herelle read his "defense"

to the meeting of the Society of Biology in Paris(14). The gist of the paper was that Twort wasdealing with something else-not bacterio-phage. He says "I have been able to find onlytwo references in the scientific literature thatcould pertain to the question of bacteriophage.The first is that of Hankin (Ann. de 1' InstitutPasteur, 1896) who states that the water ofcertain rivers of India possesses a bactericidalaction ... no doubt bacteriophage has been thecause." The second reference that d'Herelle ad-mits may pertain to bacteriophage is Twort'spaper of 1915. He points out that Bordet andCiuca have recently cited Twort's work buthave possibly misinterpreted a part of it. Al-though Bordet and Ciuca claim that Twortshowed that his lytic activity was transmissiblein series in a suspension of bacteria, d'Herelleclaims that he (Twort) made "not the slightestallusion to such a phenomenon" (14). WhatTwort does say is the following "This conditionor disease of the micrococcus when transmittedto pure cultures of the micrococcus can be con-veyed to fresh cultures for an indefinite numberof generations" (35). All Twort's other observa-tions had been made on cultures of micrococcusgrowing on agar, so there is no reason to thinkhe did this experiment in liquid cultures, asBordet and Ciuca say he did. So it is probablytrue that they did misinterpret Twort, but themisinterpretation had very little to do with thesubstance of their paper-namely, that Twort

VOL. 40, 1976

798 DUCKWORTH

had observed transmissible lysis of bacteria.This argument over a minor point, which takesup a good part ofd'Herelle's paper, was only thefirst of many spurious arguments that were tobe used by both sides in an escalating warbetween the Paris phage-ologists and the Bel-gian phage-ologists. (The ridiculous argumentsseemed to reach their epitome when, at a July1922 meeting in Glasgow, Gratia, arguing thatphage is not living says, "fire is not living yetfire is endowed with power of reproduction,"and later: "when we pour out a glass of sodawater there appear on the wall of the glasssmall round bubbles of gas, the size of whichincreases exactly as the so-called colonies ofbacteriophage, and yet gas is not a virus" [23].)The remainder of d'Herelle's 1921 paper islargely a summary of Twort's work as reportedin 1915, after which he says, somewhat dramat-ically, "'the bacteriophage- is it the cause ofthevery interesting phenomenon described byTwort?" He answers himself, "it is possible but'peu vraisemblable' " (14). His reasons forthinking so seem to be largely intuitive, al-though he does cite a difference between thethermal inactivation points of the agent thatTwort isolated and his own bacteriophage prep-arations. He ends his paper by pointing outthat "the intensity of bacteriophage action is so'brutal' that many bacteriologists ought to haveseen it in the course of their researches buthave neglected it because it was for them in-comprehensible" (14).Who were these many bacteriologists who

had seen the phenomenon but neglected it?Was one ofthem perhaps a young bacteriologistworking in Mexico to save the world fromplagues of locusts?

d'Herelle's defense of himself seems to havehad the effecti of a declaration of war on theBelgian microbiologists. A formidable ally,Andr6 Gratia, joined Bordet's forces, and inNovember 1921, a note by Gratia and Jauminwas read to the Belgian Society of Biology (25).The title was "Identity of the Phenomenon ofTwort and the Phenomenon of d'Herelle."

Gratia and Jaumin had isolated from vacci-nia pulp (the same material Twort had startedwith) an agent that had a "marked inhibitingand dissolving action on staphylococci" andshowed that the agent could be transmittedindefinitely from one culture to another. Theirnote begins by saying that they have confirmedthe observations of Twort, and furthermore,obtained a lytic principle that has permittedthem "to reproduce with Staphylococci all theparticulars of the phenomenon for whichd'Herelle claims priority and which he atri-

butes to bacteriophage." For evidence, they re-fer to earlier papers which, they say, leave nodoubt as to the identity of Twort's phenomenonand d'Herelle's phenomenon-"an identitywhich the latter author thought he could con-test." An examination of their earlier papers(21, 22) reveals that it is far from certain thatthey had proved the two phenomena were thesame. They had, indeed, isolated a lytic princi-ple -but they had isolated it from some "small,clarified areas" of their agar slants, not fromcolonies that started to grow and then becameglassy. Little attempt was made to repeat theidentical experiments of Twort. They mentionthe word "glassy" once, but more often talkabout irregular colonies as being the ones thatcontain the lytic principle. One gets the feelingthat the word "glassy" was thrown in somewhatsurreptitiously.Their November report (25), although enti-

