CLASSIFICATION OF BACTERIAL VIRUSES: CHARACTERISTICS OF THET1,D20 SPECIES OF COLI-DYSENTERY PHAGES'

MARK H. ADAMS AND EVELYN WADENew York University College of Medicine, New York, New York

Received for publication January 27, 1955

Coliphage Ti of Demerec and Fano (1945) hasbeen found to be closely related serologically todysentery phage D20 of Burnet (1933). A com-parative study of the properties of these twophage strains has been made in order that thedistinctive characteristics of the T1,D20 speciesmay be better defined. It is unfortunate that of alarge number of enteric phages studied, thesetwo were the only strains found to be in thisserological group. Burnet had found only oneother phage, strain G, to be serologically relatedto D20, and strain G is no longer available. Thispaper is one of a series devoted to the develop-ment of a systematic classification of the bacterio-phages (Adams, 1953). It is hoped that suchstudies will indicate those taxonomic criteriawhich are of greatest value in phage classification.

MATERIALS AND METHODS

Phage strains Ti and D20 are well known, andtheir properties have been reported in numerouspublications to which reference will be made.Phage T1 has also been known as phage alphaand phage P28. Conditions affecting the adsorp-tion rate of phage Ti to its host cell have beeninvestigated by Puck, Garen, and Cline (1951).Various studies on the physical properties ofphage Ti have been collected and discussed byPollard (1953). Most of the technical details ofmethods used in the present paper have been de-scribed in a review (Adams, 1950).

EXPERIMENTAL RESULTS

Serological relationships. A valuable screeningprocedure in phage classification is the testing ofunknown phages for susceptibility to neutraliza-tion by various known antiphage sera. Neutral-ization of one phage by the antiserum to anothersuggests a close phylogenetic relationship. How-ever, this must be checked by numerous other

1 Aided by a grant from the National Founda-tion for Infantile Paralysis.

tests because of the possibility that the observedneutralization might be due to contamination ofan antiserum with unsuspected antibodiesagainst other phages. In contrast, the absence ofcross neutralization does not rule out close rela-tionship (Adams and Wade, 1954).One of the peculiarities of phage Ti is that its

neutralization by Ti antiserum does not gener-ally follow first order kinetics (Delbriick, 1945)as does the neutralization of most other phagesby their antisera. However, an exceptional anti-serum studied by Hershey (personal communica-tion) did give first order kinetics for neutralizationof phage Ti. Most anti-Ti sera are of rather lowtiter, it being seldom that such sera can be usedat dilutions above 1 :100.

According to Burnet, Keogh, and Lush (1937)phage D20 was a relatively poor antigen, but noneutralization data for this phage were pub-lished. The first anti-D20 serum that we preparedwas something of a surprise. This serum gavefirst order neutralization kinetics through 99 pecent inactivation at a serum dilution of 1/1,000.Because of this unexpected result, two more rab-bits were immunized with phage D20. The seraof these rabbits gave a close approximation tofirst order kinetics for neutralization of about 90per cent of the phage stock, after which theneutralization rate slowed markedly. One ofthese sera was quite potent with a neutralizationrate constant of 1,600 min-. These and otherresults indicate that closely related phages aresimilar in the shape of neutralization curvesusually obtained and in the neutralization titersof typical antisera. However, sera from excep-tional rabbits may differ in neutralizationkinetics from those usually observed. Similarexceptional sera have been reported for phageT5 by Fodor and Adams (1955).

Various anti-Tl sera gave strong cross neutral-ization reactions with phage D20, and two outof three anti-D20 sera neutralized phage Ti thusdemonstrating a serological relationship between

