rtimttutit) ill! It liiii I I ill II ! UNIVERSITY OF CALIFORNIA «EPARTMENT OF CIVIL ENGINEER!** BERKELEY, CALIFORNIA Civil Engineering Dept UNIVERSITY OF CAUFQKNIA DEPARTMENT OF CIVIL ENGINEERING BERKELEY, CALIFORNIA THE BACTERIOPHAGE ITS ROLE IN IMMUNITY ENGLISH EDITION THE BACTEEIOPHAGE Its Role In Immunity WITH FOURTEEN TEXT ILLUSTRATIONS BY F. D'HERELLE it Pasteur Institute AUTHORIZED TRANSLATION BY GEORGE H. SMITH, PH.D. ASSISTANT PROFESSOR OF BACTERIOLOGY AND PATHOLOGY Yale University School of Medicine PUBLISHED BY WILLIAMS & WILKINS COMPANY BALTIMORE, U. S. A. 1922 ft I 8 Engineering Library COPYRIGHT 1922 WILLIAMS & WILKINS COMPANY Made in United States of America All rights reserved, including that of translation into foreign languages, including the Scandinavian COMPOSED AND PRINTED AT THE WAVERLY PRESS BY THE WILLIAMS & WILKINS COMPANT BALTIMORE, MD., U. S. A. PREFACE In this monograph I have collected and coordinated the several notes and communications published, not only by myself but by others also, since 1917. The study of the phenomenon was undertaken without any preconceived ideas regarding the nature of the causal principle involved. Indeed, what could it matter whether the active force was a diastase, a living germ, or some new property of the bacterium? Whatever it might be, the interest would remain the same. It was only after two years spent in investigation, with the completion of some hundreds of experiments, each one more conclusive than the preceding, that I became convinced that the cause of serial transmissible bacteriolysis could be nothing other than a living organism. And it was not until this time that my first communication on the subject was presented. Some three years later the attention of others became directed to the subject and they, for the most part, in turn suggested hypotheses as to the nature of the phenomenon differing among themselves and differing fundamentally from that which I had announced and which, in view of all of the facts and phenomena involved, is the only one which is tenable. None of these investigators have considered all of the factors and facts involved in the reac- tion; instead, each has selected a particular group of facts, suffi- cient to support his thesis, and has neglected all other experi- mental data such as would render his hypothesis inadmissible. It may be added that all of these hypotheses were considered prior to my first publication, and the solution of the question required many experiments indeed. In spite of the fact that the results of these experiments have been published in various memo- randa none of those who have opposed my theory have refuted them, or even alluded to them. Naturally, in this monograph, these experiments will be presented, experiments which by them- selves refute the interpretations of bacteriophagous activity as a diastatic action, whether the agent be derived from the organism 5 6 PREFACE which is defending itself or from the bacterium which is causing the infection, both of which sources have been suggested by dif- ferent authors. It may be that the reader will find sometimes in the course of this discussion that I have multiplied evidence by repetition, or that it is necessary to make an effort to follow certain of the experiments, so that he is lost in a maze, not only of new phenom- mena for which he is as yet unprepared, but also in phenomena of extreme complexity. The difficulties of exposition of the sub- ject will readily be comprehended if we realize that up to the present time Bacteriology has been considered as a " problem of two bodies," bacterium and medium, whether the medium be the organism parasitized or a culture fluid. And this problem of the two bodies has been indeed complex. But it is of necessity much less complicated than the "problem of three bodies" with which we must now be concerned, where we must recognize the interactions between the medium — culture medium or organism parasitized, — the bacterium parasitizing this medium, and the ultramicrobial bacteriophage parasitizing the bacterium. Pan's, July 1, 1921. PREFACE TO THE ENGLISH EDITION The present English edition of the monograph which presents the results of my investigations on the Bacteriophage is not a simple translation of the French edition, which, appearing in October, 1921, was one of the series of monographs of the Pasteur Institute. The bibliography has been extended, and in many of the chapters new experimental evidence has been introduced, embodying work completed since the publication of the French edition. In addition, a wholly new chapter has been prepared dealing with the " Nature of the Bacteriophage," which pre- sents a statement of and an analysis of the diverse hypotheses which have been advanced in explanation of the intimate nature of the principle. Indeed, and this is of some significance, this is the only point — purely theoretical moreover — upon which dis- cussion of the subject turns. As for the experimental facts themselves, they have never been questioned, for all investiga- tors who have worked with the bacteriophage have confirmed the experimental results almost in their entirety. It is my hope that this work will be of interest to the biologist. I hope especially that it will stimulate American bacteriologists in greater and greater numbers to become interested in investiga- tive work upon the subject; one which is new and which promises to be fruitful in theoretical and practical results. And this is my wish the more since, although working for many years in France, the signatory of these lines in the quality of a Canadian, feels himself a part of the great American family. March 15, 1922 TABLE OF CONTENTS Preface 5 * * * * PART I. THE BACTERIOPHAGE Introduction 15 CHAPTER I. BACTERIOLYSIS Bacteriolysis 23 Technic for the isolation of the bacteriophage 24 Technic for enhancing virulence 29 Enumeration of the bacteriophagous ultramicrobes 31 Multiplication of the ultramicroscopic bacteriophage 34 The bacteriophage : An obligatory parasite 38 The effect of the condition of the bacteria 40 The influence of the medium 42 Culture of the bacteriophage on solid media: Isolated colonies 45 Effect of the concentration of bacteria in the medium; inhibitory effects of the products of lysis 49 Effect of external physical conditions 53 Effect of antiseptics upon lysis 54 Soluble substances of the bacteria 57 The ultramicrobial bacteriophage: An endoparasite 58 Bacteriolysis under the microscope 61 CHAPTER II. THE BACTERIOPHAGE AND THE BACTERIUM Virulence of the bacteriophage 66 Evaluation of the degree of virulence 69 Resistance of the bacterium 70 The origin of secondary cultures 73 Instability of mixed cultures 76 The characters of mixed cultures 78 The resistant bacterium 84 Microscopic observations 89 The acquisition of resistance 91 Production of antilysins by bacteria 92 Multiple cultures 94 9 10 CONTENTS CHAPTER III. VIRULENCE OF THE BACTERIOPHAGE Multiple virulence 96 Persistence of virulence 97 Bacterial species attacked 103 Bacillus dysenteriae Shiga 104 Bacillus dysenteriae Hiss 105 Bacillus dysenteriae Flexner 105 Bacillus dysenteriae "X" 106 Bacillus coli 107 Bacillus typhosus 107 Bacillus paratyphosus A 108 Bacillus paratyphosus B 108 Salmonella (hog cholera) 108 Bacillus typhi murium. ". 108 Bacillus proteus 109 Bacillus gallinarum (Klein), B. gallinarum (Moore); paragalli- narum 109 Bacterium diphtheriae 110 Staphylococcus 110 Bacterium of barbone 110 Bacillus pestis Ill Bacillus of flacherie Ill Bacillus subtilis Ill Vibrio choLerae Ill CHAPTER IV. THE BACTERIOPHAGOUS ULTRAMICROBE Morphology 113 Vitality 115 Susceptibility to different agents 116 Unicity of the bacteriophage 120 The lysins of the bacteriophage 123 Opsonic power of the lysins 125 CHAPTER V. THE BACTERIOPHAGOUS ANTISERUM Complexity of the antibodies 130 Antibodies to the bacteria 132 Antibodies to the bacterial toxins 132 Antibodies to the bacteriophagous ultramicrobes 134 Antibodies to the lysins 138 Incidental conditions resulting from the existence of the bacteriophage. 141 CHAPTER VI. THE NATURE OF THE BACTERIOPHAGE The nature of the bacteriophage 144 CONTENTS 11 The possible hypotheses 146 Experimental proofs of the living nature of the bacteriophage 149 Refutation of the hypothesis of Kabeshima 153 Refutation of the hypothesis of Bordet and Ciuca 154 Refutation of the hypothesis of Bail 158 Refutation of the hypothesis of Salimbeni 159 Conclusions 159 PART II. THE ROLE OF THE BACTERIOPHAGE IN IMMUNITY Introduction 163 CHAPTER I. THE BACTERIOPHAGE IN DISEASE Choice of diseases to study 173 Bacillary dysentery 176 Colon bacillus infections 189 Typhoid fever and the paratyphoid fevers 189 Avian typhosis 204 Hemorrhagic septicemia of the buffalo (barbone) 217 Bubonic plague 224 Flacherie 227 Conclusions 228 CHAPTER II. THE BACTERIOPHAGE IN THE HEALTHY INDIVIDUAL The bacteriophage in healthy man , 229 The bacteriophage in the horse 233 The bacteriophage in the fowl 236 The bacteriophage in diverse animals 237 Conclusions 239 CHAPTER III. IMMUNIZATION BY MEANS OF THE BACTERIOPHAGE Immunization against avian typhosis 242 Immunization against barbone 248 Immunization against dysentery 262 CHAPTER IV. THE BACTERIOPHAGE AND IMMUNITY Summary and conclusions 272 BIBLIOGRAPHY. . 283 PLATE 1 MULTIPLICATION OF THE ULTRAMICROBIAL BACTERIOPHAQE A bacterial suspension is inoculated with one-billionth of a cubic centi- meter of a filtrate containing the Bacteriophage. From time to time one- fiftieth of a cubic centimeter is taken from this suspension and spread on agar. After a period of incubation of the agar tubes the following results are obtained (reading from left to right). Tube 1. Planting made immediately after the inoculation of the bac- teriophage. A normal bacterial culture results. Tube 2. Planting made lj hours after the inoculation. A normal bac- terial culture results. Tube 3. Planting made 1\ hours after the inoculation. The layer of bacterial growth shows three colonies of the bacteriophage. Tube 4> Planting made 3f hours after the inoculation. Confluent colonies of the bacteriophage are disseminated throughout the bacterial growth. Tube 5. Planting made after 5 hours. There is no evidence of bacterial growth, the number of ultramicrobes being such that all of the bacterial elements were destroyed. 12 PLATE 1 PART 1 THE BACTERIOPHAGE INTRODUCTION HISTORICAL Examination of the scientific literature discloses but two com- munications bearing on the subject of the bacteriophage. In point of priority the first is that of Hankin.1 This author states that he detected in the waters of certain rivers of India a very marked antiseptic action, directed against bacteria in general, but against the cholera vibrio more particularly. Thus, for instance, the water of the Jumna as it leaves the town of Agra contains more than 100,000 bacteria per cubic centimeter, while at a distance of 5 kilometers further down the bacterial content is but 90 to 100. With reference to Vibrio cholerae in particular, his laboratory experiments gave results as follows. Specimen "A" represents the filtered water (filtered through porcelain); specimen "B" is the same filtered water after boiling; both specimens being inocu- lated with a culture of V. cholerae. NUMBER C )F ORGANI SMS AFTEB 0 1 hour 2 hours 3 hours 4 hours 25 hours 49 hours Specimen A Specimen B 2,500 5,000 1,500 4 000 1,000 6,000 500 10 000 0 6,000 0 10, 000 0 36,000 The germicidal action of the water of these streams could always be detected but was present to varying degrees. It is to this antiseptic action that Hankin attributes the fact that the ingestion of the water can not be incriminated as the origin of cholera. Moreover, these streams have never been the vectors of epidemics since the propagation of such outbreaks is always from downstream upward. 1 L'action bactericide des eaux de la Jumna et du Gauge sur le vibrion du cholera. Ann. de Tlnst. Pasteur, 1896, 10, 511 . 15 16 INTKODUCTION m states that the antiseptic principle is destroyed by boiling and he considers himself warranted in affirming that it is a volatile substance. To my mind there is no doubt that this antiseptic action ought in reality to be assigned to the bacteriophage. The second publication is that of Twort2 entitled "An Investi- gation on the Nature of the Ultramicroscopic Viruses." In the course of his experiments upon the filtrable virus of vaccinia this author obtained on certain of his agar slants inoculated with the glycerinated vaccinal pulp, a culture of a micrococcus of which certain colonies presented a vitreous and transparent aspect. The micrococcus had been replaced by fine granules. At other times he obtained a film of growth showing spots composed of the same vitreous material. These colonies slowly spread over the entire culture, the micrococcus everywhere being transformed into granules. When a pure culture of the micrococcus was touched with a platinum wire which had previously been in contact with the vitreous material, a spot of the same nature developed and extended gradually over the whole surface. The action was feeble on cultures previously killed. The vitreous substance, when diluted, passed through a porcelain filter, for a drop of the filtrate transformed a normal healthy culture into one of the vitreous appearance. The transformation process began in isolated points and rapidly extended over the surface. However, if some portion of the normal culture never came in contact with the filtrate, the healthy growth regained the advantage and extended over the vitreous stratum but without effecting its destruction. The material of transparent and vitreous nature maintained its activity for at least six months. It resisted a temperature of 52°C. but was destroyed at 60°C. Twort obtained similar results with an organism of the colon group, isolated from the intestinal mucosa of a dog affected with Hundeseuche, and with a large bacillus not belonging to the colon group isolated from the intestinal contents of an infant suffering from diarrhea. In both cases the material transformed the nor- mal culture into matter of a vitreous transparent aspect. 2 An Investigation on the Nature of the Ultramicroscopic Viruses. Lan- cet, 1915, ii, 1241 (December 4). INTRODUCTION 17 This author reviews the different hypotheses which he consid- ered as possible causes in effecting the transformation of the cul- tures. The substance of vitreous appearance, he says, certainly contains an enzyme which is destroyed at 60°C. On the other hand, the material is susceptible of continued cultivation, since it can be indefinitely transplanted on a culture of the micrococ- cus. May it be a cultivable enzyme? May it be living proto- plasm of indeterminate form? May it be an ancestral form of the micrococcus which can not be cultivated in this form but which incites the normal micrococcus to take this form of regression? Or, may it be an enzyme secreted by the micrococcus itself which produces thus its own destruction, with the formation of a new quantity of destructive enzyme? May the substance of vitreous appearance be composed of a filtrable virus which may be itself the virus of vaccinia or a filtrable virus entirely without patho- genicity derived from the air and which enters the micrococcus cultures by passing through the cotton which closes the tubes? Twort did not choose from these several hypotheses which he formulated, although it seemed to him most probable that the vitreous substance was produced by the organism, the micrococ- cus, itself. He indicated further, that he had no idea as to the relation which might exist between the bacillus or the micrococ- cus, the vitreous material, and the disease. The phenomenon observed by Twort has been thus emphasized because it involves a serial activity, and because, on the other hand, it is hardly probable that the cause is a bacteriophage. Indeed, it has been proved that with a true bacteriophage active against Staphylococcus albus the phenomenon described by Twort can not be reproduced. The very peculiar characters, and indeed, the characteristics presented by the phenomenon observed by Twort render confusion of this principle with the bacteriophage impossible. The difference in thermal death points, amounting to 15°C., between the two principles (the bacteriophage becomes inactive only at about 75°C.) is alone sufficient to differentiate them. According to some experiments which I have performed the facts observed by Twort may be ascribed to a fragmentation of the bacteria; it is only necessary to use the ultramicroscope to see that the "vitreous substance' ' is composed of very minute cocci. 18 INTRODU CTION However that may be, the intensity of the bacteriophagous action is sometimes of such violence that it must have been ob- served by many bacteriologists in the course of their investiga- tions even though the nature of the phenomenon and its mechan- ism were not understood. For example, I have been informed that in Haffkine's laboratory, it has been noted several times that cultures of the plague bacillus in bouillon underwent clarification, the medium becoming perfectly limpid within the space of a few hours. Not knowing the reason for this curious phenomenon the cultures were termed " suicides." For such reactions the bac- teriophage was certainly the cause. Another observation of the same nature is reported by Eliava, who, being in charge of the examination of the water of the Koura river at Tiflis, noted the following phenomenon. The suspected water under examination was added to a peptone solution. After a few hours of incubation a specimen taken from the surface of the medium for microscopic examination showed very numerous vibrios of normal morphology. Planted upon agar, this speci- men yielded upon incubation a growth of dull appearance which microscopically appeared to be a culture of vibrios. Twelve hours later, in so far as the peptone water was concerned, all trace of the vibrios had disappeared. This experiment, repeatedly performed, always gave the same result; it was impossible to secure a culture of the vibrio. Although starting a normal de- velopment, later, within a few hours, the vibrio had disappeared. This phenomenon remained unexplained until the findings with reference to the bacteriophage were published. In fact, it is certain that a large number of bacteriologists, indeed, it may be said all bacteriologists, — and reasons for this statement will appear in the course of this discussion, — have accidentally encountered this strange phenomenon. Seen at times in a fluid medium, at other times on a solid medium, such reactions have repeatedly been observed but their study has been neglected since their importance was not recognized. FUNDAMENTAL EXPERIMENT The experiment which served as a point of departure for sub- sequent work was as follows. An adult, suffering with a severe INTRODUCTION 19 dysentery (B. dysenteriae Shiga) was under treatment in the Pasteur Hospital. Each day an examination of the feces from this patient was made by the inoculation of bouillon with some of the fecal material. After incubation at 37°C. for over night the growth was filtered through a Chamberland filter. To a second tube of broth, previously inoculated with a culture of the Shiga bacillus, twelve drops of the filtrate were added and the culture so treated was returned to the incubator. Throughout the period of the infection, all of the tubes, prepared each day in the same manner, yielded normal growths of the dysentery bacillus. One day, examination of the tube prepared the day before showed no growth, and investigation showed that the patient presented symptoms indicative of marked improvement. Definite convalescence rapidly followed. The bouillon which had been inoculated with both culture and filtrate was to all appearances sterile, and to this was again added a suspension of Shiga bacilli taken from a young agar culture. The inoculation was sufficiently heavy to present a definite turbidity, but after incubation for twelve hours it was again clear. The excreta from which the filtrates were prepared contained, then, a prin- ciple which dissolved the dysentery organisms. When a drop of the lysed culture was added to a young bouillon culture of the Shiga bacilli this culture in turn became dissolved. In the same way, several successive passages were accomplished, introducing each time a drop of the culture previously lysed into a fresh culture of the Shiga strain. Instead of losing in potency through such passages it increased in lytic capacity; the dissolv- ing action being accomplished more and more rapidly. From this it was evident that the lytic principle derived from the ex- creta was capable of cultivation in series. When a very minute amount (0.00,001 cc.) of one of these lysed cultures was added to a young broth culture of the Shiga bacillus and then this mixture was tested immediately and again after incubation periods of one, two, and three hours, a drop of the material being inoculated on to agar slants, the latter showed after incubation very interesting characteristics. In the first tube thus inoculated the agar was covered by a normal film of dysentery bacilli, but with two circular areas about 2 mm. in 20 INTRODUCTION diameter entirely free of any evidence of bacterial growth. The second tube, inoculated with material taken one hour after the admixture of culture and lytic agent, presented six of the clear plaques. In the third tube there were about 100; and finally, on the fourth there was no apparent growth. Here was new evidence that the lytic principle actually multi- plied, and furthermore, that this principle actually existed in particulate form. The element from which the lytic phenomenon originated was composed of masses which were deposited upon the agar in definite points. Each mass was capable of multi- plication since, independent of the action in series, it yielded a colony. It could be considered as nothing other than a ferment or a living being parasitic on the bacteria. But it is impossible to comprehend a soluble ferment — a diastase — as multiplying in the form of granulations and concentrating its activities in limited, clearly defined points. As we will see in the course of this work all the experiments, without a single exception, are in accord in showing that the principle acts as a virus; and, indeed, as a virus which presents all the chief characteristics of organisms, including the fixation of complement with an antiserum. In employing the word "microbe," or better "ultramicrobe," following the happy expression of Calmette, I give to this word its true meaning: "minute living being," without any suggestion as to what kingdom it may belong. Is it a bacterium, a proto- zoon, or a yeast? That must be ignored. Its dimensions are too small to permit the determination of this question by direct microscopic observation. May it be a cell, infinitely small, an organite, derived from a superior organism; a cell indefinitely cultivable in series in vitro at the expense of bacteria and main- taining itself as an autonomous being? This is hardly probable, but it is never permissible to reject a priori any conception which accords with the known facts. Experiment has shown that this lytic principle, which has been termed Bacteriophagum intestinale or bacteriophage, is a particle which proliferates at the expense of bacteria; and, as a result, is capable of assimilation and is indefinitely cultivable in series in vitro in the form of a filtrable substance. It behaves like living matter because assimilation and reproduction are fundamental characteristics of life. INTRODUCTION 21 This introductory discussion should not be concluded without a word on the subject of the term "Bacteriophage," a term which has been criticized. The suffix "phage" is not used in its strict sense of "to eat," but in that of "developing at the expense of;" a sense that is frequently used elsewhere in scientific termi- nology. Certain protozoa, for example, are parasitized by the Nucleophaga which develop within the interior of the nucleus. This is precisely the interpretation to be given the term "phage" in the word "Bacteriophage." CHAPTER I BACTERIOLYSIS Bacteriolysis. Technic for Isolating the Bacteriophage. Enhancement of Virulence. Technic for Enumeration. Multiplication at the Expense of the Bacteria in a Fluid Medium. The Bacteriophage; an Obligatory Parasite. Effect of the Condition of the Bacterium. Effect of the Medium. Cultivation on Solid Media; Isolated Colonies. Effect of the Concentration of Bacteria in the Medium. Destructive Action of the Secretory Products of the Bacteriophage. Effect of External Physical Conditions. Effect of Antiseptics. The Soluble Bacterial Substance. The Bacteriophagous Ultramicrobe ; an Internal Parasite. Bacteriolysis under the Microscope. BACTERIOLYSIS It is expedient to define, at the beginning of this work, what is meant by the word "bacteriolysis" for it is a scientific term used in a somewhat equivocal manner. The term "autolysis" was introduced into science byJacoby as a substitute for the word "autophagy" which had previously been employed to designate the process of softening; the tendency toward a liquefaction, more or less marked, such as is produced by a yeast isolated upon a nutrient medium. The term autophagy, which presumed nothing as to the final condition of the process, is more suitable certainly, than that of autolysis. Etymologi- cally the latter signifies auto-dissolution, whereas as a matter of fact, the process of autolysis, as it occurs with bacteria and yeasts even if prolonged for several months, never results in a complete cellular dissolution. The end product is a semifluid mass, which, examined microscopically, shows cellular debris along with a greater or less number of cells more or less profoundly modified. The degree of disintegration depends somewhat upon the type: of bacterial cells employed. Autolysis, then, is characterized, not by an actual dissolution, but by a disintegration, a cellular fragmentation with a partial dissolution of certain elements. Even in the most favorable cases, when the autolysis is considered 23 24 THE BACTERIOPHAGE complete, it is attended by the formation of an amorphous mass insoluble in the fluid. If the significance of the term "lysis" is indeed exact in autolysis where there is a partial dissolution, it is by no means the same in " bacteriolysis" as this is understood by many authors. Treatises dealing with immunity speak of " bacteriolytic sera," realizing explicitly that the reactions be- tween antigen and antibody never consist of a digestion. How then, under these conditions, is it possible to have a dissolution? And, in fact, none is ever observed. The action manifested by the bacteriophage is wholly different. It comprises phenomena of which the final result is a digestion such as leads to a total dissolution of the bacterial bodies. It is a bacteriolysis in the true sense. At first I considered desig- nating this new phenomenon by a new word, " bacteriophagy" for example, since the word bacteriolysis is often employed to designate processes differing entirely from dissolution. But since the use of too frequent neologisms might distract the reader I have felt that the abuse which has been made of the term " bacteriolysis" may be only a transitory one and that the passage of time will soon cause to be forgotten the phenomenon of so- called bacteriolysis without a dissolution of the bacteria. To summarize: the phenomena which we will consider have nothing in common with that which is usually designated by the terms " lysis" and "bacteriolysis." Here, the term "lysis" should always be taken in its strict etymological sense of a com- plete dissolution. A bacterial culture in bouillon or a suspension of bacteria in a fluid where bacteriolysis, as we understand it, takes place, completely clears, without residue. The fluid be- comes as limpid as it was prior to its inoculation with culture. TEGHNIC FOR THE ISOLATION OF THE BACTERIOPHAGE All bacteriological laboratories possess the equipment required for the isolation and cultivation of the bacteriophage. For isola- tion a filtering apparatus is indispensable. (I have employed the model of Martin, with Chamberland L2 and L3 bougies.) For culture media a peptone bouillon, or such a bouillon incorporated into a 2 per cent agar, is adequate. BACTERIOLYSIS 25 The active bacteriophage may be sought for in materials which require some preliminary treatment, since in their natural physi- cal state they may not lend themselves readily to nitration. The following types of material may be examined as a source of the bacteriophage. 1. A sterile fluid; sterile in the sense in which the word1 is usually employed; for example, blood, or an organic fluid col- lected aseptically. With such no treatment is necessary. 2. The material to be examined may be a clear liquid but not sterile. With this, filtration will eliminate the bacteria while the bacteriophage passes through into the filtrate. 3. The material may show a homogeneous turbidity; as a bacterial culture. Here, direct filtration results in an early occlu- sion of the pores of the bougie. Thus, it is desirable to resort to a preliminary filtration. The following method of treatment is most satisfactory. Provide a funnel with a folded niter paper sufficiently large to receive at one time the entire volume to be filtered. Fill the filter with water to which has been added a small amount of infusorial earth. When the water has passed through, the paper is left coated with a thin layer of the infuso- rial earth, thus rendering the paper less permeable. Through this the material to be examined is filtered prior to filtration through the bougie. 4. The material may be a fluid holding in suspension organic particles, or it may be matter more or less solid in nature. This is the type of substance most frequently examined; such as fecal 1 In the course of this work I find myself frequently in difficulty in the exposition of facts because of certain expressions which have been appro- priated to describe certain conditions. I shall apply the word sterile to a medium which contains no visible microscopic organisms or organisms capable of cultivation upon artificial media. An ultrasterile medium is one which contains no ultramicrobes. A substance containing the virus of measles, for example, is sterile but not ultrasterile, since it is still capable of transmitting measles although it contains nothing visible or cultivable. Media containing the bacteriophage are likewise sterile in the bacteriologi- cal sense of the word, since they are perfectly limpid and since the germ which they contain can not be cultivated alone upon artificial media of any kind. But such a medium is not ultrasterile, for it does contain a principle which will grow at the expense of bacteria, just as the virus of measles will grow at the expense of higher organisms. 26 THE BACTERIOPHAGE material more or less fluid, pasty, or solid; or excreta admixed to a greater or less degree with earth, organic debris, etc. In such a case it is necessary to disintegrate as completely as possible the material to be examined. To effect such a disintegration the most simple procedure consists in care- fully suspending the material in bouillon, about 5 gm. to 50 cc. of the medium, and incubating this suspension at 37°C . for from twelve to eighteen hours. The bacterial fermentations which ensue, because of the diverse organisms introduced into the medium, lead to a sufficient disintegration. Upon removal from the incubator the material may be treated, as indicated above, by nitration through infusorial earth and a bougie. If the material under examination contains the bacteriophage and has been subjected to filtration, the ultramicrobe will be found in the filtrate. The methods of purification outlined above are applicable to two purposes, (A) the detection of a bacterio- phage active toward a given bacterial type, and (B) to test the activity of a bacteriophage, either upon diverse organisms or against a single bacterial strain of indeterminate type. A. The first case is the more simple and will be considered first, taking as an example the detection of a bacteriophage active against B. dysenteriae Shiga. The day before the test is to be made an agar slant is inoculated with the dysentery strain. From this fresh culture, on the day of the test, four tubes of peptone broth are inoculated.2 To the first of these tubes is added one drop of the filtrate, to the second, ten drops, and to the third, two cubic centimeters. One tube, simply inoculated with the dys- entery organism, serves as a control. The tubes are incubated at 37°C. After twelve to eighteen hours one of several results may be observed : the three tubes (in addition to the control tube) may all show a turbidity due to the growth of the dysentery organism, only one or two of the tubes may be turbid, or the three tubes may be clear. * As will be shown later this bouillon should be alkaline in reaction. The ordinary neutral bouillon (I have always used by preference the bouillon of Martin) to which is added 6 cc. of N/l NaOH per liter is per- fectly satisfactory. This is, moreover, the degree of alkalinity most frequently employed in bacteriological work. BACTERIOLYSIS 27 1. All three tubes are turbid. From such a result it may not be concluded that an active bacteriophage is not present, for if a complete lysis of the bacteria is taken as the only criterion for determining its presence the bacteriophage will, in a majority of cases, be overlooked. Lysis is but a single fact in the midst of a very complex group of phenomena. If the three tubes are turbid take about 0.02 cc. from each of the tubes by means of a platinum loop and spread over the surfaces of three tubes of slanted agar. If, after incubation, these tubes present normal cultures of the dysentery bacillus the result of the test is negative. That is, the original material did not contain an active bacterio- phage for B. dysenteriae Shiga. The presence of the bacteriophage in active form is indicated by an abnormal appearance of the growth as it develops on the agar. In accordance with the num- ber of bacteriophagous ultramicrobes present the aspect of the culture will vary. The layer of bacillary growth may show one, or several, circular areas where the surface of the agar appears devoid of growth. Or, the culture may appear broken up, or corroded, as a result of the confluence of the areas. Indeed, there may be only fragments of culture or even isolated colonies remaining. When the number of ultramicrobes is still greater the slant may be free of any evidence of bacterial growth. As will be shown, each strain of bacteriophage is endowed with an individual degree of virulence, the word " virulence" to be taken in its true meaning, that is, "ability to multiply at the expense of the parasitized being." Certain races of the bac- teriophage multiply rapidly, others increase but slowly. The first possess a high degree of virulence toward the bacterium provided for their development; the second possess but a feeble virulence. We will elsewhere return to this subject of the viru- lence of the bacteriophage. It is mentioned here simply to ex- plain the reason why strains of the bacteriophage show varia- bility in growth when isolation is attempted. The diameter of the clear areas, varying with the individual strain of bacterio- phage from a fraction of a millimeter (the bacterial growth appears as though sprinkled with pin point areas) up to 4 to 5 mm., gives a measure of virulence; the larger the area the higher the viru- lence. We shall see that whatever the virulence of a particular 28 THE BACTERIOPHAGE strain of the bacteriophage when it comes from the organism it can be enhanced in vitro. 2. The first or first and second tubes only give a culture of the dysentery bacillus. Here the filtrate contains a bacteriophage of average or high activity. It is only necessary to proceed as is indicated in the following case to secure areas of bacteriophagous growth on agar. 3. All three tubes remain clear. This indicates the presence of a bacteriophage of extremely high activity. Confirmation is simple. It consists in taking three tubes of bouillon, in adding to them a suspension of young bacilli taken from an agar slant in a concentration to give a slight turbidity. Introduce into each of these tubes a drop of the fluid from each of the three tubes which had remained clear, shake, and then immediately distribute a loopful of each upon the surface of an agar slant. Both sets of tubes are incubated. After twelve to eighteen hours the three bouillon tubes will be limpid; the three agar slants will present the appearance already described for cultures of B. dysenteriae admixed with the bacteriophage. B. dysenteriae Shiga has been taken as an example, although whatever may be the bacterial type against which an active bacteriophage is sought the technic for isolation remains essen- tially the same. The medium is inoculated with the bacterium in question, as, for example, with B. pestis if a bacteriophage active for the plague bacillus is sought. B. Instead of determining if a given material contains a bac- teriophage active for a certain bacterium, it may be desirable to ascertain if a bacteriophage which has been isolated possesses an activity for several bacterial types at the same time. In this instance three tubes with each of the bacterial types to be investigated may be prepared. Thus, to investigate the activity of an intestinal bacteriophage against B. dysenteriae Shiga, B. dysenteriae Flexner, B. dysenteriae Hiss, B. typhosus, and B. coli, five series of three tubes are prepared. The first set is inocu- lated with B. dysenteriae Shiga, the second with B. dysenteriae Flexner, the third with the Hiss strain, and the fourth and fifth sets with B. typhosus and B. coli respectively. To the first tube of each series one drop of the filtrate is added, to the second, ten BACTERIOLYSIS 29 drops, and to the third, two cubic centimeters. With each set the procedure is that already indicated for the Shiga organism. In routine work a single tube can be used in place of the three tubes, and to this ten drops of the filtrate is added, but with this simplified technic the danger lies in the fact that a bacteriophage of weak activity may not be detected.3 TECHNIC FOR ENHANCING VIRULENCE It has already been stated that a bacteriophage may be present even though it is unable to induce the slightest macroscopic evi- dence of the lysis of a bacterial suspension. Indeed, this is the situation most frequently encountered in the process of isolation However, it is, as a rule, easy to increase the activity of such a bacteriophage. One of the following methods suffices: 3 Mention may be made here of a method of Bordet and Ciuca, and it is upon this procedure, moreover, that these authors have based their theory of hereditary lysis. They inoculate a guinea pig intraperitoneally at three or four different times at intervals of a few days with a culture of B. coli. The day after the last injection, according to them, it is only necessary to wash out the peritoneal cavity, whereupon the principle giving rise to "hereditary lysis" is found in the exudate. From their first communication on the subject it is evident that they consider this observation a specific example of a general law — which indeed would be one of the sine qua non conditions for the validity of their theory — that an injection of any bac- terium causes the organism to respond with the production of a principle which gives birth to the phenomenon of lysis in series. They stated that they would shortly announce the results secured with diverse bacteria but this communication has never appeared . I have tried without success, as have several other investigators, to duplicate the results described by Bordet and Ciuca. In reality, in this experiment, there has been a passage of the anticoli bacteriophage, which, as experiment shows, is normally present in the guinea pig intestine, into the peritoneal cavity as a result of the irritation induced by the inoculations. After its appearance in the peritoneum it multiplies there by virtue of the B. coli inoculated. The correctness of this interpretation is vouched for by the fact that the results reported by Bordet and Ciuca can be secured if, a few hours before the intraperitoneal injection of bacterial culture, the bacteriophage active for this bacterial type is given the animal per os. The active bacteriophage found in the intestine passes through into the peritoneal cavity. The method of Bordet and Ciuca gives results only by accident, only when an active bacteriophage was previously present in the intestine. Thus, all of the conclusions of these authors fall ipso facto. We will return else- where to this question (Chapter VI, Nature of the Bacteriophage). 30 THE BACTERIOPHAGE When the agar inoculation has shown that a bouillon suspension contains an active bacteriophage this suspension is filtered through infusorial earth and then through a bougie. A slightly turbid suspension is prepared, using the bacterial strain against which the bacteriophage has shown some activity, and into this suspen- sion are introduced some four or five drops of the filtrate. After an incubation period of twenty-four hours at 37°C., if lysis has not been produced, this second bacterial suspension is filtered as before and a third suspension is inoculated with four or five drops of the filtrate. Such transfers are continued until evident lysis occurs. During the process it is easy to verify the presence of the bacteriophage in each passage, and to detect any increase in virulence, simply by spreading the successive cultures on agar slants. Comparison of the cultures secured with each passage reflects the degree of virulence. For example, the agar growth obtained from the first passage shows a culture growth with ten plaques, the second passage shows 100, with the third the layer of bacillary growth is broken up with an abundance of the areas, while with the fourth passage only a few isolated colonies of bacteria are seen. It can be readily seen that the virulence of the bacteriophage, that is, its ability to develop at the expense of the bacteria, increases with each transfer until a point is reached where lysis of the suspension is obtained. Successive transfers can be made upon agar slants, taking the material from a tube showing the clear areas. With a platinum wire material can be removed from the bacterial growth bor- dering on a plaque and inoculated on a sterile slant. A sec- ond, third, and fourth (or as many as may be desired) transfer from agar to agar can be made. When a condition is reached where the agar growth shows only fragments of bacterial culture the surface of this tube is carefully washed off and filtered through infusorial earth and a bougie. In the filtrate is found a bacterio- phage sufficiently active to produce lysis of a bouillon suspension. As we shall see, the bacteriophage is not destroyed at 65°C., that is, at a temperature above the thermal death point of most non-sporulating bacteria. Thus, instead of filtration the appli- cation of heat may be employed. Heating at 58 to 60°C. for thirty minutes will kill the bacteria and not harm the bacterio- BACTERIOLYSIS 31 phage. However, filtration has always appeared to give more satisfactory results. In a later chapter, under a paragraph en- titled "Multiple Cultures" a third procedure will be considered. In certain cases the virulence of the bacteriophage can be in- creased in vivo. A guinea pig is injected intraperitoneally with two cubic centimeters of the filtrate containing the bacteriophage whose virulence it is desired to increase and with a few cubic centimeters of the bacterial culture against which the bacterio- phage is active. After twelve to eighteen hours, ten cubic centi- meters of sterile bouillon is injected into the peritoneum and a few minutes later the peritoneal exudate is removed by puncture with a trocar. The fluid is collected in a few cubic centimeters of citrate solution and after a few hours' incubation the material is filtered (infusorial earth and bougie) . This filtrate frequently shows that a bacteriophage is present which is significantly more virulent than that which was introduced into the guinea pig. Usually it is relatively easy to increase the virulence of a weak strain of the bacteriophage, but at times it may become very difficult, particularly when working with strains active against the Gram-positive cocci. In such cases it is necessary to effect a great number of passages, and there is considerable risk of losing the bacteriophage altogether, particularly during the first few passages. I might cite as an example an anti-staphylococcic strain with which Eliava was forced to make passages during four months in order to obtain sufficient virulence to induce complete lysis of a suspension containing 500 million staphylo- cocci per cubic centimeter. ENUMERATION OF THE BACTERIOPHAGOUS ULTRAMICROBES A trace of a filtrate containing a bacteriophage very active for a given bacterium introduced into a broth suspension of this bacterium causes a lysis of the organisms there present within a few hours. The medium becomes as clear as broth which has never been inoculated. A trace of the lysed suspension intro- duced into a new suspension similar to the first causes a similar lysis, a trace of this second lysed suspension introduced into a third tube reproduces the same phenomenon, and so on. Dur- ing the past three years with a single strain daily passages have 32 THE BACTERIOPHAGE been made, sometimes two or three passages on a single day, introducing in each transfer about 0.001 cc. of the last tube lysed into a fresh suspension. After more than 1500 passages the lysed suspension of the last tube was as active, even more active, than the filtrate which served to start the phenomenon in the first tube of the series. A suspension once lysed does not contain any living bacteria. On the other hand, the amount of bacteriophagous ultrami- crobes introduced to start the lysis is increased, since the new lysate is as active as was the preceding one. The lysed suspen- sion has become what can be called, literally, a culture of the bacteriophage. It has been previously stated that the inoculation of agar with a bacterial suspension to which has been added a small amount of fluid containing the active principle gives a bacterial culture studded with clear areas. This observation suggests a means of determining with approximate exactness the phenomenon of multiplication of ultramicrobes, since each area undoubtedly indicates the point at which, during the inoculation, there was deposited an active element, — an ultramicrobe. It is only neces- sary to work with measured quantities to ascertain the number of active germs in a fluid. From an abundance of experiments a single one showing the method of counting is taken. To avoid repetition it may be stated that "suspension of B. dysenteriae Shiga" is to be under- stood. A series of tubes is prepared, once for all, by any standard pro- cedure, containing B. dysenteriae Shiga suspensions of the follow- ing counts: 100, 200, 250, 300, and 400 million bacilli per cubic centimeter. These suspensions are stabilized by the addition of a small amount of formol and the tubes are sealed with the blowpipe. The suspension to be subjected to lysis should be taken by preference from a young growth on agar. A concentrated sus- pension is prepared by adding one or two cubic centimeters of bouillon to the agar slant and allowing the tube to remain inclined for a few minutes in such a way that the whole bacterial growth is under the fluid. With shaking, a perfect suspension BACTERIOLYSIS 33 is secured. With a pipette a certain quantity of this concentrated suspension is added, drop by drop, into a tube of bouillon until the turbidity corresponds to that of the 250 million control sus- pension. This approximation is adequate for routine practice, giving a suspension which contains about 250 million bacilli per cubic centimeter. This is the suspension to be used. A young broth culture can be utilized, but the other suspension, more accurately adjusted, is to be preferred. Experiment /. To a tube containing 10 cc. of a suspension of B. dysen- teriae Shiga is added 0. 00,002 cc. of a culture of the bacteriophage ten days old, i.e., a Shiga suspension that has been lysed for ten days. The tube is shaken to ensure even distribution and with a tared platinum loop 0.01 cc. of the liquid is removed and spread as uniformly as possible, by rubbing, over the surface of an agar slant. This tube is incubated at 37°C. After eighteen hours it presents a Shiga culture studded with 51 plaques. The calculation is simple. The 10 cc. of suspension received 0.00,002 cc. of the bacteriophage culture, or 0.00,000,2 cc. for each cubic centimeter of medium. The amount of material taken for planting on agar was 0.01 cc. This contained, therefore, 0.00,000,002 cc. of the original bacteriophage culture. Upon agar this 0.01 cc. gave 51 clear areas, or 51 colonies, each one of which developed from a single bacteriophagous germ deposited upon the surface. Fifty-one germs for 0.00,000,002 cc. represent 2,550,000,000 germs per cubic centimeter. This is, therefore, the content of the culture of the bacteriophage which was used to inoculate the bacterial suspension. The technic for counting the ultramicroscopic bacteriophage hardly differs from that used in counting ordinary bacteria. With the latter isolated colonies are secured on a surface other- wise sterile. With the ultramicrobe, clear areas representing colonies are obtained superimposed upon a layer of bacterial growth. Counting can not be effected otherwise, since the bac- teriophagous ultramicrobe is only able to develop upon the bac- teria which constitute its culture medium. Each plaque represents a colony of the bacteriophage. This is indisputable, for if the centre of such an area is touched with the point of a drawn-out capillary pipette and this pipette is dropped into a suspension of dysentery bacilli, the culture, when 34 THE BACTERIOPHAGE planted upon agar, reveals characteristic clear plaques. If, in the bouillon, lysis is allowed to proceed for some hours, there are more plaques. The plaque, then, to all appearances sterile, is in reality a colony of the bacteriophage. MULTIPLICATION OF THE ULTRAMICBOSCOPIC BACTERIOPHAGE The multiplication of the bacteriophage in the course of its activity can be followed by the method of counting. With a definite suspension of Shiga bacilli, always 250 million per cubic centimeter, two extreme cases are very interesting: (1) that which occurs when a large number of ultramicrobes are introduced,4 and (2) what transpires when but few, or only a single one, is inoculated. 1. Mass inoculation. Ten cubic centimeters of a suspension of Shiga organisms are inoculated with 0.04 cc. of the culture of the bacteriophage. The culture of bacteriophage contains 3000 million germs per cubic centimeter. Observed macroscopically from time to time, it will be seen that the turbidity gradually increases up to about the third hour,5 and from that time the liquid clears little by little. Between the fourth and sixth hours from the time of the inoculation the suspension has become limpid. 4 As already stated, study of the bacteriophage is always the study of a complex problem, because it is essential to consider the mutual actions and reactions of three variable factors — medium, bacterium, and bacterio- phage. The complexity is rendered more difficult to express clearly because of the lack of suitable words. Here, for example, what shall we term the act of introducing a definite quantity of the bacteriophage into a bacterial culture? Obviously the term "inoculation" can be employed, but the medium has already been inoculated with the bacterium; thus an equivocal or ambiguous meaning may result . The proper term would be ' 'contamina- tion" but unfortunately this term has already been appropriated in bac- teriology as a synonym for "pollution". The term "inoculation", then, must be employed. A culture medium may be "seeded" with bacteria, and then "inoculated" with the bacteriophage. • Frequently it will be observed, and the conditions for this reaction are not exactly determined, that the dissolution of the bacteria is preceded by a very marked agglutination. This reaction is noted particularly when the bacterial culture is inoculated with a relatively large amount (twelve drops for example) of a bacteriophage of average virulence. BACTERIOLYSIS 35 If, immediately after the inoculation, and at regular intervals for six hours, a loopful of the liquid is planted on agar, all the tubes remain sterile. One would readily think that the absence of colonies of Shiga means that they have been killed with the first contact with the inoculated bacteriophage. But this is not the case for if instead of planting agar with the suspension as such, it is previously diluted to 1 : 1000 with bouillon, and then this dilution is planted on agar numerous colonies of B. dysenteriae develop. The bacilli have by no means been killed. If the transfer of the undiluted suspension to agar remains sterile it is simply because the agar has been planted simultaneously with many living Shiga bacilli and also with a large number of the ultramicrobes. Indeed, the bacilli have commenced to multiply in the nutritive medium but the bacteriophage has not been inactive. Finding the bacilli which they parasitize within reach, the ultramicrobes act upon them, reproduce, and inhibit the bacillary growth. In the other case, with the suspension diluted to 1 : 1000, the bacilli and the bacteriophagous germ, a thousand times less numerous, are separated by spaces sufficiently great so that immediate interaction is less readily accomplished. This allows the bacilli which are outside of the immediate neighborhood of an ultramicrobe to multiply and to form colonies. A second dilution (1:1000) of the suspension, made after the latter has been incubated for an hour, more often remains sterile when transferred to agar and only rare colonies develop. After two hours of incubation, cultures on agar always remain sterile. At this time all the bacilli contained in the suspension have been attacked and none of them are able to reproduce. 2. Inoculation with few anti-microbial elements. Six tubes of the Shiga bacillus suspension are inoculated with the bacteriopha- gous culture (containing 3000 million per cubic centimeter) in such a way that each tube receives one six-millionth of a cubic centimeter. When incubated, four give normal cultures of B. dysenteriae and all subcultures on agar give normal growths. They are therefore without interest to us. The other two, how- ever, which have each received probably one, certainly not more than two ultramicrobes, show the following picture: The sus- pension becomes more and more turbid. After two hours at 36 THE BACTERIOPHAGE 37°C. the opacity is about two times as great as at the beginning; after three hours, it is about two and one-half times as great; after four hours, about three times; and then it begins to diminish, so that after five hours the density is about twice as great as at the beginning of the incubation. This clearing continues gradu- ally, so that after fourteen hours the culture is almost entirely clear. If immediately after the inoculation with the bacterio- phage and then every thirty minutes, 0.02 cc. of each of these two suspensions is transferred to agar slants, these tubes will show, after incubation, the following: Plantings after one-half, one, one and one-half, and two hours yield normal growths of B. dysenteriae Shiga. After two and one- half hours the subcultures show three plaques in one tube, five in the other (average four). Therefore, after two and one-half hours the inoculated suspension contains 2000 bacteriophagous ultramicrobes. After three hours the tubes show five and four respectively. Hence, there has been no material increase between two and one-half and three hours. The three and one-half hour plantings show nine and five areas (average seven). The number of bacteriophagous ele- ments has slightly increased. After four hours, the agar tubes show 101 and 111 plaques re- spectively (average 105). After four hours, therefore, the num- ber of ultramicrobes is between fifty and sixty thousand. After four and one-half hours, the counts are 145 and 160 (average 152), indicating that the suspension contains 75,000; a number but slightly different from the count after four hours. After five hours, the agar tubes are sterile. When diluted to 1 : 1000 in a suspension of Shiga bacilli and transferred immediately to agar in the same way, the tubes give four and six areas. Thus, it appears that after five hours the suspension contains about 1,500,000 bacteriophagous germs. From this it is readily apparent that the multiplication of the ultramicrobes is extremely rapid, and, what is most remark- able, it is associated with successive jumps, each augmentation being separated by an interval of about seventy-five minutes. We will refer elsewhere to this experiment when we consider the mode of reproduction of the bacteriophage. BACTERIOLYSIS 37 This experiment shows that it is only necessary to have a single bacteriophage present in a bacterial suspension to produce a complete lysis of the bacteria, provided the strain of bacteriophage is of maximum activity. Needless to say, such experiments have been repeated many times, always with results comparable to those cited. Indeed, this statement holds for all of the experiments reported in this monograph — all have been repeated. The attention of investigators should be called to this particu- lar point, namely, that once a strain of bacteriophage of suffi- cient virulence has been obtained, the end results of the experi- ments are always the same, without exception — an increase in the number of ultramicrobes inoculated and a complete lysis of the bacterial suspension. Repeating the same experiment several times with the same strain of bacteriophage, employing always the same conditions of medium and temperature, the proliferation of the ultramicrobe progresses in the same manner and lysis is effected in the same length of time. But if one is making a comparative study of different strains, although all are endowed with high activity, certain differences are noted. With one strain complete lysis will be obtained after three and one-half hours (this is the shortest period thus far observed), with another only after fourteen hours, all the conditions being the same. In a word, and this observation likewise applies to all of the experiments here reported, the phenomenon always proceeds as has been indicated. The time alone may vary. The ultramicroscopic bacteriophage is a living being, and as such, the processes which it carries out can not go on with the regularity of a diastatic action. Experiment II. Here is, to cite an example, an experiment conducted with another strain of bacteriophage. It will be noted that lysis is effected much more quickly than in the instance given above. The suspension of Shiga bacilli is made in bouillon previously warmed to 38°C., and the suspension is inoculated with 0.00,01 cc. of the culture of bacteriophage. The macroscopic appearance showed that : — After two hours the suspension is three times as turbid as at first. After two and one-half hours it is about three and one-half times as turbid. After two and three-quarter hours it is about three times as turbid as at .first. After three hours the turbidity is hardly apparent. 