tled so as to lead one to believe they had conclu-sively reconciled Twort's observations withd'Herelle's, is, in actual fact, an attempt toidentify a lytic principle, a phage, they hadisolated from vaccinia pulp (not necessarilywhat Twort had isolated) with a staphylococcalbacteriophage they had isolated previously andthat they believed to be a multiplying enzyme.It becomes apparent that had Twort claimed tohave discovered a "living germ-an obligatorybacteriophage" (16, 17) as d'Herelle had, andnot an enzyme that could reproduce itself, hiswork would have remained undiscovered foreven longer than it did. There seems to be littledoubt that it was because Twort favored thehypothesis of the Belgian group that he wasdiscovered. (This view has also been expressedby P. Fildes [20].)

THE CONTROVERSY GROWSIn the age of molecular biology, when it is

known rather precisely what (although notwhy) a virus is, the hypothesis of Bordet's groupseems, to say the least, unsophisticated.d'Herelle's criticism does not seem unreasona-ble:

The leucocytic exudate induces the 'nutritivevitiation' which results in the lysis of the bacte-ria in the first tube of the series, but in thefollowing ones the same effect will be produced,no longer by the leucocytic exudate, which willof necessity have disappeared in the first tubesbecause of the dilution, but by the bacteriallytic ferment alone. Bordet and Ciuca seem tofind this substitution of cause entirely logical,when in reality it is contrary to all that weknow (17).

BACTRIUOL. REV.

WHO DISCOVERED BACTERIOPHAGE? 799

However, it is probable that the merits of theidea (or lack thereof) had very little to do withthe heat generated by the controversy betweend'Herelle and Bordet. Whether or not all scien-tists can be said to be largely motivated by thepossibility of widespread recognition, it is cer-tainly true that, in the controversy overwhether agents of bacterial lysis were "livinggerms" or "multiplying enzymes," there wereno tiny egos involved. d'Herelle, who is nowacknowledged to have been, for the most part,correct (19), at least as regards virulent phage,may have been particularly abrasive.

In writing of d'Herelle shortly after hisdeath, one of his associates, Pierre Lepine, said:"Felix d'Herelle was a curious and magneticpersonality. Lively, sometimes irascible, hewas a faithful and sure friend. If he sometimesgave to those who knew him but little theimpression of a difficult character, those whoknew him well have not ceased to feel his affa-bility and joyous ardor in his work" (29).That he was easily aroused in his own de-

fense there is no doubt. A mere glance at theTable of Contents of his first book (16) revealsthat he had no patience with those who disa-greed with him. Among the subtitles of hischapter, "The Nature of the Bacteriophage" (achapter added in 1922 for the English edition),are the following: experimental proofs of theliving nature ofthe bacteriophage; refutation ofthe hypothesis of Kabeshima; refutation of thehypothesis of Bordet and Ciuca; refutation ofthe hypothesis of Bail; refutation ofthe hypoth-esis of Salimbeni; conclusions. In his conclu-sions he states, "all of the authors who haveadvanced hypotheses other than that of an ul-tramicrobe parasitizing the bacteria have for-gotten that a hypothesis ought always to ac-count for the entire mass of facts."

In the summer of 1922, d'Herelle, Twort, Bor-det, and Gratia were invited to speak at a Bri-tish Medical Association symposium on thebacteriophage held in Glasgow. d'Herelle spokefirst-and longest. He says again that manyauthors have neglected the whole assembly offacts-a remark sure to win him few friends.He presents four hypotheses on the nature ofbacteriophage and proceeds systematically toeliminate all but one: the one to which "I haveheld since my first publication" (15).

Twort's presentation was very much a reiter-ation of his 1915 paper in which he tries veryhard to walk the thin line between Bordet andGratia on the one hand and d'Herelle on theother. He views his original results as "evi-dence not only against [d'Herelle's] view thatthe lytic agent is a living organism, but also

against the view of Bordet and Ciuca" (36). Ford'Herelle, anyone that was not with him wasagainst him. Hence, in his second book pub-lished in 1926 (17), the hapless Twort comesunder attack. In his first book d'Herelle hadmentioned Twort's paper as bearing on the sub-ject of bacteriophage (16). In 1921 he had alsoallowed that Twort's observation might bear onthe subject of bacteriophage (14). But in his1926 book he states that "as is evident, thephenomenon observed by Twort and the phe-nomenon of bacteriophagy present two entirelydifferent aspects ... those authors who havelikened the phenomenon of Twort to bacteri-ophagy have restricted themselves to affirma-tions only, without offering any supportingproof whatever." He states that Twort refusedto discuss the issue with him publicly, both in1922 and in 1923 and that "it is difficult to avoidthe conclusion that if Twort refrains from sucha discussion it is simply because the facts re-vealing the dissimilarity of the two phenome-non are indisputable."