253

on June 6, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

MARK H. ADAMS AND EVELYN WADE

TABLE 1Neutralization kinetics with anti-Ti and anti-D20

sera. Initial neutralization velocity constantsat the serum dilution stated

k Values and Serum DilutionsSerum Used

Phage Ti Phage D20

Anti-Ti serum no. 10 at 1/100 10 at 1/100380

Anti-Tl serum no. 2.5 at 1/20 1.4 at 1/2079

Anti-D20serum 6atl/100 64 at 1/1,000no. 100

Anti-D20 serum 3 at 1/100 1,600 at 1/5,000no. 89

Anti-D20 serum 0 at 1/10 115 at 1/100no. 88 99 at 1/300

95, 74, 75 at1/1,000

2.3 D log Po/Ptt min

these two phages. None of the cross reactingsystems had first order neutralization kinetics.However, the initial slopes of the neutralizationcurves were used to calculate initial velocityconstants to serve as an estimate of the relativeneutralizing potencies of the various antisera. Itmay be seen from the results in table 1 that theTi antisera neutralized phages Ti and D20equally, whereas the anti-D20 sera were far moreactive against D20 than against Ti. Similar non-symmetrical relationships have been notedwithin other phage species (Adams and Wade,1954; Fodor and Adams, 1955).Host ranges. As may be seen from table 2, the

TABLE 2Host ranges of TI and D20. For comparison the

host range of T5 is also included

Host Ti D20 T5

B + + +B/1, tryptophan - + +B/1,5 _ + -B/D20 + - +Shigella paradysenteriae (2) - +Shigella sonnei - +

+ indicates plaques are formed when a fewhundred phage particles are used per plate.- indicates no plaques under the same condi-

tions.

TABLE 3The efficiency of plating of Ti and D20 phage

stocks on various indicator strains

Plaque Counts with0.1 ml of a 10- Dilution

Host Bacterial Strain of Stocks of Phage

Ti D20

B 120 7B/1,5 0 5B/D20 138 0S. paradysenteriae (type 2) 0 88

host ranges of phages Ti and D20 are quite dis-tinct from each other. Phage T5, which is notrelated to phage Ti in a phylogenetic sense, isincluded because it is similar to Ti in host rangebut differs from D20. These results demonstrateagain that host range is not of value in showingrelationships among phages at the species level.Bacterial strain B/D20 is a mutant of strain Bisolated by plating the latter with an excess ofD20 phage and subculturing a surviving colony.It is a suitable indicator for phage Ti in thepresence of D20. Neither Ti nor D20 was ableto attack any of a considerable number of sal-monella strains, and no hosts for Ti were foundamong the shigella strains tested.The efficiency of plating (EOP) of phages Ti

and D20 on several indicator strains of bacteriais recorded in table 3. The EOP of D20 on strainB and on a number of mutants of strain B wasonly about 0.1 as compared with plaque countson Shigella paradysenteriae. The EOP was thesame whether the D20 phage stock had beengrown on strain B or on the shigella host, indicat-ing that the EOP is not a phenotypic propertyunder host control. Attempts to find mutants ofD20 with increased EOP, or environmental con-ditions which would increase the EOP of phageD20 on strain B, have not been successful. Wehave been unable to find host range mutants ofD20 capable of plating on B/D20, or mutants ofTi able to plate on B/1,5, S. paradysenteriae, orShigella sonnei. Host range mutants of Ti platingon B/1 tryptophan are well known.The low efficiency of plating of phage D20 on

strain B is due to a very slow rate of adsorption.It was impossible to measure the rate of phageadsorption even with bacterial concentrations of109/ml and adsorption periods as long as 2 hoursat 15 C. For this reason it was not possible to doone step growth experiments with phage D20 on

254 [VOL. 70

on June 6, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

CLASSIFICATION OF BACTERIAL VIRUSES

strain B, or to prepare phage stocks of D20using strain B as host in broth. Starting with107B/ml and 109D20/ml of broth culture at37 C, the bacteria grew as if the phage were notpresent, mass lysis never occurred, and the phagetiter remained constant. Changing the salt con-centration in broth did not improve the situation.Stocks of phage D20 on B as host could, however,be prepared by the agar layer method (Swan-strom and Adams, 1951). Plaques of D20 onstrain B were variable in size with a preponder-ance of very tiny plaques as might be expected ofphage particles which adsorb very slowly.