38 THE BACTERIOPHAGE In this experiment the lysis was almost entirely accomplished within a space of fifteen minutes, that is, during the period of time between two and three-quarters and three hours after the inoculation. A single ultramicroscopic bacteriophage is, therefore, adequate to provoke lysis. If successive dilutions of a culture of the bac- teriophage are prepared, one drop of the culture into a tube of sterile bouillon, one drop of this first dilution into a second tube, a drop of the second into a third, and so on, and if into each of these dilutions a fixed quantity of a concentrated Shiga culture is introduced, lysis is secured in all tubes which have received at \ least one ultramicrobe. This is usually the first four tubes of the series. The remaining tubes will show a normal growth of the Shiga bacillus. Since we have been able to make counts of the bacteriophage and recognize the rapidity with which even a single ultramicrobe can proliferate and bring about lysis, these observa- tions are self-explanatory. Without this means of investigation one woulpl be liable to commit a serious error and to conclude that the sterile bouillon of the first four tubes had contained a culture of the bacteriophage other than that introduced in pre- paring the dilutions. Incidentally, this error has been committed by certain authors. In reality, while there has been a dilution, the diluted culture was active just so long as there was to be found a single ultramicrobe. Recognizing the number of ultramicrobes in a lysed suspension, which has, in effect, become a culture of the bacteriophage, and the value of the dilution, it can be mathematically determined whether a bacterial suspension inoculated with such a dilution will undergo lysis or not. This test has been performed by ex- periment more than a hundred times with very diverse strains of the bacteriophage. THE BACTERIOPHAGE: AN OBLIGATORY PARASITE Whatever may be the medium employed, in the absence of a bacterium for which the bacteriophage is active, multiplication of the ultramicrobes never takes place. And this remains true even if inoculated into a medium containing, instead of living bacteria, organisms that have been killed by any procedure what- BACTERIOLYSIS 39 soever. All experiments have been uniformly negative in attempt- ing to obtain multiplication of the bacteriophage by contact of the ultramicrobe with bacteria killed by age, by heat, by chloro- form, by thymol, by the essences of cinnamon and mustard, by alcohol, by bichloride of mercury, by phenol, by sulfuric and hydrochloric acids, and by iodine. With such suspensions no action whatever is secured; — no lysis and no culture of ultrami- crobes. The bacteriophage is an obligatory parasite, multiplying only at the expense of living bacteria. An experiment of the following nature is interesting in that it shows that the bacteriophage will attack only normal bacteria. The bacteria are suspended in a medium containing an antiseptic in a quantity so small that the bacteria will only be killed after a time exceeding that required for the lytic process of the bacte- riophage. In such a case the bacteria are not affected by the bacteriophage and the latter fail entirely to 'proliferate. The antiseptic selected was, by intention, one without action on the diastases, — sodium fluoride. In a one per cent solution of sodium fluoride in bouillon the Shiga bacilli are still cultivable after thirty-six hours, a time more than adequate for the bacteriophage to manifest its lytic activ- ity and to multiply. Furthermore, if transfers in series are made with such a suspension, inoculating the first tube with a drop of the bacteriophage culture; the second, after incubation for twenty- four hours, with a drop of the first; the third with a drop of the second, and so on, it will be found that the bacteriophage disap- pears with the transfer from the second to the third tubes of the series (this can be confirmed by placing a drop of each tube into a suspension of normal bacilli). Controlling this procedure with a second series, using pure bouillon or even sterile water, it will be found that here likewise the bacteriophage will not be present in the third or fourth tube. This is, as will be seen later, simply a result of dilution. The bacteriophagous germ, therefore, can not be cultivated in a suspension containing fluoride, in sterile bouillon, or in pure water. Moreover, it has been shown that it is solely because of the medium into which it is introduced that the bacillus is not sub- 40 THE BACTERIOPHAGE ject to attack. For after a twenty-four hour stay in fluoride bouillon a normal culture will be secured by transfer from this medium into ordinary bouillon, and this culture is normally lysed by the bacteriophage. This will be discussed later, accepting for the moment that here is a medium in which the Shiga organism remains alive for at least thirty-six hours, in which the bacteriophage likewise remains alive, which exerts no inhibitory action on the diastases, but in which the bacteriophage fails to multiply. THE EFFECT OF THE CONDITION OF THE BACTERIA These experiments have been conducted so as to determine the effect of the state of the bacterium upon the lytic process. Since lysis is the result of the multiplication of the ultramicrobes the lytic process will not be complete unless all of the bacteria present in the suspension are capable of being attacked. Instead of taking a suspension prepared from a young culture we may inoculate the bacteriophage into a fifteen-day old broth culture. A clearing of the medium, a partial lysis, results but a certain degree of turbidity remains. Nevertheless, it is possible to continue to use such a medium, making as many passages as may be desired. Some tube of the series when planted on agar or in bouillon will remain sterile, and a drop of this tube inoculated into a suspension of young bacilli will produce perfect lysis. In the old culture, then, the bacteriophage multiplies normally but does not produce lysis, or at least, the lysis is incomplete. What is the explanation of this reaction? To answer this it is sufficient to compare the results of counting the total number of bacilli existing in an old culture (this can be done by the method of counting cells) with the results secured by counting the viable organisms only (done by the plating method). For a confirma- tion of this type, a Shiga culture in Martin's bouillon is made, incubated for fourteen hours, and allowed to stand at laboratory temperature for fifteen days. The total count of bacillary bodies will be about 625 millions; that of the viable bacilli, that is, those capable of yielding colonies when transferred to agar, will be about two millions in each half cubic centimeter of culture. Now, as we have seen, the bacteriophage is only able to develop at the BACTERIOLYSIS 41 expense of living bacteria, these being the ones which are lysed. In the old suspension which we have mentioned in which there is only about one organism in three hundred which is capable of being dissolved, it can readily be comprehended that if lysis of a suspension be taken as a criterion, the bacteriophage appears to be without action in such a culture. It is useless to work with old cultures. In a broth culture of Shiga, after only twenty-four hours of incubation, as has been shown, about one-third of the organisms present are incapable of producing colonies when planted on agar. If, on the other hand, an agar slant culture is utilized, almost all of the bacteria are liv- ing after twenty-four hours at 37°C. A twenty-four hour bouillon culture will, then, remain slightly turbid when the lytic process is accomplished, while a suspension made in broth from a young agar culture containing the same number of bacteria will be perfectly limpid when the lysis is achieved. In this last case all of the bacteria were living and susceptible to the attack of the bacteriophage. It is for this reason that it is preferable to effect cultures of the bacteriophage in a suspension rather than directly into a bouillon culture. Certain bacteria give a homogeneous growth in a young cul- ture in bouillon but when taken from agar they can be suspended only with difficulty. B. pestis is such an organism. When work- ing with such bacteria it is preferable to have the bacteriophage act on a broth culture in the following manner. A bouillon tube is lightly inoculated with the bacterium. When the culture has clouded, the bacteriophage active for this bacterial strain is introduced and at the same time the culture is diluted with an equal volume of sterile medium. This dilution should be made before the bacteriophage has had time to multiply sufficiently to parasitize an appreciable number of bacteria. Thus, the bacterial culture at the time of " departure," that is, when the bacteriophagous organisms are sufficiently abundant, will consist almost entirely of young bacilli, readily subject to attack. It has been demonstrated that the products of bacterial growth as found in an old culture, products which, as is well-known, inhibit the development of bacteria (as in the so-called "vac- cinated" media) are without effect upon the lytic phenomenon. 42 THE BACTERIOPHAGE Experiment III. Cultures of B. dysenteriae Shiga, aged fifteen days and eighteen hours respectively, are centrifugalized. The sediment from the first culture is suspended in the supernatant fluid of the second, and the sediment of the second culture in the fluid of the first. Both suspensions thus formed are inoculated with a drop of a culture of the bacteriophage. The suspension consisting of "old" bacilli and "young" medium remains turbid ; that of "young" bacilli and "old" medium becomes perfectly limpid after seven hours. Thus, while the products of bacterial metabolism are not in- hibitory for the lytic process, the products of lysis, as we will see, exert quite a different action. These products are the result of the activity of the ultramicroscopic bacteriophage, and as such, they impede its activity. THE INFLUENCE OF THE MEDIUM It is evident from the foregoing experiments that the true cul- ture medium of the bacteriophage is the living bacterium. The nature of the fluid in which the bacteria are suspended is without direct influence upon the culture of the bacteriophage, provided only the bacteria remain living in it throughout a sufficient period of time and provided it does not alter the constitution of the bacterial cell. Experiment confirms this statement. The only additional condition regarding the medium is that it be alkaline in reaction. Lysis will not take place in a medium of acid reaction.6 Experiment IV. Peptone water (containing 25 grams of Chassaing pep- tone and 5 grams of NaCl per liter) is neutralized to phenolphthalein. The medium is then frankly alkaline to litmus. It is then distributed into tubes, 10 cc. to each. Hydrochloric acid is added to each tube in dilutions to form an increasing scale of acidity. All of the tubes are inoculated with a concentrated suspension of Shiga, sufficient to yield a normal suspension of 250 millions per cubic centimeter. Finally, each tube is inoculated with 0.001 cc. of a culture of the bacteriophage. After twenty-four hours the numbers of bacteriophage in the several tubes are determined by the method previously described. 6 Certain commercial peptones contain glucose in appreciable quantity, hence their use may be attended by failure. BACTERIOLYSIS 43 TUBE REACTION TO PHENOLPHTHALEIN APPEARANCE OF THE SUSPENSION AFTER TWENTY-FOUR HOURS NUMBER OP ULTRA- MICROBES PER CUBIC CENTIMETER 1 0 Very slight clouding 400, 000, 000 2 - 2 Very slight clouding 500, 000, 000 3 - 4 Limpid 500, 000, 000 4 - 6 Limpid 1,250,000,000 5 - 8 Limpid 2,750,000,000 6 -10 Limpid 1,000,000,000 7 -12 Limpid 1, 000, 000, 000 8 -14 Slight turbidity 250, 000, 000 9 -16 Turbid 500, 000 10 -18 Turbid 1,000,000 11 -20 Turbid 500, 000 12 -22 Turbid None Neutrality to litmus corresponds to about — 16, thus it may be concluded that the bacteriophage ceases to grow when the medium presents even the slightest acidity. The ultramicrobial elements detected in the slightly acid tubes — those where lysis is not produced — is not an indication of multiplication. They are simply the ultramicrobes which were inoculated, still living. When the medium is decidedly acid even these elements are destroyed. The following experiment demonstrates this still better and further confirms the fact that the bacteria constitute the only culture medium for the bacteriophage.7 7 As will be shown in a later chapter, bacteria in general are destroyed after exposure for a very short time in physiological saline. On the other hand, the different strains of dysentery bacilli present a variable resistance. Certain races are no longer cultivable in bouillon after an exposure of five to six hours, others remain cultivable up to 45 to 50 days. Experiments involving lysis in physiological saline ought to be performed with strains of the last type. Moreover, it is necessary to employ an extremely active bacteriophage, capable of exerting its action in the minimum length of time — if possible, one capable of producing complete lysis in about three hours. A saline suspension of a bacterial strain of weak vitality may be already sterile after four to five hours, that is to say, the great majority of the bacteria are dead before the lytic process is able to be effected . And the bacteriophage is without action on dead bacteria. 44 THE BACTEBIOPHAGE Experiment V. One hundred cubic centimeters of a suspension of young Shiga bacilli are prepared in 0.85 per cent saline, neutral to litmus. This suspension is inoculated with five drops of an earlier saline culture of the bacteriophage, and the material is divided into five portions, 20 cc. in each tube. The first of these remains as already prepared. The second is ren- dered alkaline by the addition of NaOH in a quantity sufficient to give an alkalinity equivalent to 10 mgm. of NaOH per liter. The remaining tubes are also rendered alkaline, the third having a reaction equal to 25 mgm. of NaOH per liter, the fourth equal to 50 mgni., and the fifth equal to 100 mgm. After incubation for eighteen hours the first tube shows its original tur- bidity, tube 2 is cloudy, and tubes 3, 4, and 5 are almost perfectly clear. In these the turbidity is such that it can just be detected and is due to the fact that a certain number of the bacilli had died in the saline before they were attacked. When the lysis is completed in the last three tubes, to tube 1, which maintained its original turbidity, NaOH is added in an amount equal to 100 mgm. per liter. Twelve hours later lysis has taken place; the suspension has been transformed so that it possesses a transparency comparable to tube 5. It has been demonstrated that this degree of alkalinity by itself is without effect on the Shiga bacilli, since they develop normally in bouillon containing as much as 500 mgm. of soda per liter. From the cleared tubes all subcultures remain sterile. The bacteriophage can be indefinitely cultured in series in a slightly alkaline saline solution. Indeed, the salt CNaCl) itself can be dispensed with. Serial cultures have been maintained in chemically pure water containing only 25 mgm. of soda per liter. Certain experiments with a synthetic medium, capable of main- taining a culture of the Shiga bacillus, were performed and are not without interest. Growth of the dysentery bacilli and com- plete lysis by the bacteriophage are secured in a medium of the following composition: water, 80 cc.; sodium chloride, 1 cc.; potassium phosphate, 1 cc.; sodium phosphate, 1 cc.; and aspara- gine, 3 cc., all in 10 per cent solutions. The medium is rendered alkaline in accordance with the nature of the experiment to be performed. In bouillon a relatively high alkalinity does not interfere with lysis. Experiment VI. Five tubes, each containing 10 cc. of bouillon ren- dered alkaline to— 8 are seeded with a suspension of the Shiga bacillus and then inoculated with 0.02 cc. of a culture of the bacteriophage. A N/10 solution of NaOH is added; to the second tube 0.5 cc., to the third 1 cc., to the fourth 1 .5 cc., and to the last 2 cc. The first four tubes are perfectly lysed after eighteen hours, only the fifth remains clouded. BACTERIOLYSIS 45 It may be affirmed, therefore, that, aside from the matter of alkalinity, the composition of the medium with respect to its nutritive properties, exerts no influence on the development of the bacteriophage. From the moment when it has at its disposi- tion living and normal bacterial cells, against which it is active, it multiplies — at the expense of these bacteria which constitute its sole culture medium. CULTURE OF THE BACTERIOPHAGE ON SOLID MEDIA! ISOLATED COLONIES It has been stated that the bacteriophage shows the formation of obvious colonies on agar, and that in order to obtain them it is only necessary to inoculate a broth suspension of the Shiga organism very lightly with a culture of the bacteriophage and to distribute a drop of this suspension on agar. After incubation the covering of bacillary growth presents a number of areas free of all apparent culture. If the inoculation of the bacteriophage has been massive the surface of the agar appears sterile. Let us consider the characters of these cultures. When the surface of the agar remains bare because of the large number of bacteriophagous organisms and maintains this appear- ance indefinitely it has become unsuited for the cultivation of the Shiga bacillus. When inoculated at such a time with a cul- ture of this bacillus, even in a very abundant sowing, the slightest development cannot be detected. The medium is, however, normal for another bacterium. If inoculated with the cholera vibrio, for example, the growth will be as luxuriant as if planted upon fresh medium. Hence, if B. dysenteriae Shiga does not grow it is only because the bacteriophagous organisms remain on the surface of the agar and exercise their dissolving action on the bacteria deposited thereon. This is readily confirmed. If we take a tube of agar which has remained apparently sterile after having been inoculated with a suspension of the bacteria containing a bacteriophagous culture, and if the surface of the medium in such a tube is washed with a few drops of sterile bouillon and to this is added a fresh suspension of bacteria, this suspension will be lysed within a few hours. 46 THE BACTERIOPHAGE It sometimes happens, especially when using agar somewhat dried out, that a few colonies of Shiga are obtained, always lo- cated at the extreme edge of the layer of agar. We will return to this extremely interesting particular in the discussion of sec- ondary cultures. If, instead of a continuous covering of the bacteriophagous growth the ultramicrobes are deposited in limited areas — and this is readily accomplished by placing drops of culture on the sterile surface of a tube of agar, or again, by drawing lines over the surface with a platinum loop dipped in the culture of bacterio- phage, and after the tubes have remained inclined for a few hours in the incubator to secure drying — we find that the areas impreg- nated with the bacteriophagous culture remain free of Shiga bacilli, but that these organisms grow, on the contrary, perfectly well on the parts not covered by the bacteriophage. When in the suspension planted upon agar the number of bacilli is infinitely great and the number of the ultramicrobes is sufficiently small, the bacteriophage culture as individual units is distributed over the surface of the agar, and under such cir- cumstances the bacterial layer will appear studded with apparently sterile areas. These areas, or plaques, have a circular form with a diameter of from 1 to 5 mm. The plaques are in general of the greatest extent when the suspension is somewhat weak although sufficiently concentrated to give a continuous layer of growth rather than isolated colonies. On such a tube the areas are larger as the subjacent medium becomes thicker, that is, toward the bottom of the tube. Upon a Petri dish, where the agar layer is of essentially the same thickness throughout, all of the plaques of a given culture are of approximately the same diameter. As will be seen, the area of the plaque bears a relationship to the virulence of the bacteriophage which causes it. If a tube or plate presenting plaques is held in the incubator at 37°C., or at an entirely different temperature, no change occurs in the plaques; their diameter remains indefinitely what it was at first. They are never covered or encroached upon by the bacterial culture. At no time does there exist within the extent of the plaque, whatever its size may be, microscopically visible bacterial cells. The plaque is always rigorously sterile. BACTERIOLYSIS 47 As soon as the culture is well developed, as after 18 to 24 hours of incubation, if .the centre of such a plaque is touched with a platinum wire and this is immersed in a culture of Shiga bacilli the bacteriophage develops in this suspension and the latter is lysed after a few hours. The plaque, although sterile, is not ultrasterile; it is in fact a colony of the bacteriophage. Furthermore, if a trace of the bacillary growth at the periphery of a plaque is taken with a platinum wire and seeded on agar it remains sterile and inoculation into a bacterial culture shows that the bacteriophage is present there also. But when the bacillary layer is taken, not at the immediate edge of the area, but at a distance of two millimeters from it, for example, and planted, the tubes show the growth of a normal culture. The bacteriophage is not found. If the culture showing the plaques is returned to the incubator and the tests are repeated three or four days later, that is, culturing the bacillary growth at a distance of two millimeters from a plaque onto agar and into a suspension it will be found that the bacteriophage is there present at that time. The bacteriophage has, therefore, gradually invaded the bacillary layer. This invasion is always slow — proceeding more and more slowly as time pro- gresses— so that the ring invaded, even after several months, amounts to a zone but a few millimeters wide. Beyond the limits of this zone the Shiga organisms remain cultivable just as long as they do in a normal control culture without the bacteriophage. The question immediately arises as to why the bacteriophage does not invade the entire layer of bacterial growth. For this there are two reasons. The bacteriophage attacks the bacterial cell most readily when the bacterium is young. When placed upon agar the bacteriophagous organisms find themselves located in the immediate vicinity of bacilli which reproduce actively as soon as they are deposited upon a nutrient medium. They find then, within their range, very young bacilli distributed in a very thin layer over the agar. Lysis is thus possible and the apparent sterility of the plaque results. But beyond this zone invaded by the bacteriophage during the first few hours the bacilli develop freely forming a layer of increasing thickness comprised of or- ganisms of increasing age. In other words, a thicker and thicker 48 THE BACTERIOPHAGE layer of bacilli always becoming more and more resistant to lysis develops. This can be readily demonstrated by direct experi- mental proof. If the agar surface in a Petri dish is heavily seeded with a Shiga culture and at some point on this a drop of the culture of the bacteriophage is placed, and after a three-hour incubation period another drop of the bacteriophage is placed on the surface and this same process repeated after six, twelve and twenty hours, with continuous incubation of the plate during the intervals, it will be found fifteen hours later that the areas upon which the first three drops were placed have remained sterile — no bacillary growth has taken place. At the point where the fourth drop was placed, that is, after the culture had been incubated for twelve hours, there is a thin layer of growth composed of dead bacilli. The area where the drop of bacteriophage was placed after twenty hours presents an appearance practically normal. These five spots, then, represent the diverse aspects of an isolated colony of the bacteriophage, as from the centre to the periphery. The second reason is of a more general nature, representing a phenomenon common to the majority of cultivable organisms. The colonies of the bacteriophage act absolutely like colonies of those bacteria which, except for organisms such as B. proteus, never progressively invade the surface of solid media. Thus, if the Shiga bacillus is inoculated upon agar in an amount suitable to yield isolated colonies, after 18 to 24 hours, each colony will be from two to four millimeters in diameter, the largest colonies to be found at the points where the medium has the greatest depth, that is, toward the bottom of the tube. Such colonies increase in size but very slowly, always more and more slowly as time progresses, and even after two months, the zone of increase will not be greater than a few millimeters. From the bacteriologi- cal point of view it is not peculiar, as has been suggested, that the bacteriophage does not invade the entire bacterial layer. It must be borne in mind that the bacteriophage, far from being dissimilar to other cultivable organisms, behaves, when in iso- lated colonies, exactly like a colony of bacteria. Why does not the bacterial colony continue to increase in size and to invade the entire surface of the medium? Because the BACTERIOLYSIS 49 soluble substances resulting from the vital activity of the bac- teria diffuse into the agar and these substances constitute an actual specific antiseptic which limits the culture. The medium is " vaccinated" around the colony. The deeper the agar layer, or the farther the colonies are separated, the greater the volume of the substratum capable of diluting this antiseptic substance, and the larger will be the colony. The situation is precisely the same with the bacteriophage; the more scattered the colonies and the deeper the substratum, the greater the diameter. Direct ex- perimentation proves the correctness of this interpretation, and that the soluble substances elaborated during the lytic process, — substances resulting from the vital activity of the bacteriophage, — inhibit the vital processes, delay growth, and prevent the ac- complishment of total lysis. EFFECT OF THE CONCENTRATION OF BACTERIA IN THE MEDIUM; INHIBITORY EFFECTS OF THE PRODUCTS OF LYSIS In all of the experiments involving the action of the bacterio- phage in a liquid medium which we have up to the present con- sidered the bacterial suspension has contained approximately 250,000,000 organisms per cubic centimeter. What occurs if the concentration of suspended bacilli is varied between 50 and 500 millions? The end result will always be the same — a complete lysis — and even the time required for this lysis will not greatly vary for the inoculation of a given quantity of the bacteriophagous culture, such is the rapidity of develop- ment of these organisms. Let us consider the two extreme cases. Experiment VII (A.) The inoculation with the bacteriophage is massive, that is, 0.02 cc. In such a case the difference in time is most marked. Number of bacilli Lysis is complete per cc. in 50,000,000 4| hours 100,000,000 4£ hours 200,000,000 5 hours 250,000,000 5 hours 300,000,000 5£ hours 400,000,000 6 hours 500,000,000 8 hours (B.) The inoculation with the bacteriophage is very weak, in the follow- ing experiment, 0.00,000,1 cc. 50 THE BACTERIOPHAGE Number of bacilli Lysis is complete per cc. in 50,000,000 14* hours 100,000,000 14* hours 200,000,000 14* hours 250,000,000 14* hours 300,000,000 16 hours 400,000,000 16 hours 500,000,000 18 hours These results are readily understood. In the first case the number of ultramicrobes is very great from the start, all of the bacteria, or by far the greater part of them, are immediately attacked, and this quickly arrests the development of the bac- terial culture. In the second case, as a result of the small num- ber of ultramicrobes inoculated, very few of the bacteria are at once attacked, and those which remain unharmed are free to develop, so much the more as the suspension is the more dilute. To be convinced of this it is only necessary to observe the sus- pensions and to compare their relative opacity from time to time. All become more and more turbid during the first few hours after the inoculation, and in five or six hours after the inoculation they all present a comparable opacity, corresponding to approximately 650,000,000 bacilli per cubic centimeter. In a word, whatever may be the original titre of the suspension at the time when it is inoculated with a limited number of the bacteriophagous or- ganisms, the latter must always operate on a suspension of about 650,000,000 bacilli per cubic centimeter, since in all cases the bacteria reproduce until they attain this number. Hence, lysis will always take place within very nearly the same length of time. A suspension of young bacilli, containing about 500,000,000 bacilli per cubic centimeter is completely lysed by the action of a bacteriophage of maximum activity. Beyond this figure the medium never entirely clears, and remains the more cloudy the more concentrated the suspension, regardless of the number of ultramicrobes inoculated. A suspension containing 1,000,000,000 bacilli per cubic centimeter, for example, will never be completely lysed, whether it is inoculated with a few individual bacteriophag- ous organisms or whether it is inoculated with some thousands of millions. However, in working with suspensions which are ex- BACTERIOLYSIS 51 tremely heavy it is found that the bouillon transplants made on to agar after eighteen to twenty-four hours are always sterile. The bacilli have been killed but not completely lysed. Experiment VIII SUSPENSION OF BACILLI (MILLIONS PER CUBIC CENTIMETER) INOCULATED WITH CULTURE OF BACTERIOPHAGE ASPECT OF SUSPENSION AFTER TWENTY-FOUR HOURS CC. 5,000 0.1 Turbid 2,000 0.1 Cloudy 1,000 0.1 Slightly cloudy 500 0.1 Clear After incubation for 8 hours all cultures appeared the same. The inhibitory force which interferes with lysis is due to the accumulation of the soluble products resulting from the lytic process, that is, to the activity of the ultramicroscopic bacterio- phage itself. In this respect the action of the bacteriophage is in accord with a phenomenon common to all cultivable micro- organisms. Experiment IX. A bouillon suspension containing 250,000,000 bacilli per cubic centimeter is inoculated with 0 . 001 cc . of a culture of the bacterio- phage. The next morning, or after fourteen hours, lysis is complete. A count of the bacteriophage shows that there are 1,600,000,000 per cubic centimeter. At this time a concentrated bacterial suspension is added to the lysed suspension in such concentration that the titre amounts to 250,000,000 per cubic centimeter. Seven hours later the medium is again limpid, and a count shows the presence of 2,100,000,000 ultramicrobes. This second lysis completed, the bacterial content is again restored. This time lysis is hardly accomplished in 48 hours, indeed, at this time the bouillon is not quite clear. A count gives 2,400,000,000 ultramicrobes per cubic centimeter. At this time, then, the medium contains in each cubic centimeter the dissolved substance of 750,000,000 bacilli. The suspen- sion is made up to a concentration of 250,000,000 once more (for the fourth time). Eight hours later the culture has cleared somewhat but remains decidedly cloudy. The count shows 2,600,000,000 ultramicrobes. Inocula- tions upon agar and into broth remain sterile. It is plainly to be seen, therefore, that the more concentrated the medium becomes in dissolved substances the more marked becomes the inhibition and the less active the culture of the bacteriophage. 52 THE BACTERIOPHAGE The quantity of the bacteriophagous germs inoculated into the suspension is without influence on the final result. Experiment X. The following tubes containing suspensions of the Shiga bacillus are prepared. Tubes 1 and 5 contain 250 million per cubic centimeter Tubes 2 and 6 contain 500 million per cubic centimeter Tubes 3 and 7 contain 1000 million per cubic centimeter Tubes 4 and 8 contain 2000 million per cubic centimeter Tubes 1, 2, 3, and 4 are inoculated with 0.001 cc. of a culture of the bacte- riophage. Tubes 5, 6, 7, and 8 are inoculated with 0. 5 cc. of the same culture of bacteriophage, or, 500 times as much as in the first set. After eight hours tubes 1 and 5 are limpid. After fourteen hours tubes 1, 2, 5, and 6 are limpid. After four days tubes 1, 2, 5, and 6 are still limpid, tube 3 is very slightly cloudy, tubes 4 and 7 are cloudy, and tube 8 is turbid. It is thus apparent that lysis is not affected by the number of the ultramicrobes inoculated. We will see in a subsequent chapter which treats of the resistance of the bacteria to the bacteriophage, that contrary to what one would a priori suppose, lysis of a culture is even more perfect when the amount of the bacteriophage added to the suspension is rather small. Quite aside from the quantitative relationships, a suspension may vary in " quality.'7 One may work with bacteria of different ages, or with organisms of different strains. In so far as differ- ence in strains is concerned, the course of the phenomenon remains essentially the same — at least in so far as the Shiga bacillus is concerned. We shall see that it is not the same with certain other bacteria, B. typhosus for example. With reference to the question of the age of the bacilli subjected to the action of the bacteriophage we have already seen that the younger the bacillus the more readily it is attacked. This dif- ference is due solely to the state of the bacillus itself and not to the soluble substances resulting from its activity; — such substances as result in a " vaccination" of the medium, to use a common ex- pression. The bacteria vaccinate the medium for themselves through the products of their activity. The bacteriophage does the same thing. But the soluble products resulting from their respective activities have nothing in common. BACTERIOLYSIS 53 EFFECT OF EXTERNAL PHYSICAL CONDITIONS The presence or absence of oxygen has no effect upon the course of the phenomenon. The rapidity of multiplication of the ultra- microbes and the duration of the lytic process are the same in aerobiosis as in anaerobiosis. On the other hand, as would naturally be expected, the effect of the temperature is marked. Experiment XI. Three tubes containing a suspension of the Shiga bacil- lus are inoculated, each tube receiving 0.00,000,01 cc. of the bacteriophage culture. These tubes are placed at different temperatures; the first at 8°, the second at 22°, and the third at 37°C. The suspension held at 8°C. shows no lysis after twenty-four hours, but after 15 to 16 days lysis is complete. The number of ultramicrobial ele- ments at this time is 180,000,000 per cc. as compared with 20 to 25 per cc. when they were introduced. The suspension kept at 22°C. shows that multiplication commences after three hours. At this time the count is 75 ultramicrobes per cubic centi- meter. After five hours the count is 8,000, after eight hours 190,000, and at twenty-five hours lysis is complete and the count is 780,000,000 per cubic centimeter. The suspension held et 37°C. shows 210 ultramicrobes after two hours, 10,000 after three and one-half hours, 200,000 after five hours, and 1,700,000,000 per cubic centimeter after thirteen hours, with a complete lysis. Between 37 and 41°C. the course of the reaction does not show appreciable variation. Between 41 and 44°C. the lysis is less and less complete. For this there are two reasons; the develop- ment of the ultramicrobes is less and less active, and the number of bacilli killed in consequence of the elevation of temperature is greater and greater. As a result the number of organisms ca- pable of being attacked and dissolved are less and less numerous. However, serial cultivation of the bacteriophage at 44°C. is still possible. The ultramicrobe can be cultivated at temperatures higher than those supported by B. dysenteriae. With the latter growth ceases at 43°C. In effect, the optimum temperature for the bacteriophage is the same as that for the bacterium, as is but logical, since all experimental work demonstrates that the more closely the bac- terial cell approaches normal so much the better is it attacked. 54 THE BACTERIOPHAGE EFFECT OF ANTISEPTICS UPON LYSIS To complete the macroscopic study of the phenomenon it may be well to consider what effect different substances that may be added to the suspension may have upon the lytic process. As we will see with reference to the properties of the bacterio- phage, although it does not present a resistance as marked as some of the ultramicrobes to chemical and physical destructive agents, it is, nevertheless, less susceptible than the majority of cultivable organisms. From the particular point of view of lysis, we must recall that the action of antiseptics is complex. The bacteriophage is only able to grow at the expense of living bacteria and all action ex- erted on the bacteria of a suspension are reflected in the phenom- enon of lysis, even if these actions are weak or wholly lacking upon the bacteriophage itself. The special resistance of the bacteriophage to antiseptics does not modify the lytic process. Antiseptic substances may be introduced into the suspension in amounts sufficiently weak to render their effect on the bac- teria negligible and thus fail to alter the course of the phenomenon. If, on the contrary, the amount of antiseptic is such that it is capable of recognition, the bacteriophage may be unable to multi- ply for lack of normal bacteria and lysis is prevented. In this last case it is easy to see that the conditions of the experiment are the same as if the bacteriophage was placed in the presence of bacteria previously modified by the antiseptic in question. And it has already been shown that under these conditions neither the growth of the bacteriophage nor the lysis resulting therefrom is accomplished. The experiment previously presented showed that the bac- teriophage failed to multiply in bouillon suspensions of Shiga con- taining one per cent of sodium fluoride, even though the bacteria remained alive during a period of time amply sufficient for com- plete lysis to be effected in the presence of normal bacteria. Glycerine acts in a different manner. In high concentrations this substance prevents the growth of the bacteria, but its activity is inhibitory rather than strictly antiseptic. BACTEKIOLYSIS 55 Experiment XII. Tubes of bouillon, containing glycerine in the fol- lowing concentrations, are seeded with a drop of an eighteen-hour bouillon culture of B. dysenteriae Shiga. The results secured with the different con- centrations are: Bouillon + 5 per cent of glycerine : Very abundant growth Bouillon + 10 per cent of glycerine : Weak growth Bouillon + 15 per cent of glycerine : Very slight growth, with sedi- ment Bouillon + 20 per cent of glycerine : Clear medium, with abundant sediment of bacteria Bouillon + 25 per cent of glycerine : Clear medium, slight sediment Bouillon + 30 per cent of glycerine : Clear medium, slight sediment Bouillon + 35 per cent of glycerine : Clear medium, trace of sediment Bouillon + 40 per cent of glycerine : No growth whatever All subcultures made from these tubes at the end of forty-eight hours give, in normal bouillon, normal growths. B. typhosus is somewhat more sensitive to the action of glycerine. Even in a medium containing 10 per cent the growth is insignificant. Bacteria suspended in glycerine bouillon, even in a concentra- tion of 25 per cent, remain alive for at least forty-eight hours; that is, throughout a time amply sufficient for the bacteriophage to develop and to effect lysis, as was the case with the fluoride medium. But the following experiments show that the culture of the bacteriophage in the glycerine medium is absolutely nor- mal while it is entirely lacking in the fluoride medium. I emphasize the fact of the growth of the bacteriophage and the lysis which is the result of this growth in a glycerine medium, for it will be necessary to return to these experiments when we review the various proofs concerning the living nature of the bacteriophage. Experiment XIII. Tubel. Prepare a suspension of B. dysenteriae Shiga, 250,000,000 per cubic centimeter, in bouillon containing 35 per cent of glycerine. Inoculate with 0.02 cc. of the bacteriophage culture. Nor- mal lysis occurs in eight hours. A control suspension of the Shiga bacilli in the glycerinated medium, but without the bacteriophage, yields positive subcultures up to the seventh day. Tube 2. Shiga bacilli, 250,000,000 per cubic centimeter, are suspended in bouillon with 50 per cent glycerine. The medium is inoculated with 0. 02 cc. of the bacteriophage. Normal lysis takes place in ten hours. The control suspension, without the bacteriophage, is cultivable up to the forty-eighth hour. 56 THE BACTERIOPHAGE Tube 3. The Shiga bacilli, in the same concentration, are suspended in bouillon with 60 per cent of glycerine. Portions of this suspension are inoculated with various amounts of the bacteriophage culture, as follows: a. With 0.5 cc. of the bacteriophage culture. With this amount lysis is complete in eight hours. 6. With 0. 02 cc. of the bacteriophage culture. Complete lysis is obtained in nine hours. c. With 0.0001 cc. of the bacteriophage culture. No lysis results. The ultramicrobes inoculated, too few in number, have not had time to develop before the bacilli are dead. A control suspension, without the bacterio- phage, gave positive cultures only up to the 18th hour. From this it appears that in glycerine broth the bacilli remain normally susceptible to attack by the bacteriophage just as long as they are living. Although the bacteria die when suspended in the glycerine medium it can not be assumed that the glycerine acts as a true antiseptic, that is, that it modifies the bacterial protoplasm. No one would contend that sodium chloride in weak concentration is an antiseptic despite the fact that non-spore-forming bacteria suspended in normal saline survive for but a very short time: twenty-four to forty-eight hours in the case of the Shiga bacillus. The experiments on lysis in the glycerinated media are, more- over, of great interest in that glycerine, in very high concentra- tions, sterilizes cultures of the bacteriophage. Substances without action on the bacterial cells are, in general, without influence on lysis. This, for example, is the case with normal serum, ascitic fluid, urine, and 2.5 per cent sodium chloride. Calcium chloride, on the other hand, has a very marked inhibi- tory effect, and potassium chlorate delays lysis. In low concen- trations magnesium sulfate and the phosphates of sodium and potassium favor lysis. This is particularly observed in the case of strains of the bacteriophage of feeble activity. When sugars which are not fermented by the bacterium against which the bacteriophage is active are added to the suspension lysis is not modified. The addition of fermentable sugars is without effect if the inoculation of the bacteriophage is massive, but if the inoculation is weak lysis does not occur or remains incomplete, depending upon the amount inoculated. BACTERIOLYSIS 57 The cause for these results is very obvious. The bacteriophage is very sensitive to acidity and with a minimal inoculation the bacteria begin to develop, to attack the sugar, and to render the medium acid before the ultramicrobes are present in ade- quate numbers to effect lysis in the time at their disposal. SOLUBLE SUBSTANCES OF THE BACTERIA Lysis of the bacteria by the bacteriophage is complete, with- out residue. In the last analysis this lysis can only be in the nature of a diastatic action. No bacterial activity is caused simply by the presence of the organisms as such, but rather by virtue of the secretory products which they elaborate. The ul- tramicrobes necessarily secrete some of these lytic diastases — the lysins — which liquefy the substances constituting the bacterial body. This point will be considered further. The chemical aspect of the reaction has been investigated but little since such studies naturally fall within the field of the chemist. All that may here be stated is that whatever may remain in solution in the clear medium when lysis is complete, it is not protein in nature, as is indicated by the reaction to heating. The viscosity of a suspension containing 500,000,000 Shiga bacilli per cubic centimeter differs from that of the bouillon used in preparing the suspension. A volume of bouillon giving nor- mally 100 drops gives but 97 when the suspension is lysed. The decomposition of the substances derived from the bacterial bodies continues certainly for some time after the lysis. This can be shown for the Shiga bacteriolysate by the fact that if a rabbit is inoculated with the material immediately after the lysis is completed, it is killed by a dose essentially the same as the minimal lethal dose of the suspension. The toxicity of the bac- teriolysate falls very rapidly; a week after the lysis it is markedly diminished, and in fifteen days the material is practically non- toxic. On the other hand, it is well known that the Shiga endo- toxin is very stable. Obviously then, this endotoxin is rapidly destroyed by the bacteriophage. In a later chapter we shall see that the lysin secreted by the bacteriophage can be obtained by precipitation with alcohol. 58 THE BACTERIOPHAGE THE ULTRAMICROBIAL BACTERIOPHAGE: AN ENDOPARASITE We have already considered the mode of action of the bacterio- phage from the point of view of its macroscopic characteristics. Various types of experiment allow us to penetrate somewhat further into the more intimate nature of the phenomenon. Experiment has demonstrated that the bacteriophage is ca- pable of development only at the expense of living bacteria, since these provide its sole culture medium. This cultivation appar- ently takes place within the interior of the bacterial cell, and ultramicroscopic observation shows that this is indeed the case. But first, let us consider some of the experiments which permit us to recognize the manner in which the infection of the bacteria is accomplished. Attempts have been made to effect a culture of the bacterio- phage of the Shiga bacillus in a filtrate derived from Shiga or- ganisms grown in bouillon for various lengths of time — 1, 7, 14, and 21 days — but all have failed. Two counts of the number of ultramicrobial elements, the one made immediately after the inoculation, the other after incubation for a week at 37°C., have given exactly the same number of germs. Thus, the bac- teriophage is entirely incapable of multiplication in a medium containing only the secretory products of the bacterium. The bacterial body itself is essential. If the bacteriophage actually proliferates within the interior of the bacterial cell, the ultramicrobes inoculated into a suspension ought, before all multiplication, to disappear from the fluid. In fact, each ultramicrobe ought first to penetrate a bacterial cell, to multiply there and to reappear in the fluid only when this cell is destroyed. It is easy to verify this hypothesis. Experiment XIV. The following suspensions are prepared : (1) 100 cc. of a suspension of the Shiga organisms containing 250,000,000 bacilli per cubic centimeter. This is inoculated with 0. 25 cc. of a culture of the bacteriophage. (2) 100 cc. of a suspension of the cholera vibrio, containing 250,000,000 per cubic centimeter. This also is inoculated with 0.25 cc. of the same culture of anti-Shiga bacteriophage. (3) 100 cc. of bouillon containing only 0.25 cc. of the same bacterio- phage. BACTEKIOLYSIS 59 The material of all three flasks is incubated at 37°C. Immediately after the inoculation, after thirty minutes, and again after 1 hour, 20 cc. are taken from each of the three flasks and centrifuged at 4,000 revolutions per minute for ten minutes. There are thus nine tubes which have been centrifuged. From the supernatant fluid of each of these 0.02 cc. is taken and introduced into other tubes containing suspensions of the Shiga bacillus, and the counts of the ultramicrobe are made by plating 0. 02 cc. of each of these nine tubes on six plates of medium. In this way an average of the counts can be ob- tained. The results of these counts indicate the number of ultramicrobes remaining in the medium, since those which have penetrated the bacterial cells before the centrifugation have been thrown down with the cells during this procedure and in consequence are to be found in the sediment. The results of the counts are as follows : Tube 1. Shiga suspension plus bacteriophage. a. Counts of the material made immediately after the preparation are 214, 193, 187, 221, 229, and 183 plaques. The average is 204, representing 5,000,000 germs per cubic centimeter in the original suspension immediately after inoculation. 6. Counts on the suspension after incubation for 30 minutes are 3, 7, 4, 6, 6, and 3 plaques. The average is 5. This indicates that there are 125,000 bacteriophagous germs in the suspension thirty minutes after the inoculation. That is to say, of each 41 ultramicrobes inoculated 40 have disappeared. A direct count of the suspension without centrifugation gives 5,000,000,000 elements per cubic centimeter. It is therefore certain that the ultramicrobes which have disappeared from the fluid during the cen- trifugation have gone down with the bacteria. And, as we will see in the two control experiments, in the absence of Shiga bacilli this sedimentation of the bacteriophage does not occur (at least, when centrifuged at a speed of 4000 revolutions). c. After 1 hour, the count, made as before on the supernatant fluid gives an average of 8 plaques, or 200,000 ultramicrobes per cubic centimeter, a number essentially the same as that secured after thirty minutes. At this time a count of a suspension which has not been centrifuged gives 6,500,000, a number very close to that secured immediately after the inoculation. It should be noted that in the hypothesis formulated with regard to intrabacterial growth, each colony forming within the interior of a bac- terium gives rise to but a single plaque ; just as in a bacterial count made by the same method a whole clump of bacteria seeded upon agar will yield only a single colony. d. Counts made upon the suspension with and without centrifugation after one and one-quarter hours of incubation give the same number of ultramicrobes, about 90,000,000. The inoculated organisms have therefore increased from 5 to 90 millions; the increase being in a proportion of about 60 THE BACTERIOPHAGE 1 : 18. And this increase has taken place in apparently a very abrupt man- ner, only to be explained as a result of the liberation of actual colonies containing an average of about 18 germs. We will see by ultramicro- scopic examination that the lysis of a parasitized bacterium takes place brusquely, by bursting. Tube 2. Control. Suspension of V '. cholerae plus the bacteriophage. Counts made immediately after inoculation of the bacteriophage give: for the centrifuged material 201, for the non-centrifuged, 211 plaques. After thirty minutes the counts are: for the centrifuged, 210; for the non- centrifuged, 216. After one hour the counts are: for the centrifuged, 203; for the non- centrifuged, 199. After one and one-half hours the non-centrifuged suspension gives 207. Tube 3. Control. Sterile bouillon plus the bacteriophage. The counts immediately after the inoculation are: for the centrifuged, 206; for the non-centrifuged, 210. After thirty minutes the corresponding counts are: 201 and 211. After one hour, the counts are: 203 and 206. After one and one-half hours the non-centrifuged medium contains 198. As is evident, in the absence of bacteria capable of being attacked, noth- ing happens. The ultramicrobes remain inert in the liquid. The nature of the multiplication taking place in the presence of the Shiga bacillus does not permit of any doubt on the follow- ing points. 1. After a contact of thirty minutes at 37°C. the ultramicrobes have almost entirely disappeared from the fluid; they are fixed by the bacteria. After one hour the situation is essentially the same. 2. After one and one-half hours there is an abrupt increase in the number of the bacteriophagous ultramicrobes. 3. The fixation is elective; it does not occur with V. cholerae, for example, for which the bacteriophage in question is without action. From this it may be concluded that the culture of the ultrami- crobes takes place within the interior of the bacillary body, and all the other observed facts support this conception based upon experiment. We can now understand the cause for the successive jumps noted in the culture of the bacteriophage, mentioned pre- viously in connection with the multiplication of the germs. Each of the ultramicrobes inoculated penetrates to the interior of a bacillus and there multiplies up to the time when the bacillary BACTERIOLYSIS 61 body bursts. This liberates the colony of ultramicrobes which have been formed in the bacterial protoplasm. A confirmation of this fact is obtained by examination of the phenomenon under the ultramicroscope, and by a study of the temporarily inhibitory action of an anti-bacteriophagous serum. It has been shown that the successive increases in number are separated by intervals of approximately seventy-five to ninety minutes. On the other hand, a complementary experiment, conducted in the same fashion, but centrifuging the suspension at ten minute intervals during the first half hour, has shown that very few of the bacteriophagous germs are fixed during the first ten minutes, although they are almost all fixed after twenty minutes. The union, therefore, requires about a quarter of an hour. Given the rapidity of multiplication of the ultramicrobe, and the time consumed in effecting each successive increment, it can readily be calculated that a single bacteriophage within a bacterium produces a colony varying in number from fifteen to twenty-five individuals; and it does this within the space of an hour or an hour and a quarter. BACTERIOLYSIS UNDER THE MICROSCOPE The anti-Shiga bacteriophage is always taken as an example in considering the action on dysentery bacilli. It has been seen that when the inoculation with the bacterio- phage is massive all the bacteria are attacked at the beginning, in other words, their multiplication is abruptly arrested. After two to three hours the medium commences to clear little by little and becomes completely limpid a short time later. If, on the contrary, the inoculation is minimal, the few ultramicrobes inocu- lated only affect an equal number of bacteria. The great majority remain unaffected and multiply as they would in a normal medium. But the ultramicrobes likewise multiply, following a progression more rapid than that pursued by the bacteria, so that within a few hours their number becomes equal to, or greater than, that of the bacteria. This is the time when macroscopic lysis commences. 