D'HERELLE IS DISCREDITEDd'Herelle's 1926 book was notable in another

respect. It was in this book that d'Herelle firstpublished in detail the story of his finding bac-teriophage in Mexico. He had actually men-tioned seeing plaques with his locust bacteriain a chapter added to the English edition of hisfirst book in 1922 (16), although he had men-tioned it in his historical introduction to theFrench edition written in 1921. It seems thatthe rediscovery of Twort had a great deal to dowith the rekindling ofd'Herelle's memory. Dur-ing the time between the publication of theFrench edition and the English edition, Bor-det's and Gratia's attacks had begun in earnestand, by 1925, when d'Herelle wrote his secondbook, the attacks had escalated. With the at-tacks came increasing insistence that Twortand not d'Herelle had first discovered bacterio-phage. Now d'Herelle's recounting of his firstseeing the plaques in 1910 became considerablymore convincing.

It is my hypothesis that with d'Herelle's in-creasing insistence that he was the sole discov-erer of bacteriophage, and the sudden appear-ance of the story of his 1910 work, came theaccusation that he had known of Twort's work.It can certainly be said that, by introducing thestory of his 1910 observations, d'Herelle actedas a guilty man. If he had truly not known ofTwort's work, would he not have been apolo-getic upon discovering that Twort had earliermade a similar observation? Some will say that

VOL. 40, 1976

800 DUCKWORTH

only a guilty man would go to such lengths toprove his innocence. One need not, however,invoke guilt to explain d'Herelle's claim to havediscovered phage before Twort. d'Herelle was atthis time an extremely beleaguered man.Whether or not Twort had seen the effects ofphage first, it was d'Herelle who, by the sheervolume of his work, had brought the phenome-non to the attention of the world and had maderesearch with bacteriophage one of the mostexciting fields to work in during the 1920s. Hedeveloped the quantitative plaque assay andquantitative dilution technique and, by use ofthese, described the process of bacteriophagegrowth, which is remarkably "close to the pic-ture we have today" (19). In spite ofthis, he wascruelly and viciously attacked. His triumphhad evaporated - the d'Herelle phenomenonhad become the Twort-d'Herelle phenomenon-not even the d'Herelle-Twort phenomenon. At-tempts were made to make him look somewhata fool (23). For a man who was extremely proudof his associations with the tradition of Pasteurand of his discussions about phage with AlbertEinstein (17), this must have been bitter medi-cine indeed. Furthermore, d'Herelle was a verysincere man who, I have no doubt, honestlybelieved that Twort had observed somethingelse. In this case, it would be immaterialwhether or not he had seen Twort's paper.

DID D'HERELLE DISCOVER PHAGE IN1910?

The role of communication in the establish-ment of priorities not withstanding, the ques-tion does remain-did Twort or did d'Herelleobserve bacterial viruses first? Did d'Herellesee "taches vierges" - clear spots - on the cul-tures of the locust coccobacillus? He did, in-deed, report on a disease of locusts that heobserved in the Yucatan peninsula of Mexico.This was in a note presented to the Academy ofScience in 1911 (10). But he does not mentionthe clear spots. Nor does he mention them inhis very scholarly papers on the locust coccoba-cillus published in 1914 (11, 12). That he doesnot mention them, does not, of course, mean hedid not see them. Remember the many bacte-riologists who had seen the phenomenon but,not knowing what it was, had neglected it (14).Unfortunately, the one person who could tell usthe true story lies buried near his home on theoutskirts of Paris. Because d'Herelle was ofCanadian birth he was incarcerated by the Ger-mans during their occupation of France. Hewas freed at the end of the war, but was a sickman. He died on February 22, 1949. Twort andGratia both died the next year.