Size and morphology. The size of phage D20 wasdetermined by Elford and Andrewes (1932) byultrafiltration through gradacol membranes.They found the filtration end point at a porediameter of 60 m, from which they calculated aparticle diameter of 20-30 m,u. On the basis ofvarious studies, Lea (1946) decided that the con-version factors used by Elford were too small.Using Lea's revised factor (0.83), the size de-duced from the filtration data would be 50 m,u.Phage Ti has been pictured repeatedly in elec-tron micrographs, perhaps the best picturesbeing found in Williams and Fraser (1953). Theseauthors found for air dried preparations: head,65 m,, tail, 150 by 10 mMi; for frozen driedpreparations: head, 50 m,u, tail, 150 by 10 mu.Through the kind cooperation of Dr. DavidSlautterback we have obtained electron micro-graphs of air dried preparations of phage D20.The phage particles have a spherical head about65 m,u in diameter and a tail 150 by 10 m,. Thepictures of D20 particles were indistinguishablefrom those of air dried preparations of phage Ti.The target size of phage Ti for ionizing radiationshas been determined by several authors. ForX-rays it was found to be 38 m,u in diameter byLuria and Exner (1941) and by Watson (1950).For deuterons the target size was found to be 28m,u by Pollard and Forro (1949). The target sizefor ionizing radiations is considerably smallerthan the true size of the phage head, as is gener-ally true for phages above 50 m,u in diameter. Onthe basis of quantitative studies of the effects ofionizing radiations on the properties of phage Ti,Pollard (1953) has drawn a rather detailed pic-ture of the structure of this phage. All that onecan say for this diagram is that evidence to proveit wrong is not available.Both phages Ti and D20 produce large

plaques, 10 mm in diameter, when plated by the

agar layer method on appropriate host bacteria.Photographs of Ti plaques have been publishedby Demerec and Fano (1945) and of D20 plaquesby Burnet (1933). As noted previously, theplaques of D20 phage on strain B are rathersmall and uneven in size due to the very slowrate of adsorption on this host.

Susceptibility to inactivation. Similarities in thesusceptibility of various phage strains to inactiva-tion were shown to be correlated with serologicalrelationship by Burnet (1933), who suggested theuse of sensitivity to inactivation as a taxonomiccriterion.

(a) Sensitivity to ultraviolet inactivation.Phages were subjected to irradiation with ultra-violet of 2537 A in a phosphate buffer usingphage T2 as a biological calibration of irradiationdosage. A plot of the logarithm of the survivorsas a function of the dosage has an initial upwardconcavity for phage Ti as noted by Dulbecco(1950) and for D20 also. The curve becomes linearafter inactivation of about 50 per cent of thephage population, permitting use of the first ordervelocity constant to characterize the inactivationrate. The inactivation curve for D20 is similar tothat for T1, the rate for T1 being 0.12 and forD20 being 0.14 of that for T2 under the sameconditions.

(b) Sensitivity to heat. Heat stability was de-termined by diluting the phage in nutrient brothcontaining 10-3 M CaCl2 at 70 C and removingsamples at intervals for assay. Inactivationobeyed the kinetics of a first order reaction, thevelocity constants being 1.3 min-' for Ti and 1.8min-1 for D20.

(c) Photodynamic action of methylene blue.This is one of the taxonomic criteria originallyproposed by Burnet (1933). The conditions forinactivation are described in detail by W7elsh andAdams (1954). After a slight initial lag the in-activation kinetics were first order. The velocityconstants were 0.75 min-' for Ti and 1.05 min-'for D20, and under the same conditions 0.12 forT2 as a standard. These experiments were carriedout by James N. Welsh.

Latent period. The latent period of intracellularvirus multiplication may be defined as theminimum period between adsorption of a phageparticle to the host cell and the lysis of the hostcell with liberation of phage progeny as deter-mined in the one step growth experiment. Thelatent period under defined environmental con-ditions is a remarkably constant characteristic of

2551955]

on June 6, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

256 MARK H. ADAMS A

a phage strain. The latent period for phage Tiin nutrient broth at 37 C using strain B as hostis 13 min (Delbruck and Luria, 1942). Becauseof its slow adsorption it was not possible to de-termine the latent period of phage D20 on strainB; however, its latent period on S. paradysen-teriae was easily shown to be about 20 min. Thelatent period may also be determined by follow-ing the lysis of a broth culture of the bacteria byturbidimetric methods after infection with anexcess of phage. Under these conditions lysisbegan after the same latent periods as observedin the one step growth experiment, indicatingthat the phenomenon of lysis inhibition (Doer-mann, 1948) does not occur with phages TIand D20.