1. Let us consider the first case, that of the massive inocula- tion. If we take from time to time a drop of the suspension 62 THE BACTERIOPHAGE up to the point when lysis is complete, spread these drops on slides and stain, either with the Gram stain, with carbol-thionin, or by the Romanowsky-Giemsa method (all staining methods give essentially the same picture), results such as the following are secured. A suspension of Shiga bacilli, 250,000,000 per cubic centimeter, is inocu- lated with 0.1 cc. of a culture of the bacteriophage and incubated at 37°C. After fifteen minutes it appears as a culture of normal bacilli. After thirty minutes it appears essentially the same, except that a few of the bacilli are poorly stained. After forty-five minutes about 10 per cent of the organisms stain poorly. Between one and two hours, the number of bacilli which stain badly continues to increase, and after 2 hours only a rare cell can be found which has taken the stain normally. At the same time, amorphous debris and granulations, derived most certainly from the bacteria already lysed, are seen. Similar material is seen very abundantly in old normal cultures of the Shiga bacillus. These granulations dissolve more slowly than the remaining portions of the bacterial protoplasm. Finally, and this is a most important point, spherical forms, more or less ellipsoidal, of variable dimen- sion, always rare, measuring 4 to 7 by 3 to 5 n may be detected. We will see in a moment to what they are due. There are occasional bacillary forms, well-stained, having a length of from 8 to 12 p. Between the second and third hours the amorphous debris considerably augments and the bacillary forms rapidly disappear. A few spherical forms are still to be seen. After four hours lysis becomes more and more complete. Only a single poorly stained bacillus will be found in two or three fields. Gradually the formless debris disappears, and, in turn, the granules. After thirty-six hours nothing whatever can be distinguished in stained preparations. With the ultramicroscope at no time can there be seen elements other than the bacilli (whose number gradually diminish, to disappear entirely in about two hours) and the extremely fine granules. It can hardly be said that the latter represent formed elements. At the beginning the bacilli present a normal appear- ance. After forty-five to sixty minutes fine granules are seen, ever becoming more and more abundant, within the interior of the bacterial cells. The number of bacterial cells containing granules also rapidly increases with a corresponding diminution in the number of normal bacilli. The most interesting part of the observation8 is that within one and one-quarter to one and 8 First noted by P. Jeantet. BACTERIOLYSIS 63 one-half hours after the beginning of the process the bacilli begin to swell, and the spherical bodies, containing a variable number of granules (averaging from 15 to 20) are, in comparison with a normal bacillus, from 3 to 5 p in diameter. If the spherical bodies are observed with care it is seen that after a variable length of time, sometimes amounting to only about ten minutes, an actual bursting takes place, consuming but a fraction of a second. Immediately afterward, in the place of the spherical body there remains a slight cloudy floccule that slowly dissolves, thus liberating the fine granules. These spherical bodies are particularly abundant at the time when the lytic process is at its maximum rate. There can be no question concerning the nature of these bodies; they are bacilli which, operated upon by a force which can only be internal, take at first a globoid form and then rupture. This is the more certain since at times one can witness the rupture of swollen bacilli, even before they have assumed a spherical contour. This observation provides direct proof that the ultramicrobe develops and exerts its action within the bac- terial cell. Destruction of the bacilli would be an entirely dif- ferent process if the dissolving action were exerted on the exterior. The spherical form and the bursting prove beyond possible con- tradiction that the operating force is internal. What do the fine granules that can be seen under the ultrami- croscope represent? While nothing can be affirmed with abso- lute assurance there is nothing to preclude the supposition that they represent the ultramicrobes, basing this upon the compara- tive examination of cultures in which the number of ultramicrobes has previously been counted by plating upon agar. By such a procedure it is found that in taking two cultures presenting a great difference in count, a parallelism is always to be noted be- tween the counts and the number of granules observed. It would likewise be well to recall what we have already seen with reference to the multiplication of the germs, that this mul- tiplication appears to take place in successive jumps (which correspond to the rupture of a large number of parasitized bacilli) and in which the number of ultramicrobes liberated after one and one-quarter to one and one-half hours corresponds to about eighteen germs to each single one inoculated. And we will see 64 THE BACTERIOPHAGE that the number of granules consequent upon the rupture of a cell amounts to between 15 and 25. There is, therefore, a great probability that the granules are actually the ultramicroscopic bacteriophagous organisms. 2. We may consider the second case, that of a minimal inocula- tion. In this case the medium becomes more and more turbid before lysis actually commences. A suspension of Shiga bacilli, containing 250,000,000 per cubic centimeter is inoculated with 0. 0001 cc. of a culture of the bacteriophage, a very active strain being selected. After 30 minutes the medium has its original turbidity; essentially that of a normal culture of the Shiga bacillus. After one hour the original turbidity is still maintained . When smeared and stained all the bacilli are of normal shape, but an occasional form stains poorly. After two hours the culture is about twice as turbid as at first. There is amorphous debris in the bottom of the tube. All of the bacilli appear to stain as normally. Many of the bacilli (about two in every three) are about four times the normal length, that is, of the bacilli used to seed the culture, and there are all intermediary forms. Oval and spherical forms are relatively numerous, but they are always fewer than would be expected from a comparative ultramicroscopic examination. These forms are indeed very fragile and are particularly liable to destruction during fixation upon the slide so that their demonstration in stained preparations requires great care. After three hours the suspension is slightly cloudy. The bottom of the tube is covered with fine debris without definite form, with, from place to place, great amorphous masses and numerous granules resembling those encountered in very old cultures of normally grown Shiga bacilli. Only a single spherical form can be detected in a ten-minute search. Each field may contain a dozen large bacilli, well stained. After four hours the turbidity is very slight. There is somewhat less material in the bottom of the tube, and this shows only a single poorly stained bacillus to a field. After six hours the medium is limpid. There is still less deposit in the bottom of the tube and it is with difficulty that a single poorly stained bacillus may be found in searching 25 fields. After eighteen hours nothing at all can be seen in the preparation. As is evident, the aspect of this preparation differs but little from that seen in the former case, the only departure being that the bacilli which have grown immediately after inoculation, be- fore the action of the bacteriophage becomes operative, present abnormally large forms. BACTERIOLYSIS 65 A comparable ultramicroscopic examination in the two cases shows that in the last, where the inoculation was made with a bacteriophage which was extremely active, at the time when lysis occurs with greatest intensity, that is, between two and three hours after the inoculation, the spherical forms were present in greatest numbers. There were as many as two to three to a field, and their rupture was readily observed. When lysis is once terminated the most careful search fails to reveal such forms. It is here fitting to recall an observation already made which should be noted by those wishing to investigate the subject. When a simple diastatic action is operative it proceeds with uni- form rhythm when under identical conditions. This is not the case here. Up to the present time more than a hundred different strains of the anti-Shiga bacteriophage have been isolated and no two of them have been found to conduct themselves in an exactly identical manner. The final result is always as has been indicated, the phases of the phenomenon always progress in the same order, but the time of the reaction will vary. With one strain of the bacteriophage complete lysis is obtained in three hours, with another, only after twelve hours. The phases follow each other in one case four times more quickly than in the other. Another point which should be remembered is that all that which has been said up to the present time has been in reference to bacteriophagous strains which were extremely active; that is to say, strains capable of producing complete lysis. A summary of the foregoing shows that, in so far as the micro- scopic observations are concerned, there is no time when one can distinguish in stained preparations, whatever the magnification, microorganisms other than B. dysenteriae Shiga. Aside from the bacilli one can see only formless cellular debris becoming more and more abundant with the more complete destruction of the bacteria, the debris later dissolving gradually. Ultramicroscopic examination indicates that the ultramicroscopic bacteriophagous germs multiply within the interior of the bacilli, and this observa- tion is corroborated by all experiments. Such examination also indicates that very probably the bacteriophagous elements are represented by the very fine granules which can be observed, first in the interior of the bacilli, and later in the ambient fluid. UNIVERSITY OF CALIFORNIA DEPARTMENT OF CIVIL ENGINES BERKELEY. CALIFORNIA CHAPTER II THE BACTEKIOPHAGE AND THE BACTERIUM Virulence of the Bacteriophage. Measure of Virulence. Resistance of the Bacterium. Secondary Cultures. Instability of Mixed Cultures. Characteristics of Mixed Cultures. Resistant Bacteria. Acquisition of Resistance. Production of Anti-lysins by the Bacteria. Multiple Cultures. VIRULENCE OF THE BACTERIOPHAGE In the preceding chapter we have taken note of the complexity of the phenomenon under study; a complexity resulting from the fact that there are simultaneously three elements which react, the one upon the others, — the medium, the bacteriophage, and the bacterium. Moreover, up to the present we have considered only the simplest case that may be presented, a bacteriophage of maximum virulence before which the bacteria always succumb. Often, the issue is very different, and for two reasons. The activity of the bacteriophage is not fixed, it varies along a scale extending from an action barely capable of detection to a lytic power most intense. The bacterium on its part is not passive; it defends itself. We have already stated that the activity of the bacteriophage is a true virulence, in the exact meaning of the word, "the ability which a micro-organism possesses to develop within the body of a host and there to secrete toxic substances." Just as for each pathogenic bacterial type there is a scale of virulence, so also for each strain of bacteriophage there is a certain virulence. It is possible to exalt or to attenuate the virulence of a given bacterium and the same can be accomplished with the bacterio- phage. Finally, just as higher forms when parasitized by a bacterium defend themselves and are capable of acquiring an immunity to this bacterium, so in like manner the bacterium attacked by a bacteriophage does not remain passive, but strug- gles, and may either be destroyed or acquire an immunity. All the vicissitudes of a conflict between an animal and an attacking 66 THE BACTERIOPHAGE AND THE BACTERIUM 67 bacterium are duplicated in the struggle between the parasitic bacteriophage and the attacked bacterium. The resemblance is complete. It is only a matter of descending a degree in the order of size in the beings concerned. It is also a property of living beings to never be the same at any two moments of their existence. If the phenomenon of serial transmissible bacteriolysis which we are considering were of purely diastatic nature, the activities as they unfolded would follow a fixed plan; for if the active element was not varied in quantity its quality would in all cases be constant. But we have seen that quite the contrary is the case. The phenomenon is independent of the quantity of the active element employed. The dominating feature is the quality of this element. Such are precisely the characteristics of vital activities. A poison acts in accord with its mass; a bacterium, with its virulence. Experiment has already shown that a bacteriophage but weakly capable of attacking an organism is susceptible to increase in potency through successive passages in contact with the bac- terium which it attacks. In order to recognize the differences presented between different strains of the bacteriophage it is preferable to work with strains freshly isolated from the organism. Experiment XV. (A) . Ten cubic centimeters of a suspension of Shiga bacilli are inoculated with 1 cc. of a filtrate made directly from the feces of a patient with dysentery. The suspension is held at 37°C. Counts of the ultramicrobes, made at different times during the incubation, give the following results when 0.01 cc. is plated on agar. When plated immediately, there develop 16 plaques, representing 1,600 ultramicrobes per cubic centimeter. The filtrate from the feces therefore contained 16,000 per cubic centimeter. After one and one-quarter hours, the count is 40 plaques, or 4,000 per cc. After two and one-half hours, a 1 :10 dilution gives 42 plaques, or 42,000 per cubic centimeter. After three and three-quarter hours, a 1 :100 dilution gives 18, or 180,000 per cubic centimeter. After five hours, a 1:1000 dilution gives 4, or 400,000 per cubic centi- meter. After fourteen hours, the lysis is not complete, the medium is cloudy and becomes more and more turbid, so that after forty-eight hours it is very turbid. Here there is an abundant culture, but lysis is never complete. The bacteria have, then, acquired a certain resistance which has allowed them to reproduce in spite of the presence of the bacteriophage. 68 THE BACTERIOPHAGE A result of this kind is usual when the filtrate is prepared from a stool taken shortly before the manifestations of convalescence appear. (B) Ten cubic centimeters of the Shiga suspension are inoculated with 1 cc. of the filtrate prepared from the feces from the same dysentery patient, but collected 24 hours later, the patient now being convalescent. Counts of this mixture give: When plated immediately, no plaques, or less than 100 ultramicrobes per cubic centimeter. Thus, the filtrate contained less than 1,000 per cc. After one and one-quarter hours the plating shows no plaques. After two and one-half hours there are 9 plaques, or 900 ultramicrobes per cubic centimeter. After three and three-quarter hours, in a 1:10 dilution, there are 27 plaques, or 27,000 per cubic centimeter. After five hours, a 1:1000 dilution shows 13 plaques, representing 1,300,000 per cubic centimeter. In this last experiment (B) the ultramicrobes were present in the filtrate in very small numbers, certainly less than 1000 per cubic centimeter, that is, there were less than one-sixteenth as many as in the filtrate of the first preparation (A). Nevertheless, the suspension was completely lysed in ten hours and the fluid remained sterile indefinitely. It should be noted that the multiplication of the ultramicrobes was much more rapid in the second experiment than in the first, and that this corresponds exactly with the idea of a greater viru- lence. These experiments show also that the number of ultrami- crobes inoculated is without effect upon the intensity of the phenom- enon, but that the important thing is the quality of the bacterio- phage, that is, its virulence. It is significant that the two strains of bacteriophage under discussion were derived from the same patient, but were taken at an interval of twenty-four hours. It is the same bacteriophage whose virulence has been increased in vivo. It would be possible to cite a great number of experiments of the same order. On each page of this text facts will be found that show that the essence of the phenomenon is the virulence of the bacteriophage, a virulence extremely variable, exalted, or at- tenuated, or indeed absent for a given bacterium according to the conditions at the moment obtaining. This extreme variability observed especially in vivo is due to a variety of conditions. It is less in vitro, where we are able, within certain limits, to secure a relative stability. THE BACTERIOPHAGE AND THE BACTERIUM 69 EVALUATION OF THE DEGREE OF VIRULENCE As will be shown in Part II of this monograph, it is necessary to the study of the processes of immunity associated with the presence of the bacteriophage, to be able to measure as exactly as possible the degree of virulence possessed by each strain of the ultramicrobe. The intensity of the action upon a bacterial sus- pension, or on a culture, in a liquid medium gives an indication of the virulence. A strain of maximum activity causes complete lysis in a relatively short period of time, varying between three and thirty hours. A less active strain causes only a partial lysis. This method of evaluation is, however, very crude and subcultures upon agar provide more precise determinations, particularly when dealing with strains which are but slightly active. If we introduce into a bacterial suspension a drop of a filtrate containing a bacteriophage active for the bacterium in the sus- pension, and if we plate upon agar a drop of this material after variable periods of incubation, it is possible to follow the multipli- cation of the ultramicrobes. It is only necessary to count the isolated colonies, which assume, as we have seen, the form of circular plaques. If working with two or more strains of the bacteriophage, it is thus easy to follow the relative rapidity of their multiplication, and by the same fact, to measure their respective powers of growth at the expense of the bacteria para- sitized, that is to say, their virulence. The extension of the plaques furnishes a second measure of the rapidity of the multiplication of the ultramicrobes. Each plaque, representing a colony, results from the extension into the culture of the descendants of a single ultramicrobial element, deposited during the plating, at the expense of the bacteria in its environ- ment. The more rapid the multiplication of the bacteriophage the more rapid the extension of the plaque. Thus, the diameter of the plaque permits a valuation of the degree of virulence of the bacteriophage which produces it. Experiment XVI. The relative virulence of four strains of an anti- typhoid bacteriophage taken from a single patient convalescent from typhoid fever (Jeanne Del ) at different periods during this convales- cence is determined. 70 THE BACTERIOPHAGE A suspension of B. typhosus containing 250,000,000 bacilli per cubic centi- meter is prepared and distributed into four sterile tubes, 10 cc. to the tube. Each of these tubes is then inoculated with 0.0001 cc. of a filtrate; the first with a filtrate prepared on the first day, tube 2 with that prepared on the second day, etc. After shaking, 0. 1 cc. is taken from each tube and plated, as usual, on an agar plate, taking care that the agar layer in all the plates is of the same depth simply to make all the conditions comparable. After incubation the following results are obtained : 1. The filtrate prepared from the stool taken on the first day of the appearance of the bacteriophage shows 16 very small plaques, pin-point in size. 2. The filtrate derived from the stool of the second day after the appearance of the bacteriophage shows 31 plaques, having a diameter of less than 1 mm. each. 3. The filtrate made from the stool of the third day gives 52 plaques, with diameters of about 2 mm. 4. The filtrate prepared from the stool of the fourteenth day shows 42 plaques, each with a diameter of less than 1 mm. Strain 3 is by far the most virulent, a conclusion that is supported by the fact that in its isolation it induced a total lysis of the bouillon culture of the typhoid bacillus. Strains 2 and 4 are much less virulent. Only by a dozen passages was it possible to effect an enhancement in virulence sufficient to give the same result. And at that time, when plated on agar the plaques had a diameter of 2 mm. These experiments demonstrate very well that with equal virulence the plaques on the surface of a medium are approxi- mately equal in size. The process of measuring virulence by counting the plaques and thus determining the rate of multiplica- tion is certainly more exact than is observation of lysis in a fluid medium. It is, unfortunately, too complicated to be applied in routine practice when a large number of strains must be examined, as is the case when working with patients. RESISTANCE OF THE BACTERIUM In the first of the two experiments just cited (Experiment XV, A) we have seen that the bacterium was successful in developing in spite of the presence of the bacteriophage. The virulence of the bacteriophage, then, although constituting a most important factor in the phenomenon is not the only consideration. The bacterium is capable of resistance. THE BACTERIOPHAGE AND THE BACTERIUM 71 Up to the present time we have considered only the case where lysis was complete and permanent, and it has been specifically stated that the phenomenon assumes this form only when the bacteriophage possesses a maximum virulence and acts upon a limited quantity of suspension — 10 to 20 cc. In spite of the ful- fillment of these conditions it sometimes happens that a suspension which has been lysed in a normal manner with a perfectly limpid appearance, will some days later become turbid. Microscopic examination shows that the turbidity is due to multiplication of the bacteria, and tests of the biologic activity prove that this cul- ture is composed solely of bacteria of the same species as was used in preparing the suspension upon which the bacteriophage was acting. According to the virulence of the strain of the bacteriophage being tested the number of tubes in which this reaction takes place, that is, the development of this secondary culture,1 is more or less great. Inoculations on agar or in bouillon of lysed suspensions, in which secondary cultures later develop, remain sterile up to the time that the secondary culture is formed. This does not often occur until five or six days after the lysis, sometimes even later. Experiment XVII. A suspension of Shiga bacilli, containing 250,000,000 bacilli per cubic centimeter, is inoculated with 0.001 cc. of a culture of the bacteriophage. Normal lysis takes place in five hours, with the medium perfectly limpid. The lysed suspension is planted on agar and in bouillon 1, 2, 3, 4, 5, 6, and 7 days after the lysis is completed. All the plantings remain sterile. On the eighth day the lysed suspension is slightly clouded. On the ninth day a drop is inoculated into broth and on to three tubes of agar. Two of the agar tubes remain sterile, the third shows four small colonies. The broth tubes give a culture agglutinated in the sediment. The resistance of diverse strains of a single bacterial species is not constant. Each strain appears, on leaving the organism, to be possessed of an individuality which is rapidly effaced by- successive cultivations upon an artificial medium. 1 In order to facilitate exposition I have called a "secondary culture" one growing again in a lysed suspension ; a "mixed culture," the inoculation into a nutritive medium of a "secondary culture" with the coexistence in the medium of bacteria and bacteriophage. 72 THE BACTERIOPHAGE In the following experiment the most powerful strain of the bacteriophage yet isolated is made to act upon two different races of the Shiga bacillus. One of these bacterial strains has been for a long time under artificial cultivation, being used by the Pasteur Institute for the inoculation of horses in the production of anti- dysentery serum (type strain). The other was recently isolated from the stool of a patient with dysentery (strain Jerv.). Experiment XVIII. (A) Twelve tubes of the suspension of the type strain of the Shiga bacillus are each inoculated with 0.001 cc. of a culture of the bacteriophage. This latter has been carried on for a great number of generations always at the expense of a single bacillary strain. In all twelve tubes lysis is perfect, with complete clearing in four hours. After three days at 37°C. one of the tubes is slightly cloudy, the others are clear, (Five other experiments, each consisting of 12 tubes, with the same strain of bacteriophage and the same bacillus give the following results: — tubes showing secondary cultures in each set, 0, 2, 0, 3 and 1. There develop, then, 7 secondary cultures in the 60 tubes, or 12 per cent.) (B) Twelve tubes of suspension were prepared with the strain Jerv., a strain with which the bacteriophage in question had never been in contact. Each of these tubes is inoculated with 0.001 cc. of the same culture of bacteriophage as that used in the preceding experiment (A). Seven of the 12 tubes give secondary cultures. The results from five other experi- ments with the same strains are, 9, 5, 10, 5, and 6 secondary cultures, or 70 per cent. A week later 12 cultures of the Jerv. bacillus are inoculated from one of the previous tubes that had remained clear. From these, 5 secondary cultures are secured. A further passage made after another week, gives 4 secondary cultures in the 12 suspensions. After another week, a fourth passage, still taking the bacteriophage from a perfectly limpid culture, yields but one secondary •culture among the twelve inoculated. (C) At the beginning of convalescence in the dysentery case (Jerv.) a bacteriophage was isolated which was tested in the same manner both on the type Shiga strain and on the Jerv. strain. This last was derived from the patient early in the infection at a time when the intestinal bac- teriophage had manifested no activity for this organism. With the bacteriophage Jerv. on the type bacillus 4 secondary cultures develop among the 12 suspensions lysed. With the bacteriophage Jerv. on the bacillus Jerv., there are no sec- ondary cultures among the 12 tubes lysed. When repeated upon an addi- tional 12 suspensions a single secondary culture develops. It is then clear that the anti-Shiga bacteriophage is not equally active for all strains of B. dysenteriae Shiga. This fact is even more in evidence with other bacterial species, for example, with THE BACTERIOPHAGE AND THE BACTERIUM 73 B. typhosus, B. coli, B. proteus, and B. pestis. Each bacterial strain possesses an individual resistance, particularly when freshly isolated, which renders it more or less resistant to a bacteriophage accustomed to an in vitro existence. Later we will see that this resistance increases by a phenomenon of natural selection. All of the phenomena in which the bacteriophage is involved, whether taking place in vitro or in vivo (the first are only an artifi- cial reproduction of the last) are dominated by two factors, — the virulence of the bacteriophage and the resistance of the bacterium. THE ORIGIN OF SECONDARY CULTURES What is the intimate mechanism of the process that results in the formation of secondary cultures? A priori two hypotheses can be formulated. Two factors are present, a bacteriophage whose virulence may be attenuated, and a bacterium whose resistance may be augmented. Thus, are secondary cultures due to a weakening of the activity of the bacteriophage, or, do there exist in the bacterial suspension certain individual cells which acquire an immunity to the bacteriophage, thus leading to the development of a resistant race? The following experiments clearly settle the question in favor of the last hypothesis. In the chapter treating of the isolation of the bacteriophage we have seen that in the large majority of cases the strains which are freshly isolated are of too low activity to effect a complete lysis of a bacterial suspension; cases where the presence of the ultramicrobe could only be detected by the presence of plaques upon the agar slants. These same strains were able to acquire, by successive passages, a very high activity, a potency which enabled them to bring about lysis of very heavy suspensions. This method of serial passages of the bacteriophage, in which it is forced to develop in vitro at the expense of a given bacterium, corresponds exactly with the method of Pasteur for effecting an enhancement in virulence of a bacterial race by repeated passage through a given animal species. This single experiment, repeated a considerable number of times, — in fact, each time that a bacteriophage of low virulence is isolated from the body — shows that secondary cultures are not produced by a simple diminution in the virulence of the bacterio- 74 THE BACTERIOPHAGE phage. Indeed, there is, on the contrary, an enhancement with each passage, even if macroscopic lysis is not to be seen. For this the following experiment offers direct proof: Experiment XIX. The contents of a tube that gave a secondary culture (as described on page 72) is filtered through infusorial earth and a bougie. Twelve tubes of a Shiga suspension are inoculated, each receiving 0.001 cc. of the filtrate. Perfect lysis is seen in all tubes, and in all but one the lysis is permanent. This single tube again becomes turbid after 4 days. From this it is clear that the bacteriophage has not lost in virulence, and that secondary cultures can not be ascribed to a change in that direction. The bacteriophage remains virulent, coexisting with bacteria which have become resistant. The secondary cultures, then, are the result of an adaptation undergone by the bacterium which acquires an immunity to its parasite. It has already been shown that the number of ultramicrobes inoculated is without influence on the appearance of secondary cultures. The conflict is not one of numbers; it is rather a strug- gle in which the significant factors are virulence on one side and ability to resist on the other. Experiment XX. A suspension of B. dysenteriae, 250,000,000 per cubic centimeter, is distributed into 6 tubes and these are inoculated with vari- able quantities of the same bacteriophage culture. The following results are obtained: TUBE AMOUNT OP BACTERIOPHAGE CULTURE INOCULATED RESULTS CC. 1 0.1 Normal lysis, secondary cultures 2 0.02 Normal lysis, no secondaiy cultures 3 0.004 Normal lysis, no secondary cultures 4 0.002 Normal lysis, secondary cultures 5 0.0002 Normal lysis, no secondary cultures 6 0.00002 Normal lysis, no secondary cultures The tubes yielding secondary cultures are distributed at ran- dom throughout the series, showing no fixed relationship to those tubes in which lysis was permanent. THE BACTERIOPHAGE AND THE BACTERIUM 75 Experiment XXI. This experiment shows the serial activity of the bac- teriophage together with the appearance of secondary cultures. Each tube of the series is prepared with a suspension of B. dysenteriae, 250,000,000 per cubic centimeter, and into each is introduced 0.001 cc. of the lysed suspension of the preceding tube. Transfers are made after twenty-four hours, that is, at a time when lysis is complete. A FRESH SUSPENSION RECEIVED THE MATERIAL INDICATED July 8 July 9 July 10 July 11 July 12 July 13 July 14 July 15 July 16 July 17 0.001 cc. of bacteriophage culture 0.001 cc. of suspension lysed on July 8 0.001 cc. of suspension lysed on July 9 0.001 cc. of suspension lysed on July 10 0.001 cc. of suspension lysed on July 11 0.001 cc. of suspension lysed on July 12 0. 001 cc. of suspension lysed on July 13 0.001 cc. of suspension lysed on July 14 0. 001 cc. of suspension lysed on July 15 0. 001 cc. of suspension lysed on July 16 Permanent lysis Permanent lysis Secondary cultures in 3 days Permanent lysis Permanent lysis Permanent lysis Secondary cultures in 4 days Permanent lysis Permanent lysis Permanent lysis Certain salts, when added to the suspension in very minute quantities, 0.1 mgm. to 10 cc. of culture, favor the development of secondary cultures. The salts of lead (nitrate and acetate) and of silver (nitrate and sulfate) act in this way. The soluble phos- phates and magnesium sulfate appear to be without action. With a single strain of bacteriophage and a given strain of bacil- lus the development of secondary cultures is, in general, more frequent when the suspension is prepared from agar cultures several days old than when made from fresh cultures. At first thought it appears strange that when secondary cultures develop with a strain of bacteriophage of high potency, they ap- pear in some tubes and not in others. The following experiment offers an explanation for this. Experiment XXII. Two flasks, each containing 200 cc. of a B. dysenter- iae suspension (250,000,000 per cubic centimeter) are inoculated with 0.04 cc. of a culture of the bacteriophage (the same strain as that used in the preceding experiments). Immediately after inoculation the contents of the first flask is distributed into 20 tubes, 10 cc. to each. In all of these lysis takes place normally, being permanent in 19, showing a secondary 76 THE BACTERIOPHAGE culture in 1. The second flask is portioned out the next day, that is, after lysis is completed, 10 cc. being placed in each of 20 tubes. None of these become turbid. When this second part of the experiment is repeated, 18 remain clear, and 2 tubes yield secondary cultures. Each flask of suspension contained 50,000 million bacilli, and the above experiments show that of this number but one or two were capable of acquiring an immunity to the very active bacterio- phage. It is these "immune" bacilli which give rise to organisms that enjoy the same degree of resistance. Secondary cultures, then, have their origin in the operation of the phenomenon of natural selection, whereby some bacilli show a greater aptitude than others to the acquisition of a resistance to the bacteriophage. The phenomenon of secondary culture formation is governed by the individual properties of the two admixed organisms, — bacterium and bacteriophage. Against a single strain of bac- terium the less virulent the bacteriophage the greater will be the proportion of secondary cultures, or, in other words, the greater is the number of bacilli in the suspension capable of acquiring a resistance. Against a given strain of bacteriophage the different strains of a single bacterial species are not endowed with an equal resistance. With certain strains secondary cultures will be the rule, with others, the exception, and with still others, they will never occur. We will shortly see the reasons for this variation; at present we may say that the degree of resistance possessed by a bacterium to a bacteriophage is, for a given bacterial species, in direct relation to the degree of virulence which this bacterial strain possesses for the higher organism which it is capable of invading. INSTABILITY OF MIXED CULTURES Mixed cultures result from a state of equilibrium between the virulence of the bacteriophage and the resistance of the bacterium. But these two factors are by nature variable and vary in intensity from one time to another, being influenced by the circumstances of the moment. This equilibrium can be interrupted experi- mentally in either direction, so as to favor either the one or the other of the factors. THE BACTEBIOPHAGE AND THE BACTERIUM 77 For example, if we place a small quantity of a secondary culture in normal saline, or preferably in 30 per cent glycerine bouillon, that is to say, in a medium which interferes with the reproduction of the bacteria and which exerts no destructive action on the bacteriophage, the equilibrium is disturbed in favor of the bacteriophage. The ultramicrobe is very sensitive to the action of acids, and if transfers from a secondary culture are made upon glucose agar it is found that the bacterium reacts upon the sugar, acidifies the medium, and breaks the equilibrium in favor of the bacterium. The bacteriophage, not being able to exert its parasitizing action, will be eliminated after a few transfers. Still another method of separation, the most practical of all, consists in employing the method used by Eliava and Pozerski (described further on) for obtaining cultures of resistant bacteria free from the bacteriophagous ultramicrobe. It is only necessary to make a few passages on agar slants, culturing each time from the extreme upper margin of the agar layer where the medium is somewhat desiccated. An ultrapure bacterial culture can also be obtained by the use of quinine, since this substance has a higher antiseptic activity for the bacteriophage than for the bacterium. We have seen that in the case of a bacteriophage but slightly virulent the addition of the filtrate in which it is present to some bouillon does not impair the development of the inoculated bacteria; it is only by spreading cultures on agar that we can detect the presence of the bacteriophage through the plaques which develop there. To increase the virulence of such an inactive strain we have seen that it is necessary to make several successive passages of the bacteriophage along with the bacterium. Between each passage it is essential either to filter the mixed culture through a bougie to separate bacteria and ultramicrobes or to heat the mixture to 60°C. to destroy the bacteria, while leaving the bac- teriophage unharmed. What is the basis for this technic? The elimination of bacteria which, because of contact with the bac- teriophage, are defending themselves and are acquiring a certain degree of resistance, a resistance which permits them to subsist in spite of the progressive increase in virulence of the bacteriophage. 78 THE BACTERIOPHAGE With each passage, therefore, we force a bacteriophage of increas- ing virulence to act upon an organism of normal resistance, that is, upon a bacterial suspension lacking resistance at the moment when it is placed in contact with the bacteriophage. Briefly stated, this technic is simply a method of opposing the develop- ment of resistant bacteria by natural selection. THE CHARACTERS OF MIXED CULTURES Bacteriophage of low virulence The means whereby the equilibrium obtaining in mixed cultures can be disturbed have been mentioned. It is of interest to allow the struggle to proceed naturally and to note its issue in a medium where each organism is dependent upon its own resources, that is, to permit the natural selection of such bacteria as are most apt in the struggle. To witness this, it is only necessary to make re- peated transfers in broth without intermediary filtration or heating. In cases where the bacteriophage is of low virulence in vitro the bacterium usually triumphs; its resistance increases little by little, the more vigorous bacilli survive and multiply, and the point is reached where the ultramicrobes no longer find bacteria suscep- tible to invasion. When this occurs the bacteriophage ceases to multiply and gradually becomes eliminated from the culture, until only a normal culture of bacteria remains. In some cases the state of equilibrium is more stable and the mixed cultures are able to continue throughout a large number of passages. Often these mixed cultures show cultural abnormalities and a partial lysis. The medium may become turbid only to become somewhat cleared later and finally to revert to a turbid condition. Experiment XXIII. Bouillon is inoculated with a mixed culture (B. dysenteriae — bacteriophage) taken from an agar slant planted thirteen months previously with a secondary culture. The macroscopic appear- ance passes through the following stages: after forty-eight hours, uniform turbidity; after five days, almost completely cleared; after thirteen days, uniformly turbid; after nineteen days, slightly cloudy with some sedimen- tation; after one month, uniformly turbid with sedimentation. This is the final appearance and at this time the bacterium and the bacteriophage coexist in the medium. THE BACTERIOPHAGE AND THE BACTERIUM 79 Transplants into bouillon continue to give mixed cultures, cloudy, with less marked but definite changes in appearance. These alterations in appearance are separated by intervals of only a few hours. This same experiment is performed with another strain of anti-dysen- tery bacteriophage, the inoculation being made from a colony on a secondary culture two months old. After three days there is uniform turbidity. In five days the medium is almost limpid. After eleven days it is very cloudy and after eighteen days it is clear with a slight sediment. All subcultures remain sterile and the medium contains a very active bacteriophage. In the mixed cultures with changing appearance the struggle between the bacteriophage and the bacterium inclines first in favor of one contending force and then to the advantage of the other. The final issue is at times in favor of the bacteriophage, at times in favor of the bacterium, and the number of transfers necessary to bring about a final issue is extremely variable. In vitro, the struggle generally ends with the bacterium the victor. In Part II of this text we will see that in vivo, with mixed cultures showing fluctuations, the issue of the struggle is deter- mined in some measure by the superior organism (that is, the animal body) in which the contending forces are operating. Bacteriophage of very high virulence With a bacteriophage of very high virulence secondary cultures are relatively rare, and when they appear they offer a very charac- teristic aspect, at least in so far as B. dysenteriae is concerned. The medium remains perfectly limpid, the bacterial culture ap- pears agglutinated, multiplying slowly in the bottom of the tube or deposited on the walls. These agglutinated masses may attain a size as large as the head of a pin and they can not be dissociated by shaking. With other bacteria the agglutination is less marked. Subcultures from these agglutinated mixed cultures give, indefinitely it appears, mixed cultures always presenting the same appearance.2 2 As one of the consequences, very important from the practical point of view, we will see that numerous so-called pure bacterial cultures are to be found in most laboratories which are in reality mixed cultures, contaminated from the time of their origin with a bacteriophage. 80 THE BACTERIOPHAGE Two different strains of anti-dysentery bacteriophage have been preserved for over two and a half years and during that time they have undergone more than 100 successive passages. Never- theless, during this period, repeated tests have shown in these cultures the constant coexistence of extremely virulent ultrami- crobes and of bacteria completely refractory to several very viru- lent strains of the anti-dysentery bacteriophage. In these mixed cultures there is a stable equilibrium between the virulence of the one and the resistance of the other, and in such cultures the changes in appearance previously noted will never be observed. They might, indeed, be spoken of as "sym- biotic cultures,"3 for the bacteriophage can not be cultivated in series unless it multiplies and it can not multiply unless it para- sitizes bacteria. Moreover, it is only necessary to disturb the equilibrium in favor of the bacterium, by such means as have been mentioned, to cause a rapid disappearance of the ultrami- crobes, rendering them henceforth incapable of cultivation. All of the bacteria present in an agglutinated culture are to be found in the agglutinated clumps, none are free in the medium. Subcultures into bouillon from the clear fluid always remain sterile, no colonies develop when transferred to agar, and micro- scopic examination fails to reveal any formed elements. All the bacteria there present have assembled in the agglutinate. The clear fluid contains only the extremely virulent ultramicrobes, as may be proved by the inoculation of a bacterial suspension which quickly becomes lysed. On the contrary, as we have seen above, a bouillon subculture made from the agglutinate always results in the growth of a mixed culture. When a mixed, agglutinated culture is inoculated into a pure culture of the bacteriophage, that is, into a suspension previously inoculated and which has undergone complete lysis, the growth consists of an agglutinated culture, just as though the inoculation had been made into fresh sterile bouillon. 3 This is possible if we interpret symbiosis in a broad sense, as Noel Bernard has. The definition of symbiosis given by this author applies admirably to mixed cultures: "An intermediary condition at which two antagonistic organisms arrive, with an equilibrium of their forces, toler- ating each other in a prolonged common existence." THE BACTERIOPHAGE AND THE BACTERIUM 81 If some of the agglutinate, even if washed, is introduced into a suspension of B. dysenteriae lysis takes place and the suspension becomes perfectly clear within five to six hours. Four or five days later, however, the agglutinated masses begin to appear and gradually increase in size. The ultramicrobes contained in the agglutinate used as inoculum provoke the lysis of the normal bacilli of the suspension, bacilli which are non-resistant, and then later the resistant agglutinated bacilli in their turn reproduce and the result is that which would have been secured had they been inoculated into fresh sterile bouillon. All stages intermediary between these two extremes may be obtained; cloudy mixed cultures presenting the appearance of a normal bacterial culture where the equilibrium is essentially unstable; cultures in agglutinated form in the presence of a per- fectly limpid fluid, representing a state of stable equilibrium. The medium may be more or less cloudy with the bacterial masses more or less compact, sometimes having but little density forming a coagulum. The type of the mixed culture bears a relationship to the virulence of the bacteriophage and to the resistance of the bacterium. Hence, the appearance of the mixed culture may be as variable as is the variability in the properties of the two organ- isms which are present. Mixed colonies on agar Instead of seeding the mixed cultures into broth they may be inoculated on to a solid medium. Often the agar will remain sterile, indicating that the equili- brium has been disturbed in favor of the bacteriophage. As has been said, this reaction may occur with inoculation into broth, but it is more frequent when agar plantings are made. It appears that the bacteria on agar are more readily attacked than when in a fluid medium. For this there may be several reasons, the prin- cipal one doubtless being the proximity of the bacteria. The first phase of the struggle is certainly associated with the phenomenon of chemotaxis. In a fluid medium the bacteria in suspension are separated by considerable spaces, certainly considerable when compared with the diameter of the ultramicrobe. Upon solid media, on the other hand, the bacteria actually touch each other, S2 THE BACTERIOPHAGE and the passage of the parasitic agent from one bacterium to another is readily accomplished since a strong chemotactic force is not required to bring together the two organisms in the struggle. When a culture develops on agar, the appearance of the growth may show considerable variation, determined by the relation between the resistance of the bacterium and the virulence of the bacteriophage. These two factors being inherently variables afford an infinite number of possible combinations, resulting in an infinite variation in the possible appearances on agar. At first sight, one of two principal distinguishing aspects may be present: 1. A smooth layer of culture, always located in the upper portion of the agar slant where the medium is less thick (although, it is needless to say, the mixed culture may be distributed over the entire surface of the slant) . The extent of this covering layer is variable. Sometimes there is only a fringe of bacterial growth at the extreme upper margin of the medium, the remaining portion being sterile. At other times the culture layer covers one-tenth, one-fourth, one-third, one-half, or even three-fourths of the slant, the portion remaining sterile always being the lower section of the slant where the agar is of greatest depth. The cause of this is simple. We have seen that the colonies of the bacteriophage appear as circular plaques, apparently sterile, and that they are of greatest size where the agar is deepest. The reason for this has been stated, and the present instance is but an application of this general fact. Certainly most bacteriologists will recall having seen cultures of this character without having recognized the cause. As a matter of fact, many cultures are to be found among culture collections which are in reality nothing but mixed cultures. In all cases, when one observes abnormal cultures, presenting the characteristics which have been described, one may be sure that such a mixed culture is present, that is, the culture is one which is infected with the bacteriophage. 2. The culture may consist of more or less numerous isolated colonies, even when the medium has been abundantly inoculated. These isolated colonies, in their turn, may present different appearances, associated always with the degrees of resistance and of virulence of the bacterium and the bacteriophage respec- THE BACTERIOPHAGE AND THE BACTERIUM 83 tively. Three types of colony may develop, each presenting individual characteristics. a. The colonies may be those of normal dysentery bacilli.4 These are encountered especially when working with mixed cultures derived from the inoculation of a suspension with a bacteriophage of low virulence. But even with a very virulent bacteriophage all of the colonies may appear quite normal. 6. Rare colonies, formed only of cocci, and cultivable under this form. They may be grown in bouillon, where an abundant culture of homogeneous turbidity is secured, or on agar, where the colonies appear somewhat different macroscopically from those of a normal bacillus, being more convex and more opaque. Subcultures obtained by the inoculation of these colonies are not mixed cultures; they contain only cocci, no ultramicrobes being present. The coccoid form is maintained during a number of generations and then the bacterium gradually reassumes its normal form. c. The colonies may be mucoid, refractile, difficult to dissociate, and of very diverse size — from the limit of visibility up to those with a diameter of about one millimeter. These colonies are cultivable on agar and reproduce colonies of the same form. They are mixed colonies, and in them the simultaneous presence of both elements, bacterium and bacteriophage, can always be demonstrated. Even when abundantly seeded upon agar these colonies never give a smooth layer of growth but always isolated colonies, more or less abundant, and always of variable size. Among the bac- teria of the inoculum but few are able to form colonies. There is always a state of unstable equilibrium between the two elements present: the bacterium with its resistance, and the bacteriophage with its virulence. The bacterium forms, or does not form, a colony according to the accidental predominance of one or the other of these factors. This is especially to be observed when agar is seeded with the agglutinated masses, for however abundant may have been the planting only very rare isolated colonies, all of the mucous type, develop. 4 B. dysenteriae Shiga is simply taken as an example; all other bacteria give mixed cultures and mixed colonies showing quite similar appearances. 84 THE BACTERIOPHAGE The cultures secured by the inoculation of the mucous colonies on differ- ent media show the following reactions: In agar stabs: small lenticular colonies about the needle track. In gelatine : as in agar, the resistant bacteria remain alive and cultivable for at least eleven months. In the case of the Shiga dysentery organisms this represents a viability at least ten times as great as that of the normal bacillus. In gelatine stabs: large opaque colonies with opaque centers. On glycerine potato (prepared as for the cultivation of B. tuberculosis}'. very rare colonies on the potato, very abundant growth in the fluid at the bottom of the tube. In milk: not coagulated in ten days. In litmus milk: turns to mauve after two months. On coagulated serum: no growth. In neutral red: no change in two months, either on agar or in bouillon. In litmus milk (Petruschky) : acid after ten days and remains acid. When the mucous colonies are suspended and heated to 60°C. they are not cultivable, for then the culture contains only the living very virulent ultramicrobe which is not killed until a tem- perature of about 75°C. is reached. Reinoculated into bouillon, the refractile, mucous, mixed colonies yield two types of culture, (a) mixed cultures showing changes in turbidity, and (6) aggluti- nated cultures, which, as we know, always depend upon the degree of virulence of the bacteriophage and the capacity of resistance of the bacterium, factors which regulate the appearance of the culture. We have seen that if an agglutinate, taken from a mixed culture in stable equilibrium, is introduced into a suspension, a lysis of the suspension is followed by a growth of the agglutinate. The same thing transpires if an abundant seeding is made on tubes of slant agar having a growth of the Shiga bacillus. First, plaques appear, and then after three or four days a mucous colony develops in the center of each plaque. In both instances the bacteriophage acts upon the normal non-resisting bacteria and dissolves them, then the refractory bacilli multiply as they would have done on a sterile agar or in bouillon. THE RESISTANT BACTERIUM From that which has preceded it may be deduced that the acquisition of resistance by a bacterium is reflected in a marked THE BACTERIOPHAGE AND THE BACTERIUM 85 change in morphology. We have seen that certain colonies on agar are composed of bacteria presenting the coccoid form, other colonies presenting the refractile appearance and a mucous con- sistency. Coccoid form The following experiments are interesting for they allow of the appearance of the coccus form with a later return to a normal morphology — the first in the colony itself at a few days interval, the second in the course of successive passages. Experiment XXIV. A Petri dish is heavily inoculated from an agar cul- ture of B. dysenteriae and is placed in the incubator at 37°C. for about four hours. A drop of the bacteriophage culture is then placed in the centre of the plate. The strain of bacteriophage should be one of average activity, that is, one capable of regularly causing complete lysis of a bacterial suspension but with which secondary cultures usually develop. (With too virulent a strain the area where the drop was placed remains sterile indefinitely.) The plate is returned to the incubator. After eighteen to twenty-four hours a layer of culture composed of normal dysentery bacilli develops, showing in the centre a spot devoid of growth, apparently sterile. After thirty-six to forty-eight hours, the spot becomes covered with extremely fine colonies, which, when examined microscopically are composed of cocci only. These cocci are of different sizes, from 1 to 4 /z in diameter, arranged in irregular forms, — in diplo- and in tetrad groupings. Two days later microscopic examination still shows cocci, but among them are bacillary forms in great number. Subcultures on to agar always give iso- lated colonies, each colony always reproducing with the same appearance and with the same sequence of forms, — first a coccoid culture, then a mix- ture of cocci and bacilli. These cultures always contain, moreover, bacter- iophagous ultramicrobes. Eliava and Pozerski have indicated a method for obtaining the coccoid, resistant bacteria free from admixture with the bacterio- phage.5 These bacteria perpetuate themselves under this form for a certain number of generations. Return to the bacillary form occurs gradually and after about fifteen transplants the culture acts as a normal dysentery organism sensitive to the bacteriophage. 8 The method permits the purification of a mixed culture by the elimina- tion of the bacteriophage. To accomplish this it is only necessary to make transfers on agar with the mixed culture, taking for inoculum in each passage material from the top margin on the agar, as near the edge as possible. 86 THE BACTERIOPHAGE To a suspension of bacteria in bouillon i's added 0. 