CONCLUSIONSWe can see that, as Thomas Kuhn predicted,

"research into the history of science makes itharder, not easier, to answer questions likewhen was oxygen discovered? Who first con-ceived of energy conservation?" (28). Who dis-covered bacteriophage? The question is unan-swerable. On the other hand, we can see thatthere is no evidence that d'Herelle "stole"Twort's discovery. As Twort's paper seems tohave been generally unnoticed until 1921, thereis no reason to think that d'Herelle had seen itin 1917. Even Bordet and Ciuca said in 1921that d'Herelle had "not known" of Twort's work(5). So it seems quite logical to conclude that,indeed, he had not; and that, if Twort did dis-cover bacteriophage in 1915, then d'Herelle didso independently in 1917, if not in 1910. We canrightly question whether or not Twort actuallydid "discover" bacteriophage at all, since hecertainly did not conclude that what he hadobserved was a bacterial virus. It was Bordet,in favoring Twort's multiplying enzyme theory,who made the priority claim for him. However,since Twort did consider the idea of a bacterialvirus, and since temperate phage may appearto have some of the properties of multiplyingenzymes, it would seem ungracious and unfit togive him no credit.

It is tempting to do a "Mertonian" analysis asto why the dispute regarding the true discov-erer of phage was so heated. It could be saidthat in refusing to credit Twort, even when hiswork was brought to light, that d'Herelle vio-lated one of the institutional mores of science-that of humility. Hence he aroused the moralindignation of the scientific community andhad to be censured. But even in their firstpublication, "Remarks on the History of Re-search Concerning Transmissible BacterialLysis," the tone of Bordet and Ciuca was any-thing but conciliatory. They, in a sense, threwdown the gauntlet. It seems more likely thatthe fight was over territory-who will be theking?Yet even if we cannot decide who was the

true discoverer of bacteriophage, or add signifi-cantly to the sociology of priority disputes, wecan draw some conclusions regarding the effectof these disputes on the rapid advancement ofscience. The field of bacteriophage research re-mained in a dire state of disunity well into the1930s. The discovery of colicins and lysogenicbacteria served only to heighten the alreadyexisting confusion. (See Andre Lwoffs excel-lent review in this same journal in 1953 for adiscussion of this era [30].) In spite of an abor-tive attempt by Burnet and McKie in 1929 (6) toreconcile the two aspects of baceriophage phe-

BACTERIOL. REV.

WHO DISCOVERED BACTERIOPHAGE? 801

nomena, the temperate aspect and the virulentaspect, Gratia in 1932 had to, in what seems arather desperate attempt, resort to a belief inspontaneous generation to try to explain all thevarious observations he and others had made.Said he:

in truth, the 'pasteurien theory' [that sponta-neous generation of life cannot occur], indisput-able for bacteria, is perhaps not valid for 'pre-cellular' forms of life and I recall here again theexample of fire. This, in some ways resemblinglife, was in the eyes of our ancestors consideredas being only able to come from a pre-existingflame ... today, everyone knows that not onlycan fire be transmitted from one flame to an-other, but that it can be 'born' from the frictionof a match or the 'travail' of fermentation (24).

One cannot but think that had a veritablewar not been started between d'Herelle, cham-pion of the virulent phage, and Bordet andGratia, defenders of the temperate phage, thetrue nature of bacteriophage would have beenrealized earlier, and spontaneous generationcould have been. left to rest in peace.

ACKNOWLEDGMENTSI am immensely grateful to Gunther Stent who, in

his excellent book on bacterial viruses (32), first"introduced" me to Felix d'Herelle and who wrote tome several times about the discovery of phage. I amalso very grateful to Andr6 Lwoff for his communi-cations about d'Herelle and for help in trying tolocate d'Herelle's daughters. I also thank Peggy Jor-dan and Stanley Laham for help in reading theFrench, Ann M. Duckworth for the copy of ScienceNews, Pat Crowley and the J. Hillis Miller HealthCenter Library for locating many of the journals,and Harry Paul for several stimulating discussions.I also thank my husband, Alistair M. Duckworth,for removing several dangling participles and otheritems of that general nature.

LITERATURE CITED1. Anonymous. 1949. Felix d'Herelle (obituary). J.

Am. Med. Assoc. 140:907.2. Baldry, P. E. 1965. The battle against bacteria.

Cambridge University Press, Cambridge.3. Beierinck, M. W. 1898. Ueber die Mosaikkran-

kheit des Tabacks (in English). Phytophathol.Classics 7:11-26.

4. Bordet, J., and M. Ciuca. 1920. Le bacterio-phage de d'Herelle, sa production et son inter-pretation. C. R. Soc. Biol. Paris 83:1296-1298.

5. Bordet, J., and M. Ciuca 1921. Remarques surl'historique des recherches concernant la lysemicrobienne transmissible. C. R. Soc. Biol.Paris 84:745-747.

6. Burnet, F. M., and M. McKie. 1929. Observa-tions on a permanently lysogenic strain ofB.enteritidis gaertner. Aust. J. Exp. Biol. Med.Sci. 6:277-284.