Citrate sensitivity. The inhibition of plaque de-velopment by citrate added to the plating agarwas one of the physiological properties proposedby Burnet (1933) as a taxonomic criterion. Whenphage growth is inhibited by citrate, it is usuallydue to a nutritional requirement for calcium ion.Burnet reported that plaque formation by D20phage was not inhibited by 1.5 per cent citrate,whereas certain other phages were inhibitedcompletely by as little as 0.5 per cent citrate. Wereported in a previous publication (Adams, 1949)that 1 per cent citrate in a broth medium pre-vented lysis of host bacteria by phage Ti. Thepresence of 1 per cent citrate in the plating agardid not prevent plaque development but did re-duce the efficiency of plating. The reason for thedecreased efficiency of plating and the failure oflysis in citrate broth is not completely clear al-though it may well be due to the abortive infec-tion that occurs with phage Ti when divalentcations are lacking (Puck, 1949; Ruegamer, 1954).

TABLE 4Effect of citrate on plating efficiency of phages ondifferent hosts. The basal agar contained 0.04 M

sodium citrate (about 1%)

Plaque counts

D20 on S.Ti on B paradysen- D20 on B

tersae

Control ............ 165 91 120Citrate ............ 202,221 28,37 0,0

* The D20 inoculum used with strain B ofEscherichia coli was 10 times larger than theinoculum used on the shigella host.

ND EVELYN WADE [VOL. 70

In confirmation of the earlier experiments we havefound that plaque formation by phages TI andD20 on efficient host bacteria is not prevented bythe inclusion of 1 per cent sodium citrate in theagar medium although the efficiency of plating issometimes decreased. However, if the phage isplated on a host on which adsorption is poor,plaque development may b3 completely pre-vented. This may be seen from the experimentsrecorded in table 4. The efficiency of plating ofD20 on S. paradysenteriae is reduced to about 0.3by citrate, whereas the efficiency of plating ofthis phage on strain B is reduced from about 0.1in the absence of citrate to less than 0.001 withcitrate. These results are entirely consistent withthe hypothesis of Puck that absence of divalentcations results in a "sensitized" phage particlewhich is inactivated on adsorption to the hostcell. A slow adsorption rate gives a longer periodof time during which the divalent cations candissociate from the phage particle. These experi-ments indicate that calcium ion is not a specificrequirement for the multiplication of thesephages as it is for T5 and many other phages.

Multiplicity reactivation. The reactivation ofultraviolet inactivated phage particles as a resultof multiple infection of bacteria was demon-strated by Luria (1947) for phages T2, T4, T6,and T5. In a footnote in a later paper (Luria andDulbecco, 1949) it was stated that some reactiva-tion by multiple infection takes place with phageTi but with a much lower frequency than withthe T-even phages and T5. Bresch (1950) statedthat he had been able to detect a weak multi-plicity reactivation with phage Ti although theonly experiment presented failed to show evidenceof multiplicity reactivation. We therefore at-tempted to demonstrate the phenomenon ofmultiplicity reactivation with phages Ti andD20, under conditions in which it is readilydemonstrable with T2.A stock of phage Ti assaying 2 X 1010/ml was

irradiated with ultraviolet light until the titerdecreased to 5 X 106/ml. A culture of strain B innutrient broth was grown to a titer of 1.3 x108/ml. To 0.9 ml of the bacterial culture wasadded 0.1 ml of irradiated phage, adsorption wascontinued for 5 min at 37 C, and the mixture wasthen assayed for plaque forming particles. Thetube contained 2 X 109 total phage particles and1.2 X 101 bacteria per ml, an input ratio of 17P/B. The input ratio of active phage to bacteria

on June 6, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

CLASSIFICATION OF BACTERIAL VIRUSES

was only. 1/24. Under these conditions at least60 per cent of the phage should adsorb if irradia-tion did not affect adsorption rate, and theaverage multiplicity of infection should be about10. The actual recovery of infectious centers inthe adsorption tube was 4.5 X 105/ml in one