01 cc.of abacteriophage culture and a drop of the material is spread on an agar slant. Frequently the medium remains sterile, as has been shown above. Sometimes a slight fringe of culture is secured, always at the upper margin where the medium is somewhat dried out. Material from this fringe is planted on a tube of sterile agar and after incubation a culture layer studded with plaques is found. A third transfer is made from the upper portion of the tube, and this is continued until an apparently normal culture is secured. That is to say, until the growth develops without plaques. At that time, the culture consists of resistant bacteria, free of bacteriophagCj and appearing coccoid in morphology. In such cultures resistance is maintained during a certain num- ber of passages. Then it gradually decreases and it is observed that this resistance is associated with the coccoid form, for the lengthening of the bacterial elements affords an indication that resistance to the bacteriophage is decreasing. The coccoid form may be secured in another manner. Certain of the agar tubes seeded with a secondary culture show after a very long time — one or two months, or even more — a colony situated near the top of the agar. This colony increases in size slowly and may attain a diameter greater than one centimeter. It is formed solely of cocci. There can be no doubt but that it consists of modified bacteria since successive subculturings yield normal bacillary forms. Atypical colonies have been secured with B. dysenteriae (Shiga, Flexner, and Hiss), with B. coli, B. typhosus, and the paraty- phoids. As is seen, the coccoid form is most certainly a resistant form of the bacterium, and the return to bacillary form taking place gradually, affords an index of the decrease in resistance. This morphologic transformation is accompanied by a profound change in the properties of the bacterium. Coccoid cultures are not agglutinable by a specific serum. This is also true of secon- dary cultures and mixed cultures iD general. Restoration of agglutinability is coincident with loss in resistance and with the return to normal morphology. In this connection it has been shown that strains of B. typhosus, inagglutinable when derived from the body, are at the same time composed of bacilli which are resistant to the anti-typhoid bac- THE BACTEKIOPHAGE AND THE BACTERIUM 87 teriophage, and further, that when, by subculture on agar, they lose their resistance they also become agglutinable. Inagglu- tinability appears to be a property of the bacteria which resist the bacteriophage. The vitality of resistant bacteria is much greater than that of normal bacilli. For example, with the Shiga dysentery organism whose vitality is weak (there are but few strains cultivable after one month, none among the numerous strains with which I have worked have remained alive without subculturing for more than two months), all of the colonies on agar of the resistant Shiga substrain are still cultivable after eighteen months. The virulence of bacteria which are resistant to the bacterio- phage is likewise considerably greater than that of normal bacilli. Whatever may be the nature of the resistant bacteria, and whatever may be their form, there is no doubt but that they can return to the normal form with normal properties. They then behave as the bacteria of the same species which were used to prepare the initial suspension upon which the bacteriophage acted. In a word, there can be no question either of an accidental contamination or that they are visible forms of the bacteriophage. With a coccoid culture of B. dysenteriae Shiga the following biologic reactions have been effected, reactions which indicate that these cocci conserve the general properties of the normal dysentery bacillus: a. The injection into a rabbit of such cultures causes the death of the animal with paralysis and intestinal lesions identical with those observed in animals killed by the inoculation of typical dysentery organisms. b. Rabbits immunized with carefully graded doses of such cultures are protected against a surely fatal dose of typical dysen- tery organisms. c. The serum of rabbits which have been treated with injections of coccoid cultures contain an amboceptor which will fix com- plement in the presence of normal bacilli. Zoogleic form Microscopic examination of the agglutinate formed in liquid media by bacteria which are endowed with a high resistance, and also of colonies which are mucoid on agar, show that the 88 THE BACTERIOPHAGE bacteria (their morphology will be discussed shortly) are sur- rounded by a mucous material. They are actual zoogleic colonies. As has been said above, there can be no doubt as to the nature of these bacteria, since all biologic reactions show that they react as did the bacteria of the same species which were used to prepare the original suspension. Moreover, in every instance, they revert to normal form. It is only necessary to eliminate the bacterio- phage, the cause of the transformation. It is thus very obvious that the resistance of the bacterium to the action of the bacteriophage profoundly modifies its form. Resistance accompanies a transformation into the coccus form, and when the bacterium, having to defend itself against an extremely virulent bacteriophage increases its resistance, there is formed a mucous capsule which certainly functions by hindering the penetration of the ultramicrobes into the bacterial body. Encapsulated bacteria thus enjoy a refractory state. The transformation is accompanied by modifications in the properties of the bacterium; increase in viability, enhanced viru- lence, and inagglutinability by specific sera. The resistance persists only as long as the bacterium must needs resist the action of the bacteriophage. In the absence of ultramicrobes6 the resistance gradually falls, somewhat more rapidly when it was not very marked. In extreme cases it disap- pears after about thirty transfers on agar. This represents a very considerable number of generations. We have seen that it is experimentally possible to render a bacterium resistant to the action of the bacteriophage and we have indicated different methods which enable us to reproduce this phenomenon at will. It is not an artificial phenomenon but is a reproduction of a natural process which takes place within the organism. Each strain of a species of a pathogenic bacterium, recently isolated from the body of an individual, offers a different resistance to the action of the bacteriophage, and this resistance, as we know, may extend to an absolutely refractory state. These differences of resistance arise, as will be demonstrated later, in hereditary 6 Suitable means have been indicated for purifying a mixed culture by eliminating the bacteriophage. THE BACTERIOPHAGE AND THE BACTERIUM 89 factors which originated in the struggle which took place between the ancestors of this bacterium and the bacteriophage within the body of the infected animal. As is the case of the resistance acquired experimentally, this naturally acquired resistance disap- pears gradually with repeated culture on laboratory media. Thus, different strains of a single species of bacteria tend to become, by a gradual process, uniform, and after a sufficient number of transplantations all are equally sensitive to the action of the bacteriophage. It is worthy of note that the degree of virulence of a bacterium is strictly in relation with its degree of resistance to the bacterio- phage. We will have occasion to demonstrate this frequently when we come to consider the relation between the bacteriophage and the infections due to bacteria. MICROSCOPIC OBSERVATIONS In the agglutinated and the zoogleic colonies certain atypical forms are encountered, which are also to be found, although less abundantly, in mixed cultures in general. Microscopic examination of preparations stained with thionin or by the Giemsa method shows a variety of forms depending somewhat upon the age of the colonies. Up to the second or third day the pleomorphism is considerable. Together with the typical bacillary forms and the cocci there are elongated bacilli, some as long as fifteen micra, with all intermediary lengths; some straight, others curved at all angles. Some are clubbed, yeast-like. Large and small granules are present together with the de"bris derived from the destruction of the different forms. All stages inter- mediary between intact forms and amorphous de*bris are repre- sented. What is the exact significance of these various forms? They are forms of involution or forms assumed by the bacterium in the development of resistance. That is all that It is possible to say with certainty. However, some of these forms give the impression that the organism in question is reproducing by sporogony. In old colonies but very few of the bacillary and coccoid forms are to be seen. 90 THE BACTERIOPHAGE The subcultures on agar can be multiplied but the appearance is always the same, and no matter which of the colonies is selected the biologic tests mentioned above always show that they are related to the original bacterium. Further, the possibility of reversion to the bacillary form in pure culture when grown on a glucose agar medium demonstrates conclusively that contamina- tion has not taken place. How may these facts be explained? Early in the consideration of the phenomenon the hypothesis presented itself that in these secondary cultures a visible form of the bacteriophage was pres- ent. But experiment has shown that this interpretation was false. A second idea has developed which unfortunately has not been completely studied for lack of time (the subject of the bacteriophage is so far-reaching that it has seemed more essential to utilize the available time on experiments dealing with the func- tion of the bacteriophage rather than with those of the question of morphology) but is presented only that it may be considered by those who are particularly competent to engage in morphologic studies. The necessity of bacterial intervention for the production of asci by certain fungi7 is a condition clearly recognized in mycology. This same situation may intervene among bacteria which are resistant to parasitism.8 This hypothesis is perhaps the less improbable since, according to Schaudinn, sexuality is a fundamental characteristic of living matter. It is observed with B. butschlii and with B. sporonema (in addition to the usual mode of reproduction by transverse 7 For example, Ascobolus furfuraceus (Moliard, 1903), a fungus of the genus Willia (Sartory, 1902), and an aspergillus growing on the banana (Sartory, 1920). 8 The following suggests that under the influence of the bacteriophage non-spore-forming bacteria may give rise to filtrable forms. I have noted, although rarely, that a filtrate obtained by passing a secondary culture through a Chamberland bougie (L2 and even Ls) becomes turbid after some days. Each time that this has been noted the turbidity has been due to the growth of a resistant bacterium such as was present in the secondary culture prior to the filtration. The conditions under which this phenomenon occurs have not been ascertained, thus the observation is simply mentioned without emphasis being placed on its interpretation. THE BACTEEIOPHAGE AND THE BACTERIUM 91 division) where there is noted an autogamous reproduction, which implies, at least, a rudimentary sexuality. In autogamy there are necessarily differentiated elements which fuse. According to Schaudinn the loss of sexuality in bacteria is an indication of a degenerative change resulting from the adaptation to a para- sitic existence. In accordance with the hypothesis suggested above, there would be a return to sexual reproduction as a means of defense, under the influence of the parasitism to which the organisms have been subjected. Such sexual reproduction would be, then, extremely frequent in vivo, as frequent indeed as a strug- gle between the bacterium and a bacteriophage occurs in the body. In Part II of this monograph it will appear that this struggle is continuous. THE ACQUISITION OF RESISTANCE How can this acquisition of immunity by a bacterium be explained? Numerous experiments have shown that if a certain quantity of a slightly active culture of a bacteriophage is introduced into a relatively heavy (1000 to 2000 million per cubic centimeter) suspension of bacilli, the ultramicrobes, readily demonstrated at first by the presence of plaques on plantings on agar, disappear from the medium after an interval of time varying from one hour to two or three days, and that they can not later be dem- onstrated. Subcultures give normal cultures of bacteria. On the other hand, we have seen that with a very virulent bacterio- phage the ultramicrobes disappear from the fluid between ten and twenty minutes after introduction into a suspension, but that they reappear in about twenty times as great a number in from one to one and a half hours later — they have multiplied within the interior of the bacteria. In the case of a bacteriophage of low virulence it seems, therefore, that penetration of the bacteria takes place but that multiplication can not be effected. The bacterium resists and the ultramicrobe is actually destroyed in vivo. These parasitized bacteria which "recover" acquire by this an immunity. We will see later that they are even capable of producing antilysins. Another fact has been sometimes observed which shows that certain bacteria are able to become "carriers." As has been said, 92 THE BACTERIOPHAGE heavy suspensions which are inoculated with a filtrate containing a relatively avirulent bacteriophage give after a few hours abso- lutely normal cultures on agar, free of plaques. If serial trans- plants are made of these cultures, inoculations made in such a manner as to yield an even layer of growth, it sometimes happens that after a certain number of transplants, two to four, a very definite plaque appears, which is indeed a colony of the bacterio- phage. This is evidenced by the fact that successive passages from this plaque yield a very active bacteriophage. From whence could this ultramicrobe have so suddenly come? The ultrami- crobe had remained alive within a bacillus, and at a given moment, it overcame the resistance of the latter and multiplied. Its viru- lence being increased, the young germs were able to parasitize the neighboring bacilli and form a colony. Any other explana- tion seems impossible, since, immediately after the inoculation of the bacteriophage, seeding upon agar shows plaques characteristic of the presence of virulent bacteriophagous germs, then these germs completely disappear, the bacteria, however, remaining sensitive to the action of a more active bacteriophage, for perfect lysis is secured if the suspension is inoculated with a trace of a very active culture of the bacteriophage, and finally, the active bacteriophage reappears after a series of subcultures on agar in the course of which all the bacillary cultures have been normal. This germ can only be one of those which had disappeared. The fact, demonstrated by experiment, of the penetration of virulent ultramicrobes into the bacteria, warrants us in thinking that this ultramicrobe (but slightly virulent) has been preserved in a latent living state within the interior of the bacterium. At a given moment the resistance of the bacterium is broken down and in- fection results. PRODUCTION OP ANTILYSINS BY BACTERIA By its mere presence the bacteriophage is not able to dissolve the bacterium. In the following chapter we will see that it acts through the secretion of lysins which, however, can be isolated free from the ultramicrobe. The following experiment shows that the resistance of the bacterium to the action of the bacteriophage, which amounts to THE BACTERIOPHAGE AND THE BACTERIUM 93 an actual immunity in the true sense of the word, is in great part due to the fact that the bacteria secrete antilysins which neutralize the lysins. Experiment XXV. A very active strain of bacteriophage, active for B. dysenteriae, is diluted in sterile bouillon, 0.05 cc. to 10 cc. of the medium, and 0.05 cc. of this dilution is introduced into a second 10 cc. of medium. In addition, a very concentrated suspension of B. dysenteriae is prepared, — a suspension representing a twenty-four-hour slant agar culture in 6 cc. of sterile bouillon. This suspension is inoculated with 0.05 cc. of the second dilution of the bacteriophage culture and incubation at 37°C. for four days follows. At the end of this time it is evident that the suspension is not lysed (because of the too great quantity of bacilli) but when a drop is planted on agar the medium remains sterile. Therefore the bacteriophage has multiplied. This heavy suspension is filtered through a bougie. To each of three tubes containing 10 cc. of a weak suspension of dysen- tery bacilli is added — to the first, 0.05 cc. of the filtrate from the heavy suspension, to the second, 0.05 cc. from the first tube, and to the third, 0. 05 cc. from the second. After an incubation period of twenty-four hours it is seen that there is no lysis in tube 1, but lysis is complete in tube 2, while the suspension in tube 3 is still turbid although it clears after forty- eight hours. But one conclusion is possible. Since lysis did not take place in the first tube, in spite of the presence of a large number of virulent ultramicrobes, and since it was produced in the other two which received infinitely less, there must have been in the filtrate some substance which inhibited the lytic action of the bacteriophage. This was not manifested in the last two tubes because of the dilution. It is likewise because of dilution that the inhibiting substance did not manifest its action on agar. It has already been noted, in the course of the earlier experiments, that lysis was more often perfect if a suspension was inoculated with a minimal amount of the bacteriophage than if the inocula- tion was massive. The cause for this fact is now clear. This experiment is in all respects identical with that described in a preceding paragraph, except that there it was accomplished by introducing a small number of ultramicrobes of low virulence into a large number of bacilli. Under these circumstances the bacteriophage is overcome and destroyed by the bacterium. In this last experiment, on the contrary, we have substituted for the small number of germs of low virulence, ultramicrobes of high 94 THE BACTEBIOPHAGE virulence. These ultramicrobes develop slowly, but at the same time the very numerous bacteria have had time to adapt them- selves and to elaborate a defensive mechanism to the lysins se- creted into the medium. They have produced an antilysin which paralyzes the action of the bacteriophage. Moreover, it has been shown that these antilysins are specific, for those secreted by the dysentery bacillus are without inhibiting action on the lysis of B. pestis. The bacteria, in general, react like higher organisms. B. tetani, for example, secretes a toxin, and an animal which would be killed by a large dose of this toxin reacts to a small dose by the produc- tion of an antitoxin which will neutralize the toxin. The bacterio- phage secretes a lysin and the bacterium, which could not resist a rapid attack, produces, when placed in favorable conditions, an antilysin which neutralizes the action of the lysin. We will further see that the organism reacts to the injection of the bacteriophagous lysins, which in effect, are nothing but foreign bodies of the same nature as the toxins,9 and, indeed, in a manner quite analogous to that of the bacteria, namely, by the production of specific antilysins. MULTIPLE CULTURES It remains for us to consider multiple cultures. In the following chapter we will see that the virulence of a strain of bacteriophage is rarely limited to a single bacterial species. What forms do secondary cultures take when a bacteriophage is forced to act upon a suspension comprising different bacterial species? We will confine ourselves to the most simple case, — a bacteriophage reacting upon two species of bacteria. Let us take as an example a strain of bacteriophage very virulent for B. dysenteriae Shiga and but slightly virulent for B. coli, as evidenced in the following experiment. Experiment XXVI. Three tubes of bouillon receive respectively 0.01, 0.1, and 1 cc. of a known anti-Shiga bacteriophage. The three tubes are then lightly planted with B. coli. Normal cultures develop in the three tubes. Platings on agar give few plaques. Each of the three cultures is 9 The organism does not defend itself against a toxin because it is toxic but because it acts as a foreign colloid in the body. THE BACTERIOPHAGE AND THE BACTERIUM 95 re-inoculated into fresh bouillon. Normal B. coli cultures develop. Transfers to agar give two plaques for the first tube, none for the other two. The culture yielding the two plaques is again re-inoculated. A nor- mal culture develops. The bacteriophage has been eliminated. This strain of anti-Shiga bacteriophage possesses, therefore, an extremely feeble virulence for the strain of B. coli under test. Experiment XXVII. To 10 cc. of bouillon is added 1 drop of a concen- trated suspension of Shiga bacilli (this should give a slight turbidity equal to about 50,000,000 bacilli per cubic centimeter) and 1 drop of an equally concentrated suspension of B. coli. This double suspension is then inocu- lated with 0.01 cc. of the anti-Shiga bacteriophage used in the above experiment. After twenty-four hours there is a slight turbidity. A new passage into a double Shiga-colon suspension is made. Perfect lysis takes place after eleven hours. The lysed suspension is then introduced, in a quantity of 0.04 cc., into a simple suspension of B. coli. Lysis is complete in seven hours. The ultramicrobes have developed at the expense of the Shiga bacilli, and thus being maintained in the medium they have gradually acquired a virulence for B. coli. In the intestinal tract a bacteriophage never finds itself in the presence of but a single bacterial species. And this experiment permits us to comprehend the process of acquisition in vivo of virulence by a bacteriophage for a given bacterium. CHAPTER III VIRULENCE OF THE BACTERIOPHAGE Multiple Virulence. Persistence of Virulence. Bacterial Species Attacked. MULTIPLE VIRULENCE A strain of bacteriophage freshly derived from the body is rarely active against a single bacterial species. Usually it at- tacks a certain number of species at this time, and possesses for each of these a variable virulence. It may be objected that this by no means opposes the concep- tion of a plurality of species in the genus Bacteriophage, each species acting against a determined bacterium. This question will be considered in detail in a later chapter. For the time being, however, it will only be stated that all experimental work, par- ticularly the work with the complement fixation reaction, favors the idea of unicity, whatever may be the origin of the bacteriophage. There is but one bacteriophage, but as isolated from the or- ganism, there is an infinite number of strains, each possessing the capacity to attack diverse bacteria. Such a strain may possess, for example, a very high virulence for B. dysenteriae Hiss, an average virulence for B. coli, a low virulence for the Shiga dys- entery strain, a very weak activity for B. paratyphosus B} and none1 for the other intestinal organisms tested. Another strain may be very active for B. coli and B. typhosus, but slightly active for B. dysenteriae Hiss, and inactive for the other bacteria tested. 1 It is evident that when it is stated that a given strain of bacteriophage lacks virulence for such and such a bacterium it must be understood "a virulence such as may be demonstrated by the present technic." Early in my investigations it was possible to detect only those strains possessing a considerable activity; all others were unnoticed. The technic has since been improved, but at the present time it is most certainly not perfect. 96 VIRULENCE OF THE BACTERIOPHAGE 97 It is manifestly impossible to make a complete analysis of a strain of bacteriophage, for to do so would necessitate a determina- tion of its activity against all known, and even against unknown, bacterial species. With a given filtrate prepared from the in- testinal contents, we can affirm that a bacteriophage is found there at the moment of testing because of its activity manifested toward a given bacterium. On the contrary, it is not possible to conclude that none is present simply because the tests were negative. Investigating the activity of the intestinal bacterio- phage in a filtrate from the feces of a healthy person a negative result has been obtained in testing against the intestinal bacteria toward which it was expected an activity would be evidenced. The investigation was extended to the most varied bacterial types, and finally a strain of bacteriophage was isolated active against an organism of the Salmonella group (bacillus of hog cholera). This strain of bacteriophage was cultivated in series and an active bacteriophage was thus secured. A given strain of bacteriophage will vary from time to time, either in the body, as can be demonstrated by isolating the bac- teriophage each day from the feces of a patient during the course of the disease and during convalescence, or it may vary in vitro, as has been shown in the preceding chapter. All combinations of virulence are possible, both as to quantity and to quality; that is to say, in the extent of the action against varied bacterial species, and in the intensity of the action for each of the bacteria attacked. It can be readily seen, in view of the infinite number of combinations possible, that two strains of bacteriophage identical in all respects can not exist. PERSISTENCE OF VIRULENCE The faculty which a strain of bacteriophage possesses to return to parasitism with a bacterium persists throughout a very great number of passages in vitro along with a bacterium of another species. For example, in 1916 a bacteriophage was isolated which was extremely active for B. dysenteriae Shiga, of but average viru- lence for B. coli, and but slightly active for B. typhosus and the paratyphoid organisms. This strain, which has been used in many experiments, was subjected during the years 1916, 1917, 98 THE BACTERIOPHAGE 1918, and 1919 to a large number of passages, somewhat more than 1200, always with the dysentery bacillus. Nevertheless, early in 1920 experiment showed that it had an average virulence for B. coli and a very low activity for B. typhosus. The action on the typhoid bacillus of a bacteriophage which had received more than a thousand passages with B. dysenteriae is evidently weak. This can be demonstrated by spreading on agar, a procedure which permits the formation of characteristic plaques. If we introduce into a tube of bouillon about ten drops of an anti-dysentery bacteriophage and then a small amount of typhoid culture, we secure, after incubation for eighteen to twenty- four hours, a culture of B. typhosus which appears normal, but if a drop is seeded upon agar a few plaques are obtained. The following observation, made by G. Eliava, is of the same nature, but more typical, for there is a crossed reaction on bac- teria but remotely related. A strain of bacteriophage isolated from the pus of an abscess (we will see that under certain cir- cumstances the intestinal bacteriophage may pass into the cir- culation) and very active against Staphylococcus aureus, appeared, even after a series of more than 100 passages with the staphylo- coccus, to be endowed with a degree of activity for the dysentery bacillus. Experiment XXVIII. Ten cubic centimeters of a Shiga suspension is inoculated with 0.25 cc. of a filtered anti-staphylococcus bacteriophage. When plated immediately on agar a normal culture of B. dysenteriae develops. After incubation at 37°C. for twenty-four hours, 0.02 cc. of this suspension plated on agar gives 4 plaques. This virulence for B. dysen- teriae was increased by successive transfers in association with this organism. Such experiments allow us to penetrate further into the phe- nomenon of virulence of the bacteriophage. We have introduced into the suspensions, either of B. typhosus or of B. dysenteriae, several hundred millions of ultramicrobes, all virulent for the Shiga bacillus in the first case, for the staphy- lococcus in the second. Each plaque on the agar represents a colony derived from a virulent ultramicrobe; virulent in one case for the typhoid bacillus, in the other for B. dysenteriae. We have also seen that of the several hundred millions of ultramicrobes VIRULENCE OF THE BACTERIOPHAGE 99 only a few were endowed with a latent virulence sufficient to parasitize the new bacterium offered them and to multiply at its expense. The persistence of latent virulence is the appanage of certain particularly apt individuals. With each passage, at the expense of the new bacterium, these individuals multiply, their virulence becomes enhanced, and finally, after a sufficient number of pas- sages, a complete lysis of the suspension is secured. With these facts as a basis, an attempt has been made to adapt to the dysentery bacillus a strain of bacteriophage active for the staphylococcus, although the ultramicrobe at first appeared de- void of all action on the dysentery strain. In this attempt, ten drops of an active anti-staphylococcus bacteriophage were inocu- lated into a double suspension containing in each cubic centimeter ten million staphylococci and ten million B. dysenteriae. After twelve passages (each passage being separated by a bougie filtra- tion and twelve drops of the filtrate inoculated into a fresh staphylococcus-dysentery suspension) the bacteriophage very actively attacked the dysentery bacillus. This property de- veloped very abruptly during the eleventh passage. After a series of twenty passages made in conjunction with Shiga alone the bacteriophage did not possess any activity for the staphylo- coccus, and it has been impossible to cause it to reassume such activity. Up to the present time it has been impossible to accomplish the inverse experiment, that is, to cause a bacteriophage active for the Shiga bacillus to acquire a virulence for the staphylococcus. This inequal persistence of latent virulence of the bacteriophage against diverse species of bacteria may be explained. The in- testinal bacteriophage maintains itself in the intestinal tract at the expense of the different intestinal bacteria. The bacterio- phage is therefore, in reality a normal parasite of the colon- typhoid-dysentery group and an accidental parasite of the other bacteria against which it acquires virulence in the same environ- ment, as a result of conditions at present undetermined. It can be understood then, that a bacteriophage active, when derived from the body, for an organism rather remote from the colon- typhoid-dysentery group, for example, a staphylococcus, retains 100 THE BACTERIOPHAGE for a very long time, probably indefinitely, the capacity to attack a bacterium of the enteric group, and this in spite of very many passages in conjunction with the bacterium accidentally attacked. On the contrary, a bacteriophage possessing, when derived from the body, a virulence for a bacterium accidentally attacked, loses more or less rapidly this virulence if it is maintained in vitro with a bacterium normally attacked, that is to say, at the expense of an organism of the colon-typhoid-paratyphoid-dys- entery group. It may be objected that all these facts may be interpreted, not as a persistence of a latent virulence but as a persistence, through the course of successive passages, of several species of the bac- teriophage. In other words, there has been a "contamination" of the anti-dysentery bacteriophage by an anti-typhoid bacterio- phagous ultramicrobe in the first case cited, and a " contamina- tion" of the anti-staphylococcus bacteriophage by an anti-dys- entery bacteriophage in the second. Both experimentation and mathematical reasoning show that such an interpretation is false. In the first example cited, it has been shown that the number of ultramicrobes virulent for B. typhosus did not vary during the course of the passages. The number of plaques obtained on agar by inoculation of a suspension of typhoid bacilli inoculated with the anti-dysentery bacteriophage is essentially the same, although the bacteriophage has been subjected to fifty, one hundred, five hundred, or a thousand passages at the expense of B. dysenteriae. If the ultramicrobes capable of attacking the typhoid bacillus were an "impurity" their number should diminish gradually in the course of the transfers, each passage being a dilution, and they should quickly disappear. If one calculates the extent of the dilution, after a thousand passages, of the filtrate which served as the original inoculum for the bacteriophage in question, a figure of such magnitude is obtained that a persistence of anti-typhoid bacteriophagous germs, throughout the series of successive dilutions, is mathematically impossible. One can calculate readily up to the thousandth passage (each passage consisting of the inoculation of 0.001 cc. of filtrate into ten cc. of bouillon). The value of the dilution in VIRULENCE OF THE BACTEBIOPHAGE 101 the thousandth tube of the series is represented, in cubic kilo- meters, by the figure 103982. To appreciate this incommensurable figure, it is sufficient to say that with only the twenty-second passage, one drop of the original filtrate taken from the feces has been diluted in a number of cubic kilometers of liquid expressed by the number 1070. That is to say, by a number of which the logarithm has for a characteristic 70; which would be a cube of liquid of such size that it would require a billion centuries for a ray of light to pass through from edge to edge.2 The action is not, therefore, to be explained as a persistence of anti-typhoid bacteriophagous germs through a series of suc- cessive cultures. According to the conception of Maurice Nicolle a bacterium may be considered as a mosaic of properties. Each of these properties: resistance to heat, vitality, virulence for such and such an animal, etc., is susceptible, within a single individual, of continuous variation. Within a bacterial culture, and at any given moment, no two individual bacteria can be found possessing exactly the same properties. This conception, demon- strated by daily experience, applies moreover to all living beings. Variation, that is, the property of adaptation, is an attribute of life and of life exclusively. Like all living beings, the bacterio- phage adapts itself continually, and in any culture the ultrami- crobes which compose it do not all possess exactly the same prop- erties. Some are susceptible of rapid adaptation toward a given bacterium, others toward another organism. A bacteriophagous ultramicrobe is a mosaic of properties. 2 It is evident that the same mathematical reasoning demonstrates that the bacteriophage is itself a living being. If one would wish to explain lysis as due to the presence of a lytic diastase in the intestinal contents (or, what actually amounts to the same thing, the presence of a co-ferment or a catalyzer in the intestinal tract capable of activating a pro-diastase contained in the bacterium), the diastase or the catalyzer or the co-ferment would be quickly eliminated by the dilution. If we suppose possible the persistence of one of these principles, in spite of the dilution which approxi- mates infinity, and its presence at each point in an incommensurable amount in the liquid, we are endowing this principle with the metaphysics of ubiquity. Any conception of transmissible serial bacteriolysis which does not admit as the origin of the phenomenon an autonomous living being, ends in a mathematical absurdity. 102 THE BACTERIOPHAGE When a bacteriophage is virulent, as it comes from the organ- ism, for several bacteria at the same time, as is the usual case, it is apparent that the virulence present for each of these bacteria is subject to variations with time. This is true, however, the virus may be preserved, whether it is kept, sealed in tubes, in the original intestinal contents or whether it is preserved in the form of filtrates prepared from the fecal material. When kept in vitro certain strains of bacteriophage lose their virulence for a bacterium, toward which they were active when derived from the body, much more rapidly than do others. Experiment XXIX. Typhoid patient Mor Examination of the stool was made at the beginning of convalescence. On August 20th, 1918, the stool was treated according to the method described for securing the bacteriophage. The filtrate is distributed in 0.5 cc. amounts in suspensions of the following bacteria : — B. dysenteriae Shiga, B. typhosus, B. paratyphosus A, B. paratyphosus B, and B. coli. After 24 hours of incubation these suspensions were planted on agar with the following results : B. dysenteriae Shiga Sterile B. typhosus Sterile B. paratyphosus A Numerous plaques B. paratyphosus B Numerous plaques B. coli Sterile Specimens of the feces and of the filtrate were preserved in sealed tubes. On January 22nd, 1919, that is, after 5 months, these materials were exam- ined again: SUSPENSION RES ULT Freshly prepared filtrate Original filtrate B. dysenteriae Shiga. Sterile Sterile B. typhosus Normal culture Normal culture B . paratyphosus A Normal culture Normal culture B. paratyphosus B Numerous plaques Numerous plaques B. coli Numerous plaques Numerous plaques In this material the virulence of the bacteriophage for B. dys- enteriae and for B. paratyphosus B remained unaltered during the five months, it diminished for B. coli, and disappeared entirely for B. typhosus and B. paratyphosus A. VIKULENCE OF THE BACTERIOPHAGE 103 It should be noted that the result was the same whether the bacteriophage was preserved directly in feces or in the filtrate, that is, in bouillon. Likewise, it is significant that the degree of virulence has no influence on the preservation or the disappear- ance of the virulence. It was strong for B. typhosus and became negative, it was weak for B. paratyphosus B} yet this remained intact. In the absence of passages, simply as an effect of old age, the virulence of the bacteriophage varies then with time, and indeed in a different manner for the diverse bacteria attacked. It be- comes attenuated more quickly for some than for others, and for this no general rule can be fixed. We have seen above, with another strain, that after four years and in spite of passages in contact with the dysentery bacillus, the virulence for B. typhosus persisted. The last experiment cited is not only interesting then, in that it shows an attenuation of virulence associated with the lapse of time, but also in that it gives evidence that the loss does not occur in equal degrees for all of the bacteria attacked by one and the same bacteriophage. BACTERIAL SPECIES ATTACKED Let us now consider the different bacteria for which active strains of the bacteriophage have up to the present been isolated.3 For each of these only the peculiarities of the reaction as they are 3 In certain cases the bacteriophage can serve for the identification of bacteria, as the agglutination reaction is used. In order to apply the test a bacteriophage must be employed which has been subjected to numerous passages at the expense of a particular bacterial type so that the accessory virulences may be attenuated as far as possible. For example, all strains of bacteria which are lysed by a strain of bacteriophage that has been cultivated together with Shiga bacilli, are certainly B. dysenteriae Shiga. With certain species, B. pestis for example, for which the specificity of the bacteriophage appears high, diagnosis by means of the bacteriophage is particularly conclusive. In this last connection it may be said that since the publication of the French edition of this text, a bacteriophage has been encountered active for B. pestis and equally active for the bacillus found in pseudotuberculosis of the guinea pig. Thus, it is necessary to recognize the lack of an absolute specificity, limiting somewhat the value of the reaction as applied to the identification of bacterial species. 104 THE BACTERIOPHAGE encountered in dealing with the bacterium in question will be mentioned. As far as general characteristics are concerned, all are similar, that is, what has been recorded in preceding chapters regarding the method of isolation, the mode of action, the variable virulence, the enhancement of virulence by passage, the resistance of the bacteria, and secondary cultures, applies to all strains of bacteriophage and to all species of bacteria attacked. In all of the experiments mentioned up to the present time the dysen- tery bacillus has been taken as an example, but it has been shown that all such experiments may be repeated with identical results with any strain of bacteriophage active for a definite organism. B. dysenteriae presents practical advantages in experimentation. It is easy to isolate a very active bacteriophage for this organism and bacteriologists can readily procure strains and repeat these experiments without conducting a long series of preliminary investigations. B. dysenteriae Shiga For this organism it is particularly easy to isolate a very active strain of the bacteriophage. The bacteriophage opposed to this bacillus exists, it may be said to be normally present, in the in- testinal tract of numerous animals, the horse and domestic fowls m particular. It is likewise frequent in man and may acquire a high virulence, not only in convalescence from an attack of dysentery, but in recovery from a variety of pathologic conditions. Up to the present time about 200 strains have been isolated, without finding any two of exactly comparable virulence and of equal extent in their action on the related bacteria of the colon- typhoid-dysentery group. Among these 200 strains, one only, and that of a moderate activity when isolated, has failed to act upon any other bacteria of the group. A strain of bacteriophage active for B. dysenteriae Shiga is usually active for B. coli and for B. dysenteriae Flexner and Hiss. From the point of view of the bacteriophage the Shiga type of dysentery bacilli represents a homogeneous species, a bacterio- phage active for one strain being equally active for all others. A bacteriophage very active for one strain of bacilli, at the ex- pense of which it has passed through a number of passages, may VIRULENCE OF THE BACTERIOPHAGE 105 be less active for a freshly isolated race, but after four or five passages in contact with this strain a virulence is acquired equal to that possessed for the first one. We have seen that the strains of Shiga bacilli which resist the action of the bacteriophage are extremely toxic, possessed of a great vitality, inagglutinable by a specific serum, and actively ferment maltose.4 B. dysenteriae Hiss An anti-Hiss bacteriophage is frequently found in the nor- mal intestine. A bacteriophage showing an activity for any member of the colon-typhoid-dysentery group frequently shows a virulence, more or less pronounced, for the Hiss bacillus. B. dysenteriae Hiss represents a homogeneous species from the point of view of bacteriophagous activity. Secondary cultures reinoculated into litmus sugar media do not ferment the sugars in the same way as do normal bacilli. Media containing glucose, maltose, and mannite become acid after ten days; those containing lactose, levulose, saccharose, and also glycerine remain alkaline. After a month the lactose, saccharose and levulose media remain alkaline. Secondary cul- tures, and also mixed cultures, give the indol reaction but do not react on either neutral red or lead acetate. The resistant bacilli are inagglutinable, have a high viability, and are more virulent for man. In Part II of this monograph we will consider a case of B. dysenteriae Hiss septicemia in which the bacillus was re- sistant to the action of the bacteriophage. B. dysenteriae Flexner The anti-Flexner bacteriophage is found in the normal intestine of vertebrates as frequently as are the other strains of bacterio- phage.5 4 Pottevin has shown that the normal Shiga bacillus definitely, although somewhat weakly, ferments maltose. Resistant bacilli ferment the sugar much more energetically. 6 The presence of an active bacteriophage in pathologic conditions is not considered here. This phase will be discussed in Part II of this text. 106 THE BACTERIOPHAGE All strains isolated, active for the Flexner bacillus, have like- wise been active for B. coli, although with some, activity for other varieties of dysentery bacilli was lacking. With reference to the bacteriophage, Flexner bacilli constitute a homogeneous species. Resistant bacilli ferment glucose, levu- lose, maltose, and mannite. They do not ferment lactose, do not blacken lead acetate in an agar medium, and do not react on neutral red. They form indol. They are inagglutinable by a specific serum and possess a high viability. The atypical character of certain strains of B. dysenteriae when freshly isolated from the organism may surely be ascribed to their resistance to the bacteriophage. Elsewhere we will con- sider a typical case. Furthermore, this observation is of general significance, applicable not to dysentery bacilli alone. B. dysenteriae "X" During the course of these investigations a very great number of specimens of feces, derived from patients with intestinal dis- turbances, have been examined. And in many cases of gastro- enteritis in adults as well as in infants a bacillus having the fol- lowing characteristics has been isolated: When inoculated on litmus sugar agar media it fails to ferment any of the sugars tested (lactose, glucose, levulose, saccharose, maltose, mannite, galactose). It causes no change in lactose and maltose Barsiekow medium, but this medium containing glucose and mannite is turned red. It is agglutinated by con- valescent serum in titres of 1:100 to 1:500, is not agglutinated by anti-Flexner or anti-Shiga sera. With a serum which agglu- tinates the Hiss strain to 1:2500 the "X" strain is agglutinated in dilutions of 1:200. It is non-motile, is morphologically like the other dysentery organisms, is Gram-negative, and is toxic for rabbits. Several strains of bacteriophage active for this bacillus have been isolated. This bacteriophage is constantly present in the intestine in convalescents who have shown B. dysenteriae "X" in their stools during the infection. Strains have also been recovered from the intestinal tracts of healthy animals, both man and other animals. The "X" bacillus constitutes a homo- geneous species as regards the bacteriophage. VIEULENCE OF THE BACTERIOPHAGE 107 Certain strains of bacteriophage active for B. dysenteriae "X" were likewise active for other species of dysentery bacilli, others were virulent for only one or two among them. When maintained for several generations at the expense of B. dysenteriae "X" they almost completely lose their activity for other dysentery organisms. B. coli An anti-coli bacteriophage is extremely frequent in the feces of normal vertebrates and invertebrates, but only exceptionally is it found possessed of any considerable virulence. On the other hand, in recovery from the most varied pathologic conditions very active strains can be isolated. B. coli, particularly when recently isolated, constitutes a heterogeneous species as regards the bacteriophage. In the presence of a bacteriophage possess- ing a very high virulence for certain coli strains, other races are hardly touched, some are even absolutely refractory. When taken from the body, B. coli always shows a degree of resistance. In the intestine it forms with the bacteriophage a mixed culture. On artificial culture media the resistance decreases very slowly with successive transfers. With a colon organism of maximum resistance, that is, one which is completely refractory, the colonies on agar are large, white, fluent, exactly like those of the bacillus of Friedlander. B. typhosus Quite frequently a strain of bacteriophage showing a slight activity for B. typhosus can be isolated from the normal intestine. The isolation of a very active strain is exceptional since such are found only in convalescents. A single strain of bacteriophage may show a very great variation in virulence for different races of B. typhosus, certain races being entirely resistant to a given bacteriophage although they may be very susceptible to other strains of the bacteriophage. In such a case there is a natural resistance, a true natural immunity, a condition which can be demonstrated not only for the typhoid bacillus but also for B. coli, the paratyphoid bacilli, B. proteus, etc. It is because of such reactions that these organisms are spoken of as belonging to 108 THE BACTERIOPHAGE species which are not homogeneous as regards the bacteriophage. Typhoid bacilli may acquire, in vivo or in vitro, a resistance to the action of the bacteriophage, that is, they may possess an acquired immunity (this must not be confused with natural im- munity) and they become inagglutinable as well as possessed of an enhanced virulence for experimental animals. As has been shown for B. dysenteriae, repeated culturing on agar progressively lowers the resistance to the bacteriophage, and coincidently restores agglutinability. B. paratyphosus A A bacteriophage showing virulence for this bacillus is relatively frequent in the normal intestine. As regards the action of the bacteriophage, B. paratyphosus A strains form a more homogeneous species than do the typhoid bacilli. Just as with B. typhosus, the paratyphoid A organisms may acquire a resistance to the bacteriophage, may become inagglutinable, and may show an increased virulence. B. paratyphosus B A bacteriophage for this organism is very frequent in normal stools. The resistance of B. paratyphosus B places this bacillus intermediary between B. typhosus and B. paratyphosus A. Re- sistant bacteria are inagglutinable and are of high virulence. Mixed colonies on agar, in which the bacterium has acquired a high resistance for the bacteriophage, present a viscid appearance resembling B. Friedldnder. Salmonella (hog cholera) One strain of bacteriophage active for this organism has been isolated from a normal man. When derived from the body it possessed an average virulence. B. typhi murium The anti-paratyphoid B strains of the bacteriophage are some- times endowed with virulence for B. typhi murium. Very active strains have been isolated from the intestinal tracts of white and VIRULENCE OF THE BACTERIOPHAGE 109 gray rats which were rendered experimentally resistant to the disease caused by the ingestion of cultures of the bacillus. The resistant bacilli are very virulent and can be used for the destruc- tion of gray rats, a large proportion of which resist the action of the ordinary virus. It is possible that human infection may be feared because of this increased virulence. The transitory appearance of such a bacteriophage in the blood of several infected white rats has been demonstrated. Such rats were resistant to infection. B. proteus Two strains of bacteriophage very active for this bacillus have been isolated from the stools of two infants having a gastro- enteritis. The virulence of these strains was tested against a dozen strains of B. proteus of different origins. Only three of the strains tested were affected by these strains of the bacterio- phage, the same three in both cases. The other nine proteus strains were non-susceptible. Included in this last group were two strains of B. proteus Xi*. The lysate secured through the interaction of a bacteriophage on a proteus suspension, is, immediately after the lysis, extremely toxic for rabbits. Indeed, they are killed within a few hours by the subcutaneous injection of but half of a cubic centimeter. After ten days the lysate loses its toxicity. B. gallinarum (Klein); B. gallinarum (Moore); B. paragallinarum Characterization of these bacilli will be reserved for the chapter in which avian typhosis is discussed. With the exception of pathogenicity for man B. gallinarum presents all the characteris- tics of B. typhosus, including agglutinability to the titre of the serum with an anti-typhoid serum. There are, as we will see, at least three different species of paragallinarum organisms. The bacteriophage active for B. gallinarum is not effective with all species of paragallinarum, nor is the anti-paragallinarum bacteriophage active for the other races. B. gallinarum is a very homogeneous species. The anti-gallinarum bacteriophage is constantly present in the intestinal tracts of fowls which resist 110 THE BACTEEIOPHAGE infection. It has been isolated from the blood of three fowls which were recovering from the infection. Outside of epizootic foci it has been found in the intestine of healthy animals. Bad. diphtherias Two strains of bacteriophage active for only atoxic strains of Bact. diphtheriae have been isolated from the feces of two horses immunized by the injection of cultures of diphtheria bacilli. This observation is only mentioned. Lack of opportunity has prevented further examination of these strains, a study which certainly offers much of interest. Staphylococcus A bacteriophage active for the Staphylococcus has been isolated under the following circumstances. A guinea pig bit my left index finger and on the next day an inflammation appeared which persisted for three days. On the fourth day there was an accumu- lation of pus, of which about ten drops were secured. When planted directly upon agar there developed one colony of Sta- phylococcus albus and six of Staphylococcus aureus. The remainder of the pus was mixed with twenty cubic centimeters of bouillon and placed in the incubator for 24 hours, then filtered through infusorial earth and a bougie. After five passages at the expense of Staphylococcus albus lysis was secured. At first the bacterio- phage isolated did not show any activity in vitro for Staphylococ- cus aureus. It has been possible to develop a virulence for this organism only after about fifty passages in a mixed Staphylococ- cus aureus and Staphylococcus albus suspension, following the technic previously indicated for the enhancement of the latent virulence of an anti-staphylococcus bacteriophage toward B. dysenteriae Shiga. Bacterium of barbone About thirty strains of bacteriophage for this organism have been isolated, of which twelve were extremely active. Their activity was comparable when tested against different bacterial strains, some derived from Italy, others f rom Indo-China. We VIRULENCE OF THE BACTEEIOPHAGE 111 will return to a consideration of this bacteriophage later, in the chapter dealing with barbone (hemorrhagic septicemia of the buffalo) . When isolated from the body all the strains presented an average or feeble virulence toward the different intestinal bacteria. After about a dozen passages at the expense of the bacterium of bar- bone these accessory virulences became markedly attenuated. B. pestis Twelve strains of bacteriophage active for B. pestis have been isolated. Eleven of these were secured from the excreta of rats in the different villages of Indo-China where plague was epidemic. The twelfth was derived from the feces of a patient convalescent from plague. This is the only strain which has been maintained. This strain was also active for the bacillus of pseudotuberculosis of guinea pigs. Bacillus of flacherie The bacillus in question was isolated from the bodies of silk- worms which had died of a disease presenting the characteristics of flacherie in the breeding establishments in Indo-China. The bacteriophage is frequent in the intestine of the healthy worms among a contaminated stock. The activity of this bacteriophage was the same for each of the three strains of the bacillus tested, isolated from three different breeding places. B. subtilis One strain of bacteriophage active for this bacillus was secured in the stools of a patient with dysentery. Having but rarely tested the virulence of the bacteriophage toward B. subtilis it is impossible to say if this virulence is frequent or exceptional. Vibrio cholerae Among about one hundred cases of cholera studied in Indo- China it was possible to observe but one following recovery. In this last, in spite of daily examination of stools, in but a single specimen taken at the beginning of convalescence has a bacterio- 112 THE BACTERIOPHAGE phage active for the vibrio been found. This gave about fifty plaques when planted on agar. In spite of many attempts it has been impossible to cultivate it by serial transfers. None of the fatal cases yielded a bacteriophage. The diversity of the bacterial types against which, up to the present time, virulent strains of bacteriophage have been iso- lated, suggests the idea that the activity of the bacteriophage may be manifested toward any bacterial species whatsoever. CHAPTER IV THE BACTERIOPHAGOTJS ULTRAMICROBE Morphology. Viability. Susceptibility to Different Substances. Unicity of the Bacteriophage. Lysins of the Bacteriophage. Opsonic Power of the Lysins. MORPHOLOGY The bacteriophagous ultramicrobe is of extreme tenuity. In a medium containing the bacteriophage the ultramicroscope reveals only some very minute brilliant points. Probably each of these points represents an ultramicrobe, particularly since their abundance, in greater or lesser numbers, corresponds some- what with the counts made upon agar. Its tenuity is such that a medium containing several thousand million ultramicrobes per cubic centimeter appears perfectly limpid. The ultramicrobe is, however, resident in a definite mass, since each element is deposited on agar in distinct points and this mass must be appre- ciable since the ultramicrobes spontaneously sediment in the course of time. Experiment XXX. A culture of an anti-dysentery bacteriophage is filtered through a bougie and allowed to stand without moving in a cup- board for eleven months. At the end of this time, specimens of the cul- ture from the surface and from the bottom of the tube are taken with capillary pipettes. The count of the superficial layers showed 280,000,000 per cubic centi- meter. The count of the deeper layers showed 2,900,000,000 per cubic centi- meter. The ultramicrobe can be sedimented, although incompletely, by centrifugation at very high speed. Experiment XXXI. Twenty-five cc. of the bacteriophage (antidysen- tery) are filtered through a bougie and are centrifuged in a Jouan appara- tus for 30 minutes at 12,000 revolutions per minute. Counts show the following: 113 114 THE BACTERIOPHAGE per cubic centimeter Before centrifugation 1,750,000,000 Surface 50,000,000 )n\Bottom 3,700,000,000 Dialysis through collodion membranes of various permeabilities gives a rough approximation of the size of the bacteriophagous ultramicrobe. As a test, one part of horse serum was mixed with three parts of a culture of the bacteriophage, and the mixture was subjected to dialysis. Whenever the albumin passed through the filter the bacteriophage passed also, and when the permeability was such that the albumin was held back, so also was the bac- teriophage. The bacteriophage then, passes through when the molecule of serum albumin passes and is retained when the latter is held back. It remains for physicists to more exactly determine its true size, and this determination will be of more interest since the bac- teriophage is the only ultramicrobe with which such measurement is actually possible, since it is the only one where the elements can be counted. Thus, it may serve to clear up an important point touching the constitution of organized matter. If one calculates the ultramicrobe as being one one-hundredth of a micron in diameter, it ought to contain about twenty molecules of albumin and five or six atoms of sulfur. Physicists have de- termined the size of the pores in the most dense collodion mem- branes as being not greater than two millionths of a micron. But the ultramicrobe of avian plague penetrates such a membrane. Each element can not be greater than one five-thousandth of a micron in diameter, hence it would be composed of one-tenth of a molecule of albumin. On the other hand, an ultramicrobe is indeed a markedly complex organism, capable of adaptation, possessing the faculty of secreting toxins, — the diastases, — having in a word, the characteristics of living matter. This in itself implies a relatively complex organism. We find ourselves, then, cornered by an absurdity, for it is impossible to conceive of a complex organism formed of a single molecule, much less of the tenth part of one. It will be much more simple to admit that it is impossible to understand under what aspect life is present in the ultramicrobe and under what form the matter composing it exists. BACTERIOPHAGOTJS ULTRAMICROBE 115 It is only since the discovery of the bacteriophage that it has been possible to affirm that each ultramicrobe is a material mass capable of multiplication in the form of like masses. Thanks to it, our ideas regarding viruses have acquired some degree of pre- cision. The study of the bacteriophage by physicists would offer findings of extreme interest, for it is the only virus demon- strated by experiment to exist in particulate form, and with this alone is it actually possible to fix dimensions, thanks to the possi- bility of recognizing the number of elements present in a liquid. VITALITY The bacteriophagous ultramicrobe is extremely resistant toward the majority of destructive agents, a property which it shares, moreover, with other ultramicrobes. The vitality is very great. Filtrates or cultures containing the bacteriophage are still active after six years when preserved in a sealed tube. However, not all of the germs present in a culture show the same degree of resistance. After preservation for four years a culture which originally contained two thousand million ultramicrobes per cubic centimeter contains only about one hundred millions of living organisms. Such vitality is not exceptional, for certain bacteria, not spore-forming, show a re- sistance of the same order. For example, cultures of B. coli are still cultivable after ten years or so, and here also the different bacilli of a culture do not offer the same resistance, for the num- ber of those which survive becomes smaller and smaller with time. If a culture of the bacteriophage is allowed to evaporate slowly at room temperature it is found that living germs may be found in the few drops of syrupy fluid remaining in the bottom of the tube. Indeed, certain bacteria act in the same way. On the contrary, living organisms are no longer to be found after twelve months in glucose bouillon cultures, although they may still be alive in lactose bouillon. In fecal material preserved at room temperature in sealed tubes for thirty-four months (September, 1915, to July, 1918) one may recover the living bacteriophage, as active as at the beginning. This experiment has been performed successfully with four specimens of feces from convalescent cases of dysentery. 