7. Compton, A. 1949. Felix d'Herelle (obituary).

Nature (London) 163:984.8. Davis, B. D., R. Dulbecco, H. N. Eisen, H. S.

Ginsberg, and W. B. Wood. 1967. Microbiol-ogy. Harper and Row Publishers, New York.

9. Delbruck, M. 1942. Bacterial viruses. Adv. En-zymol. 2:1-32.

10. d'Herelle, F. 1911. Sur une epizootie de naturebacterienne sevissant sur les sauterelles aumexique. C. R. Acad. Sci. Paris 152:1413-1415.

11. d'Herelle, F. 1914. Le coccobacille des saute-relles. Ann. Inst. Pasteur 28:280-387.

12. d'Herelle, F. 1914. Le coccobacille des saute-relles. Ann. Inst. Pasteur 28:387-419.

13. d'Herelle, F. 1917. Sur un microbe invisible an-tagonistic des bacilles dysenterique. C. R.Acad. Sci. Paris 165:373-375.

14. d'Herelle, F. 1921. Sur l'historique du bac-teriophage. C. R. Soc. Biol. Paris 84:863-864.

15. d'Herelle, F. 1921. The nature of bacteriophage.Br. Med. J. 2:289-293.

16. d'Herelle, F. 1922. The bacteriophage; its role inimmunity (English translation). Williamsand Wilkins, Baltimore.

17. d'Herelle, F. 1926. The bacteriophage and itsbehavior (in English). The Williams & Wil-kins Co., Baltimore.

18. d'Herelle, F. 1949. The bacteriophage. Sci.News 14:44-59.

19. Ellis, E. 1966. Bacteriophage: one-step growth,p. 56. In J. Cairns, G. S. Stent, and J. D.Watson (ed.), Phage and the origins of molec-ular biology. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

20. Fildes, P. 1951. Frederick William Twort 1877-1950. Obituary Notices Fellows R. Soc. 7:505-517.

21. Gratia, A. 1921. Preliminary report on a Staphy-lococcus bacteriophage. Proc. Soc. Exp. Biol.Med. 18:217-218.

22. Gratia, A. 1921. L'autolyse transmissible du sta-phylocoque. C. R. Soc. Biol. Paris 85:25-26.

23. Gratia, A. 1922. Concerning the theories of theso-called bacteriophage. Br. Med. J. 2:296-297.

24. Gratia, A. 1932. Antagonisme microbien et "bac-teriophage." Ann. Inst. Pasteur 48:413-437.

25. Gratia, A., and D. Jaumin. 1921. Identitk duphenomene de Twort et du phenomene ded'Herelle. C. R. Soc. Biol. Paris 85:880-881.

26. Joklick, W. K., and D. T. Smith (ed.). 1972.Microbiology. Appleton-Century-Crofts, NewYork.

27. Kabeshima, M. 1920. Sur le ferment d'immunitebacteriolysant. C. R. Soc. Biol. Paris 83:471-478.

28. Kuhn, T. 1972. The structure of scientific revo-lutions; 2. University of Chicago Press, Chi-cago.

29. Lepine, P. 1949. Felix d'Herelle (obituary).Ann. Inst. Pasteur 76:457.

30. Lwoff, A. 1953. Lysogeny. Bacteriol. Rev.17:269-337.

31. Merton, R. K. 1957. Priorities in scientific dis-covery: a chapter in the sociology of science.Am. Sociol. Rev. 22:635-659.

VOL. 40, 1976

802 DUCKWORTH

32. Nicolle, P. 1949. Felix d'Herelle (obituary).Presse Med. 57:350.

33. Stent, G. S. 1963. The molecular biology of bac-terial viruses. W. H. Freeman Co., San Fran-cisco.

34. Stent, G. S. 1966. Introduction: waiting for theparadox. In J. Cairns, G. S. Stent, and J. D.Watson (ed.), Phage and the origins of molec-ular biology. Cold Spring Harbor Laboratory,

BACTERIOL. REV.

Cold Spring Harbor, N.Y.35. Twort, F. W. 1915. An investigation on the na-

ture of ultra-microscopic viruses. Lancet2:1241-1243.

36. Twort, F. W. 1922. The bacteriophage: thebreaking down of bacteria by associated filter-passing lysins. Br. Med. J. 2:293-296.

37. Twort, F. W. 1949. The discovery of the bacte-riophage. Sci. News 14:33-34.