experiment and 4.2 X 105/ml in another experi-ment in comparison with 5 X 105/ml expectedif no reactivation had taken place. Phages relatedto T2 or T5 would have shown a very large in-crease in plaque count under these conditions.A similar experiment carried out with irradi-

ated phage D20 and S. paradysenteriae as hostalso showed no evidence for multiplicity reactiva-tion. In this experiment the input ratio was 20phage particles per bacterium, and control ex-

periments with unirradiated phage indicatedbetter than 90 per cent adsorption under theseconditions. We cannot conclude from these ex-

periments that multiplicity reactivation does notoccur with these phages because we have notcarried out an exhaustive study of conditionsaffecting the phenomenon. We were unable,however, to demonstrate multiplicity reactiva-tion with phages Ti and D20 under conditions inwhich this phenomenon is very readily demon-strated with phages related to T2 and T5. Thisis a very marked difference in behavior which we

believe to be of taxonomic significance.

Mixed infection. The results of mixed infectionhave been proposed as a taxonomic criterion forestimating the degree of relationship between twophage strains (Adams, 1953), genetic recombina-tion being evidence of relationship within thespecies category. That genetic recombinationoccurs among mutants of phage Ti has beenabundantly demonstrated by Bresch (1953).

It would seem likely that genetic recombina-tion between Ti and D20 could be demonstratedif a suitable common host could be found formixed infection experiments, but this technicaldifficulty has not been surmounted as yet.

DISCUSSION

The experiments described in this paper havedemonstrated that phages Ti and D20 are (1)serologically related, (2) morphologically indis-tinguishable, and (3) quantitatively similar inphysiological properties. These results justify theconclusion that these two phages are closely re-

lated phylogenetically and should be classified as

two strains in the same species of bacteriophage.In spite of the close similarity in most properties,these two strains may be readily distinguished bydifferences in host range and serological speci-ficity.The properties of the T1,D20 species of bac-

teriophage are summarized in table 5, which in-

TABLE 5A comparison of the distinctive properties of four species of bacteriophage

Property T1,D20 T3,D44 T2,C16 TS,PB

Size, head m,u* 50 47 95 by 65 65Size, tail m,u*........................ 150 by 10 15 by 10 100 by 25 170 by 10Ultraviolett ......................... 0.12-0.14 0.1-0.3 0.5-1.0 0.8Heatt ............................ 1.3-1.8(70 C) 0.03-0.3(60 C) 0.7(70 C) 0.1-0.7(70 C)Photodynamic action§ ............... 0.8-1.1 1.9-2.1 0.12-0.16 2.7-2.8Latent period, min.................. 13-20 13-20 20-25 40-50Calcium requirement ................ No No No YesLysis inhibition.......... No No Yes NoMultiplicity reactivation............. No No Yes YesSerological group (Burnet)II Group 5 Group 2 Group 11 ?Host genera....... Escherichia Escherichia Escherichia Escherichia

Shigella Shigella Shigella ShigellaSerratia Salmonella Serratia

Salmonella

* Size in electron microscope (Williams and Fraser, 1953).t Rate of inactivation relative to that of phage T2 as 1.0.$ First order velocity constants K min-' at indicated temperature.§ First order velocity constants under standard conditions; data from Welsh and Adams (1954).11 For Burnet's taxonomic criteria see Burnet (1933).

2571955]

on June 6, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

MARK H. ADAMS AND EVELYN WADE

cludes for comparison the corresponding proper-ties of the T3,D44, the T2,C16, and the T5,PBspecies of enteric phages. Examination of thistable will reveal that the properties of the T1,D20species are very similar to those of the T3,D44species. In contrast, these two species differ mark-edly in properties from the T2,C16 and theT5,PB species. The most obvious difference be-tween the T1,D20 species and the T3,D44 speciesis the length of the tail.No information is available at present as to

whether morphological changes of this order canoccur readily in phage phylogeny or not. It is notinconceivable that a single mutation might resultin amputation of the tail, or it might be the resultof a slow evolutionary change. The similaritiesbetwveen the T1,D20 and the T3,D44 species sug-gest the possibility that these two species mightbe closely enough related to be included in thesame genus. However, suitable data must beaccumulated on a great many other phage strainsbefore definite suggestions can be made withrespect to criteria for taxonomic categories abovethe species level.