116 THE BACTEBIOPHAGE In a dry state the bacteriophage is resistant for a long time. A fragment of sterile filter paper is saturated with a drop of a bacteriophage culture (anti-dysentery), dried in the air, and pre- served in a sealed tube for six months at room temperature. After this time the piece of paper is introduced into a suspension of B. dysenteriae and normal, although delayed, lysis is obtained. The ultramicrobes have, therefore, survived. Another ultrami- crobe, that of the tobacco mosaic, has the same property. It remains alive for two years in the dried leaves. It is indeed, unnecessary to search for examples of bacteria as resistant as the ultramicrobes. The cocco-bacillus of locusts is a non-sporu- lating bacillus, but in the cadavers of locusts dead of the disease which it incites in these insects (cadavers dried over sulfuric acid, pulverized, and preserved in sealed tubes for three years). I have shown that the cocco-bacillus remains alive and virulent, for this powder, seeded into bouillon, gives normal cultures viru- lent for the locust. SUSCEPTIBILITY TO DIFFEKENT AGENTS1 Physical agents: Effect of temperature At the beginning of my experiments I stated that the tempera- ture of destruction of the bacteriophage is about 65°C. Shortly after this Kabeshima in an early report cited 70 to 75°C., and in a later note 70°C. Very recently Gratia and Jaumain have noted that the lethal temperature showed considerable variability; 61 or 62°C. for the lytic principle acting on the staphylococcus and 65°C. for that acting on the colon bacillus. For testing this point I had taken as a criterion the ability or inability of material subjected to different temperatures to pro- duce lysis of a bacterial suspension. Moreover, this has been the method adopted by the other investigators who have consid- ered this question. In view of these contradictory results, the effect of temperature has been reconsidered, taking as a criterion, not lysis of a suspension in a fluid medium, but the action of a culture on solid media, a procedure much more delicate. 1 The experiments dealing with the effects of temperature have been made in collaboration with E. Pozerski. BACTERIOPHAGOUS ULTRAMICROBE 117 Experiment XXXII. In the following experiments the culture of bac- teriophage under test, previously filtered through a bougie, is taken up in capillary pipettes, sealed at both ends, and completely submerged in a water-bath maintained at the temperatures indicated in each experiment. In each series of experiments 8 tubes with culture are maintained for thirty minutes at temperatures of 60, 62, 64, 66, 68, 70, 72, and 75 °C. Anti-Shiga bacteriophage Two drops of the culture from tubes maintained at 60, 62, 64, and 66°C., when introduced into suspensions of Shiga bacilli, cause complete lysis in less than fourteen hours. The tests repeated with a second strain of Shiga bacilli give identical results. The bacteriophage heated to 68 and 70°C. causes lysis with one strain of Shiga bacilli but not with the other. When heated to 72 and 75°C. the bacteriophage fails to cause lysis. One drop of each of these suspensions, which had received the bacterio- phage cultures previously maintained at 68, 70, 72 and 75 °C., and which had not been submitted to lysis, are planted on slant agar. After incuba- tion, all of the cultures, except the last, which is normal, show plaques characteristic of the presence of the bacteriophage. Serial passages may be effected, thus permitting the enhancement in virulence of the bacteriophage attenuated by the action of temperature. After two such passages, with the ultramicrobe heated to 68 and 70°C., and after three passages with that which was heated to 72°C., lysis in liquid media is obtained. Comparable experiments have demonstrated that the bacteriophagous ultramicrobes active for B. dysenteriae Flexner, B. dysenteriae Hiss, B. coli, and B. paratyphosus B, act in a quite similar manner. With the bacterio- phage active for B. paratyphosus A attenuation begins at about 64°C. (at least with the strain tested). With that virulent for B. typhosus attenua- tion is already apparent at about 62°C. In all cases, when heated to 75°C. the bacteriophage is completely inactive, either actually destroyed or attenuated to such an extent that its presence can no longer be detected. In all these instances the bacteriophage shows a recuperative power, the virulence being restored when the temperature to which the virus has been subjected is not higher than 72°C. Anti-staphylococcus bacteriophage Attenuation of this bacteriophage is already manifest after heating to 60°C. Subcultures of suspensions which have not been lysed show that it is a simple attenuation, for, even with suspensions inoculated with a bac- teriophage previously held at 72°C. for thirty minutes, plaques are obtained characteristic of the presence of an active bacteriophage. Moreover, two passages suffice to restore the original virulence to cultures heated to 62, 64, 66, and 68°C. After heating at 70 and 72°C. the attenuation of virulence does not disappear until after six passages. When heated to 75°C. the bacteriophage is deprived of all activity. 118 THE BACTEKIOPHAGE It may be concluded from these experiments that all strains of the bacteriophage react to temperature in the same manner. When heated above 60°C. they are attenuated more or less rapidly according to the bacterial species upon which they operate. All are completely killed, or at least paralyzed, at about 75°C. Chemical agents The bacteriophagous ultramicrobe will attack bacteria either in the presence or absence of oxygen, or indeed in an atmosphere of either nitrogen or hydrogen. Antiseptics It is interesting to study the action of antiseptics for these substances do not act in the same manner on certain ultramicrobes as on ordinary bacteria. While very sensitive to the action of certain antiseptics, they are very resistant to others. A culture of anti-dysentery bacteriophage in physiological saline2 is distributed into four tubes. One serves as a control. The second receives mercuric chlorid to a concentration of 1:200, the third receives copper sulfate to a concentration of 1:100, and to the fourth phenol is added in a concentration of 1 : 100. After contact with these substances for three days the bacterio- phage is living in all the tubes. After four days it is killed in the tubes containing the mercuric chlorid and the copper sulfate. After seven days it is killed by the carbolic acid. It remains alive in the control tube. It is not killed after contact for a week in a fluid saturated with essence of thyme or of cloves, but its lytic action is not mani- fested there. The same results are secured in media containing chloroform or sodium fluoride (Bablet). Eliava and Pozerski have determined with precision the lethal limits, in 24 hours, of concentrations of free H and OH ions. The zone compatible with life lies between pH 2.5 and 8.54, corre- sponding approximately to an acid 1/160 N, and a base 1/260 N, whatever may be the acid or alkali tested. 2 Bouillon is not suitable for this work because of the precipitates which form upon the addition of certain antiseptics. These give rise to entirely false results. BACTERIOPHAGOUS ULTRAMICROBE 119 It is difficult to make a direct comparison with the limits of resistance of other micro-organisms, the bacteriophagous ultra- microbe is at present the only one for which such determinations have been made.3 The action of glycerine is very interesting. The bacteriophage remains alive for at least two years in a fluid composed of equal parts of glycerine and bouillon or physiological saline. A sus- pension of bacteria in such a medium, a medium in which the bacteria are unable to reproduce, is lysed as perfectly as in or- dinary bouillon, yet in a higher concentration of glycerine the bacteriophage is destroyed. Bablet has in fact shown that when 0.5 cc. of bacteriophage is added to 9.5 cc. of glycerine the ultramicrobe is killed in six days. It may be well to recall that glycerine constitutes the best medium for the conservation of the toxins and diastases. The other known ultramicrobes resist, in general, the action of glycerine. We have seen that a suspension in a glycerine medium may be lysed by the bacteriophage, and, as always, the lysed culture becomes a culture of the bacteriophage. If we allow such a cul- ture to evaporate slowly at room temperature we will finally have a residue composed of glycerine, all the water being evaporated. Under such conditions, the bacteriophage becomes adapted to its environment and remains alive in the glycerine residue, al- though it is killed if transferred directly from a bouillon culture to concentrated glycerine. Many instances are known of the adaptation of bacteria to antiseptics; and adaptation is a func- tion of living matter. 8 1 may cite, for example, the following findings which have been reported, not on the zone of life, but on the zone of growth. G. Dernby (Ann. de 1'Inst. Pasteur, 1921, 35, 277) gives the following figures: Staphylococcus pH 4.8 to 8.1 B. subtilis pH 4.5 to 8. 5 B. proteus pH 4.4 to 8.4 B. coli pH 4. 4 to 7.8 The bacteriophage is, therefore, extremely sensitive to the action of bases and acids, since its fatal limit of alkalinity is the same as the limit for growth of ordinary bacteria ; its fatal acid limit is not far distant. It is, therefore, more sensitive than bacteria to concentrations of free H and OH ions. This is further evidence of its living nature. 120 THE BACTERIOPHAGE Eliava and Pozerski have shown that the neutral salts of qui- nine exert an antiseptic action on the bacteriophage, in three per cent solution killing it in thirty minutes, in one per cent, in a few hours. Emetine hydrochlorate and saponine in the same con- centrations are without action. This susceptibility to the anti- septic action of quinine is singular in a germ which is otherwise relatively resistant to antiseptic activities. It is hardly possible to deduce that the bacteriophage is protozoan in nature, for quinine exerts antiseptic properties toward many bacterial species, although on the contrary, it is without action on the diastases and toxins. When a culture is treated with acetone the albuminoid materials of the bouillon are thrown out of solution, and this precipitate encloses the bacteriophagous virus. The greater portion of the virus is, however, destroyed. The bacteriophage reacts like the spore-forming bacteria, which are found, living, in the precipi- tate. In this connection a curious thing has been noted. Ordinarily acetone is considered a sterile fluid, but it is not necessarily so, since it has been found that several containers of acetone have been contaminated by B. subtilis. Alcohol gives the same precipitate, but the bacteriophage, con- trary to that which happens with the virus of the tobacco mosaic, is killed in less than forty-eight hours in 90 per cent alcohol. The precipitate, as we will see, contains the secretory products of the bacteriophage. UNICITY OF THE BACTERIOPHAGE In the preceding chapters detailed experiments have been given which show that whatever the bacteria attacked, the ul- tramicrobes which attack them belong always to the same species. We will return to other proofs shortly. A single statement, grouping these experiments will be given here. First. Usually a single strain of bacteriophage will attack several species of bacteria at the same time. Second. A strain of the bacteriophage, continued through more than a thousand passages in vitro, always in conjunction with the same bacterial strain, namely, B. dysenteriae Shiga, attacks B. typhosus and B. coli. BACTBRIOPHAGOUS ULTRAMICROBE 121 It has likewise been shown that a bacteriophage active for the staphylococcus and maintained through more than one hundred transfers with this staphylococcus still possessed virulence for the dysentery bacillus. And it is also possible to effect, in vitro, the adaptation for this Shiga bacillus of a strain of bacteriophage active only for Staphylococcus aureus. The staphylococcus and B. dysenteriae are bacterial species but remotely related and the crossed reaction constitutes an irrefutable argument in favor of the unicity of the bacteriophage. Third. An antibacteriophagous serum, the properties of which will shortly be considered, contains an amboceptor specific for the bacteriophage, as is demonstrated in the complement fixa- tion reaction of Bordet-Gengou, and this amboceptor is the same for all species of bacteriophage — the anti-dysentery bac- teriophage, the anti-plague bacteriophage from man, the anti- plague bacteriophage from the rat, and the anti-barbone bac- teriophage from the buffalo, all fix complement in the presence of serum from a rabbit treated by repeated injections of cultures of the anti-dysentery bacteriophage.4 For this particular ex- periment strains of bacteriophage were selected which failed to show a crossed reaction in vitro with regard to the different bac- teria attacked. The complement fixation reaction is specific with respect to species differentiation. The proofs of the unicity of the bacteriophage are therefore multiple. There is but a single bacteriophage, common to both man and animals, capable by adaptation of acquiring a virulence toward all bacterial species. As we have seen in the earlier chapters the bacteriophagous ultramicrobe can not be cultivated in any artificial medium. It is an obligatory parasite, capable of reproduction only within living cells. Moreover, this is the case with all known ultrami- crobes. The single one making the exception to this rule, the Asterococcus of pleuropneumonia is hardly longer to be consid- ered as an ultramicrobe, since it has been shown to be perfectly 4 Since the publication of the French edition of this text Bruynoghe and Maisin have confirmed this fact. They have also shown that fixation of complement is also to be obtained with an anti-staphylococcus bacterio- phage under the conditions already mentioned. 122 THE BACTERIOPHAGE visible in stained preparations, as a bacterium of minute size. All the parasitic ultramicrobes are intracellular parasites, for the lesions which they produce are, in all cases, characterized by protoplasmic inclusions or alterations in the nuclei of the cells. The bacteriophagous ultramicrobe differs from the other known ultramicrobes only in its elective action for unicellular organisms. The others act in multicellular organisms. It would be indeed strange that of all living organisms the bac- teria alone should enjoy the privilege of absolute immunity. Such an immunity must have seemed remarkable even before the dis- covery of the bacteriophage, before ultramicrobes sufficiently small to parasitize them had been recognized. The ultramicrobe is in diameter certainly 2000 times smaller than a bacterium of average size, in volume nearly 2000 million times less. In size, one of these ultramicrobes is, to a bacterium, as the bacterium is to a large fly. It should be remembered, however, that although up to the present time parasitism of bacteria has not been recognized we have for a long time observed and studied many parasites which incite infectious disease among the protozoa. Several examples will be found cited among the works of Metchnikoff.5 It may be well to mention a study of Dangeard6 entitled "Sur les parasites du noyau et du protoplasma," for the facts disclosed by this investigator offer certain analogies to those presented in the preceding chapters. But there are these differences, namely, the parasite of Dangeard attacks a protozoan, and its dimensions are such that it can be readily observed microscopically and therefore classified. The observations of Dangeard deal with an Oomycete, Nucleo- phaga amoebae Dangeard, which parasitizes the nucleus of Amoeba verrucosa Ehr. The Amoeba verrucosa has a large, doubly-con- toured, spherical nucleus, and also a nucleolus, likewise spherical, whose diameter is about two-thirds that of the nucleus. The substance of the nucleolus is very dense and stains with great s Legons sur la pathologic comparee de V inflammation, Paris, 1892, Masson & Cie. L'immunite dans les maladies infectieuses , Paris, 1901, Masson & Cie. 6 Sur les parasites du noyau et du protoplasma, Le Botaniste, 1894/95, 4, 199-248. BACTEEIOPHAGOUS ULTRAMICBOBE 123 intensity with various nuclear staining reagents. Between the nucleolus and the nuclear membrane is a space filled with the nuclear fluid. The zoospore of Nuckophaga amoeba first penetrates the proto- plasm of the amoeba but it never develops there; it passes into the nucleus through Ijie membrane which it perforates, most cer- tainly through the aid of a dissolving diastase. Dangeard has demonstrated the portal of entrance of the parasite as a minute circular opening, as though made by a punch, persisting after the entrance of the parasite. After its penetration into the nu- cleolus the parasite resembles a refractile corpuscle, increasing slowly in size in proportion as the nuclear substance disappears. When this nuclear material has been utilized completely the entire interior of the nucleus is filled and the membrane is distended. At this time the nucleus of the parasite, up to the present time single, actively divides and when sporulation is effected there are about one hundred regularly spaced nuclei. About each of these nuclei a zoospore organizes, and a sporangium is thus formed, containing distinct, rounded corpuscles, which contain nuclei at the time of sporulation. Frequently a single amoeba is parasitized by two or perhaps several zoospores, and in such cases each develops separately and gives birth to a distinct sporangium. When the sporangium reaches maturity the protoplasm of the amoeba disintegrates, the sporangium ruptures, freeing the young zoospores, and these become distributed throughout the medium, ready to parasitize the healthy amoebae in their neighborhood. It is evident that I have not made any comparison between Nucleophaga amoeba and Bacteriophagum intestinale, and that these observations are mentioned simply because there is a cer- tain resemblance between the two phenomena of destruction, that of the amoeba and that of the bacterium. THE LYSINS OF THE BACTERIOPHAGE7 It is obvious that the bacteriophage is unable, merely by its presence, to dissolve a bacterium. This action can only be accomplished through the agency of lytic diastases. 7 The experiments dealing with the lysins have been performed in col- laboration with G. Eliava. 124 THE BACTEBIOPHAGE In a culture of bacteriophage the lysins which effect the solu- tion of the bacteria ought to remain in solution when lysis is com- pleted. On the other hand, the ultramicrobe does not resist treatment with alcohol. Therefore, in order to obtain lysin it is only necessary to subject the culture of bacteriophage to the classic procedure for the separation of diastases. If we mix one volume of bacteriophage culture (anti-dysentery) with nine volumes of 96 per cent alcohol, after contact for 48 hours the precipitate which is formed is well compacted and the super- natant fluid may be decanted. The precipitate, which contains the lysins admixed with all the substances of the medium precipi- table by alcohol, is almost completely soluble in saline.8 Experiment XXXIII. Precipitate a culture of anti-dysentery bacterio- phage with alcohol and dissolve the precipitate in a quantity of 0.8 per cent saline equal to the original volume of the culture. Mix equal parts of this solution and bouillon and add a B. dysenteriae suspension sufficient to give a slight turbidity. As a control, prepare a tube containing an equal volume of bacilli suspended in a medium half bouillon and half saline. Place these tubes in an incubator at 37°C. After twenty-four hours the control is turbid, the bouillon containing the lysin is slightly cloudy. Plantings on agar from the two tubes give normal bacillary growths. After forty-eight hours the control presents the same appearance and agar inoculation gives a perfect growth. The culture containing the lysin is slightly cloudy and inoculations on agar give only isolated colonies. A count shows that there are 22 times less living bacilli in the last culture than in the control tube. After three days the appearance is the same as after forty-eight hours. After four days the bacteria begin to develop a resistance to the action of the lysin, the medium becomes cloudy and inoculations on to agar again give a film of growth. At no time does one obtain on the agar the plaques character- istic of the presence of the bacteriophage, and the action is not continued in series. The alcohol precipitate therefore contaias a lytic diastase, free of living ultramicrobes. The dissolving action, although definite, is weak; but on the contrary as we will see later, the lysin manifests itself by an extremely powerful opsonic action. 8 This manipulation should be carried out aseptically, since filtration of the fluid through a bougie considerably weakens its activity. BACTERIOPHAGOUS TJLTRAMICBOBE 125 OPSONIC POWER OF THE LYSINS In the following experiments the opsonic power has been de- termined by the method of Wright and Douglas, making a mix- ture of one part of the fluid of which the opsonic action is to be measured, one part of a suspension of leucocytes, and one part of a suspension of the bacteria against which the opsonic effect is to be determined. The mixture is aspirated in a capillary pipette which is sealed and placed in the water-bath at 38°C. for fifteen minutes. The contents of the pipette are then spread on a slide, stained, and examined. As reagents we have taken: Guinea pig leucocytes, a culture of an anti-Shiga bacteriophage, and a suspension of Shiga bacilli. Experiment XXXIV 1. Control. Leucocytes, bacilli, ordinary bouillon 100 leucocytes phagocytize 36 bacilli Opsonic index = 1 2. Leucocytes, bacilli, culture of bacteriophage two years old 100 leucocytes phagocytize 692 bacilli Opsonic index = 19.2 3. The same mixture, except the bacteriophage culture is diluted 1:250 100 leucocytes phagocytize 156 bacilli Opsonic index = 4.3 4. Leucocytes, bacilli, culture of bacteriophage six days old 100 leucocytes phagocytize 1510 bacilli Opsonic index = 41 . 9 5. The same mixture, except the bacteriophage culture is diluted 1 :250 100 leucocytes phagocytize 146 bacilli Opsonic index = 4.1 6. The same mixture, except the bacteriophage culture is heated at 60 QC. for 30 minutes 100 leucocytes phagocytize 728 bacilli Opsonic index = 20.2 7. The same mixture, except the bacteriophage culture is heated and diluted to 1 : 250 100 leucocytes phagocytize 101 bacilli Opsonic index = 2.7 In the mixtures 2, 4, and 6, the indices recorded represent a minimum. Many leucocytes contain such a number of phagocytized bacilli that counting is impossible. Since in the above counts only those cells 126 THE BACTERIOPHAGE which did not contain masses of bacteria have been included, the actual index is therefore somewhat higher.9 The opsonic action of a culture of the bacteriophage manifests itself with such rapidity that it is improbable that the opsonic power can be exercised directly by the ultramicrobes. We have seen, in fact, that the bacteria are parasitized only after an ap- preciable lapse of time, — ten to twenty minutes. Experiment XXXV. Leucocyte suspension, Shiga suspension, and anti-Shiga bacteriophage culture are mixed in equal parts. After various periods of incubation drops of the mixture are examined showing : TIME INTEKVAL NUMBER OF BACILLI IN 100 LEUCOCYTES INDEX Immediately after mixing 197 5.4 After 2^ minutes 362 10.0 After 5 minutes 372 10.3 After 7^ minutes 440 12.2 After 10 minutes 824 23.0 After ten minutes some of the leucocytes are so completely filled with bacilli that counting is impossible. The figure given is a minimum based only on leucocytes in which masses of bacteria were not present to interfere with enumeration. The opsonic power must be exercised, not by the ultramicrobes, but by the lysin contained in the culture, as the following ex- periment proves. Experiment XXXVI. Two milligrams of the alcohol precipitate (of which we have spoken above), still moist, are dissolved in 10 cc. of 0.8 per cent saline. A mixture is made of equal parts of this solution, sus- pension of Shiga bacilli, and leucocytic suspension. After 15 minutes at 38°C. microscopic examination of stained preparations shows that 100 leucocytes have taken up 536 bacilli (as certain leucocytes contain masses rendering a count impossible the figure is a minimum). The index is 14.9. The opsonic power of cultures of the bacteriophage is, there- fore, due to the lysin secreted by the ultramicrobes, to the lysin which remains in the culture once the bacteria have been dis- 9 It will be recognized that the opsonic indices obtained with sera are far below these secured with cultures of the bacteriophage. With the former an index of 2 is exceptional. BACTERIOPHAGOUS ULTBAMICROBE 127 solved. It is interesting to note its action on bacteria resistant to the action of the bacteriophage. Experiment XXXVII. Mix equal parts of a suspension of Shiga bacilli resistant to the action of the bacteriophage, an anti-Shiga bacteriophage culture, two years old, and a suspension of leucocytes. After fifteen minutes, 100 leucocytes have ingested 8 bacteria. The index is thus 0.22, or 90 times lees than with normal bacilli (experiment XXXIV, 2). Prepare a similar mixture, but with a bacteriophage culture six days old. Here, 100 leucocytes have phagocytized 13 bacilli. The index is 0.38, or 108 times less than with normal bacilli (experiment XXXIV, 4). Another mixture is made, using the lysin solution, 100 leucocytes have phagocytized 19 bacilli. The index is 0.53, or, 28 times less than with normal bacilli (experiment XXXVI). From this it is clear that bacteria which resist the bacteriophage also resist phagocytosis. The same experiment has been performed with a strain of the antibarbone bacteriophage and the bacterium of barbone. The results are comparable. Experiment XXXVIII (A) 1. Control. Mixture of equal parts of leu- cocyte suspension, bouillon, and suspension of the bacterium of barbone. After fifteen minutes at 38°C. there are no bacteria in 100 leucocytes. 2. Mixture of equal parts of leucocyte suspension, the suspension of the bacterium of barbone, and an anti-barbone bacteriophage culture, 8 months old. After fifteen minutes 100 leucocytes have ingested 109 bacteria. 3. The same mixture, except that the bacteriophage culture is diluted 1 : 250. One hundred leucocytes have phagocytized 52 bacteria. 4. Mixture of one-third leucocyte suspension, one-third bacterial sus- pension, and one-third solution of the alcohol precipitate of a recent culture of the anti-barbone bacteriophage (2 mgm. of precipitate in 10 cc . of saline) . One hundred leucocytes have phagocytized 239 bacteria. (B) A mixture is made of equal parts of leucocyte suspension, culture of the bacterium of barbone, and a fresh (four days old) culture of anti- barbone bacteriophage. During incubation at 38°C. drops taken for examination show: Immediately, in 100 leucocytes there are 30 bacteria. After two and one-half minutes, in 100 leucocytes there are 139 bacteria. After five minutes, in 100 leucocytes there are 201 bacteria. After seven and one-half minutes, in 100 leucocytes there are 271 bacteria. After ten minutes, in 100 leucocytes there are 269 bacteria. In the control mixture made with bouillon no bacteria were phago- cytized. 128 THE BACTERIOPHAGE Here it is impossible to calculate the opsonic indices, since no phagocytosis occurred in the control mixture. The indices are infinity. The bacterium of barbone which resists the action of the bac- teriophage is also resistant to phagocytosis. Experiment XXXIX. Prepare a mixture of one-third leucocytic sus- pension, one-third of the same culture of anti-barbone bacteriophage as that used in the preceding experiment, and one-third of a suspension of the bacterium of barbone resistant to lysis. After fifteen minutes at 37°C. 100 leucocytes have ingested 3 bacteria, that is to say, 90 times less than with normal bacteria. The strain of anti-dysentery bacteriophage employed in the experiments previously described manifests a definite, although feeble, lytic action against B. typhosus. The following experi- ments show that it also exerts a definite opsonic action on this bacillus. Experiment XL. 1. Mix equal parts of bouillon, leucocyte suspension, and B. typhosus suspension. After fifteen minutes at 38°C. 100 leucocytes have phagocytized 68 bacilli. The opsonic index is 1 . 2. Mix equal parts of anti-Shiga bacteriophage culture, leucocyte sus- pension, and typhoid suspension, After fifteen minutes 100 leucocytes have ingested 203 bacilli. Opsonic index = 3. 3. Mix equal parts of leucocyte suspension, typhoid suspension and lysin solution (the same one as that employed in the experiments with the dysentery bacillus). After fifteen minutes 100 leucocytes have ingested 109 bacilli. Opsonic index = 1.6. The lysin possesses a property which is indeed peculiar. When used in the complement fixation reaction as an antibody it func- tions as an amboceptor. The experiment cited below is taken from among many others which gave identical results. Experiment XL1. Antigen: This is prepared according to the method of Maurice Nicolle. One loopful of an agar culture of B. dysenteriae Shiga is suspended in 4 cc. of saline. This suspension, heated at 100°C. for five minutes, then cooled, serves as antigen. Antibody: An alcohol precipitate of a culture of anti-Shiga bacteriophage taken into solution in a quantity of saline equal to the original volume of the culture acts as antibody. BACTEKIOPHAGOUS ULTBAMICROBE 129 The complement is fresh guinea pig serum, titrated. The hemolytic system is the usual anti-sheep system. TUBE ANTI- GEN ANTI- BODY COMPLE- MENT SALINE HEMO- LYTIC SYSTEM RESULT CC. CC. CC. CC. i p CC. 1 0.5 0.2 0.2 1.6 f g 1 + + 2 0.5 0.4 0.2 1.4 ® £ 1 -J- -j- 3 0.5 0.5 0.2 1.3 *i g 1 + + + 4 0.5 0.6 0.2 1.2 .3 ^ 1 + + + 5 0.5 — 0.2 1.8 1 1 1 Complete hemolysis 6 — 0.6 0.2 1.7 || 1 Complete hemolysis, rapid 7 — — 0.2 2.3 1 Complete hemolysis 8 — — — 2.5 J I 1 + + + + By itself, the lysin does not fix complement, on the contrary, as shown by tube 6, it rather activates hemolysis. There is, there- fore, an antibody associated with the lysin which fixes the com- plement. It is difficult to reach a conclusion regarding this curious ex- periment. Later investigations will show that there is a relation between the amboceptor of Bordet and the lysin of the bacteriophage, since it acts as two different principles, acting in an identical manner, in so far as complement fixation is concerned. All of these experiments show that the bacteriophagous ultra- microbe secretes a principle, precipitable by alcohol, resisting a temperature of 58°C., and persisting for several months in the cultures of the bacteriophage. This principle, aside from its solvent action, exercises a powerful opsonic action upon the bac- teria for which the ultramicrobe from which it is derived possesses a virulence. The opsonic activity appears proportional to the virulence of the bacteriophage for the bacterium under consid- eration. Bacteria which have acquired a resistance to the bac- teriophage are equally resistant to phagocytosis. In addition, from another viewpoint, we will see that they possess an increased virulence. CHAPTER V THE BACTERIOPHAGOUS ANTISERUMI Complexity of the Antibodies. Antibodies to the Bacteria. Antibodies to the Bacterial Toxins. Antibodies to the Bacteriophagous Ultrami- crobes. Antibodies to the Lysins. Incidental Conditions Resulting from the Existence of the Bacteriophage. COMPLEXITY OF THE ANTIBODIES The phenomena here involved are exceedingly complex. It is known that when the body is injected with a bacterial culture it responds with the production of diverse principles which are grouped under the name "antibodies." Some of these act upon the bacterial bodies: the agglutinins, amboceptors, opsonins; others, the antitoxins and antiferments, neutralize the secretory products of the bacteria formed in the culture fluid injected. When a mixture of two bacterial cultures is injected, the body responds with a duplicate series of antibodies. This takes place when a culture of the bacteriophage is injected. A culture of the bacteriophage is composed, as we know, of a culture or a suspension of a bacterium lysed by the action of the bacteriophage directed against and endowed with virulence for this bacterium. The bacteriophagous germs inoculated have multiplied at the expense of the bacterial bodies found there and when lysis is terminated the bacterial substance is dissolved in the medium. A culture of the bacteriophage is, then, a com- plex medium which contains: a. The substance of the bacterial bodies in a dissolved state. 6. The bacterial toxins (exo- or endotoxins). c. The bacteriophagous ultramicrobes which have developed at the expense of the bacteria. 1 The experiments performed on the bacteriophagous antiserum have been made in collaboration with G. Eliava. 130 BACTERIOPHAGOUS ANTISERUM 131 d. The products resulting from the activity of the bacterio- phage, which we have grouped under the name of "lysins," and which remain in the medium after the lytic process is completed. Does the bacteriophage attack the bacterium by means of a single diastase or through a combination of diastases? At this time it is impossible to say, and indeed, it would not materially affect the question with which we are concerned. There is still another category of substances present in the culture. We have seen that the bacteria do not remain passive to the action of the bacteriophage, and that this defense is ac- companied by the production of an anti-diastase — an anti-lysin — which is likewise to be found in the medium. This should, then, stimulate the formation of anti-anti-lysins. These have not been investigated, and it is only sugested that they may possibly be present in the serum of immunized animals. As an example of an antibacteriophagous serum we will take the antibacteriophage-Shiga serum. This is particularly interest- ing because of the potent endotoxin of the dysentery bacillus. A rabbit is injected with four doses of a culture of the anti- dysentery bacteriophage2 that is to say, of a lysed culture of B. dysenteriae Shiga, amounting to two, four, six, and eight cubic centimeters, with an interval of six days between each injection. The rabbit is bled fifteen days after the last injection. Theoreti- cally, the serum of this rabbit ought to contain the following anti- bodies: a. Antibodies to the bacteria: Amboceptor and agglutinin. b. Antibody to the bacterial toxin: Antitoxin. c. Antibodies to the bacteriophagous ultramicrobe : Ambo- ceptor and agglutinin. d. Antibody to the lytic diastase of the bacteriophage: Anti- lysin. Let us see if the antibodies present in such a serum actually correspond with those which theoretically should exist. 2 As we will see in regard to barbone of the buffalo, the serum of an animal which has received a single and minimal injection of a bacteriophage culture does not present the antibacteriophagous property, or, at least, if it exists, it is not detectable. 132 THE BACTERIOPHAGE ANTIBODIES TO THE BACTERIA The dysentery bacilli are agglutinated by the antibacterio- phagous serum, and this serum contains also an amboceptor which permits the fixation of complement. The presence of such antibodies is inevitable and is obtained by the injection of any material containing the dysentery bacilli, living or dead, intact or dissolved. The presence of such antibodies is without especial significance. ANTIBODIES TO THE BACTERIAL TOXIN In the present case these antibodies should neutralize the dysen- tery endotoxin. The serum of an animal prepared by the in- jection of dysentery bacilli, living or dead, contains an antitoxin.3 On the other hand, a culture of B. dysenteriae lysed by the bac- teriophage contains a toxin, for if experimental animals are in- jected with such a culture a short time after lysis the animals die as though they had received a lethal dose of the toxin of Nicolle. The serum of an animal treated with such cultures ought to con- tain an antitoxin. This can be verified. Experiment XL1I. A mouse receives by subcutaneous injection a lethal dose of the dysentery toxin prepared by the method of Nicolle, and at the same time 0.5 cc. of the bacteriophage-Shiga antiserum. A second mouse receives the same amount of toxin and 0.5 cc. of an anti-dysentery serum. A third mouse receives a lethal dose of the toxin only. The first mouse dies in about thirty hours after the injection, the second lives, the third dies four days after the injection. The bacteriophage-Shiga antiserum is therefore not antitoxic; indeed, on the contrary, it is definitely sensitizing. Let us con- sider this singular phenomenon further. 3 The antidysenteric serum furnished by the Pasteur Institute is derived from horses treated by injections of dysentery toxin secured according to the method of Rowland, as modified by Maurice Nicolle. The bacterial bodies are ground with anhydrous sodium sulfate, the powder obtained is dried in the air, and dissolved in water at the time of injection. The turbid fluid thus obtained is centrifuged, and the clear supernatant portion is used for the injection. The serum neutralizes the endotoxin, as animal experimentation shows. BACTERIOPHAGOUS ANTISERUM 133 Experiment XLIII. Four mice receive subcutaneously a dose of toxin equal to one-tenth of the lethal dose. The first is held as a control. Two others receive 0.2 cc. of the bacteriophage-Shiga antiserum, the last 0.1 cc. of this serum. The first remains perfectly well indefinitely. The two which received the 0. 2 cc. dose of serum die after 40 hours, and the last one after fifty-four hours. This experiment proves that the bacteriophage-Shiga anti- serum sensitizes the animal to the action of the toxin. It should be stated here that whatever the number of lethal doses of the toxin of Nicolle injected into a mouse, death never occurs before the fourth day. Here, when the antibacteriophage serum is added to the toxin, even to a dose below the minimal lethal dose, death takes place within forty-eight hours. Instead of toxin, let us take living dysentery bacilli and see the effect of the antiserum on injections of this nature. Experiment XLIV . Four mice receive subcutaneously a dose of dysen- tery bacilli equal to one-fifth the lethal dose. The first mouse is held as a control. The second receives, subcutaneously, 0. 2 cc. of the antibacterio- phage-Shiga serum, the last two 0. 1 cc. of this serum. The control animal lives, showing nothing abnormal. Those which at the same time received the serum die in seven to nine days after the injection, after showing during the last twenty-four hours a paralysis of the posterior extremities. In general, mice do not show this symptom after the injection of B. dysen- teriae; only the rabbit shows this particular symptom. The result is clear-cut; in all cases the antibacteriophage-Shiga serum is sensitizing. This is, incidentally, the first example of an anti-immunizing serum. It is possible that the antibacteriophage-Shiga serum actually contains an antitoxin but that this is masked by the presence of a powerful "sensibilisine." This is the more plausible, for we shall see in the chapter dealing with immunization, that rabbits which have received but a single minute injection of a culture of the antidysentery bacteriophage are effectively vaccinated against the effects of the toxin. The sensibilisine develops in the animal only after the second injection. This condition is not peculiar to the case of dysentery ; we find it again when we consider immuni- zation against barbone. This phenomenon of sensitization invites much further investi- gation, which will permit, without doubt, an extension of our 134 THE BACTEKIOPHAGE knowledge of the nature of antitoxic immunity. One thing already appears certain, namely, that the bacteriophage must play some r61e in the manifestation of this immunity, since the antibacterio- phagous serum sensitizes to the toxins. It is probable that this sensitization of the animal must be associated with the inhibiting power with which the antibacteriophage serum is endowed, an action which we will shortly consider. ANTIBODIES TO THE BACTERIOPHAGOTJS ULTRAMICROBES Agglutinins It has been impossible to demonstrate definitely the presence of agglutinins, although experimentation indicates that their presence is probable. In all cases, if present, they are weakly active. Experiment XLV. When a culture of the bacteriophage is centri- fuged for 15 minutes at 3000 revolutions per minute no trace of sediment appears; it requires a speed of at least 12,000 revolutions to obtain an appreciable amount . A culture of the bacteriophage is mixed with one-tenth of its volume of antibacteriophage serum and centrifuged for ten minutes at 3000 revolutions. Counts of the ultramicrobes in the sediment and in the supernatant fluid (after energetic shaking of the latter for five minutes) shows that there are about three times as many germs in the sediment as in the supernatant fluid. Is this agglutination? It is possible, but not certain. The experiment is not sufficiently clear-cut to permit an affirmative answer. The formation of agglutinins in the serum of treated animals is subject to great variation, dependent upon the bac- terial species injected. With Vibrio cholerae and with B. typhosus for example, potent agglutinins are secured; with certain coli- form bacilli and with B. Friedldnder they are usually so weak that their action is almost inappreciable. We know nothing regard- ing the formation of agglutinins in the case of the pathogenic ultramicrobes. Amboceptors The test for an amboceptor specific for the anti-dysenteric bacteriophage, detectable by the complement fixation reaction BACTERIOPHAGOUS ANTISEBUM 135 of Bordet and Gengou, can not be made. In effect, the culture of the anti-dysentery bacteriophage is only a suspension of ul- tramicrobes in a liquid containing the dissolved substance of the dysentery bacilli, so that the antibacteriophage-Shiga serum con- tains two amboceptors, one specific for the dissolved substance, the other for the ultramicrobe, and it is impossible to separate the two actions. A fixation of complement would always be obtained without the possibility of knowing with which of the two antigens it had been effected. However, the question can be solved in another manner; a manner which is very conclusive. We have considered in the preceding chapters the experiments which demonstrate the unicity of the bacteriophage. If the hypothesis based on these experiments is true the amboceptor present in the antibacteriophage-Shiga serum ought to fix comple- ment with all ultramicrobes and the detection of an amboceptor ought to be possible, for working with a culture of bacteriophage other than the antidysenteric there would be nothing to inter- fere with the reaction. A culture of antiplague bacteriophage, for example, is a suspension of ultramicrobes in a fluid containing the substance of B. pestis. The only possible common element in a culture of antidysentery bacteriophage and in a culture of antiplague bacteriophage is the bacteriophagous ultramicrobe. And the single "anti" element contained in an antibacterio- phage-Shiga serum capable of exercising an action toward such a plague culture can only be an amboceptor for the single element common to all cultures of the bacteriophage; the bacteriophagous ultramicrobes themselves. The complement fixation reaction has been conducted utiliz- ing as antibody the antibacteriophagous-Shiga serum; as anti- gens, cultures of bacteriophage for dysentery, plague of human origin, an anti-plague strain from rats, and an anti-barbone strain from the buffalo. Experiment XLVI. Fixation of complement. Antigen: culture of anti-Shiga bacteriophage containing 2,000,000,000 ultramicrobes per cubic centimeter. Antibody: antibacteriophage-Shiga serum. With the antigen in an amount of 1 cc. and antibody in quantities of 0.2 and 0.1 cc. fixation is positive. 136 THE BACTEKIOPHAGE Moreover, the antigen by itself fixed complement. Experiment XLVII. Antigen : culture of anti-Shiga bacteriophage con- taining 100,000,000 ultramicrobes per cubic centimeter. Antibody: antibacteriophage-Shiga serum. Positive fixation was secured with mixtures containing the antigen in 1 cc. amounts and the serum in 0. 2 and 0.1 cc. quantities. The antigen by itself did not fix complement. Experiment XLVIII. Antigen: B. dysenteriae Shiga. Antibody: antibacteriophage-Shiga serum. Fixation of complement occurred. Experiment XLIX. Antigen: culture of the anti-Shiga bacteriophage. Antibody: an antidysentery serum. Complement is fixed. Experiment L. Antigen: culture of the anti-Shiga bacteriophage. Antibody: antibacteriophage-Shiga serum. The antigen, antibody, and complement are incubated together for 1 hour at 37°C., and then the hemolytic system is added. The following protocol shows the results. TUBE ANTI- GEN ANTI- BODY COM- PLE- MENT 1:20 SALINE HEMO- LYTIC SYSTEM RESULT CC. CC. cc. CC. CC. 1 0.75 0.2 0.2 0.35 1 + + + + 2 0.50 0.2 0.2 0.60 1 + + + + (optimum) 3 0.25 0.2 0.2 0.85 1 + + + 4 0.75 0.1 0.2 0.45 1 + + + + 5 0.75 0.3 0.2 0.25 1 + + + + 6 0.75 — 0.2 0.55 1 Complete hemolysis 7 0.75 0.2 0.3 0.25 1 + + + (excess complement) 8 0.75 0.2 0.4 0.15 1 -)- + (excess complement) 9 — 0.3 0.2 1.00 1 Complete hemolysis Experiment LI. Antibody: antibacteriophage-Shiga serum. Antigens: Four different antigens are employed, as follows: I. Anti-Shiga bacteriophage, 1,500,000,000 ultramicrobes per cubic centimeter. II. Anti-barbone bacteriophage, 250,000,000 ultramicrobes per cubic centimeter. III. Anti-plague bacteriophage, a strain derived from the rat, 450,000,000 ultramicrobes per cubic centimeter. IV. Anti-plague bacteriophage, a strain derived from a case of plague in man, 700,000,000 ultramicrobes per cubic centimeter. The results secured with these four different antigens are shown in the following table. BACTERIOPHAGOUS ANTISERUM 137 ANTI- GEN ANTI- BODY NORMAL RABBIT SERUM COM- PLE- MENT 1:20 SALINE HEMO- LYTIC SYSTEM ANTIGENS I II ill IV CC. CC. CC. CC. CC. CC. 1 0.2 — 0.2 1.1 1 + + + + +++ ++++ ++++ 1 0.1 — 0.2 1.2 1 + + + + -}--f +-H+ ++++ — 0.2 — 0.2 2.1 1 CH CH CH CH 1 — — 0.2 1.3 1 CH CH CH CH — — — 0.2 2.3 1 CH CH CH CH 1 — 0.5 0.2 1.3 1 + + + + = Complete inhibition CH = Complete hemolysis Experiment LII. The antibody and the antigens are the same as those used in the preceding experiment . The following table presents the results. ANTI- GEN ANTI- BODY NORMAL RABBIT SERUM COM- PLE- MENT 1:20 SALINE HEMO- LYTIC SYSTEM ANTIGENS I II ill IV CC. CC. CC. CC. CC. CC. 1 0.05 — 0.2 1.25 1 + + + + + + + + + + + ++++ 1 0.025 — 0.2 .30 1 + + + + _|__j_ _!__{_ ++++ 1 0.005 — 0.2 .30 1 -f-f-f C H C H CH 1 — 0.05 0.2 .25 1 C H C H C H C H 1 — 0.025 0.2 .30 1 C H C H C H C H 1 — 0.005 0.2 .30 1 C H C H C H C H - 0.05 - 0.2 2.25 1 Complete hemolysis — — 0.05 0.2 2.25 1 Complete hemolysis 1 — — 0.2 1.30 1 C H C H C H C H — - - 0.2 2.30 1 Complete hemolysis ~ ~ mm — 2.50 1 No hemolysis It will be noted (experiment LI) that the antigen slightly fixed complement with normal rabbit serum. This is not strange since the bacteriophage is a normal inhabitant of the intestine. The antibacteriophage serum contains, therefore, an amboceptor specific for the bacteriophage, whatever the strain may be, what- ever the species of bacteria attacked, and whatever the animal species from which it is derived. There is but one bacteriophage. 138 THE BACTERIOPHAGE ANTIBODIES TO THE LYSINS4 We have already seen that it is possible to obtain lysins without admixture with viable bacteriophagous ultramicrobes by pre- cipitating a culture of the bacteriophage with alcohol. Since growth of the bacteriophage takes place within the in- terior of the bacteria the ultramicrobe can effect its penetration only by corroding the wall of the bacterium in order to make way for its passage, and it is evident that it can do this only by means of a lysin. If the antibacteriophage serum contains an antilysin it is evident that the penetration will be retarded or even pre- vented by the presence in the medium of the neutralizing serum. And this delay or prevention, according to the amount and po- tency of the serum, will be associated with a delay or prevention of the lysis, since the ultramicrobes will be unable to penetrate the bacterial cells. The antibacteriophage serum will assure, in a word, the protection of the bacteria, without actually exercis- ing any action on the vitality of the virus itself. This is, in fact, what is actually observed. The antibacteriophage serum used in these experiments pos- sessed a considerable inhibiting power; 0.00001 cc. added to 10 cc. of a suspension of dysentery bacilli — this is one-millionth of the final volume — markedly retarded lysis. With 0.001 cc. lysis was prevented. Obviously, this fact might be interpreted as a destruction of the bacteriophage pure and simple. But this would be indeed strange in view of the fact that a serum has never destroyed a bacterium in vitro, even in the presence of comple- ment. I know that this affirmation is contrary to universal opinion touching the bacteriolytic action of antisera, but in Part II of this work I will show the foundation for it by an ex- periment which permits of no doubt. In so far as the action of an antibacteriophage serum upon lysis is concerned, the following experiment shows that the bac- teriophage is not destroyed; its action is only inhibited. 4 To repeat what is meant by lysin : The aggregate of the secretions of the bacteriophage, without prejudging that they operate as a diastase only, or as a collection of diastases, as is more probable. It has been shown in an earlier chapter that the bacteria, like higher organisms, react to the lysins by the production of antilysins. BACTERIOPHAGOUS ANTISERUM 139 Experiment LIII. Prepare a mixture of equal parts of antibacterio- phage-Shiga serum and culture of anti-Shiga bacteriophage. Allow them to remain in contact for five days. They are placed under conditions such that if the serum exerted a destructive action on the ultramicrobes its effect could not fail of manifestation . After these five days of contact, three tubes, each containing 10 cc. of sterile bouillon, are seeded with a drop of a bouillon culture of Shiga bacilli. To the first of these tubes add a drop of the mixture of &ntiserum-bacteriophage; to the second add a drop from this first tube after it has been shaken ; and to the third add a drop from the second. We have them a series of three tubes, planted with B. dysenteriae Shiga, containing decreasing concentrations of the serum-bacteriophage mixture. After twenty-four hours at 37°C. normal cultures of Shiga are obtained in the three tubes, and plantings on agar likewise give normal cultures. Up to this point it looks as though the lytic principle has been destroyed. Continue the experiment. Replace the three tubes in the incubator and twenty-four hours later it is seen that lysis has commenced in the first tube of the series. Agar inoculation from this tube remains sterile. The last two, on the contrary, still contain a normal culture of B. dysenteriae. Return the tubes again to the incubator. After twenty-four hours, that is, three days after the beginning of the experiment, lysis takes place in the last two tubes. All subcultures on agar remain sterile. The bacteriophage is therefore not destroyed by the antibac- teriophage serum; its power is simply inhibited for a time. The experiment further shows that the action is truly inhibi- tive, acting upon the entire number of ultramicrobes. In other words, the delayed lysis is not due to the revival of certain ultra- microbes particularly resistant, since lysis is produced even in the last tube which has received, as a result of successive dilu- tion, an infinitesimal quantity of the germs. The presence of an antilysin in the antibacteriophage serum allows us to obtain further information regarding the nature of the virulence of the ultramicrobe. Experiment LIV. To a suspension of the bacterium of barbone add a drop of a bacteriophage culture active against this organism and then 10 drops of an antibacteriophage-Shiga serum, that is, a quantity of serum completely inhibitive of lysis in a suspension of Shiga bacilli. Prepare also a control without the serum. In both tubes lysis proceeds normally and in a parallel fashion. Here, then, the serum has exerted no inhibitory action. 140 THE BACTERIOPHAGE The inhibitive effect is exercised only against the strain of bacteriophage which has been used to inject the animal in the preparation of the antiserum.5 On the other hand we know that strains of the bacteriophage differ one from the other only in the virulence which they have acquired, by adaptation, for this or that bacterium. It follows that the lysin secreted by the bac- teriophage is different for the different bacteria attacked, since the antilysin neutralized only the lysin of the strain which has served in the treatment of the animal furnishing the serum. Each bacterial species requires for its lysis, then, the production of a specific lysin. Virulence for a given bacterium is, therefore, in the last analy- sis, the power possessed by the bacteriophage to secrete a lysin specific for this bacterium. The bacterial species are divided into groups. In each group the species which compose it present certain common character- istics. For example, we have the colon group, the typhoid group (B. typhosus and the paratyphoid bacilli), the dysentery group (the Shiga, Flexner, and Hiss types, etc.). These three groups are indeed closely related one to another. On the other hand, we see the Pasteurella group (bacteria of the diverse hemorrhagic septicemias, chicken cholera, barbone, etc.), the staphylococcus group, and so on. Each bacterial species requires that it be attacked through the secretion of a specific lysin and the differ- ence between these specific lysins will be slight in passing from one bacterial species to another in the same group or in a neigh- boring group. The bacteriophage adapts itself rapidly. A single strain is in fact generally active for all the bacteria of the group, or for the organisms of the most closely related ones. On the contrary, adaptation is difficult of acquisition in passing from a bacterium of one group to an organism of a remotely related group. Furthermore, the bacteriophage normally parasitizes the intestinal bacteria which constitute its habitual culture medium. 6 It is evident that if the antibacteriophage-Shiga serum is tested against closely related bacterial types, forms for which the bacteriophage would have a certain activity, an inhibitive effect more or less pronounced will be noted. BACTERIOPHAGOUS ANTISEBUM 141 It retains for a long time its hereditary faculty of attacking the organisms of the colon-typhoid-dysentery group, even if it is cultivated for many generations at the expense of another bac- terial species. INCIDENTAL CONDITIONS RESULTING FROM THE EXISTENCE OF THE BACTERIOPHAGE I wish to note certain incidental consequences which spring from the facts which we have considered. Although accessory, these consequences are of some practical significance and it may be well to mention them. Because of the ubiquity of the bacteriophage and its con- stant presence in all living beings, and because of the resistance which the bacteria are able to oppose to its action, and which gives birth to the phenomenon of mixed cultures, it is henceforth necessary to verify the purity of bacterial cultures from the point of view of possible contamination not only by another bacterial species, but also by an ultramicrobe. We will see, for example, that in B. coli pyelonephritis, the pathogenic agent is always a colon bacillus which is resistant to the bacteriophage. If one plants the urine from a case of pyelone- phritis on agar for the purpose of isolating the etiological agent one is always liable to find mixed colonies of the colon bacillus and the bacteriophage. Subculture from these mixed colonies will give mixed cultures, indefinitely cultivable in this form. Such cultures are usually considered pure, for they contain no other organism visible under the microscope. They are, how- ever, contaminated by the bacteriophagous ultramicrobe; they are not "ultrapure," and investigations undertaken with such mixed colon-bacteriophage cultures may furnish peculiar results, especially if they are used in immunological experimentation. It is not to be assumed that I have mentioned an exceptional case, far from it, as the two following examples demonstrate. In two different instances, both accidental findings, I have demonstrated that cultures of the colon bacillus isolated origi- nally from cases of cystitis in the hospital, were in reality mixed cultures. In both instances it was easy to isolate from these cultures a very active bacteriophage. 142 THE BACTERIOPHAGE Here is another instance of the same order. Recently investi- gating a bacteriophage active for the streptococcus of gourme of horses, Doctor Forgeot sent me three different strains of Strep- tococcus equi, taken from the culture collection of the Central Veterinary Laboratory. Of these three strains, only one was pure. The two others were in reality mixed cultures of the streptococcus and the bacteriophage. These two examples suffice to give an idea of the great num- ber of mixed cultures which must actually exist among stock cul- tures. Not only are the intestinal bacteria subject to contamina- tion by the bacteriophage, but bacteria in general, for we will see that the bacteriophage does not remain restricted to the intestinal tract, but passes into the circulation and exercises its action in the different organs. It is therefore quite essential before undertaking any experi- mental work to ascertain if the bacterial culture involved is not in reality a mixed culture, composed of a resistant bacterium and a bacteriophage. One must be sure that the culture is not only pure, but ultrapure. The mutations noted by different authors may most certainly be attributed to the frequency of these mixed cultures. And the confirmation of these reports has been lacking for the very simple reason that the verification has been attempted, not with strains contaminated by the bacteriophage, but with cultures really pure. The experimental results have thus very naturally differed. In this regard I might say that it is indeed singular, in view of the unique character of certain conceptions concerning the bacteriophage, that not a single author has yet traced to the fact that bacterial cultures are frequently contaminated by the bacteriophage, the conclusion that the bacteriophage takes origin spontaneously in these cultures. In reality, as we see it, in con- nection with the genesis of these cultures, each time that there is a contamination by the bacteriophage, it is in the form of a mixed culture. The bacteriophage exists in the culture from the beginning, that is, from the time of isolation. In the experiments touching immunity, the contradictory results may likewise well reside in the presence of a bacteriophage in the cultures employed in the experiments. We will see in the BACTERIOPHAGOUS ANTISEBUM 143 second part of this work, the role which the bacteriophage plays in immunity, and it is evident that the experimental results in immunological investigation will be entirely different if the bac- terial strain employed is, or is not, contaminated by the bacterio- phage; whether it is a resistant strain or a normal strain. In a word, the idea of the existence of the bacteriophage im- poses the obligation of always verifying the bacterial cultures with a view to determining that they are, not simply pure, but ultrapure; and this under penalty of obtaining entirely false ex- perimental results as a result of the possible presence of the bacteriophage. CHAPTER VI THE NATURE OF THE BACTERIOPHAGE Nature of the Bacteriophage. The Number of Possible Hypotheses. Experimental Proofs of the Living Nature of the Bacteriophage. Refu- tation of the Hypothesis of Kabeshima. Refutation of the Hypothesis of BordetandCiuca. Refutation of the Hypothesis of Bail. Refutation of the Hypothesis of Salimbeni. Conclusions. THE NATURE OF THE BACTERIOPHAGE All of the facts which have been recognized up to the present time and which have been recorded in the preceding chapters have been confirmed by all authors who have investigated the question. The phenomena themselves have never been the sub- ject of controversy, and because of their definiteness, it might be said because of their violence, and because of the facility with which they can be reproduced, they can not be controverted. The discovery of the bacteriophage was associated with a study of a disease of locusts, in which I for the first time noted in the intestine of the insects which resisted the infection a principle antagonistic to the action of the pathogenic cocco-bacillus; a principle which could be demonstrated by its effects but which of itself could not be isolated. With this suggestive observation as a basis, I systematically sought for a comparable principle in the intestinal contents of patients with enteric infections. Finally, during the year 1915, in studying an epidemic of dysentery which prevailed in a squadron of cavalry stationed in the neighborhood of Paris, I noted the phenomenon of plaques in the cultures on agar tubes. Shortly after, from the stools of a patient under treatment in the Pasteur Hospital I was able to isolate by filtra- tion the antagonistic principle. The study was continued dur- ing the rare moments which my duties as Chief of the Laboratory Service for the Preparation of Vaccines permitted. (More than twenty million doses of vaccines were furnished by the Service to the Allied Armies during the war.) I tried particularly to 144 NATURE OF THE BACTERIOPHAGE 145 determine the nature of this principle, and it was only after I had considered all possible a priori hypotheses, multiplying the control experiments which had demonstrated experimentally with certainty that the principle could be only an autonomous living being, — a filtrable microbe parasitizing the bacteria, — that I resolved to publish, in 1917, the first communication announc- ing the discovery of an ultramicrobe parasitizing the dysentery bacillus. In this report I gave the principal characteristics of the virus and indicated the role played by it in the course of the disease. From 1917 to the end of 1919, I continued these in- vestigations alone, extending them to other diseases, and it was only in December 1919 that Kabeshima, working in my labora- tory with strains which I had furnished him, published results which confirmed mine. Since that time, investigations on the bacteriophage have multiplied, and rare indeed are the labora- tories which are not interested in the question. It may seem strange, at first sight, that I should have been able to work alone on this question for such a long time; a cir- cumstance which has permitted me to establish the facts in their entirety, and the relation between them; to demonstrate their importance from the point of view of immunity, and to accom- plish this before any other communication appeared. In this I have been favored by circumstances and even by the strangeness of the facts themselves, which at first excited, not merely astonish- ment, but incredulity, even among my most friendly colleagues, who were not loath to consider me a visionary. This time has passed. The facts are recognized to be correct. The most recent work appears to affirm the curative properties of cultures of the bacteriophage. The only point at issue is my conception of the nature of the active principle. I may be permitted to make a few remarks upon the subject of this discussion. All authors, without exception, who have formu- lated an hypothesis regarding the nature of the bacteriophage have adopted a method of reasoning that is somewhat peculiar. None of them have taken the trouble to review the experiments that I had accumulated in favor of the living nature of the bac- teriophage during the years that I had alone been occupied with this question ; experiments moreover, which in no instance are open 146 THE BACTERIOPHAGE to question. Each of them has taken simply a particular fact, suited to support his thesis, and has neglected entirely the great group of experimental facts which render this hypothesis inad- missible, forgetting that that which accords with experiment is, for a theory, the sole and indispensable criterion of its truth. But the strangeness of their procedure is not restricted to the interpretation of their experimental findings, but extends to the experiments themselves. These experiments correctly performed react against their hypotheses. I have called attention to these errors and Kabeshima seems to be converted by the evidence, for he has published nothing for two years. As for Bordet, he does not maintain that his fundamental experiment, called that of leucocytic exudates, may be repeated. He has recognized that the specificity of the bacteriophage, a condition sine qua non for his hypothesis, is contrary to fact, but he nevertheless continues to support a hypothesis thenceforth without foundation. THE POSSIBLE HYPOTHESES The number of possible hypotheses is limited, and after a con- sideration of these fundamental hypotheses it is only necessary to select that which accords with the observed experimental facts which have been contradicted by no one. These hypotheses were carefully reviewed prior to all publication in an attempt to determine the nature of the principle which I had discovered. Three fundamental hypotheses can be formulated and any other view must be a modification or a combination of one or more of these three. Discussion of these three must necessarily dispose of any subsidiary hypotheses that may be advanced. What, then, are these three fundamental hypotheses which comprise all that can possibly be formulated? First hypothesis The bacteriophage is derived from the superior organism in its reaction to the bacterial invasion by the production of a prin- ciple which provokes the destruction of the bacterium. This first hypothesis admits of two solutions. 1. The principle in question is a substance of diastatic nature. The single fact of the serial action of the principle is sufficient to NATURE OF THE BACTERIOPHAGE 147 reject this explanation, for such a substance would be rapidly eliminated in the successive dilutions during repeated transfers. It is thus useless to discuss this further. Up to the present time no one has recommended this. 2. The active principle derived from the organism reacting against the infection is particulate, an organic being, capable of developing afterward outside of the organism at the expense of the bacteria. This hypothesis does not constitute a scientific heresy, for it is not contradicted by any experimental fact. In the case of any hypothesis, however improbable it may appear in view of the actual state of biologic science, if it can not be experimentally demonstrated false and if it harmonizes with the demonstrated facts, it can not be rejected a priori. Moreover, Carrel has shown that it is possible to cultivate tissues outside of the organism; and in addition, Altmann has proposed a theory according to which the zymogenic granulations can be nothing but bioplasts, independent elements, having their individual existence and capable of reproduction by division in a cellular medium. The bacteriophage may be a bioplast, derived from the superior organ- ism, and capable of multiplication at the expense of the bacteria. However this may be, since this particle, this "organite," comports itself from the time when it is taken from the organism as an autonomous being capable of assimilation and reproduction, and since it acts as a being corresponding to the definition of a microbe, it must be a minute being endowed with life. We will revert to this idea in the case of the third hypothesis. Second hypothesis The bacteriophage may be derived from the lysed bacterium itself. The two subsidiary hypotheses given above may again be formulated: 1. The bacteria secrete a diastase with autolytic functions. As we will see, this is in effect the conception of Kabeshima. Bordet takes over this hypothesis with an added complication, since he explains the origin of the principle in terms of the first hypothesis, that is, a substance derived from the or- ganism, and explains the continuity in series by means of the 148 THE BACTERIOPHAGE second hypothesis, that is, to a substance derived from the bac- terium itself. This hypothesis fails before the following experimental facts: a. The bacteriophage exists in the form of particles which multiply. b. The bacteriophage does not exercise a specific action upon any single species of bacteria, but at one and the same time, the same bacteriophage may be active against several species. c. All strains of the bacteriophage, whether they be active against the staphylococcus, against the dysentery bacillus, against B. pestis, or against any other organism, belong to the same species, as demonstrated by the complement fixation reaction. 2. The bacteriophage is derived from the bacterium, but is particulate, an "organite" capable of life, conducting itself thence- forth like an autonomous ultramicrobe. This is the hypothesis of Bail. While an interpretation which comprehends an " organite" derived from the superior parasitized organism does not neces- sarily imply specificity of the bacteriophage, any hypothesis in- volving an " organite" derived from the bacterium which is sub- jected to lysis does imply a strict specificity. The last view, therefore, is inadmissible, since experiment proves that a single strain of the bacteriophage may be active toward several species of bacteria at once, and that all strains of the bacteriophage be- long to the same species. All possible hypotheses, save that of the ultramicrobe, which we will shortly examine, can only be some combination of the two preceding ones. All, then, will be subject to the same ob- jections; none will be admissible for it will be contradicted by experimental facts. Third hypothesis The bacteriophage is an autonomous organism, an ultramicrobe parasitizing the bacteria. This hypothesis is the only one which accords with all the recorded experimental facts, and it is for this logical reason that I have attached myself to it. For it, I have acquired more and more convincing evidence which I have not been able to disprove, for the numerous new facts that I NATURE OF THE BACTERIOPHAGE 149 have discovered harmonize with this hypothesis, with this hy- pothesis only, and none contradict it. Moreover, I should repeat that I have not formed any hypothesis as to the species to which the ultramicrobe belongs. I have called it Bacteriophagum intestinale, a name simply denoting its characteristic property and the place where I first found it. Is it a protozoan? Is it a bacterium? Does it belong to a kingdom which is neither vegetable nor animal? Does it arise even in another organism, a possibility which has been suggested in ex- amining the first hypothesis? These questions can be ignored. It is an ultramicrobe, a filtrable being endowed with the functions of assimilation and of reproduction, functions which characterize the living nature of beings and which pertain to them alone. That is all that experimentation is actually able to demonstrate. To try to penetrate further into its identity would be nothing but purely speculative reasoning. EXPERIMENTAL PROOFS OF THE LIVING NATURE OF THE BACTERIOPHAGE This section will, without doubt, be judged unnecessary by the reader who has followed the experiments recorded in the preceding chapters. However, it may be well to group these proofs and to comment on certain experiments which can leave no doubt, even in the minds of the most skeptical and uninformed. 1. The bacteriophage proliferates, since serial cultures can be continued indefinitely. In the action of the diastases there is always a certain proportionality. The action is the more ener- getic when the amount employed is large. With the bacterio- phage this is not true. The activity is due to the quality of the principle, not to its quantity, and this is, indeed, a property of vital activity. A diastase acts in proportion to its quantity, a bacterium in proportion to its virulence. 2. The bacteriophage presents properties analogous to those of other known living beings. Its resistance to agents of de- struction, although great, is, however, less than that of many organisms of which the living nature is unquestionable and uncontroverted. 150 THE BACTEEIOPHAGE I ought in this connection to dwell upon the action of tempera- ture, since some authors have suggested that the temperature of destruction of the bacteriophage was too high to allow a con- sideration of them as living beings. It is thus necessary to re- call some of the elementary facts which readers of this work cer- tainly ought not to ignore. The lethal temperature, as we have seen, is about 75°C. Without mentioning the living organisms from thermic sources, of which the temperature reaches up to 93°C., it is known that one may readily isolate from sewage bac- teria which develop normally at a temperature of 75°C. More- over, Duclaux has shown that the young cells of Tyrothrix tennis do not die until a temperature of about 100°C. has been reached. A lethal temperature of 75°C., far from being exceptional, must be recognized as well below that resisted by a large number of unicellular organisms. In so far as the action of antiseptics is concerned, the bacterio- phage takes a position intermediate between the bacteria in their vegetative form and the spores derived from these bacteria. More resistant than the first, they are more sensitive than the sec- ond. Compared from this point of view with other known ultra- microbes, they are definitely more susceptible than some. The virus of the tobacco mosaic, for example, will resist for several months a concentration of alcohol which will kill the bacteriophage in a short time. The virus of rabies, and that of vaccinia, remain alive in concentrations of glycerine that destroy the bacteriophage. It is to be noted that the bacteriophage presents the character- istic of being particularly sensitive to certain reagents which have absolutely no effect upon the diastases; quinine for example. As for glycerine, which destroys the bacteriophage, it constitutes the medium of choice for the indefinite preservation of bacterial toxins and the most sensitive diastases. 3. With sufficiently active strains of the bacteriophage a com- plete and permanent lysis is secured; all of the bacteria contained in a suspension are definitely destroyed. Moreover, serial passages of the bacteriophage are possible in bacterial suspensions made in fluids which do not permit the development of these bacteria; — physiological salt solution, or bouillon with forty per NATURE OF THE BACTERIOPHAGE 151 cent glycerine, for example. This proves again that the survival of a certain number of bacteria does not constitute a factor in the serial activity, for, as this factor would fail the series could not be continued. 4. On agar the bacteriophage gives colonies at the expense of the bacteria, and this permits counting the active elements. A soluble ferment, diastase or toxin, can not concentrate its action in definite points. It may be objected that the lytic diastase is provided by the bacteria themselves and that each plaque on agar represents the area where is to be found, after the inocula- tion, a bacterium particularly able to furnish the diastase under the influence of a force "X." The following experiments demonstrate that this objection is not valid. Experiment LV. (A}. Take 10 tubes. Into the first place 10 cc. of a suspension of B. dysenteriae Shiga containing 100,000,000 bacilli per cc., into the second place 10 cc. of a suspension containing 200,000,000 bacilli per cc., into the third place 10 cc. of a 300,000,000 suspension, and so on, increasing by 100,000,000 the concentration of the suspensions introduced into each tube of the series. The tenth tube, then, will have a suspension containing 1,000,000,000 bacilli per cc. Each of these tubes is then inocu- lated with an equal quantity of a very dilute culture of the bacteriophage filtered through a bougie, about 0.000005 cc. We have then a series of 10 tubes containing a more and more concentrated suspension of B. dysenteriae Shiga in the ratio of 1 : 2: 3: 4: 5: 6: 7:8: 9: 10, and an equal amount of bacterio- phage culture. The tubes are carefully shaken, and 0.02 cc. from each tube is planted upon agar slants. After incubation at 37°C., each of the 10 tubes of agar shows a culture of B. dysenteriae spotted with plaques, and the number of these plaques is practically the same for all the tubes. (B) Take 10 tubes, each containing 10 cc. of a suspension containing 100,000,000 B. dysenteriae per cc., that is, a like suspension in all 10 tubes. To the first add 1/100,000 cc. of filtered bacteriophage culture, to the second 1/200,000 cc., to the third 1/300,000 cc., 1/400,000 cc. to the fourth, and so on, each tube receiving a smaller and smaller amount of bacterio- phage culture, so that the tenth tube will contain only 1/1,000,000 cc. Thus, we have a series of 10 tubes, all containing an equal number of dysen- tery bacilli and an amount of bacteriophage culture varying according to the ratio 10:9:8:7:6:5:4:3:2:1. Shake the tubes thoroughly and plant 0. 02 cc. from each of the 10 suspensions on to agar slants. After incubation at 37°C. each of the agar tubes will show a covering growth of B. dysenteriae studded with plaques, but the number of plaques in the tubes varies with the proportion of bacteriophage culture which has been added. Prac- 152 THE BACTEKIOPHAGE tically, their number varies from tube 1, which had received the largest amount of bacteriophage culture, to tube 10, which had received the smallest, in the proportion of 10:9:8:7:6:5:4:3:2:1. These two experiments, which complement each other, demon- strate unquestionably, that the active principle is contained solely in the filtered culture of the bacteriophage; and that this active principle is composed of material elements, capable of forming colonies on agar at the expense of the surrounding bac- teria in such a way that their enumeration is possible. These material elements are capable of multiplication as is shown by the formation of colonies and by action in series. It can thus only be a living organism. This experiment by itself is sufficient to demonstrate that the bacteriophage is a " formed ferment," which in reality implies the existence of an ultramicrobe parasitic of the bacteria. 5. Eliava and Pozerski have shown that toward concentra- tions of free H and OH ions the range fatal for the bacteriophage is more limited than for the bacteria. The diastases and toxins react in a wholly different fashion. 6. Dumas, confirmed by Beckerich and Hauduroy, have iso- lated the bacteriophage from the soil and from the filtered water of streams. I have myself isolated it from sea-water. There is nothing strange in this, since it is an ultramicrobe derived from stools. This fact can not, on the contrary, be reconciled with a hypothesis of a ferment, whether the ferment be of leucocytic or other origin. 7. Diastases in solution are absorbed by the precipitates which form upon the addition of alcohol. This takes place with cul- tures of the bacteriophage, but the elements which precipitate are not the ultramicrobes themselves. The ultramicrobes are de- stroyed by the alcohol, and the principle which is precipitated will not reproduce the action in series. This possibility of ex- tracting from a culture an active principle which can only be a secretory product of the bacteriophage shows indeed that the latter can be nothing other than a living being. 8. I have shown that the bacteriophage is capable of adaptation to the harmful action of glycerine. Adaptation is the appanage of living beings exclusively. NATURE OF THE BACTERIOPHAGE 153 9. It is impossible to isolate two strains of the bacteriophage which are identical in the intensity of their action or in the scope of their activity. With a single strain the intensity of the ac- tion can be varied experimentally. Variation is an exclusive characteristic of life. 10. The lytic action is always exercised by one and the same element as is demonstrated in the complement fixation reaction. This element adapts itself to parasitism toward such and such a bacterium, and the possibility of such adaptation necessarily implies the living nature of the element which exercises it. 11. Bruynoghe and Maisin have shown that the bacteriophage is phagocytized and destroyed by the leucocytes. This fact shows that the bacteriophage is foreign to the organism, and it alone demonstrates that it can not be of leucocytic origin. The nature of the bacteriophage is not open to question; its origin alone may be disputed. Is it an entirely autonomous being, a "species", botanically or zoologically? Is it a "bioplast" capable of indefinite reproduction at the expense of living bac- teria, conducting itself as an autonomous being of which it has all the properties? It is still impossible to decide this; experiment alone will de- termine. However this may be, the two hypotheses, although they may differ as to the origin of the bacteriophage, agree as to its nature. The most probable interpretation is that the ultra- microbe is autonomous, a botanic or zoologic "species." REFUTATION OF THE HYPOTHESIS OF KABESHIMA Without doubt readers will wish to know the diverse hypotheses proposed by those who have opposed the living nature of the bacteriophage, and I will discuss them in their chronological order, although this may not be their documentary standing. Apart from the confirmation of the experiments which I alone, or with my collaborators, have effected, these discussions repre- sent about all that has been done on the question of the bacterio- phage. By indicating these and commenting on them, this work becomes completed and is a comprehensive statement of the present knowledge regarding the bacteriophage. 154 THE BACTERIOPHAGE Kabeshima, in 1920, starting on the one hand from considera- tions without significance (as, for example, the thermal death point of the bacteriophage, which he found to be too high to be applied to living beings), and on the other hand, from inexact experimental results (he announced, for example, that the bac- teriophage resisted the action of alcohol; that it was active in the presence of sodium fluoride, etc.), formulated the following hypothesis. Under the action of a proferment playing the role of a catalyzer, the autolytic diastases are activated and bring about their dissolving action. We have already considered the principal objections which render this hypothesis untenable. It fails to explain serial ac- tion, for a proferment would disappear rapidly as a result of dilu- tion in the course of passages; it does not take account of the fact that the bacteriophage presents itself in the form of autonomous particles capable of being counted; it is formally contradicted by the fact that the same bacteriophage can act on diverse bac- terial species, etc. The inadmissibility of the hypothesis of Kabeshima has been recognized, moreover, by all authors who have considered the question. KEFUTATION OF THE HYPOTHESIS OF BORDET AND CIUCA Bordet and Ciuca (October, 1920) very significantly modified the hypothesis of Kabeshima to the end of explaining serial activity. They said: "In view of the fact that the stools of patients with dysentery are rich in leucocytes, and that the lysinogenic power is only observed toward the period of convalescence, we have asked ourselves if the phenomenon of d'Herelle is not the result of a defensive activity of the organism, and particularly of an activity of the leucocytic exudate. This produces in the bacterium an hereditary nutritive vitiation, consisting in the produc- tion by the bacterium of a sort of lytic ferment, which is capable, moreover, of diffusing in the ambient fluid and as a result, reacting in the same fashion on normal bacteria of the same species." This proposition takes no account of the previously established facts. Among other facts it disregards that I had made it known some time previously that the bacteriophage was a normal in- habitant of the intestine, and that lysogenic power was also to NATURE OP THE BACTERIOPHAGE 155 be observed at times other than at the moment of convalescence. At a time considerably earlier, I had likewise indicated that a strain of the bacteriophage exercised its action not only against a single bacterial species, but against several at the same time. Like the hypothesis of Kabeshima, this of Bordet and Ciuca takes no account of the fundamental fact, already demonstrated ex- perimentally, of the existence of the bacteriophage in the form of particles which it is possible to count. Fundamentally, how does this hypothesis of Bordet and Ciuca differ from that of Kabeshima? It differs in nothing except it be in the form in which it is stated. It hinges upon an arbitrary transposition of effect and cause; a transposition upon which they lay no stress. The leucocytic exudate induces the " nutri- tive vitiation" (?) which results in the lysis of the bacteria in the first tube of the series, but in the following ones the same effect will be produced, no longer by the leucocytic exudate, which will of necessity have disappeared in the first tubes because of the dilution, but by the bacterial lytic ferment alone. Bordet and Ciuca seem to find this substitution of cause entirely logical, when in reality it is contrary to all that we know. In admitting a priori, that a liquid, indeed a filtered liquid, is able to transport with it an hereditary property, there is, it appears, an affirmation which must needs be based on experiment and not solely upon the inference of Bordet and Ciuca. And what could have been the fundamental experiments which suggested such conclusions? They say: "If one or two days after the last injection, one removes by puncture the peritoneal exudate, rich in leucocytes, from a guinea pig which had received three or four intraperitoneal injections of B. coli at intervals of a few days, one can demonstrate that this exudate, when added to normal bacteria of the same species, modifies them, and confers upon them a very pronounced autolytic power, transmissible from culture to culture." They add that they will shortly publish the results secured with other bacterial species. These results, promised more than a year and a half ago, have not yet been furnished. Furthermore, not having succeeded in reproducing the experiment with the leucocytic exudate, in spite of numerous attempts, I refuted the statements of Bordet and Ciuca, in a note published in the Compt. 156 THE BACTERIOPHAGE rend. Soc. de biol. This contradiction has remained without a reply for more than eight months — evidence that these authors themselves have not succeeded in repeating the experiment. Furthermore, the experiment with the leucocytic exudate, had it been correct, would in no case have provided proof for the non-reality of the bacteriophagous ultramicrobe, for it would not have disproved any of the experiments which demonstrate its reality, and of more significance, it accords perfectly with the idea of a parasitic ultramicrobe. Long before Bordet and Ciuca worked with the bacteriophage I had demonstrated that there was in certain cases a passage of the intestinal bacteriophage into the circulation, and that it could be isolated from the blood. Since the bacteriophage may acquire by adaptation the faculty of parasitizing any bacterial species, and since the bacteriophage is a normal inhabitant of the body of all animals, it is by no means impossible that one might experimentally succeed in provoking in the body of an animal this adaptation for a given bacterium which has passed into the circulation. We will see in Part II of this work that this is exactly the series of phenomena which occur in natural disease; and I can not see in what respect the fact of the experimental reproduction of this sequence of events will be opposed to the doctrine of an ultramicrobe parasitizing the bacteria. I may say further, that even had their experiments been cor- rect, the logical interpretation would not have been that of Bordet and Ciuca; that the bacteriophagous principle is derived from the leucocytes. In fact, if the primum movens of the bacterioly- sis transmissible in series resides in the leucocytes, it should be enough, for example, to add to a suspension of B. dysenteriae the leucocytes from a horse furnishing an anti-dysentery serum, that is, from an hyperimmunized horse, to reproduce the phenom- enon of lysis transmissible in series. This reaction would have been recognized long ago by innumerable investigators, perhaps first of all by Bordet, who for more than thirty years has in- vestigated the antibodies in the blood of immunized animals. This experiment I have in vain attempted many times, well before Bordet and Ciuca announced their hypothesis, when I was attempting to test all possible hypotheses touching the origin NATURE OF THE BACTEEIOPHAGE 157 of the principle dissolving the bacteria. Far from possessing the ability to start serial lysis, the leucocytes derived from a horse hyperimmunized with the dysentery bacillus, do not even intervene to inhibit the growth of this bacillus, whatever may be the quantity of leucocytes employed in the test. Finally, had the experiment of the leucocytic exudate been correct its interpretation would not in any case have served as a proof for the reality of an ''hereditary nutritive vitiation" trans- mitting itself by means of a liquid factor. To be tenable, an hypothesis must take into account all the facts, and that of Bordet, exactly like that of Kabeshima from which it is copied, is incom- patible with the experimental facts which I have reported. It implies, among other things, the strict specificity of the bacterio- phage. Even if one admits as possible the entirely speculative hypothesis of " hereditary nutritive vitiation" transmitted by the intervention of a liquid, one is absolutely unable to admit that this liquid is able to transmit the " hereditary vitiation' ' from one bacterial species to another bacterial species. Moreover, Bor- det and Ciuca at first maintained that there was such a strict specificity. However, in view of the accumulated evidence they have recognized that the action of the bacteriophage is not specific, but, in spite of this, they have not abandoned their conception or even offered anything by way of explanation. The accidental positive result obtained by Bordet and Ciuca in the experiment with the leucocytic exudate can readily be ex- plained in perfect harmony with the doctrine of the ultramicrobial bacteriophage. Such an explanation is, that the intestinal bac- teriophage has passed into the peritoneal cavity of the experi- mental guinea pig as a result of the irritation produced there by the injections. This has been followed by a growth of the bac- teriophage in this cavity at the expense of the bacteria injected. As we will see in Part II, the bacteriophage does not remain con- fined to the intestinal tract; it is able to enter the circulation. Moreover, the experiment of Bordet regularly becomes positive, even if the experimental guinea pig has received only one pre- liminary injection of bacteria, provided one or two cubic centi- meters of a culture of the bacteriophage active for the bacterium inoculated is administered per os a few hours before the intra- 158 THE BACTERIOPHAGE peritoneal injection. This experiment is adequate to give a correct interpretation to the accidental finding obtained by Bor- det and Ciuca. To summarize: the hypothesis of Bordet and Ciuca dealing with the nature of the bacteriophage is inadmissible, first, because it does not conform to the experimental facts; second, because it is founded upon an erroneous interpretation of an experiment which has not been repeated, and which, even if it had been cor- rect, would not have served as a basis upon which such an hy- pothesis could be founded; third, because invoking as an explana- tion an hereditary phenomenon, it is in formal contradiction to all known facts concerning the hereditary transmission of characters. Bruynoghe and his collaborators, who at the beginning of their studies adopted the point of view of Bordet, have since realized that the experimental facts do not harmonize with this hypothesis. They now support the idea of the ultramicrobe, a parasite of the bacteria. REFUTATION OF THE HYPOTHESIS OF BAIL Bail, in 1921, rejected the hypothesis of Kabeshima and of Bordet. According to him the bacteriophage existed indeed in the form of autonomous masses as I had demonstrated. It con- ducted itself as an ultramicrobe but these particles could only be constituted by the "splitter," that is to say, by particles derived from the lysed bacteria themselves. These organized particles, capable of reproduction under a filtrable form at the expense of the same bacteria, secreted a dissolving diastase. Bail adduces in favor of his hypothesis the fact that he has been able to isolate from old cultures a bacteriophage active for the dysentery bacillus of Flexner. The cause of Bail's error is easy to detect. He has dealt with mixed cultures. I have already called attention to the frequency of such cultures and I have indicated their origin. The hypothesis of Bail corresponds exactly to the second solu- tion of the second possible hypothesis discussed above. It is needless to repeat the refutation which has been presented. Let us simply recall that the condition sine qua non for the validity NATURE OF THE BACTERIOPHAGE 159 of this hypothesis would be strict specificity, and experiment demonstrates that this does not obtain. All authors are actu- ally in accord on this point and have confirmed the observation that a single bacteriophage is able to attack diverse bacterial species. REFUTATION OF THE HYPOTHESIS OF SALIMBENI The hypothesis of Salimbeni is merely mentioned. Admitting that the bacteriophage is an autonomous microorganism, he considered it a Myxomycete, presenting itself in visible forms, and visible even to the naked eye. His observations have been contradicted by all who have studied the bacteriophage. More- over, Salimbeni himself, has not continued to maintain this hy- pothesis. Possibly the observation upon which he based his hypothesis was due to the use of contaminated cultures. CONCLUSIONS All of the authors who have advanced hypotheses other than that of an ultramicrobe parasitizing the bacteria have forgotten that a hypothesis ought always to account for the entire mass of facts; that it is necessary, not only to demonstrate that it suffices to justify the phenomena, but it must also prove that these phe- nomena can not be justified if this hypothesis is abandoned or if it is modified. After all, the whole controversy on the subject of the nature of the bacteriophage is only the renewal of the old discussion which we had for a long time thought terminated. With the new ideas of diastases capable of multiplication, or of co-ferments or catalyzers playing with the metaphysics of ubiquity, or of se- cretory immunity transmissible by communicated motion; it is only taking up anew the old theory of Stahl, that "any body brought to a state of putrefaction transmits very easily this state to another body still free of corruption." This theory of the multiplication of a principle of communicating motion had been held by Liebig in his discussion with Pasteur concerning the mechanism of fermentation. Pasteur demonstrated experi- mentally that it was false, and we were lead to believe the fact definitely acquired. With vital phenomena or communicating 160 THE BACTERIOPHAGE motion, the discussion and the experiments of demonstration hinge on the same facts; we only descend a step in the order of magni- tude of the beings concerned. This same discussion may perhaps be renewed again some day if we descend still another step; if we discover, for example, a virus parasitizing the bacteriophage. The infinitely small is as conceivable as the infinitely great; we have not the right to assign limits to them. PART II THE ROLE OF THE BACTERIOPHAGE IN IMMUNITY INTRODUCTION Up to the present time investigations on immunity have been directed toward solving the following question: What are the means of defense which permit an immunized animal or one naturally refractory to resist infection? These studies have resulted in the development of diverse theories. When an animal is affected with a contagious disease of bac- terial origin, the cellular immunity, which we call "organic immunity," abstracting it from all theory as to its intimate na- ture, is only established in a variable length of time after the inception of the disease. Does the animal remain without de- fense up to the time that this organic immunity becomes effec- tive? By what phenomenon is it possible to acquire this organic immunity? All animals sensitive to an infection and exposed to it do not contract the disease. Why do some of them remain unharmed? These are the principal points upon which the experiments to be discussed have turned. As will be seen, they lead to a new chapter in the study of the means of defense against infection; and the conclusions themselves which will be derived from these investigations will not actually contradict anything in the present theories, for they apply to different states. There is, nevertheless, a particular point in the present theories of immunity to which I wish to call attention, namely, that which deals with bacteriolysis as induced by specific sera. This is certainly pertinent since it deals with the subject under dis- cussion:— the lysis of bacteria. Everyone knows the nature of the phenomenon of Pfeiffer. If one injects a suspension of cholera vibrios into the peritoneal cavity of a guinea pig previously "immunized," a transformation of these vibrios into granules is noted. Pfeiffer has suggested that the transformation into granules constitutes only the first phase of the vibriolysis; but in this he was in error, for the granules maintain this form indefinitely. 163 164 INTRODUCTION We have very frequently sought for the act of disappearance of the granules in drops taken from the peritoneal fluid, but the number of these transformed vibrios has never diminished, even after several days, and we have therefore not been able to detect the phenomenon of dissolution of the granules. In spite of all this, it is incontestible that the granular transformation is a manifestation of a very grave change to which the cholera vibrios have been subjected under the influence of the peritoneal fluid of the immunized organism.1 Is this granular transformation, as Metchnikoff states, a mani- festation of very grave lesions? The fact, established by him, that one of these granules seeded on agar gives a colony of nor- mal vibrios is not an index of a very profound alteration. Let us consider for the moment that there can not be, in any sense, a lysis of the cholera vibrios in the Pfeiffer reaction. One may try by all sorts of methods to provoke the same phenomenon in all kinds of animals with all kinds of bacteria, but without result. It can only be secured with the cholera vibrios; they alone can be transformed into granules. Metchnikoff, and later Bordet, showed that the reaction of Pfeiffer could likewise take place in vitro. To obtain a granular transformation it is only necessary to introduce the vibrios into the fresh serum of an immunized animal. Bordet further showed that this transformation was brought about by the in- teraction of two principles, amboceptor and alexin. The amboceptor, thermostabile, is specific; that is to say, it is only active toward the element against which the animal fur- nishing the serum has been immunized. It exists only in traces, or not at all, in the serum of normal animals. It develops as an effect of immunization. The alexin, thermolabile, is, it appears, common. It exists in as large an amount in a normal animal as in the immunized animal. It is fixed by any element previously acted upon by a specific amboceptor. Bordet next discovered that the blood of an animal prepared by the injection of the red blood cells of a different animal species formed specific amboceptor. If to a suspension of these cells is added a heated serum of the treated animal (thus containing 1 Metchnikoff. Uimmuniii dans les maladies infectieuses, Paris, 1901, Masson et Cie. INTRODUCTION 165 the amboceptor) and then some fresh serum from a normal animal (thus containing alexin) the phenomenon of hemolysis is obtained, — a phenomenon in which the dissolution of the red cells is simu- lated, although in reality, as Bordet himself showed, there is simply a diffusion of the hemoglobin, the stroma remains intact. Finally, Bordet showed, in an indirect manner, by means of the complement fixation reaction which bears his name, that bacteria, as well as red blood cells, absorb specific amboceptor and are thus able to fix complement. Here is where the equivocation commences. By analogy it was concluded that since, under the influence of the complement fixed by the cells in combination with the specific amboceptor, a hemolysis of the red cells was produced, so with bacteria which likewise absorb a specific amboceptor and fix complement in the same way, there must necessarily be a bacteriolysis. This bac- teriolysis has never been observed directly, and this fact is ade- quate, it seems to me, to arouse some doubt concerning the reality of the phenomenon. The following experiment, which I have repeated several times, shows that in reality the bacteria are by no means destroyed under such conditions. The result is quite the opposite. Experiment LVI. Take 4 tubes, each containing 20 cc. of 0.8 per cent saline. To each add a quantity of cholera vibrio suspension sufficient to give a count of about 1000 vibrios per cubic centimeter. The first tube remains as the control. The second tube receives 0.25 cc. of fresh guinea pig serum, after this serum has been shown by test to contain complement. The third tube receives 0.1 cc. of an anticholera serum, in which the presence of specific amboceptor has been demonstrated. The fourth tube receives 0.25 cc. of the fresh guinea pig serum and 0.1 cc. of the anticholera serum. Thus, in this tube, the vibrios are in the presence of a specific antibody and of complement. The 4 tubes are incu- bated at 37°C., and from day to day are tested to see if the vibrios are alive or dead. To this end, after twenty-four hours, the tubes are thoroughly shaken and from each of them 2.5 cc. is immediately taken and planted into a tube of sterile bouillon. The first of the tubes usually (4 times in 6) remains sterile2 while the other three give cultures of the cholera vibrio, 2 This bactericidal action of physiological saline is rather strange. Even when working with relatively concentrated suspensions, containing from 50 to 100 million bacteria (cholera vibrios or B. dysenteriae) per cubic 166 INTRODUCTION normal in the second tube containing complement only, agglutinated in the last two. After forty-eight hours transfers are again made. The first tube always remains sterile, the second is often sterile (3 times in 6), the last two give agglutinated cultures. After three days the subcultures result as follows ; the first two tubes are always sterile, the third often so (4 times in 6), the last always gives an agglutinated culture. After four days the first three tubes are always sterile. The last tube only, that is, the one containing both antibody and complement, gives an agglutinated culture. After six days the same result is obtained. The same experiment has been performed with the Shiga dys- entery strain and the anti-dysentery serum of the Pasteur In- stitute. The result was in all respects comparable. The Shiga bacillus, like the cholera vibrio, persists for a long time in the suspension containing the anti-serum and the alexin. Variations in all directions have been made in the proportions of the sera, both in that containing the complement, and in that containing the antibody, as well as in the concentration of the bacterial suspension. The results have all been essentially the centimeter, the sterilization is complete in a few hours at a temperature of 37°C. On the contrary, these organisms will remain alive for some days in tap water. And what is still more singular, is that everyone has adopted physiological saline for the preparation of bacterial suspensions, concluding a priori, that bacteria must be preserved alive for a long time in a medium spoken of as "isotonic." At the bottom of this we find a false deduction by comparison. Eed blood cells hemolyze in a few seconds in tap water, but, on the contrary, they resist hemolysis in isotonic saline solution. Thus, it is concluded, without doubt, that bacterial cells must conduct themselves in the same manner. As a matter of fact, this is absolutely contrary to what takes place. As we will see, the same reasoning has been held with regard to the so- called bacteriolysis with sera. There also, one falls into an error, and this will always be the case when an attempt is made to substitute deduction based on analogy for experimentation, especially when the elements con- cerned are as different as a bacterium and a red cell. With reference to the toxic action of sodium chloride, Loeb (Biochem. Zeitschr., 1906, 2, 81) has shown that this salt may be toxic for all the unicellular organisms living in the sea, and that this toxicity may be neu- tralized by the salts of potassium and calcium. INTRODUCTION 167 same. The bacteria always remain alive in the antibody-com- plement mixture for a much longer time than in pure physiological saline. These experiments show that cholera vibrios, or dysentery bacilli, are rapidly destroyed in saline; that they remain alive longer in the presence of a normal fresh serum containing com- plement or in an antiserum; and that they remain alive still longer in the presence of both the antiserum and complement. It is therefore evident that the bacteria, sensitized and having fixed complement, far from being subjected to lysis, are more resistant than normal bacilli. The sera termed " antibacterial" do not, in vitro at least, play any bacteriolytic role. Everything indicates that it is the same in vivo. We know that at times the antisera, derived from horses thoroughly immunized by the injection of living bacilli, are con- taminated by bacilli of the same species as those which have been injected, and that it is possible to demonstrate them by culture. Anti-rouget serum provides a remarkable example. The serum from a horse hyperimmunized by a series of injections of living culture possesses extremely marked curative and preventive properties, but it is, nevertheless, rather frequently contaminated by the bacillus of rouget, even if the serum is withdrawn some ten to twelve days after an injection. How can we reconcile this fact, which indeed is not an isolated observation, with the hypothe- sis that the immunity which it confers is, by some mechanism at present unexplained, associated with the presence of an " anti- bacterial antibody?" If it should produce bacteriolysis, it cer- tainly ought to do so in the hyperimmunized animals themselves, where the bacterium finds itself in contact with an abundance of antibody and of complement. On the other hand, sera very rich in antibody may entirely lack preventive power, and the opposite is also true. Metchni- koff and his collaborators have furnished many examples of this. If we pass to a consideration of natural immunity we readily discern that there is absolutely no parallelism between the state of the patient and the antibody content of the blood. In human typhoid fever, for example, the fatal relapses may occur when the antibody is at its maximum. Finally, in diseases which result 168 INTRODUCTION in immunity the antibody disappears a short time after the at- tack, but this does not prevent the persistence of immunity for years, or even decades, after the disappearance of the antibody. To summarize : however the serum of hyperimmunized animals or the serum of an individual affected with some infectious disease may act, it is impossible to establish any relation between the antibacterial antibodies — so-called — and immunity. These anti- bodies, like the agglutinins, can only be considered as indices of infection. It may be objected that preventive vaccination by a sensitized virus indicates a fragility in the bacteria impregnated with anti- body. This objection is in reality not an objection, for experi- mental facts show, on the contrary, that a sensitized bacterium is as virulent as a normal bacterium. I have proved this for B. typhi murium with the mouse, and for the bacterium of bovine hemorrhagic septicemia in cattle. These animals are killed in the same time and by the same dose, whether the bacteria in- jected are normal or sensitized. The possibility of injecting man, without great inconvenience, with sensitized cholera vibrios or sensitized typhoid bacilli (and it may still be said with some re- serve in this last case) does not negative these findings at all, seeing that Ferran has given tens of thousands of preventive vaccinations against cholera, using living cultures, and that Ch. Nicolle has demonstrated the possibility of vaccinating man against typhoid fever by injecting him with living, normal typhoid bacilli. In general, injections of sensitized bacteria are no more inoffensive than injections of the normal living organ- isms, and they are equivalent, since living sensitized bacteria and living normal bacilli are virulent to the same degree. One cannot avoid the conclusion that it is impossible to attribute any active role in the production of antibacterial im- munity to any actually known antibodies. All organic immunity is reduced to antibacterial immunity, assured by phagocytosis, and to antitoxic immunity, assured by the antitoxins.3 But is this organic immunity an attribute of the immunized animal only? Is it not enjoyed by a susceptible animal? All 3 Meaning by antitoxins all antibodies which neutralize a soluble toxic substance. INTRODUCTION 169 individuals exposed to an infection do not contract the disease, and to what do they owe this privilege? Once diseased, the immunity in the susceptible animal manifests itself only after a lapse of time, which in the most favorable cases is hardly less than twelve days.4 Does the animal remain without defense during this lapse of time? Finally, in this animal affected by disease, whatever may be the mechanism of organic immunity, how can this immunity be established to render this sick animal refractory? Under what influence is phagocytosis released? Under what influence do the antitoxins originate? The role which the partisans of the theory of "bactericidal humoral immunity" have desired that these simple indices of infection — the antibodies — play, is based upon a non-existent phenomenon of bacteriolysis by immune sera. Could it not in reality be played by the bacteriophage, that principle endowed with a powerful bacteriolytic action, operating upon the most varied bacteria? Could not the bacteriophage play a role in the defense of the organism, a preponderant role in the susceptible animal, and as such, deprived of all acquired immunity? In other words, does there not exist by the side of the homogeneous organic immunity, an immunity originating in the bacteriophagous ultramicrobe, and, as a result, an heterogeneous immunity? We have seen in a preceding chapter that the lysin of the bac- teriophage may possess an extraordinarily potent opsonic activity. Can not the bacteriophage play, in addition to its direct action, an important role in phagocytosis itself, in bringing about what might be called a phagocytic education? 4 Animals vaccinated by an attenuated anthrax or rouget virus are not protected against natural infection until after this period of time. I have also shown that at least twelve to fifteen days are necessary to secure an immunity following the use of an attenuated virus in bovine homorrhagic septicemia. In typhoid fever, the immunity acquired as a result of infec- tion is even longer in establishing itself, as the possibility of relapse in well-advanced convalescence shows. This lapse of time, twelve days, is therefore a minimum insofar as naturally acquired organic immunity is concerned. There is nothing to be gained here by discussing laboratory experiments concerning the development of immunity in refractory ani- mals, for such experiments are laboratory phenomena only and have nothing necessarily in common with natural conditions. 170 INTRODUCTION Finally, in dissolving the bacteria, can it not be an indirect factor in naturally acquired antitoxic immunity? These are the points which we will consider in the second part of this monograph. It would seem that the only method which ought to be followed in investigating the relation between immunity and a principle to which one may attribute a protective power ought to be founded on the observation of natural disease, and that the parallelism between the state of the patient and the presence, and the potency, of the supposed protective principle, ought to serve as the cri- terion for determining its true role. If a parallelism exists, it may be regarded in the possible relation of cause and effect, and one can then turn to the counter-test for confirmation. If a parallelism does not exist, the relation of cause and effect cannot be invoked, and the principle under consideration cannot play an active role in the processes of recovery. This is the method of investigation, the only logical one it seems to me, that I have applied in investigating the relationship between the bacterio- phage and immunity. I consider, in fact, that a theory of im- munity based only on simple observation or on comparison, always remains subject to discussion. For simple observation readily leads to error, especially when the observations are made on refractory animals and are not found to be confirmed when applied to a susceptible animal. It is certainly much easier to experi- ment in the laboratory with caged animals; to study the immunity against the cholera vibrio, for example, on a guinea pig which is resistant to the disease naturally, than to run everywhere in search of epizootics in order to study the disease in its normal environment. But common sense alone is adequate to make it apparent that the first method can prove nothing, and that only observation of the natural disease, complemented by experimen- tation on an animal susceptible to it, can give results that have an absolute value. It may indeed seem strange that we use the word " immunization" in speaking of a refractory animal, since the refractory state already represents immunity carried to its highest degree. I wish to be free of such criticism and will thus follow an order which seems to me the most logical. We will observe first, INTRODUCTION 171 natural infection and we will see if the search for the bacterio- phage and the determination of its properties at different stages of the disease and of convalescence provides results which have any relation to the pathologic condition of the patient. I have selected for this investigation different infections, enteric and septicemic, diseases of man and of animals, with which we will show that defense by the bacteriophage is a phenomenon of a general nature. Some of the diseases studied are epidemic, and we will have occasion to note the effect of the bacteriophage on the progress of the epidemic itself. If the bacteriophage is an agent of immunity, it will not appear only at the exact moment when it is most needed. It should be a normal inhabitant of the intestine. We will look for it, then, in the healthy individual, choosing subjects throughout all animal species, and this will show the generality of the presence of the bacteriophage. Finally, we will attempt the counter-test. If, in the susceptible animal the principle of antibacterial immunity resides in the bacteriophage, the administration to a susceptible animal of a bacteriophage active for a given bacterium ought to render the organism resistant to the disease caused by this bacterium. Thanks to the kindness of M. Roux, Director of the Pasteur In- stitute, and to M. Yersin, Director of the Pasteur Institute in Indo-China, I have been able to accomplish in its entirety the program which I have outlined. In France I have had the opportunity to study the role of the bacteriophage in intestinal diseases, and during the course of a year spent in the Pasteur Institute at Saigon, I have been able to verify the generality of the phenomena observed, by a study of a highly contagious septidemia, — barbone in the buffalo, — and by a disease of glandular localization, — plague. It is certain that a theory of immunity based on the bacterio- phage, that is, on an autonomous organism, is so far outside of all present opinion that it will stir up at first incredulity and will be called a "finalistic theory," — a synonym of " anti-scientific." I affirm that from my point of view this theory can not be "final- istic." "To be is to struggle, to live is to conquer," a very just statement by Le Dantec. It is all contained in a single word — 172 INTRODUCTION evolution. A being which evolves is necessarily a being which lives, which adapts itself, and which conquers. From the instant that it ceases to adapt itself — to evolve — it dies. Evolution is always conducted according to the law of least effort. The multi- cellular organisms have profited by securing for their defense the parasitism of the bacteriophage for the bacteria; which is only a chapter in the universal struggle. If, among all living beings, the bacteria alone escaped para- sitism, where would we arrive? It is very simple. One of two things would take place. Either evolution would not extend beyond the stage of the unicellular being, or evolution would be accomplished in another manner and immunity would be as- sured by other means; a simple matter of adaptation. The bac- teriophage does not exist for the defense of the superior organ- ism against the bacteria, it exists simply because in the course of evolution certain germs have parasitized others. Nothing in nature exists simply for an end, for nature is not an end. That there exist on the earth thinking beings, or that they might not have been, is a perfectly negligible incident. Is this point of view "finalistic"? But what does it matter; a scien- tific theory is true or false according to the proofs upon which it is founded. Each time that we will speak in the course of this discussion of "antibacterial immunity" it is essential to understand "anti- bacterial immunity in a susceptible individual." These observa- tions and experiments, as I have already remarked, are concerned with this and this alone. Up to the present I have paid little attention to the phenomena of immunity in the refractory animal. It indeed seems, in general, in the special type of immunity which characterizes the refractory state, that the elimination of bacteria which may gain access to the body, and which because of the refractory state are not pathogenic, is effected by phagocytosis. In this special case, defense by the bacteriophage could not possess, the greater part of the time, the opportunity to act. Phagocytosis is ac- complished too rapidly to allow the bacteriophage time to in- crease its virulence toward the bacterium which is introduced into the organism. CHAPTER I THE BACTERIOPHAGE IN DISEASE Choice of Diseases to Study. Bacillary Dysentery. B. coli Infections. Typhoid and Paratyphoid Fevers. Avian Typhosis. Barbone. Bu- bonic Plague. Flacherie. Conclusions. CHOICE OF DISEASES TO STUDY From the point of view of the study of immunity human infec- tion offers an inconvenience. Man is not available for experi- mentation; observation alone is permitted. On the other hand, the study of a human infection, such as typhoid fever or cholera for example, in a refractory animal — and they are all so — can only lead to illusory results. Study of disease in the animal, on the contrary, permits of confirmatory experimentation upon the susceptible animal itself where error is no longer unavoidable. However, this method of procedure is very complicated; the dis- ease does not come to us, we must go to it. The study of typhoid fever and of dysentery allows us to show by observation the role of the bacteriophage in the course of the disease. These same phenomena may be reproduced in the course of infection in animals, and it is possible with the latter to conduct such experiments of verification as will confirm that which simple observation has already shown. In order to ascertain the influence of the bacteriophage on the morbid state, a method which consists in investigating at random the activity of the bacteriophage in a specimen of material taken at any time whatsoever will not lead to any result. It is neces- sary to take the patient as quickly as possible after the inception of the disease and to examine the feces each day until recovery is complete. The daily findings are then plotted in a curve which is superimposed on that expressing the general state of the indi- vidual, such as the number of stools in dysentery, or the tem- perature in typhoid. A comparison of these two curves allows one to draw a conclusion. This mode of procedure necessitates 173 174 THE BACTEBIOPHAGE considerable work, but it must be applied for it is the only proce- dure which will allow of a conclusion. We have seen in the course of the preceding chapters that the virulence of a strain of the bacteriophage is rarely limited to any one particular bacterial species, but exercises in general, with variable intensity, its action on several species pertaining to the same group or to closely related groups. It would be practically impossible, in view of the length of the operations, to investigate all of the bacteria which may be attacked by a bacteriophage isolated from the stools of a patient at any given time. The procedure must, therefore, be reduced to a systematic examination in each case of the virulence of the bacteriophage toward the particular bacterium involved in a causal relationship. The virulence should be determined for a type strain of this bacterium which has been cultivated for a long time in the laboratory, for the strain derived from the pa- tient himself, and for B. coli. Eventually, the investigation may be extended to bacteria belonging to the same group or to related groups. We know that the virulence of different strains of bacteriophage for a given bacterium is far from constant. It varies throughout a scale which goes from zero to an activity such that it is suffi- cient to add only a few germs to a heavy suspension of this bacterium in order to obtain within three or four hours a complete and permanent lysis, all the bacteria being then destroyed. Be- tween these two limits, — no virulence and extreme virulence — all intermediate degrees are possible. A weak virulence we have seen, may be enhanced in vitro, but in so far as the study of im- munity is concerned, the point in which we are interested is the virulence presented by the bacteriophage in the organism at the moment of observation; or the actual virulence of the bacterio- phage contained in the filtrate prepared directly from the feces at any given time during the disease. As we have also seen, the appearance of cultures of the bacterio- phage in bouillon or on agar enables us to evaluate its virulence for the bacterium in question. In order to facilitate explanation in the further discussion of the subject we will adopt a scale of virulence fixed as follows: THE BACTERIOPHAGE IN DISEASE 175 0 = no virulence toward a given bacterium. Normal cultures of the bacterium develop in bouillon or on agar, whatever the quantity of the fil- trate from the feces which had been added. + = weak virulence. The growth in bouillon of the bacterium to which the filtrate has been added is apparently normal. Transfer of this culture to agar gives, after incubation, a culture layer showing a few minute plaques. Some of the bacteriophagous germs have therefore attacked the bacteria and have formed colonies. + + = medium virulence. The culture of the bacterium to which the filtrate has been added is almost normal in bouillon. Transfers of this cul- ture to agar give, after incubation, either a culture layer of the bacterium studded by very numerous colonies of the bacteriophage, pre- senting an appreciable surface area, or of fragments of bacterial culture because of the very great number of bacteriophage colonies, high virulence. Lysis of a bacterial suspension is obtained but secondary cultures constantly develop. The reinoculations on to agar remain sterile or give only rare colonies of the bacterium, extreme virulence. The bouillon suspension shows complete, and, in general, permanent lysis. Inoculations on to agar always remain sterile. Obviously, it would be possible to establish a more detailed scale of virulence. (Moreover, this has been done in the curves which will be given, where the interval between no virulence and extreme virulence has been subdivided into ten steps, in accord- ance with the aspect of the cultures, the number of colonies of the bacteriophage, and the size of the plaques, which bear a rela- tion to its virulence.) Practically, the appreciation is adequate with four steps, particularly in view of the fact of the extreme variability of virulence in the bacteriophage in the body of a single individual from one time to another. 176 THE BACTERIOPHAGE The expression "Shiga + + ++, Hiss +, Flexner 0, Typhoid ++, Para A 0, Para B 0, B. coli + ++" means, then, that the bacteriophage contained in the filtrate derived from the stool of an individual presents an extreme virulence for B. dysenteriae Shiga, a weak virulence for B. dysenteriae Hiss, an average viru- lence for B. typhosus, and a high virulence for B. coli, with none for B. dysenteriae Flexner or for the paratyphoids A or B. BACILLARY DYSENTERY The subjoined curves show, much better than any explanation, the relations which exist between the condition of the patient and the virulence of the intestinal bacteriophage against this pathogenic bacterium. The upper tracing gives the number of stools in 24 hours; the single line indicating stools without blood, the double line those containing blood and mucus. On the lower portion of the chart is indicated (1) by the dotted line, the viru- lence of the bacteriophage for the colon bacillus; (2) by the broken line, the virulence of the bacteriophage for the stock strain of the Shiga bacillus which had been maintained for a long time under laboratory cultivation; and (3) by the heavy line, the virulence for the Shiga strain taken from the patient himself. The five cases given as examples have been treated at the Pasteur Hospital. It has thus been possible to follow them with all necessary attention and to obtain material for examination as often as the investigation demanded; at least once, often several times, during the course of each day.1 For these examples, cases of different severity have been selected. In all of them B. dysenteriae Shiga was isolated from the stools at the beginning of the disease. 1. Germaine Mel. . . . (sixteen years, fig. 1). This was a mild case of dysentery. The patient was an inmate in an institu- tion where there were about thirty young girls. During the period from the 12th to the 22nd of July about twenty of these girls presented intestinal disturbances of sudden onset, accom- panied by a profuse diarrhea, followed by a rapid amelioration 1 1 must here thank the sisters, nurses in the Pasteur Hospital, who have with unwearying kindness provided me with the numerous specimens which have allowed me to follow the condition of the patients. THE BACTERIOPHAGE IN DISEASE 177 of symptoms. Within one or two days after the onset all had again become normal. In only one or two cases did the stools contain traces of blood. In order to establish a diagnosis the directrix was asked to send a patient to the Hospital during the earliest symptoms. Germaine Mel. . . . entered the Hospital on the 18th of July. From the first stool passed after her arrival a bacillus presenting the biochemical characteristics of the Shiga bacillus was isolated Day of the Disease eo FIG. 1. GERMAINE MEL. . . . DYSENTERY (SHIGA) B. dysenteriae from the patient- Virulence for "I B. dysenteriae, stock strain — B. coli. . after considerable difficulty. It was inagglutinable, and it was only after three passages on agar that agglutination was secured (1:500). As can be seen in the tracings, the number of fluid stools, seventeen on the first day, fell quickly during the second day to two, without medication. The intestinal bacteriophage, isolated from the fifth stool of the first day, was endowed with an extreme virulence for the bacillus 178 THE BACTERIOPHAGE of the infection, and with a somewhat lower grade of virulence for the stock Shiga strain and for B. coli. The stools of eleven of the inmates of this institution were examined. Among the number were nine who had shown in- testinal disturbances two or three days previously. Two had shown no morbid symptoms. All of those examined contained a bacteriophage with a high or extreme virulence for the Shiga strain isolated from the stool of Germaine Mel. ... as well as for the stock strain of Shiga and for B. coli. Day of the Disease ' — — - JN umber 01 stools + per 24 hours + + '• . 9 o \ 1 16 II U i? (x iri it 4 | \\ [ b I I q 0 II 1 1 \ i£ U!| jjj m ti u w 14 tr •V 11 I « M 20 1 ', - ^ S^ s j | \ ^ » — •^s. f — — — — - — — - --/ /• s s \- 1 \r? f \ '' j I / <"l \_ -j s s V \ • \ \ I - - ! \ \ i i FIG. 3. VICTOR KER (6 years) DYSENTERY (SHIGA) f B. dysenteriae from the patient Virulence for \ B. dysenteriae, stock strain - - (B.coli Stools contained blood 3. Victor Ker. ... (5 years, fig. 3). The dysentery was due to the Shiga bacillus, was of moderate severity, and was contracted by contact with the patient next discussed. When admitted to the Hospital, on the third day of the disease, the intestinal bac- teriophage already manifested an average virulence (++) for the stock Shiga strain as well as for the strain isolated from the patient. This virulence increased rapidly and maintained a high value up to the time of complete convalescence (+H — h). It then abruptly disappeared. 180 THE BACTERIOPHAGE Day of the Disease 1 t- |S \ * J \ r ^ ^ s ./ '\ , *• K'. 21 \ \ s c .... ' \ f \ ,: "j V •-• \ 1 ' I •^ : ) ' 1 FIG. 4. JEAN KER (6 years) DYSENTERY (SHIGA) {B. dysenteriae from the patient B. dysenteriae, stock strain B. coli Stools contained blood 4. Jean Ker. . . . (six years, fig. 4). This patient was a brother of the foregoing. The general condition was poor when admitted to the Hospital on the third day of the disease. There were from twenty to thirty bloody stools a day; a severe dysen- tery due to the Shiga bacillus. On the fourth day of the disease there were twenty-four bloody stools. The bacteriophage was feebly active (+) for B. coli and was inactive for the Shiga bacillus. The record shows the following: VIRULENCE OF THE BACTERIOPHAGE AGAINST DAT OF DISEASE NUMBER OF BLOODT STOOLS B. dysenteriae (patient) B. dysenteriae (stock) B. ccli 5th 23 0 + 4. 6th 13 0 _]_^_ _!__!_ + + 7th 9 0 +4-4- + + + + 8th 12 0 +4- + + + + 9th 11 0 + 4.4.4.4- 10th 12 -f _l_++4 4-4-+ llth 12 +++ 4-4-4-4- +++ 12th (4 of 6 stools without ++ + _l_4-_l_4- + blood) THE BACTERIOPHAGE IN DISEASE 181 From this time on improvement became more and more marked. The activity of the bacteriophage did not disappear after con- valescence had been established. In the first three of these cases the dysentery was mild. The bacteriophage was active at the onset, the bacterium did not acquire a resistance, and its growth was quickly suppressed. In the last case there was a struggle and the bacillus acquired a resist- ance which was finally overcome. The condition of this patient was much more serious. 5. Lans. . . . (seventy years, fig. 5). This case illustrates an extremely severe dysentery due to the Shiga bacillus. The patient entered the Hospital on the second day of the disease. In this case the struggle was prolonged, with fluctuations due to the mixed cultures formed in the intestine. The condition of the patient registered faithfully the changes in the struggle. It may be noted particularly that the bacteriophage manifests a transitory activity on the eleventh day of the disease and the stools temporarily lose their bloody character. But the bacillus increases its resistance and this permits it to develop, and blood reappears in the stools. The disease is only definitely overcome at a time when the virulence of the bacteriophage is sufficiently high to dominate the resistance of the bacterium. Aside from the five cases cited as examples others have been followed, both in France and in Indo-China. Seventeen other cases differing in severity were examined daily, and twenty-nine more were observed less frequently. In all of the cases the activity of the bacteriophage was manifested in an identical manner: 1. In case of recovery, the virulence of the bacteriophage com- mences to manifest itself in a marked manner toward B. coli. 2. The virulence next extends to the type strain of the Shiga bacillus, that is to say, toward a strain which has been for a long time under artificial cultivation and which, for this reason, has been deprived of much of its resistance. 3. It manifests itself next, more or less quickly, toward the Shiga bacillus isolated from the patient himself at the onset of the disease.2 2 Obviously it is necessary to preserve this strain without replanting. The isolated colonies obtained on the original plates are planted on several 182 THE BACTERIOPHAGE Day of the Disease £_j ; II • 1 FIG 5. LANS (70 years) DYSENTERY (SHIGA) f 71 Jvewtvrlno (rn-n-i fVia T->oliVnt Virulence for \ B. dysenieriae, stock strain (B.coli Stools contained blood ; it ! i \ s\ f <. 1 1 \ s;i ^i 1 1 C — i ) 1 1 ^^ ..•- • en "jj .^* e:i <^ i ^~ ,,-- 5; > 1 1 ,- =- <[" Ih- ... v._ t;l ^ ) \ •• 1 3 ; I a i | \ JO ( II \ ^ ) 1 E ?rl ^ xx ; V) } V .^1 i \ Si \ \ S i 4] 7 3 C / =c t y 1 ,• 5.1 § 5 s 2 1 <•' y;i / i ^x. i jjjj \ ! 5l r ! y' . 1 r:i II s r 1 sn J /~ •• 1 iii I ? '. 1 Vj| I 1 ^ i ^ 1 ^*~^_ .,-1 ^ I 1 . ^^ I si ^J jj s| ^^ 1; jj /*"" | CT- / | ,--- ~l \ | 5S ^ H_ g| X^ - 1 ^ 3| f | •^1 > 1 -^ ,, — r 7JI _^ ^ -S t" <^ \ - | ^ — | i 2J — — ' ^_ Ii T>«J >-^! 1 1 -• 1 D ->! ill ^ 1 DC .„- - *> 1 ii — •=! •*TTj hi -! 1 ' S il > £ 2 - ol aS^ndouaioBff S{oo^s jo aaquin^j THE BACTERIOPHAGE IN DISEASE 183 4. In all cases the fluctuations in the virulence, as well as the fluctuations in the resistance of the bacteria, parallel the state of the patient, and the onset of improvement coincides with the moment when the virulence of the bacteriophage dominates clearly the resistance of the bacterium. We thus see reproduced in vivo the same mode of action as that observed in vitro; per- manent and complete lysis, mixed cultures with negative trans- fers, mixed cultures with alternations in the dominating force. In Indo-China an opportunity was afforded to follow four fatal cases of bacillary dysentery in natives. At no time during the course of the infection did the intestinal bacteriophage show a trace of activity for the Shiga bacillus, either for the stock strain or for that isolated from the stools of the patients. A last case offers an especial interest, for it shows that it is not only in vitro that the bacteria are able to become refractory to the action of the bacteriophage. Although this may occur in vivo these cases must be very rare, even exceptional. Alix Desp. . . . (fifty-six years). The patient entered the Pasteur Hospital on September 26, 1919. At the time of ad- mission there was a profuse mucous diarrhea with thirty to forty stools a day. Examination of the intestinal contents gave an almost pure culture of a dysentery bacillus presenting atypical characters, as follows: Non-motile bacillus. Gram negative. Indol positive. No black- ening of lead acetate agar. No change in neutral red media. Litmus sugar agar media not fermented with any of the sugars. In Barsiekow's medium, maltose and lactose are unchanged, glucose and mannite are turned red. After six transfers on agar it agglutinated to the titre (1 :6000) with a Hiss agglutinating serum, to 1:400 with an anti-Flexner serum of which the titre was 1:6000, and was not agglutinated at 1:20 by an anti- Shiga serum. In spite of these atypical characters it is, then, a Hiss strain possessing weak fermentative properties. When secured from the body this bacillus was not affected by a bacteriophage very virulent for a normal Hiss bacillus, but it agar tubes and a portion is taken from these tubes for the tests conducted during the course of the disease. It is well-known that resistance is attenuated by successive transplantations. 184 THE BACTEEIOPHAGE was lysed after about a dozen transplantations. It was, then, a bacillus which was refractory to the bacteriophage when taken from the body. At the same time a strain of bacteriophage was isolated from the stools of the patient. This presented the following virulences: Shiga 0, Flexner + + ; stock culture strain of Hiss + + + + ; B. coli + ; the Hiss strain from the patient +. After twelve sub- cultures of the Hiss strain from the patient the virulence of the filtrate was again tested. Perfect lysis was secured, showing that the bacillus had lost its resistance by transfers on agar. The bacteriophage of the patient is active to a maximum degree against a stock strain of the Hiss bacillus but it is only slightly active for the individual strain causing the infection, with which it forms, in vitro, mixed cultures indefinitely cultivable. There was likewise in the intestine of the patient a mixed culture of the bacteriophage and the refractory Hiss strain. In spite of every care and repeated injections of anti-dysentery serum the patient became more and more weak; the temperature oscillated between 38° in the morning and 40°C. in the evening; the number of stools gradually increased and became uncountable on about the thirtieth day; and at about this time the patient fell into a marasmic condition, the temperature stayed at about 38°C. and death occurred on the thirty-fifth day. Bacteriologically, the stools, tested each day, showed an al- most constant bacterial flora. The pathogenic bacillus was always abundant, often in almost pure culture, and presented the char- acteristics described. The virulence of the bacteriophage in- creased continuously until the fifteenth day when it became fixed, showing: — Shiga + + + + ; Flexner + + + + ; Hiss + + ++, B. typhosus + + + ; B. paratyphosus A + + + ; B. paratyphosus B + + + ; B. coli + + + + ; bacillus of the patient 0 (completely refractory) when freshly isolated, + -f- + after fifteen transplants. At autopsy3 there was isolated from the contents of the colon, from a fragment of mucous ulceration, from the liver, from the spleen, and from the heart blood, a Hiss dysentery bacillus, pre- senting the same characteristics as that which had been isolated 3 Performed by L. Ge"ry, whom I thank for the specimens he was kind enough to send me. THE BACTERIOPHAGE IN DISEASE 185 at the beginning of the disease . From all the organs a bacteriophage was isolated presenting the same characters as that which had been isolated from the stools and whose virulence has been indicated. This case, altogether exceptional (I believe that it is the first case reported of a B. dysenteriae Hiss septicemia) is very interest- ing for it shows in an unquestioned manner the role that the bac- teriophage plays in the defense of the organism. In all of the cases examined heretofore we have seen, either recovery starting from the time when the bacteriophage had acquired sufficient virulence to dominate the pathogenic bacillus, or death in the case of the failure of adaptation. In this last case, the bacteria developed a refractory condition, the bacteriophage was overcome and remained without action whatever its virulence may have been. The barrier thus being lacking, the bacteria developed freely and invaded the entire organism. The patient succumbed to a septicemia with the Hiss bacillus. This exceptional case provides us with new information. A bacterium is pathogenic for a given organism if it secretes sub- stances toxic for the cells of this organism. It is the more viru- lent the more capable it is of development at the expense of this organism. The dysentery bacilli are pathogens because of this secretion of toxic substances, for they do not invade the organism, but remain localized in the intestine and in the intestinal mucosa. Nevertheless, in the case of the woman Desp. . . . the Hiss strain was accidentally endowed with an extreme virulence, and this solely because the bacteriophage had been overcome. This suggests an idea which we will have occasion to confirm in the following chapters, — that the virulence of a bacterium at any given moment is the greater if its resistance to the bacterio- phage is at this time high. The case Desp. ... is exceptional. As a general rule death occurs in dysentery, not because of the acquisition by the bac- terium of a refractory condition, but by a failure of the bacterio- phage to adapt itself to bacteriophagy toward the pathogenic organism. In the four cases mentioned above which were fatal, a bacteriophage active forlthe Shiga bacillus could not be isolated at any period of the disease. 186 THE BACTERIOPHAGE During the course of the epidemic of dysentery which occurred in the region of Paris during the early autumn of 1918, an oppor- tunity was given to observe twenty-nine cases of benign diarrhea. In all of these cases a bacteriophage of very high or extreme ac- tivity for the Shiga bacillus was isolated from stools taken the day after the malaise. This bacillus was the cause of all the severe cases studied at this same time. Living at this time in a locality (Meulan) where several severe cases of dysentery were noted together with a large number of cases of transitory diarrhea, I examined the stools of nine per- sons who were healthy, but who lived in contact with individuals who had had dysentery. From these nine individuals a bacterio- phage of average or high activity for the Shiga bacillus was iso- lated. We have noted above that the same fact was observed in the institution where Germaine Mel. . . . had contracted dysentery. Individuals who are exposed to infection and who resist show therefore in their intestine a bacteriophage virulent for the causative pathogenic bacillus, exactly like the affected individuals who recover. As a result, in an epidemic period the simple cases of diarrhea must in reality be cases of aborted bacillary dysentery, thanks to the rapidity with which the intestinal bacteriophage adapts itself to bacteriophagy against pathogenic bacteria. And healthy individuals, living in contact with affected people, are only spared by virtue of a still more rapid adaptation occurring before mor- bid symptoms appear. We will find comparable facts in all the diseases which we will discuss. To summarize : the pathogenesis and the pathology of bacillary dysentery are dominated by two factors, operating in different directions; the dysentery bacillus as the pathogenic agent and the bacteriophage as the agent of immunity. The history of a case of dysentery is only the story of the struggle, occurring with- in the body, between these two factors, and the condition of the patient faithfully reflects the vicissitudes of the struggle. In case of a rapid enhancement in the virulence of the intestinal bacteriophage toward a pathogenic bacillus, the latter is unable to develop a resistance and is destroyed in the struggle, so that the disease aborts before the appearance of any symptoms or manifests itself only in a transitory disturbance. THE BACTERIOPHAGE IN DISEASE 187 The increase in the virulence of the bacteriophage for the in- vading bacterium may be retarded for one of two reasons : — First, as a result of unfavorable intestinal conditions. (We have seen the considerable importance, in vitro, of very slight variations in the reaction of the medium on the development of the ultra- microbial bacteriophage.) In accordance with the chemical and physical state of the intestinal contents, one bacterium is favored at the expense of another; the intestinal fermentations, and as a result, the reaction of the medium will vary according to the predominating flora. The development of the bacteriophage is then doubly influenced, first, by a change in the state of the medium itself, and second, by changes in the flora which increase or de- crease, according to circumstances, the bacterial species at the expense of which it normally develops. This of course, necessi- tates variations in virulence in response to the variation in the bacterial species. Moreover, it has been known for a long time that catarrhal diarrhea affects (provoked by the ingestion of un- digestible foodstuffs, of green fruits in particular, or by the "froid au ventre" so common in tropical countries) the incidence of certain intestinal diseases — dysentery and cholera among others. Second, as a result of a more or less marked degree of resistance to the bacteriophage of the invading bacillus. We have seen that in the course of the disease the pathogenic agent defends itself. Such a bacillus in a state of resistance, ingested by a healthy person will develop in spite of the presence of a bacterio- phage, particularly if the latter is but slightly active, whereas a non-resistant bacillus is destroyed without a struggle. In cases of bacillary dysentery, even very severe, but in which the patient improves rapidly, the bacteriophage manifests its presence in a very active manner at the outset, not only for labora- tory strains of the bacillus, but for the strain secured from the patient himself, and this takes place at the moment when the symptoms begin to improve. There may be a rapid increase in the virulence of the bacteriophage without a corresponding resist- ance in the bacterium. In cases where the disease is prolonged, two cases may be considered : 188 THE BACTERIOPHAGE 1. The bacteriophage shows no, or but slight, activity as long as the condition of the patient remains stationary. The improve- ment occurs when the activity of the bacteriophage manifests itself in an energetic manner, not only for the stock cultures of the bacillus but also for the strain derived from the patient. There has been a delay in the adaptation, then a sudden acquisi- tion of a high virulence. Recovery takes place promptly, for the pathogenic bacterium is not able to develop a resistance. 2. At a given moment of the disease the virulence of the bac- teriophage manifests a more or less energetic action on the stock bacilli, but on the contrary, it is inappreciable or but very weak on the strain taken from the patient. Here there has been a de- lay in the adaptation, since the bacteriophage has gradually ac- quired virulence for the pathogenic bacillus, but this has allowed sufficient time for the creation of a resistant race of the latter. As a result there is a struggle, and the condition of the patient reveals the fluctuations of the struggle. This conflict is particularly to be noted in cases of long dura- tion with a relapse. During the latter, especially, the virulence of the bacteriophage shows daily fluctuations. At certain times it may be extreme for the stock culture, although uniformly very weak for the strain of the infection. Recovery begins to take place at the moment when the bacteriophage shows an activity as intense for one strain as for the other. The disease has a fatal issue in two cases : 1. When the bacteriophage exerts no protective action through a lack of adaptation to the pathogenic bacillus. Here there is no struggle at all, and the bacterium develops freely. In the great majority of such cases non-adaptation is the cause of death, which then occurs quickly. 2. In certain exceptional cases the pathogenic bacterium ac- quires an almost absolute resistance, — a refractory state. And the bacteriophage, whatever the degree of virulence it acquires, remains ineffective. From this moment, when the bacterium becomes equal to the bacteriophage, the entire body is invaded and death ensues after a greater or less length of time. THE BACTERIOPHAGE IN DISEASE 189 COLON BACILLUS INFECTIONS Sometimes the colon bacillus may become pathogenic and may be encountered as the etiological agent in diverse localized infections, or even in septicemias. It at first appears strange that so common an organism, a normal inhabitant of the intestine, should at a particular time develop pathogenicity. There must be "a something" which differentiates the pathogenic B. coli from the banal B. coli. It is this which I have tried to determine. Five specimens of infected urine secured from individuals with pyelonephritis have been examined. In all of these cases not only was the colon bacillus present, but there was a mixed culture of B. coli and the bacteriophage, as shown by inoculation of the urine on agar. In one of the cases simple plating of the urine on agar gave a colon culture studded with plaques, in the other four, agar cultures made after a bouillon growth gave the same appearance. The colon bacillus possessed a high resistance, although it was not entirely refractory. Thus the struggle con- tinued in the organism. The ordinary B. coli is not pathogenic. The resistant B. coli becomes so because of its resistance to the action of the bacteriophage. The history of a morbid condition is the history of the struggle between the bacteriophage which attacks with its virulence, on one side, and a bacterium susceptible of resistance on the other. Moreover, the struggle can be continued as long as the bacterium secretes products toxic for the infected body, but in the last analysis, it is the issue of this conflict which decides the fate of the individual. We will have occasion to return to the case of pyelonephritis TYPHOID FEVER AND THE PARATYPHOID FEVERS Several cases of typhoid fever of varied severity have been studied by the same method as that employed in bacillary dys- entery. Fourteen of these were in the Pasteur Hospital for treatment, and of these the stools were examined at least once a day throughout the course of the disease and in convalescence. Fourteen more under treatment in other hospitals were followed 190 THE BACTERIOPHAGE with somewhat fewer examinations. In all the charts which follow, the following data is presented; — in the upper portion is the curve showing the temperature ; in the lower portion there are three tracings, (1) in dotted line, showing the curve of the virulence of the bacteriophage for B. coli, (2) in broken line, showing the virulence of the bacteriophage for an old laboratory strain of B. typhosus, a strain which has undergone a great many transfers on laboratory media (this same strain was used in all the cases studied), and (3) in solid line, indicating the curve of virulence Day of the Disease 3 1Q u (X 0 u mi 2 U 'i i. U(H y w a V. s ti F3 io il H u JW jt it V u •>\k. <i uj *r 40' 39' 38' V , \ -^ r ^ v — \ V ^ X \ Ti — ^ ^ 3 V . — • ^ ^ \ _ __ D 4 ... - '/* « 7- £ .-: i '--• 4 " s -. - /: J "\ \. " i 1 \ S '•'. / ;, \ / V^ < s v 1 \ - - ; -L. 1 '/ - — v - - / - - - - - - ^ \ -x • FIG. 6. MARDE Mo (55 years) CLINICALLY, TYPHOID FEVER Virulence for B. typhosus B. coli. . B. paratyphosus A — • B. paratyphosus B - B. dysenteriae Shiga of the bacteriophage for the strain of B. typhosus from the patient himself, isolated either by stool culture or by blood culture. In order to use bacilli as comparable as possible with those found in the body of the patient the strains were transplanted as infrequently as possible. In each case an agar tube was inocu- lated with a colony taken from the primary culture, and each time that a fresh culture was needed for the preparation of sus- pensions against which the nitrates containing the bacteriophage from the patient were to be tested, it was always taken from this tube. In this way, the bacteriophage throughout the course THE BACTERIOPHAGE IN DISEASE 191 of the disease was tested against a culture as nearly constant as possible, uniform especially from the point of view of the resistance of the bacterium. For the first three curves only (figures 6, 7, and 8) the organism of the patients had not been isolated (they had fevers which appeared benign) and the curves of the virulence of the bacterio- phage against the bacillus of the patient is, of course, lacking. For these three cases the virulence of the bacteriophage against Day of the Disease 41' 40' 5;' r 1 il q i3 u u L! Lk K 'I ir H !» )l 11 51 U X K Jf M J1 *i 11 t 4i ^ it «i Qf4 <|j? t 5( J! [j 51 " ) r X " ' « i ^ /v -^ JJ^ 7 \l \ ^ y f^ 1 i q " s •3 =v ~ *r - - - •:.- -•: "x .'- a. -. - -s. a "\ .- -. -, . \ V - ^, , ,* '' -• - •' *V •': " > .. ^ / \ \ ,x •" - - •' •- * \s \ / v \ \ : FIG. 7. Louis Pi (17 years) CLINICALLY, TYPHOID FEVER Virulence for B. typhosus B. paratyphosus A — B. coli B. paratyphosus B — B. dysenteriae Shiga . a Shiga dysentery strain, and against the paratyphoids A and B are given. We will select as examples cases of different severity. 1. Mild infections These were cases of typhoid fever or paratyphoid fever with a mild course. Clinically they were typhoid fever but the blood and stool cultures were negative. The curves for these three cases are given on pages 190, 191 and 192. 192 THE BACTERIOPHAGE 1. Marie Mo. . . . (fifty-five years, fig. 6). 2. Louis Pi. ... (seventeen years, fig. 7). 3. Franc. ois Jod. . . . (thirty-four years, fig. 8). In these cases the virulence of the intestinal bacteriophage was determined for B. coli, B. typhosus, B. paratyphosus A and Bt and B. dysenteriae Shiga. It is needless to comment on these observations, since examination of the curves is more instructive than would be an explanation. Day of the Disease O' >>.2 ^5 '58 <« I 9 ,: n " [3 IN U it '? '7 g J 2/. uh-i t» U ii B i! 2 ,, j, uly M sr % 1) " If J<7 •ikj r 4? . " i !\ _. x -' ••" ** -> / I \. ,y A \ / ** ? v ~ •• ~' —' "^ '\ - U. - - - sv \; 4. FIG. 8. FRANCOIS JOD (34 years) CLINICALLY, TYPHOID FEVER Virulence for • B. typhosus coli. . B. paratyphosus A - B, paratyphosus B - B. dysenteriae Shiga What is the causative bacillus in each of these three cases? It is indeed difficult to make a diagnosis by means of the bacterio- phage, which as we have seen, but rarely develops a single virulence. This virulence extends to other bacteria of the same group to a more or less marked degree, and this fact is particularly in evidence when working with the representatives of the colon- typhoid-paratyphoid-dysentery group. It appears, however, in the case of Louis Pi. ... that the causative bacillus must have been the typhoid bacillus, with Marie Mo. . . . B. paratyphosus A, and in Frangois Jod. . . . B. paratyphosus B. THE BACTERIOPHAGE IN DISEASE 193 It should be noted that in these three cases in which improve- ment was rapid, the curves representing the virulence of the bacteriophage are comparable. It is also to be noted that in all, the accessory virulences for B. dysenteriae Shiga and for B. coli are very high, and that the acquisition of virulence for B. typhosus and B. paratyphosus A and B is early and is maintained up to the beginning of convalescence. In the case of Louis Pi. . the abrupt deflection in virulence on the twenty-second day preceded a slight relapse which occurred on the twenty-fifth to the thirty-second day. This complication did not prove serious since the virulence of the bacteriophage increased gradually from the twenty-third day. 2. Severe injections The three cases cited here were serious, with both stool and blood cultures positive. All were infected with B. typhosus. 1. Renee Mar. . . . (thirty-two years, fig. 9). The bacteriophage was from the beginning virulent for B. coli, and remained so during the course of the disease, throughout convalescence, and up to the time when the patient was dis- charged from the hospital completely cured. It may be noted that the acquisition of virulence by the bacteriophage for the bacillus of the infection coincides with the first defervescence. Then this virulence became reduced and the temperature again went up. The infection is definitely overcome at the time when this virulence is again established. 2. Juliette Ou. . . . (thirty-six years, fig. 10). 3. Jeanne Del. . . . (twenty years, fig. 11). The curves for these two cases are self-explanatory. 8. Typhoid fever with relapse 1. GilberteFon. . . . (four years, fig. 12). On the sixteenth day the disease appeared ended. However, the virulence of the bacteriophage toward the bacillus of the patient disappeared before the end of the crisis and the destruc- tion of the pathogenic bacteria was not complete. Whereupon there was a relapse, very severe, which did not show improvement until the bacteriophage recuperated with a virulence sufficient to control the resistance of the bacterium. 194 THE BACTERIOPHAGE ft IL I- J5^-L_ ttj ^ \* * r si ^-N 4> en ^5 >H ji cs ^3 THE BACTERIOPHAGE IN DISEASE 195 4-1 £L 196 THE BACTERIOPHAGE / / I 1 £J I I : 9- 1 ! A. 1 1 1 i ^ 1 1 ! *i- II ! ; •£ j " ! b L H . 5 3 C ' ' i" 1 1 T x * i / 5 > 1 1 I fc- \ C^ ) i cr- f c; v i fcj )l 1 ti 1 | >• 2 ^ 1 f Ci ) 1 s. / L / ^ 3 a 2 1 5* •^ i ^i 1 •£ i ^ "\ 2 5? ^y i S' ,' ^^* 7 <^ ^^^ 1 >T f ^^*"1 t^j 1 ,^ I .-- 5: ^ I < ,^_ ^ ^- j 1 . ^ f *. ^r i • !q THE BACTERIOPHAGE IN DISEASE 197 r i I : 1 i 1 i 1 I | ! i I: i i jj i i "£ ^ 1 a I z £, ^x 1 f . ^^ 2 1 p > < j ••* I j u 1 ( J J 5 : j) 3 ^ ^ i / 1 ^^ x ~ i 1 S; s 1 / i ! P •-••'•* '( ~ 1 / i i ,,..- a = 1 / i I <~ ? ~l ^ J i 1 ^^ L. ~s -si ^"^ i 1 •v ""~^. L 1 ^1 ^^ > "^>t j j^ i O •» . y i *3 1 v i ^.,. ! £ JT i L— " 5 ^.^^ / 2;i ^V"" -^. i 1 . r su -71 I .3 f i 1 \ *5 2T / •S ^ 1 tr-) ) -1. ^j i 1 o! x^" i I n* 198 THE BACTERIOPHAGE 4. Typhoid fever of extreme severity 1. Andre"e Dess. . . . (thirty years, fig. 13). 2. Jeanne Cot. . . . (twenty-four years, fig. 14). In these two cases strains of B. typhosus were isolated at different times during the course of the disease. These bacilli presented a marked resistance to the action of a very active strain of anti- typhoid bacteriophage and lost this resistance only after about ten transfers on agar. It is to be noted that at their isolation from the organism these bacilli were inagglutinable (this fact has frequently been observed) and that they did not become agglutinable until after a series of cultures. This transitory inagglutinability is, as we have seen, associated with resistance to the action of the bacteriophage. Examination of the curves shows clearly the struggle which was carried on within the organism between the bacterium and the bacteriophage, and the repercussions of this campaign upon the state of the patient. We find then, in typhoid fever, — an intestinal infection compli- cated by a septicemia — the same facts as seen in bacillary dysentery. The virulence of the bacteriophagous ultramicrobe isolated from the stools of the typhoid patient is not limited, in general, to a single pathogenic bacillus; at one and the same time it ex- tends, in some degree, to some or all of the bacilli of the colon- typhoid-dysentery group. This fact is particularly noted in mild cases or those of average severity. In the severe cases the bactericidal action is more specific and is often limited to the specific pathogenic organism and to B. coli, the latter always being attacked. In certain very severe cases the specificity becomes such that up to the beginning of actual improvement only the bacillus isolated from the patient is attacked, whether it has been secured by stool or by blood culture, to the exclusion of other bacilli, taken either from old laboratory cultures or from strains recently isolated from other patients. It seems, then, that in the course of their struggle each of the two organisms present, — bacteriophage and bacterium — acquires an individual personality, which differentiates them from other organisms of THE BACTERIOPHAGE IN DISEASE 199 ft 200 THE BACTERIOPHAGE THE BACTERIOPHAGE IN DISEASE 201 the same species rendered banal as a result of cultivation. How- ever, this individuality is effaced by cultivation, both in the case of the bacteriophage, which, after a few passages at the expense of the bacillus to which it is sensitive may develop activity toward any strain of B. typhosus, and of the typhoid bacillus which is then able to be attacked by any strain of antityphoid bacteriophage. Typhoid fever is not a purely intestinal infection as is dysentery. In the latter it can be understood how, when all of the pathogenic bacteria of the intestine or of the mucosa, that is, those in proximity to the ultramicrobes, have been destroyed the disease ends ipso facto. In typhoid fever there is in addition a septicemia and even though the destruction of the bacilli contained in the intestinal contents is sufficient to delay the appearance of the disease or to restrain it from the beginning, it may not be adequate to overcome the infection once the pathogenic bacilli have invaded the organism. We will see later that the protective action of the bacteriophage is not limited to the intestine. The intervention of the bac- teriophage, even in the same organism, may manifest itself in different ways. In the chapter dealing with the properties of the bacteriophage we have seen that the products which it secretes are possessed of an extremely high opsonizing power. A culture of an antity- phoid bacteriophage is precipitated by the addition of four vol- umes of 96 per cent alcohol. The precipitate is allowed to remain in contact with the alcohol for forty-eight hours, a time adequate to ensure the complete destruction of all of the bacteriophagous germs. One centigram of this moist precipitate is dissolved in ten cubic centimeters of saline. In determining the opsonic index, it will be seen that under the action of the lysin the leucocytes become so loaded with typhoid bacilli that it is impossible to count the num- ber of organisms ingested. The opsonic index is certainly higher than fifty. It is possible that the lysin secreted in the intestine as soon as the bacteriophage has acquired a virulence to dissolve the typhoid bacilli may be resorbed and pass into the circulation, and thus assure the destruction of the bacilli by phagocytosis. On the other hand, the bacteriophagous ultramicrobe itself does not remain strictly localized in the intestine; at times it 202 THE BACTERIOPHAGE passes into the circulation. This has not been demonstrated in man for it is not practicable to carry out on man the repeated blood examinations which such a study requires.4 However, in paratyphoid fever in the rat induced by the ingestion of a very virulent strain of B. typhi murium a transitory appearance of the ultramicrobe in the blood has been demonstrated by cardiac puncture made between the fourth and sixth days after the inges- tion of the infectious material. All of the rats in which this phenomenon occurred were protected. At this time a bacterio- phage active for the pathogenic bacillus was present in the intestine and the rats resisted infection. In the third place we will see experimentally that the dissolved products found in the cultures of the bacteriophage provoke, after an incubation period, the development of an " organic im- munity" so potent that it borders on a refractory state. These dissolved products likewise form in the intestine of the patient, and even within the body, since it is possible for the bacteriophage to pass into the circulation in a septicemia. Twenty-eight more non-fatal cases of typhoid fever were studied in order to determine the influence of the bacteriophage on the course of the disease. Three fatal cases were observed. In these three cases, at no period of the disease could the presence of a bacteriophage be demonstrated active for B. typhosus, either for a stock strain or for the bacillus from the patient. Furthermore, examination of the strains from the intestinal contents from five individuals who had died of typhoid failed to show any activity for the typhoid bacillus. But the bacteriophage was not entirely absent, since in six of these eight cases a bacteriophage of moder- ate activity for the colon bacillus was found. This bacteriophage did not, however, show any activity for the pathogenic organisms. Death in typhoid fever results, usually, because of a failure of the bacteriophage to adapt itself for the bacteriophagy of the in- vading bacillus. May death occur because of the acquisition of a resistant con- dition by the typhoid bacillus, which protects it from the action 4 Beckerich and Hauduroy have very recently found it in blood cultures. It is essential to examine them systematically for the bacteriophage; the negative cultures in particular. THE BACTERIOPHAGE IN DISEASE 203 of the bacteriophage, as we have seen in the case of dysentery? There has been no opportunity to establish this up to the present but it is the more probable, since, in vitro as in vivo, the tendency toward resistance is certainly more marked for the typhoid bacillus than for B. dysenteriae. In any case, this cause of death is cer- tainly the exception, even in typhoid. It must necessarily accompany a septicemia when it occurs. In typhoid, as in dysentery, the investigation of the virulence of the bacteriophage is of prognostic significance. It is sufficient to verify simultaneously the virulence of the intestinal bacterio- phage of the patient toward B. coli, toward the pathogenic bacillus taken from the patient, and toward a stock culture of B. typhosus. A comparison of these three results furnishes the information desired. The detection of resistance in the pathogenic bacterium would indicate a poor prognosis, and that in proportion as the resistance is the more pronounced. The establishment of a re- fractory state in the bacterium, resulting in the formation of a mixed culture in the intestine accompanying a septicemia, im- plies a fatal outcome with a quick maturity. To summarize: in all of the cases of typhoid fever studied, whatever may have been their severity, the appearance in the bacteriophagous ultramicrobe of virulence for the pathogenic bacillus has been preceded by an increase in virulence for B. coli, which has always begun in the course of the second week and has rapidly attained great intensity. This activity is maintained during the entire course of the infection and appreciably decreases only during convalescence, sometimes even later. On the con- trary, the development of virulence for the pathogenic bacillus has varied according to the severity of the disease. In cases that were mild or of average severity the activity of the bacteriophage for this bacillus appears before the end of the second week and disappears toward the end of convalescence. The activity for B. coli and for B. typhosus is there parallel. In the severe cases the activity for the typhoid bacillus only commences to manifest itself in an energetic manner towards the beginning of definite improvement. It persists for a greater or less length of time, in some cases up to the middle of the period of convalescence. In the forms with relapse and recrudescence the struggle is complicated by the fact of the acquisition of a resistance by the 204 THE BACTERIOPHAGE bacteria, and it is only toward the decline of this relapse or of the recrudescence that the virulence of the bacteriophage is sufficient to definitely control the resistance of the bacterium. Here, the activity of the bacteriophage is maintained up to complete re- covery, that is to say, up to the moment when, because of a total destruction of the pathogenic bacteria, the ultramicrobe is no longer able to develop at their expense. In all cases, the condition of the patient faithfully registers the vicissitudes of the struggle taking place within the body between the bacteriophage and the invading bacterium. AVIAN TYPHOSIS The disease Avian typhosis is a disease affecting principally the Gallinaceae. Despite its frequency it for a long time remained undetected, confounded with chicken cholera. This last disease is, in reality, very rare. In 1919, in investigating epizootics for the purpose of testing on domestic animals, which allow of experimentation, the conclusions reached as to the role of the bacteriophage in human dysentery and typhoid, an extended focus of "chicken cholera" was found in the Department of the Aube. In the first examinations the error which had been made became apparent; it was the disease known in the United States as "fowl typhoid/ ' whose existence in France had up to that time been unrecognized. Shortly after this numerous foci throughout the surrounding territory were discovered. Fowl typhoid, which will here be called fowl typhosis, is a very interesting disease. Its study is complicated by the existence of several " paratyphoses" which resemble still more the human typhoid. The pathogenic agent, B. gallinarum Klein, studied by Moore under the name of B. sanguinarium, presents, with the exception of motility, all of the characteristics of the bacillus of Eberth (B. typhosus). It is even agglutinated to titre by an anti- typhoid serum. Aside from this type bacillus there are often found, in the same foci, bacilli presenting different agglutinative and biochemical reactions. The clinical type of the infection which they provoke does not differ from that caused by the THE BACTERIOPHAGE IN DISEASE 205 typhoid type. These differing species of bacteria have up to the present been studied only by American workers; Ph. Hadley among others, who describes B. pullorum A, B. pullorum B, B. jeffersonii, B. rettgerei, and B. pfaffi. A discussion of the distinctive char- acters of these different bacilli will not be presented here since it would not be germane to the study with which we are concerned.5 It is sufficient to know that in France in the epizootic of 1919 the most frequent pathogenic agent was of the B. gallinarum type (found in 57 of 73 examinations). Along with B. gallinarum other forms have been found: — B. pullorum A (once), B. pullorum B (6 times), B. jeffersonii (4 times), and B. pfaffi (4 times). In a single focus, of which the centre was found in the village of Trainel (Aube), a paratyphosis infection occurred due solely to B. pfaffi without admixture with bacilli of the true typhosis type. The clinical picture hardly varies whatever may be the causa- tive bacillus. A typical observation follows. On the evening of May 24 the chicken appeared perfectly well. On the morning of May 25 it remained apathetically on the ground of the poultry-yard and took no measures for its defense. The next day, toward noon, it appeared somnolent, the plumage rough, the eyes half-closed, the crest slightly violet colored. It did not eat or drink, and remained humped up "in a ball." The inspirations were deep, twenty-five per minute. There was a greenish yellow diarrhea with portions definitely yellow. The condition became worse in the afternoon. It fell on its side at about 8 o'clock and died a few minutes later. The necropsy showed the crest to be violet in color, with spots of the same nature over the skin. The liver was voluminous, congested, and presented foci of degeneration. There was a pericarditis. By direct microscopic examination the blood at first appeared negative, but a very careful search revealed three bacilli in a whole smear. The blood and tissues when cultured gave a pure growth of B. gallinarum, and this organism was also found, very abun- dantly, in the intestinal contents. 5 Readers who are interested in the subject will find much useful informa- tion in the contribution by Ph. Hadley "The Colon-Typhoid Intermediates as Causative Agents oj Diseases in Birds." Bulletin No. 174, Rhode Island Agric. Exper. Sta., 1918. 206 THE BACTERIOPHAGE Sometimes death occurs more rapidly still, in certain cases in a striking manner. Epizootics of avian typhosis have a high mortality. In 1919 foci existed throughout the extent of France. In general, the epizootic begins quickly; within the space of three or four weeks a half, three-quarters, sometimes more, of the fowls on a farm succumb. Then the disease assumes a sporadic character, only an occasional animal dying during the course of a year. The disease may disappear for a few months and then reap- pear. The annual mortality amounts to forty to seventy per cent of the population of the infected poultry-yards. Young adults are the most susceptible, then the old animals; the chicks are in general spared. Epizootics of typhosis extend rapidly over large areas; cer- tain Departments were contaminated throughout in 1919. The establishment of a new focus begins by the importation of the organism from an infected region, either through the agency of a flock of sheep or herd of cattle, or by horsemen (this last mode of dissemination was particularly frequent during the war; this explains the extension of the disease during the years 1917 and 1918). The disease rages for a few days on a farm, passes to a neighboring farm, and then extends rapidly into the surrounding villages. The pathogenic bacillus remains alive and virulent during several months in the regions where the infection has been epi- demic. In several tests it has been shown that an isolated in- fected chicken-yard, cleaned and left unoccupied for six to eight months, still contains virulent germs, for, when repopulated with chickens from a region free of the disease, the infection breaks out again within a few days among the new occupants. Avian typhosis being a disease in general but little known, I have thought it useful to consider it in some detail, since it will allow us the better to understand the facts now to be presented. The role of the bacteriophage in the course of the disease Because of the exceptional severity of the infection in avian typhosis it has been possible to follow only four cases which re- covered. In all, the picture has been identical. In the morn- ing the infected chicken remains on the ground, " balled up," THE BACTERIOPHAGE IN DISEASE 207 the feathers roughened, and with the characteristic diarrhea. The appearance is the same as in the fowls which succumb. At this stage of the infection examination of the feces gives results such as: B. gallinarum, present in abundance. Intestinal bacteriophage, virulent for B. coli -f (in 2 cases) or ++ (in 2 cases); for B. gallinarum 0 (in the four cases). The blood culture was positive in the two cases in which it was done; the blood for culture being taken aseptically by puncture of the crest. During the course of the day the condition remains the same as that shown by animals which die. This state is prolonged and the next morning the chicken still appears the same. Examina- tion of the feces at this time shows: B. gallinarum present in three cases, absent in one. Intestinal bacteriophage virulent for B. coli + + + (4 cases); for B. gallinarum + (in 3 cases) + + + (in 1 case). Towards noon, in one case, in the course of the afternoon in the three others, blood cultures were negative. In three eases a bacteriophage active for B. gallinarum was found in the blood. The blood which was ultrasterile was that of the chicken whose condition was the best at this time and which had shown no pathogenic bacilli in the intestinal tract in the morning. The presence of the bac- teriophage in the blood is extremely transitory. On the morning of the third day the animals appeared normal, they drank a great deal, ate some grain, and the diarrhea was less profuse. Examination of the feces showed: B. gallinarum absent in the four cases. Intestinal bacteriophage active for B. coli + + + (4 cases), for B. gallinarum + + -f- (3 cases) + + + + (1 case). Blood cultures were negative: no bacilli, no ultramicrobes. On the fourth day the animals were practically normal. In the four chickens which recovered the bacteriophage re- mained active for B. gallinarum for a very long time. After three months it showed the same degree of activity as at the time of recovery. In one of them, in which it has been possible to make an examination after five months, it was still as active as at first. We will see, from experimental observations that this 208 THE BACTEKIOPHAGE persistence of virulence depends solely upon the fact that the pathogenic bacillus, distributed in profusion in the exterior en- vironment, is frequently ingested by the animal and this maintains the virulence of the intestinal bacteriophage since it is able to grow at its expense. The feces of about one hundred chickens which had died of avian typhosis were examined. In no case was there a bacterio- phage active for B. gallinarum or for any of the bacillary agents of the paratyphoses. Nevertheless the bacteriophage had been present for it could be disclosed (91 times in 97 examinations) because of the activity shown for one or several species of the colon-typhoid-dysentery group. One sees clearly, then, that the lack of defense is not due to the absence of the bacteriophage, but solely to the fact that the intestinal bacteriophage remained passive because it failed to acquire a virulence for the pathogenic bacillus. To summarize: as in dysentery and in typhoid fever in human beings, the acquisition of virulence by the intestinal bacteriophage for the pathogenic bacterium is the sine qua non of recovery. Rdle of the bacteriophage in the course of the epizootic Because of the dissemination and extent of the disease it was possible to study the role of the bacteriophage in the course of the epizootic as well as in the course of the disease in the individual infected animal. Let us consider first a fact bearing on the territory involved in the epizootic. During the last three years eighty-one examina- tions have been made upon the feces of barn-yard animals, not only in France but also in Indo-China, in regions where avian typhosis had not occurred in epidemic form among the fowls for several years. In each of these examinations a bacteriophage active for one or several of the bacilli of the colon-typhoid-dysen- tery group was isolated, but in no instance has the bacteriophage shown any detectable activity for B. gallinarum. In contaminated regions the situation is quite different. As an example, observations made on a farm located at Pougy-sur- Aube may be cited, where the disease was followed very closely. The disease appeared in 1917 in July. Within the period of a THE BACTEBIOPHAGE IN DISEASE 209 month fifty-one of the ninety-eight fowls died; then the epizootic disappeared. In May, 1918 it reappeared in less violent form. Twenty-five of one hundred and four fowls died in the period from May to September, and it again disappeared. In 1919 it broke out again early in April. On the 21st of May, twenty-one of eighty had died. At this time I began my observations. On May 21, specimens of the excrement of thirty of the fifty- nine survivors were taken. Examination, made later in the lab- oratory, showed in twenty-six a bacteriophage of weak or moder- ate activity for B. gallinarum (23 were +, 3 were ++), in four it was absent. On May 22, two chickens contracted the disease. The strains taken the day before were numbered and examination showed that an active bacteriophage had not been found in these two animals. On May 23 one of the two chickens affected the day before died. On May 24, a third chicken, sick in the morning, died in the following night. Its excrement, collected on May 22, did not contain a bacteriophage active for B. gallinarum. On the morning of May 24 the chicken which had been taken sick on May 22 and which had resisted showed in its intestinal con- tents a bacteriophage of extreme activity (+ + ++) toward the pathogenic bacillus. On May 26 the fourth chicken, one of those whose f eces had not showed an active bacteriophage when examined on May 22, was affected. It resisted, and on May 28 its symptoms had disappeared. The disease disappeared suddenly and during the next three months no new cases developed. On May 30 the feces of thirty chickens were examined and the following results were obtained: Virulence for B. gallinarum; in five + + ++> in twenty-one + + + , in four + + . We see, then, on May 22, four animals among thirty in which the intestinal bacteriophage lacked activity for the pathogenic bacillus. These four animals contracted the disease during the four following days. In the twenty-six specimens collected on May 22 and showing positive results, the bacteriophage showed a relatively weak virulence. Nine days later this activity was very much greater, that is, at the time when the epizootic ceased. What, then, took place in this interval? The bird which became sick on May 22 and which resisted showed in its feces, when ex- 210 THE BACTERIOPHAGE amined on May 24, a bacteriophage endowed with a considerable activity for the pathogenic agent. Here is a second example of the same general nature, giving the results secured on farm M. . . . at Vericourt (Aube). The epizootic first appeared among the flock of twenty-five chickens in May, 1919. The first animal died on May 18. On the next day twelve specimens of excreta were collected at random. Three only contained a bacteriophage, and that of feeble activity, for B. gallinarum. From May 19 to 26 twelve birds contracted the disease and of these eleven died. One, which became sick on May 23, showed on May 25 a strongly active bacteriophage (H- + H- for B. gallinarum) and recovered. The epidemic stopped abruptly. On May 27 twelve specimens were taken at random. In all a bacteriophage active for B. gallinarum was found (in 1 + + + + , in 9 + + +-, in 2 ++). A third example may be mentioned, in which the infection was a paratyphosis.6 On October 15 strains of B. pfaffi were isolated from two specimens of blood, taken from animals which had died in a chicken-yard where for about a month there had been an infection presenting the characters of typhosis. From specimens of the feces taken from two healthy animals living in the same yard two strains of bacteriophage were isolated, one showing a low virulence (+) for B. pfaffi, the other showing no activity for this bacillus. Towards the end of the month three chickens became sick, recovered after an interval of two or three days, and then the epizootic ceased. Six specimens of feces examined at this time all showed a bacteriophage of high virulence (+ + +) for B. pfaffi. Against B. gallinarum four were inactive and two showed a weak virulence (-}-). B. pfaffi was therefore the cause, for when the epizootic broke out three months later the eighty chickens which had survived received a subcutaneous injection of 0.5 cc. of a culture of the anti-pfaffi bacteriophage and the epidemic stopped abruptly and permanently from the time of the injection. We will see later that this abrupt cessation is the rule following immuniza- tion by means of a culture of the bacteriophage. 6 These experiments were carried out with the assistance of M. Micheau, D. V. M. at Trainel (Aube). THE BACTERIOPHAGE IN DISEASE 211 These facts can be explained in only one way. A weak or moderate activity of the intestinal bacteriophage for the patho- genic bacterium is sufficient to render the animal resistant to infection. The pathogenic bacteria which are able to penetrate into the intestine are destroyed before they can multiply. But it is not the same once the disease has appeared and the organism is invaded. The animal recovers, and this is very rare in typhosis, only because of a rapid adaptation of the bacteriophage and the acquisition of a high virulence which leads to an intensive destruc- tion. This bacteriophage with exalted virulence is distributed broadcast with the excreta of the recovered or convalescent ani- mals, and continues, indeed, during several months after recovery. This bacteriophage is necessarily ingested by the other animals of the barn-yard which become, in fact, "infected" by an ex- tremely active bacteriophage and by this means acquire a complete protection against the disease, in spite of the presence of the pathogenic organism in the environment, and in spite of its fre- quent ingestion, an ingestion which serves to maintain the viru- lence of the bacteriophage. These hypotheses are not simply idle speculation, for the interpretation given to these observed facts is confirmed by ex- periments which provide in a controlled manner the natural conditions of the epizootic. Furthermore, it will be seen that the role of defense assigned to the bacteriophage is confirmed by the immunization of several thousand animals by the adminis- tration of cultures of an active bacteriophage. Before discussing these control experiments I ought to mention that, thanks to the kindness of the veterinarians of different regions invaded by typhosis, I have been able to procure numer- ous specimens of blood and excreta taken from sick chickens, from chickens which had died or from those which had recovered, de- rived from eleven different foci scattered throughout all France. This allows me to generalize from the facts that I have personally observed. Control experiments The control experiments have been conducted in Paris, that is to say, entirely outside of the epizootic area. 212 THE BACTERIOPHAGE Six chickens, procured from a region free of infection, were placed under observation. Their excreta were examined daily for ten days for the purpose of establishing the complete absence of a bacteriophage active for B. gallinarum. Chicken no. 1 then received, per os, 1 cc. of a culture of a strain of bacteriophage very active for B. gallinarum (+ + ++). Chicken no. 2 received 0.5 cc. of the same culture by subcutaneous injection. The next day examination of the feces of these two animals showed the presence of a bacteriophage strongly virulent for B. gallinarum. Therefore, the bacteriophage passed into the intestine, whether ingested or injected. This same fact has since been verified with man and with different animals. Chicken no. 1 next received per os daily for twenty-five days, 2 cc. of a bouillon culture of B. gallinarum. The active bac- teriophage persisted in the intestine with its primary virulence (+ + ++) and maintained itself up to nine days after the last dose of the pathogenic organism. Chicken no. 2, which had received nothing after the inocula- tion of the active bacteriophage ceased to show an active strain for B. gallinarum within three days after the injection. In other words, chicken no. 1, subjected to repeated reinfections, retained an intestinal bacteriophage active for B. gallinarum for thirty- four days, while chicken no. 2, not infected, for only three days. It follows that the intestinal bacteriophage remains active only if it is able to develop in the intestine at the expense of this bacterium, but in such a case it remains active just so long as this condition is fulfilled. Inversely, the presence in the intestine of a bacteriophage possessing virulence for a given bacterium indicates that this bacterium was a short time previously in the intestine. In the course of the preceding experiment chickens nos. 3 and 4 were placed in contact with chicken no. 1. They all ate and drank from the same containers, the more so since they were changed about in the pens in such a manner as to simulate condi- tions of life analogous to those of the chicken-yard. Two days after the first contact, in the case of chicken no. 3, three days after with chicken no. 4, their ex^r^ta contained a bacteriophage very THE BACTERIOPHAGE IN DISEASE 213 virulent for B. gallinarum (H — h ++). From this time on they each received each day for twenty-one days, 2 cc. of a bouillon culture of B. gallinarum. At no time did they appear sick. The intestinal bacteriophage remained active for the bacillus through- out the entire period of the administration of the pathogenic bacillus, and even longer — seven days in no. 4 and ten days in no. 3. The intestinal bacteriophage did not then disappear, for as in the case of chickens nos. 1 and 2, it remained active for one or several members of the colon-typhoid-dysentery group. But the virulence for B. gallinarum did not persist when the ingestion of cultures of this last bacillus was stopped. The experiment with chickens nos. 3 and 4 shows clearly that the bacteriophagous ultramicrobe is infectious in exactly the same sense as is the pathogenic bacillus itself, since these birds were ''contaminated" by contact with chicken no. 1. Chickens nos. 5 and 6, which had not been in contact with the other chickens, and which on repeated examinations were shown to be free of a bacteriophage active for B. gallinarum, each re- ceived per os, on some bread, a single dose of 2 cc. of a bouillon culture of B. gallinarum. Three days after the infecting meal diarrhea appeared and they died two and three days later, after having shown all of the symptoms of the natural disease. Ne- cropsy showed the presence of the same lesions. Cultures of the blood gave pure cultures of the pathogenic bacillus, which was likewise found in abundance in the intestinal contents. Chickens nos. 1, 3, and 4, which had resisted repeated inges- tions of B. gallinarum culture without showing the least incon- venience, were therefore immunized; the first as a result of the ingestion of a bacteriophage active for the pathogenic bacterium, the two others by simple association with the first. About one month after the virulence of the bacteriophage for B. gallinarum had disappeared in chickens nos. 1, 2, 3 and 4 each of them was given on each of three days 2 cc. of a culture of the bacillus. In all the intestinal bacteriophage showed a new virulence for the pathogenic organism. None of them showed the slightest trouble. In all of these experiments the infections have been made with bouillon cultures of B. gallinarum prepared directly from the 214 THE BACTERIOPHAGE blood of chickens dead of spontaneous natural infection. This is essential because of the loss in virulence of this organism which takes place under artificial cultivation. With chickens nos. 5 and 6 the ingestion of the pathogenic bacillus caused a fatal attack of typhosis. The intestinal bac- teriophage at no time manifested an activity for the causative organism. In chickens nos. 1, 2, 3, and 4, on the contrary, the ingestion of the same culture caused no disturbance and their intestinal bacteriophage which for about a month had showed no activity for the bacillus, rapidly recuperated its first activity. It had, therefore, not disappeared from the intestine, although its activity was no longer evident, but when it found itself again in contact in the intestine with the pathogenic organism it rapidly regained its potency. This " latent virulence" may be maintained for a very long time. In this connection I may recall the fact cited of a strain of bacteriophage still possessing after three years and more than 1000 passages in vitro, always with the Shiga bacillus, the power to attack B. coli and B. typhosus. It showed a weak power, but was capable of rapid augmentation by transfers at the expense of these organisms. This is exactly what this experiment shows us to take place in vivo in the chicken. Can a chicken contract typhosis in spite of the presence of an active bacteriophage in the intestine? It certainly can. As we have seen in many experiments the bacterium may develop a resistance to the action of the bacteriophage and this resistance is one of the factors comprising the virulence of the bacterium. We have then, on the one hand, the bacterium, which when in- troduced into the organism may acquire a resistance to the action of the bacteriophage ranging from zero to absolute resistance, and on the other hand, the bacteriophage, which at the same time may possess a virulence running from zero to extreme activity. Infection occurs, or does not occur, according to whether the algebraic sum of virulence -f resistance is in favor of the one or the other of the two organisms present. Once the disease has manifested itself, the virulence of the one and the resistance of the other become increased or attenuated according to the con- ditions of the moment and the aptitudes previously acquired THE BACTERIOPHAGE IN DISEASE 215 favoring the one or the other of the two germs. The sequence in which the events of the struggle unfold determine the issue. Conclusions The observations made in natural disease and the experiments which confirm the deductions which these observations suggest, show that the bacteriophagous ultramicrobe is always present in the intestine of the chicken, whether it is healthy or sick, whether it lives in a locality free of infection or in an epizootic zone. Against a definite bacterium, B. gallinarum in so far as avian typhosis is concerned, the intestinal bacteriophage may be viru- lent or avirulent, and in the first case its virulence may be ex- ercised according to a scale which passes from the smallest degree capable of detection to one of extreme activity. Virulence of the bacteriophagous ultramicrobe for B. galli- narum is only observed in an infected locality. The absence of such a virulence is equally the rule with animals which are about to die and with those which have died. In a contaminated area animals which harbor in their intestine a bacteriophage endowed with sufficient virulence for the patho- genic bacterium are by this very fact protected against the dis- ease, and they remain so, provided the actual or latent virulence is maintained at a level sufficiently high to effect a rapid destruc- tion of the pathogenic bacilli ingested. The ingestion of pathogenic bacilli at sufficiently frequent intervals constitutes the principal factor in maintaining the virulence for the given bacterium. Among the factors which contribute to diminishing the virulence or causing the virulence of the bacteriophage for a pathogenic bacterium to disappear, I would place as most significant the introduction into the or- ganism of bacteria endowed with resistance to the action of the bacteriophage. We have clearly seen this fact in the course of the experimental study of the . phenomenon of the resistance of bacteria. Another possible factor, influencing the activity of the bacteriophage is the reaction of the medium in the intestine, which may vary according to the accidental conditions of the moment, the type of food, etc. The importance of the reaction of the medium has already been shown for lysis in vitro. 216 THE BACTERIOPHAGE A bacteriophage which has lost its virulence for the pathogenic bacterium lacks the power to exercise it because of the absence of this bacterium, but it possesses nevertheless, a latent viru- lence. When placed again after a greater or less length of time in the presence of this bacterium it regains its original virulence. The fact of the habitual virulence of the intestinal bacterio- phage for B. gallinarum in the infected regions indicates the frequency of the ingestion of these bacilli, and consequently the excessive contamination of the environment by the pathogenic organism. In contaminated regions the animal in which the intestinal bacteriophage does not enjoy any activity for B. gallinarum, quickly contracts the disease. It may resist and recover, but this is the exception, occurring only when the intestinal bacterio- phage quickly acquires a virulence for the infecting bacillus. In the contrary, and usual, case the animal succumbs. In a chicken which recovers, the intestinal bacteriophage ac- quires a considerable virulence against the pathogenic bacterium and maintains this for a very long time; in fact, as long as the exterior environment remains infected. This persistence of viru- lence is maintained by the frequent ingestion of pathogenic or- ganisms, which allow the bacteriophage to multiply at the expense of the particular organism. The resistant animal disseminates in its excreta the bacteriophage of enhanced virulence; the ani- mals which associate with it become " contaminated" and by this fact they enter the same class of resistant animals as those which have recovered. Recovery of one animal in a barn-yard often marks the end of an epizootic, or its arrest for a few months. The study of an epidemic of avian typhosis shows, in a word, that the history of the contagion reflects, in the last analysis, the story of the struggle between the two agents — the pathogenic bacterium and the bacteriophagous ultramicrobe — and since this last is transmissible from individual to individual the immunity is contagious in the same sense as the disease itself. The be- ginning of an epizootic is marked by a diffusion of the bacteria, the end by a diffusion of a bacteriophage virulent for these bac- teria. We will encounter the same facts in another disease; in hemorrhagic septicemia in the buffalo. THE BACTERIOPHAGE IN DISEASE 217 HEMORRHAGIC SEPTICEMIA OF THE BUFFALO (BARBONE)7 Barbone, the disease Contrary to the diseases discussed up to this point, barbone does not present intestinal symptoms; it is of the hemorrhagic septicemia type. The pathogenic organism is a Pasteurella. Cultures of the organism in beef bouillon maintain their virulence for a considerable time — at least eighteen months. The inocula- tion of a buffalo or of a cow with 0.0002 cc. of a virulent culture kills the animal in between thirty-six and forty hours with all the symptoms of the spontaneously acquired disease. At ne- cropsy identical lesions are found and the pathogenic bacterium swarms in the blood and in the organs.8 The buffalo is par excellence the beast of burden in the culti- vation of rice-fields; it replaces the ox in all southern Asia and in the islands of the Sunda Straits. It is utilized in certain regions of Italy, in Egypt, in Hungary, and in the Balkans. Wher- ever the buffalo lives there also will be found barbone, the most terrible, without doubt, of all the contagious diseases. The re- ports indicate a mortality of from 70 to 95 per cent. I was present during an epizootic which raged in June, 1920, in the Province of Bac Lieu (Cochin-China) where among the thirty thousand buffaloes of the region ten thousand died, and I did not have an opportunity to observe a single animal which recovered. Recovery may occur, but it is certainly rare, and the mortality in Cochin-China is certainly above 99 per cent of the animals affected. The average duration of the evolution of the disease is but eighteen to twenty-four hours; rarely thirty-six. Death some- times takes place without precursory symptoms. An animal yoked to a plow stops, remains motionless for a few moments with an 7 The experiments on barbone have been performed in collaboration with G. Le Louet, Chief of the Veterinary Service in Cochin-China. 8 In two different attempts I have proved that diluted blood or macera- tions of organs (liver and lung) taken from animals dead of spontaneous infection, filtered through a Chamberland filter (L2) and inoculated in large amounts into the buffalo or into cattle do not cause the slightest dis- ease symptoms. 218 THE BACTERIOPHAGE haggard aspect and then falls as though struck by lightning. In typical cases, which can be reproduced in a perfect manner in experimental infection, the animal appears dejected, the eyes fixed, the head lowered. The temperature rapidly mounts to 41.5 to 42.5°C., the respiration, at first accelerated, becomes slowed and then dyspneic, the inspirations less and less frequent. The animal shows meteorism; it lies flat on the ground in complete lateral decubitus usually a short time before death which is pre- ceded by cramps and at times convulsions. Often tumefaction is to be observed, appearing usually in the region of the throat and extending back to the shoulder. The engorgement is produced by a gelatinous exudate of a yellow color within the connective tissue. At times the tumefaction appears in another part of the body, or it may be entirely lacking. This tumefaction, as shown in experimental infection, marks the portal of entrance of the pathogenic bacteria. Infection usually occurs by way of the digestive tract and the virus most frequently penetrates the tissues through some portion of the nasopharynx. A tumefaction on another part of the body — thigh, abdomen, rump — indicates a reinfection by the penetration of the virus through an excoriation. Examination of cadavers shows that the absence of tumefaction indicates an infection by way of the stomach and intestine. Bovines and the buffalo are equally susceptible, as was noted a long time ago by Piot in Egypt. The statistics of Indo-China indicate, it is true, that the mortality from barbone is but slight for cattle, but this is solely due to the fact that these animals are present in but small numbers in the regions where barbone rages; regions which are extremely humid and admirably adapted to the buffalo, a semi-aquatic animal. The rare cattle found some- times in such regions contract the disease and die like the buffalo, after having presented identical symptoms. The effect of low places and swamps on the contagion has been from time immemorial recognized by the natives. When it is possible, as soon as a case of barbone is detected in a neighborhood, they hasten to collect their animals and remove them to a more elevated region. It is known, moreover, that the organisms of the Pasteurella group remain virulent for a very long time in the mud of the marshes and in the slime of the streams. THE BACTERIOPHAGE IN DISEASE 219 Rdle of the bacteriophage in the disease In Cochin-China barbone is always present in sporadic form causing each year numerous small epizootics which remain lo- calized in individual villages. A localized epidemic observed in Long Huu in the Province of Gocong may serve as an example. From May 5 to 13, 1920, seventeen buffaloes died: — on May 5, one; May 7, three; May 8, two; May 9, one; May 10, two; May 11, four; May 12, three; and May 13, one. Then the epizootic stopped and not a single case was detected during the next six months. Specimens of the feces of four of these animals were collected, either before death or from the cadaver. None contained a bac- teriophage active for the bacterium of barbone. On May 13 specimens of feces were collected from healthy animals, as follows : First. From a buffalo in a stable where two animals had died, one on May 12, the other on May 13. Second. From three buffaloes in a stable where one had died on May 5. Third. From two buffaloes in a stable where two had died, one on May 8, the other on May 11. Fourth. From four buffaloes in a stable which had not been invaded. Fifth. From one buffalo, living alone in a stable located at a distance of about five kilometers from the village of Long Huu. Sixth. From eight buffaloes in the surrounding villages, from eleven to nineteen kilometers distant. Of all the specimens, those in the first, second, third and fourth groups gave a bacteriophage of weak or average activity (+ or ++) for the bacterium of barbone. An active bacteriophage was not found in the specimens from groups 5 and 6. Again on May 19 specimens were collected in Long Huu, as follows: First. From the buffalo which had furnished specimen no. 1 on May 13. Second. From two buffaloes living in a stable where three had died from May 7 to May 12. These specimens all gave a bacteriophage moderately virulent (++) for the bacterium of barbone. 220 THE BACTERIOPHAGE The animals which resisted, therefore, showed in their intestine a bacteriophage virulent for the pathogenic bacterium. The epizootic does not always remain localized in a village. At times it spreads rapidly from village to village and within a few days will extend over a very considerable territory. It is rarely possible to determine the primary focus, so great is the speed with which it spreads. The mortality then becomes con- siderable, the losses often amount to tens of thousands of animals, as has been observed many times in China, in British India, and in the Dutch East Indies. Sometimes even, as actually happened in Java, the buffalo, as a race, is practically eliminated. In the first two weeks of June 1920 the epizootic became general in the Province of Bac Lieu and in certain parts of the adjacent provinces (western Cochin-China). It was possible to examine the blood of eleven animals which died in widely scattered parts of the area invaded, and in all the bacterium of barbone was found in considerable quantity.9 The epizootic died out during the first fortnight of July. It had persisted for a month, killing a third of the animals in the district. The region of Thoi Binh was particularly affected, the loss amounting to more than fifty per cent of the buffaloes in the locality. From July 8 to 13, at the time when the epizootic was disappearing (the last animal to be affected died on July 12), twenty specimens of feces were collected. These were taken from buffaloes which had resisted the infection and which at no time showed any evidence of the typical symptoms of the disease. All of the animals examined lived on the farms of the village of Thoi Binh or in the neighboring hamlets within a radius of fifteen kilometers. Tests for the virulence of the intestinal bacteriophage against the bacterium of barbone gave the following results: 9 Bacteriological diagnosis is easy, even if the only available material is some blood or a fragment of an organ taken without any special pre- cautions in the field, as is usual in such countries. Even if the specimen is some days old it is only necessary to smear it over the shaved skin of a rabbit. If the bacterium of barbone is present the animal will die within 24 hours, and the organism will be found in pure culture in the blood, from which it may be readily isolated. This is also the best method for detecting the bacterium in soil or in fecal material. THE BACTERIOPHAGE IN DISEASE 221 FARM MORTALITY THE LAST ANIMAL DIED ON NUMBER OF A.NIMAL8 WHICH RESISTED VIRULENCE OF THE BACTERIO- PHAGE Ncau 1 3 July 7 10 + + Ngau 2 + Ngau 3 +4--h De 2 ' June 10 1 -}- Doi 1 5 June 28 4 ++++ Doi 2 ++ + La/nli 9 July 2 4 -f- \- Tran 1 July 4 2 ++++ The 3 July 11 1 +++-f H v Chanh 2 July 2 2 H-+++ Hien 0 4 ++4 + Du 0 6 + Sam 2 July 2 8 ++ P v Chanh 6 June 30 8 +-f + + Cu 1 July 2 5 ++-f So 3 July 6 5 ++ No 5 June 30 10 ++ + Phuc 8 July 3 4 + ++4- Gia 8 June 29 3 +++ Man 1 June 30 3 ++ From this it appears that the intestinal bacteriophage is en- dowed with virulence for the bacterium of barbone in all the buffaloes which the disease had spared. In the course of different trips across Indo-China, I collected forty-one specimens of feces from buffaloes, each specimen col- lected in a different village in which no buffaloes had died of barbone for at least two years. In only three of these specimens could a bacteriophage active for the bacterium of barbone be demonstrated, and in these cases it was weak (+)• Nevertheless, the intestinal bacteriophage was present in all; but although it was active for one or another of the intestinal organisms, its viru- lence was weak or lacking for the bacterium of barbone. We will see later, on the contrary, that in a contaminated area at the time when the epizootic dies out, the intestinal bacterio- phage of all of the buffaloes which escaped the disease is virulent for the bacterium, the causative agent of the epizootic. We find here, then, the same facts as in the previous disease studied; 222 THE BACTERIOPHAGE that the protection of the body in the case of barbone, a septicemic disease, is assured by the bacteriophage. In the buffaloes of a region ravaged by the disease the bacterio- phage preserves for a very long time its virulence for the patho- genic bacterium. This, the following example shows. In November, 1919, a localized epizootic of barbone occurred among the buffaloes of the village of Phuoc Thien (Province of Bien Hoa). On a farm having twenty-one buffaloes seven died — two adult animals and five aged from one to two years. The disease died out, or to speak more correctly, after this, two ani- mals recovered one after another. On the 12th of the following April, that is to say, five months later, specimens of the feces of eight of the surviving animals were collected. All contained a bacteriophage active for the bacterium of barbone (six were + + > two were +)• Two specimens of the mud of a water-hole where the animals were accustomed to remain immersed up to the neck during the hottest hours of the day were also examined. In both a bacterio- phage virulent (++) for the bacterium of barbone was found. The destruction of the pathogenic bacterium in the external medium must often be effected by the bacteriophage, for it is certain that if the bacterium of barbone has once been introduced into a water-hole by a sick animal the bacteriophage present there must destroy it. Furthermore, this fact shows one of the modes of " contagion" of the active bacteriophage. A single buffalo, in the intestine of which the bacteriophage has acquired a virulence for the pathogenic bacterium, is sufficient to "con- taminate" all the herd which frequent the water-hole. Localized epizootics are of short duration, but in spite of this we find that the pathogenic bacterium persists for several months in the ex- ternal world and that their ingestion by buffaloes is frequent, since the virulence of the bacteriophage maintains itself against this bacterium. The repeated ingestion of a bacterium is, as we have seen, essential for the permanence of the virulence of the bacteriophage toward this bacterium. The epizootic dies out, not because of an absence of pathogenic bacteria but because of the presence of a virulent bacteriophage in the intestine of all exposed animals. THE BACTERIOPHAGE IN DISEASE 223 All of the observations are therefore comparable, whether they deal with avian typhosis or with barbone in the buffalo. These epizootics of very different nature were investigated in- tentionally, that the general nature of the role of the bacterio- phage in immunity might be the better established. One may at first be quite astonished that the intestinal bac- teriophage, whose role can easily be conceived in infections with intestinal manifestations, constitutes a defense of the organism in septicemias. In reality, whatever may be the infection, the pathogenic bacterium always gets into the intestine. Let us take a localized disease, cerebrospinal meningitis, for example. We know that the initial symptom is a rhino-pharyngitis and that even healthy subjects who have been in contact with a patient often carry the specific germ in the nasopharynx. There can be no doubt but that a fair number of the meningococci present in the rhino-pharynx are swallowed and pass into the intestine. It is needless to insist on this, that, aside from a few rare exceptions to which we will later return, whatever may be the disease under consideration, the portal of entrance of the virus is either the buccal route or by way of the respiratory tract. In either case the ingestion of organisms is, it might be said, ob- ligatory. The pathogenic bacterium is always at some time in contact with the intestinal bacteriophage, this organism there- fore is thus able to adapt itself to the bacteriophagy of the bac- terium and to acquire a virulence. In the particular case of barbone the pathogenic bacterium is found freely disseminated through the exterior world in con- taminated regions. In an epizootic zone I have been able, in two different trials, to isolate it from the mud of a marsh where the buffaloes were accustomed to bury themselves. This is but natural since the bacterium of barbone is found in the intestinal tract of sick animals or of those which have succumbed. The ingestion of the pathogenic bacterium by the animals which re- main immersed for whole hours in a mire containing these or- ganisms is necessarily frequent. If the animal which ingests them has an erosion at any point in the digestive tract it is sus- ceptible to infection. Otherwise the bacteria reach the intestine and come within the range of the intestinal bacteriophage which 224 THE BACTERIOPHAGE can then acquire a virulence for the virus. If this takes place the animal is thenceforth protected from the infection and becomes a carrier of the virulent bacteriophage. A diseased animal prop- agates his disease; an animal in a resistant condition propagates his immunity. BUBONIC PLAGUE Through a lack of favorable circumstances it has not been possible to follow the evolution of the intestinal bacteriophage in man affected with plague. The few cases that have been ex- amined have all been fatal, and at no time could the intestinal bacteriophage be shown to have the least virulence for B. pestis. The activity in these cases remained restricted to B. coli. How- ever, the stools of two convalescent individuals have been se- cured and examined. According to the physicians treating the cases the material was collected on the sixth and the eleventh days after the beginning of convalescence. Examination showed, in the first case, a bacteriophage of average virulence (++), and in the second case, one of feeble virulence (+) for B. pestis. The virulence of the first of these strains has been enhanced in vitro and the bacteriophage has been maintaioed in culture. An attempt was made to find a bacteriophage active against this bacillus in the feces of twenty-two natives living in regions free of plague, but in no case could a strain be isolated. However, in view of the particular mode of infection in bubonic plague the study of its propagation in man offers only a matter of secondary interest, at least from the epidemiological point of view. We know that an epidemic of plague in man is only consequent to an epizootic among rats. That which it is interesting to study is, therefore, the epizootic, the primary cause of the epidemic. In order to attain a correct interpretation of results it is essential to follow the natural order of things. From the point of view of man the epidemic is obviously the important fact; from the point of view of nature this is but a secondary incident, for if we were able to suppress the epizootic the epidemic would cease spontaneously. From what we actually know about the epidemiology of plague, it results that all of the rats living in a city where there has been THE BACTEBIOPHAGE IN DISEASE 225 a case of plague in man are the animals which have resisted the contagion, either because they were infected and recovered, or because they remained unaffected. I have then, investigated the virulence of the intestinal bacteriophage of the rat toward B. First. Twenty-one specimens of the excrement of rats taken from towns in Indo-China free of plague were examined. The intestinal bacteriophage was found, active against one or another of the intestinal bacteria, but it never showed any virulence what- ever for B. pestis. Second. A small epidemic of plague (eleven fatal cases) occurred in the village of Bac Lieu, in the eastern part of Indo-China, during July, 1920. On the following 6th of November I procured in this town four specimens of the excreta of rats, each speci- men composed of some dozens of particles, and certainly derived from several individuals. The tests for virulence against B. pestis gave the following results: Specimen derived from a granary ...................... Specimen derived from the embarkment quay .......... ++ + Specimen derived from a decorticating mill ............ +++ Specimen procured in the house of a native ............ ++++ Those rats which have survived an epizootic, therefore, har- bor in their intestine a bacteriophage possessed of a high viru- lence for B. pestis. Plague has existed in the form of sporadic cases in the region of Phantiet, in southern Annam, for about twenty years. I obtained specimens of the excrement of rats in the infected vil- lages, each specimen being composed of the feces derived from several animals. The results of the tests for the virulence of the intestinal bacteriophage in these specimens were: Village Virulence Thien Due .............................................. ++ Hung Long ............................................. + Due Hang .............................................. + + + Due Thang ............................................. ++ Tri Long ................................................ + 4- PhuTay ................................................ + + + Cu Long ................................................ + 226 THE BACTERIOPHAGE The results are thus identical with those secured at Bac Lieu, although the virulence seems to be somewhat lower, but this can only be of relative importance since in the present case each specimen was composed of the excreta taken from several rats; the results then, indicate only an average. Is the bacteriophage present in all of the rats of an infected region or only in a certain number? At Phantiet I collected the excrement of six young rats, according to their weight, aged from three to four weeks. Examination showed that four of the speci- mens contained a bacteriophage active for B. pestis (+) while two did not. These last two animals were therefore susceptible to plague. From the results given above one may conclude that, as for avian typhosis and for barbone, the cause of the resistance against B. pestis is the presence in the organism of a bacteriophage pos- sessing a virulence for this bacillus. How is the adaptation effected in the case of B. pestis? At different times it has been noted that the bacillus has been found in the intestinal contents of victims of plague. Thus, it is pos- sible for them to be disseminated by the feces throughout the external world where they may again be ingested. The bodies of dead rats constitute another mode of dissemination. These bodies are often devoured by the surviving rats and this extends the infection. In those animals which resist and which are infected the intestinal bacteriophage is maintained virulent for the pathogenic bacillus. But observation and direct experi- mentation have shown us that a bacteriophage is only possessed of a virulence for a bacterium when the ingestions of this bac- terium are frequent. The permanence of the virulence of the intestinal bacteriophage of the rat against the plague bacillus indicates the persistence of this bacillus in the external world, at least for several months after the last human case has taken place. Moreover, the -revival of the epidemic each year in cer- tain localities, Bac Lieu for example, shows that it can not be otherwise.10 10 Demonstration of the presence of a bacteriophage active for B. pestis in the rats of a locality would in certain cases be very useful for it would indicate the presence of the bacillus in the exterior world and the possi- THE BACTERIOPHAGE IN DISEASE 227 FLACHERIE A few experiments have been made on this disease, but only for the purpose of determining if defense against infection in invertebrates is also assured by the bacteriophage. In a breeding-place in Cochin-China a certain number of silk worms died of a disease presenting all of the characteristics of flacherie. Examination of the excreta of the sick worms, as well as of the cadavers, showed the presence of a cocco-bacillus, Gram-negative, which was not present in the dejections of healthy worms. The ingestion, on mulberry leaves, of some of the culture of this cocco-bacillus reproduced the disease; eleven out of twelve worms dying in from six to eleven days after the infecting feeding. Three nitrates were prepared from the excreta of healthy worms living in the baskets where the affected worms were found. These three filtrates contained a bacteriophage of moderate or high virulence (++, + ++, + + +) for the cocco-bacillus. On the other hand, two filtrates were prepared, the one with the intestinal contents of a sick worm, the other with the intestinal contents of a worm which had died of the infection. Neither contained a bacteriophage active for the coccobacillus. These experiments have not been carried further, since the desired end had been attained. They were adequate to show that the facts observed in infectious disease in mammals were reproduced in an infectious disease of an invertebrate. From this it seems logical to conclude that the defense of the organism by the bacteriophage must constitute a general fact throughout all animals. bility of a renewal of the epidemic. Such a demonstration might also be useful in establishing a retrospective or doubtful diagnosis. Suppose a few suspicious deaths have occurred in a group some time previously. The pres- ence in the rats of the neighborhood of a bacteriophage showing a virulence for B. pestis would eliminate all doubt; the deaths were due to plague. Or, the question of the nature of a epizootic among the rats may be in question. Was the mortality due to plague? The demonstration of a bacteriophage active for B. pestis either in the dead rats or in those that have survived provides the answer. 228 THE BACTERIOPHAGE CONCLUSIONS We may limit ourselves for the moment to the following points: Whatever may be the disease considered, the picture remains the same; when a pathogenic bacterium is introduced into an organism, one of two situations develops: First. The intestinal bacteriophage shows an activity for this bacterium, the latter is destroyed before it can develop, and disease does not appear. Second. The intestinal bacteriophage is inactive, the bac- terium develops, and disease results. In the course of the disease one of two things may happen: First, the bacteriophage in contact with the pathogenic bac- terium may acquire a virulence, or Second. The bacterium may acquire a virulence, in other words, may become resistant to the action of the bacteriophage. The vicissitudes in the struggle between these two factors are reflected in the condition of the infected individual. Convales- cence begins at the moment when the virulence of the bacterio- phage is sufficient to give it, definitely, the upper hand. The disease has a fatal outcome if the bacteriophage is inactive as a result of unfavorable conditions, or if the bacterium is able to acquire a refractory state. This last situation appears to be very infrequent, at least, in the diseases studied. In epidemics we find a large scale reproduction in a community of individuals of the struggle which takes place in a single indi- vidual between the ultramicrobe and the bacterium. The bacteriophagous ultramicrobe is transmissible from one individual to another just as is the bacterium itself. The his- tory of an epidemic is, in the last analysis, the story of an infec- tion with two microorganisms. The epidemic ceases at the moment when all susceptible individuals harbor a bacteriophage active for the causative organism of the epidemic. Either the bacteriophage has acquired virulence in the body of the indi- vidual who harbors it, or this individual has been " contaminated " by a bacteriophage which has acquired a virulence in another individual for the specific bacterium involved. CHAPTER II THE BACTEBIOPHAGE IN THE HEALTHY INDIVIDUAL The Bacteriophage in Man. The Bacteriophage in the Horse. The Bacteriophage in Fowls. The Bacteriophage in Diverse Animals. Conclusions. The experiments conducted on patients and on healthy in- dividuals exposed to infection have shown that a resistance to the infectious agent accompanies the presence in the intestine of an ultramicrobial bacteriophage possessing a virulence for the causative bacterium. On the other hand, as I have shown in Part I of this monograph, there is but a single species of bacterio- phage, capable, by adaptation, of acquiring virulence for the diverse bacteria which it attacks.1 These facts being true, a question quite naturally arises. Does this bacteriophage which acquires virulence for diverse patho- genic bacteria make its appearance only at the exact moment when it is needed? Or is it, indeed, a normal inhabitant of the intestinal canal? Examination of the feces of numerous individ- uals belonging to very varied species permits an answer to this question. THE BACTERIOPHAGE IN HEALTHY MEN Variations in the virulence of the intestinal bacteriophage in healthy men have been followed. To this end, in a first series of experiments, specimens of the stool were collected every fifteen days. The study has been limited to testing the activity of the ultramicrobe against the following bacterial species: B. coli, B. dysenteriae Shiga, B. dysenteriae Flexner, B. dysenteriae Hiss, B. typhosus, B. paratyphosus A, and B. paratyphosus B. Later 1 The possibility of a germ being virulent for a great number of different beings is not an exception peculiar to the bacteriophage. It is only neces- sary to mention B. tuberculosis, to which but few animals are insusceptible. 229 230 THE BACTERIOPHAGE where indicated, the tests were extended to other bacterial species of particular interest. With the first examinations a weak activity (+), especially for B. coli, was not detected or remained doubtful, but when the tests were repeated after several months, using a more satisfac- tory technic, the bacteriophage was clearly demonstrated. As will be seen upon examining the table where the results are re- corded (table 1), some of the examinations remained negative; the bacteriophage appeared to be absent. Would it have been the same if it had been possible to test the filtrate against all of the bacteria which may be found in the intestine? An answer to this question was sought. A specimen taken on July 1st was inactive toward the eight species of bacteria routinely employed, and it was tested against a different series of bacteria, selected at random. The filtrate showed a high activity for an organism of the Salmonella (hog cholera) group. When the same experi- mental tests were repeated on December 1st this filtrate was active for B. enteritidis. These two examples are sufficient to show that the absence of the bacteriophage is only apparent. It must be remembered that the ultramicrobe is recognized only by its activity, and con- sequently its presence in a filtrate can be detected only by test- ing it against a bacterium for which it has a definite activity. On the other hand, it is not possible to carry out the examina- tion on strains of all bacteria which may be found normally or occasionally in the intestine. For this there are several reasons, the chief being the difficulty of numbers, for there is no species of bacteria which may not be found in the intestinal tract at one time or another. We have seen further that certain bacterial species are not " homogeneous" as regards the bacteriophage. B. coli is of this group. Certain strains are attacked while others remain unharmed. It would then be necessary to make tests with all varieties of a single species; a new impossibility. Finally, it is known that there exist in the intestine certain bac- teria, revealed by the microscope, which it is impossible to iso- late and cultivate. May the bacteriophage not live in commensal- ism with these bacteria, with which they may form in the in- testine ''mixed cultures?" It may even be that the impossibility THE BACTERIOPHAGE IN THE HEALTHY INDIVIDUAL 231 of isolating these bacteria may be due solely to this phenomenon of commensalism, for we have seen that the agar on which mixed cultures are planted sometimes remains free of bacterial colonies. TABLE 1 DATE VIRULENCE OF THE INTESTINAL BACTERIOPHAGE FOR B. coli B. dysenteriae Bacillus Other organisms .1 2 to Flexner 1 Typhosus