SUMMARY

The properties of phage strains Ti and D20have been compared. It is suggested that thesetwo phages should be classified as two strainswithin the same species. A comparison of theproperties of four species of enteric phage indi-cates that theT ,ID20 species is much more closelyrelated to the T3,D44 species than these twospecies are to the other species studied.

REFERENCES

ADAMS, M. H. 1949 The calcium requirement ofcoliphage T5. J. Immnunol., 62, 505-516.

ADAMS, M. H. 1950 Methods of study of bac-terial viruses. In Methods in medical research.Vol. II. Year Book Publishers, Chicago, Ill.

ADAMS, M. H. 1953 Criteria for a biologicalclassification of bacterial viruses. Ann. N. Y.Acad. Sci., 56, 442-447.

,ADAMS, M. H., AND WADE, E. 1954 Classifica-tion of bacterial viruses: The relationship oftwo Serratia phages to coli-dysentery phagesT3, T7, and D44. J. Bacteriol., 68, 320-325.

BRESCH, C. 1950 Zur Reaktivierung von Bak-teriophagen. Z. Naturforsch., 5b, 420-422.

BRESCH, C. 1953 Genetical studies on bacterio-phage Ti. Ann. inst. Pasteur, 84, 157-163.

BURNET, F. H. 1933 The classification of

dysentery-coli phages. III. Correlation ofserological classification with certain bio-chemical tests. J. Pathol. Bacteriol., 37,179-184.

BURNET, F. M., KEOGH, E. V., AND LUSH, D.1937 Immunological reactions of filterableviruses. Australian J. Exptl. Biol. Med. Sci.,15, 227-368.

DELBRUCK, M. 1945 Effects of specific antiseraon the growth of bacterial viruses. J. Bac-teriol., 50, 137-150.

DELBRUCK, M., AND LURIA, S. E. 1942 Inter-ference between two bacterial viruses actingupon the same host, and the mechanism ofvirus growth. Arch. Biochem., 1, 111-141.

DEMEREC, M., AND FANO, U. 1945 Bacterio-phage resistant mutants in E. coli. Genetics,30, 119-136.

DOERMANN, A. H. 1948 Lysis and lysis inhibi-tion with E. coli bacteriophage. J. Bacteriol.,55, 257-276.

DULBECCO, R. 1950 Experiments on photore-activation of bacteriophage inactivated byultraviolet light. J. Bacteriol., 59, 329-347.

ELFORD, W. J., AND ANDREWES, C. H. 1932 Thesizes of different bacteriophages. Brit. J.Exptl. Pathol., 13, 446-456.

FODOR, A., AND ADAMS, M. H. 1955 Geneticcontrol of serological specificity in bacterio-phage. J. Immunol., 74, 228-235.

LEA, D. E. 1946 Actions of radiations on livingcells. Cambridge University Press, Cam-bridge, England.

LURIA, S. E. 1947 Reactivation of irradiatedphage by transfer of self-reproducing units.Proc. Natl. Acad. Sci. U. S., 33, 253-264.

LURIA, S. E., AND DULBECCO, R. 1949 Geneticrecombinations leading to production ofactive bacteriophage from ultraviolet inacti-vated bacteriophage particles. Genetics, 34,93-125.

LURIA, S. E., AND EXNER, F. M. 1941 The in-activation of bacteriophage by X-rays-influence of the medium. Proc. Natl. Acad.Sci. U. S., 27, 370-375

POLLARD, E. C. 1953 The physics of viruses.Academic Press Inc., New York.

POLLARD, E., AND FoRRo, F., JR. 1949 Examina-tion of target theory by deuteron bombard-ment of Ti phage. Science, 109, 374-375.

PUCK, T. T. 1949 A reversible transformationof Ti bacteriophage. J. Bacteriol., 57,647-655.

PUCK, T. T., GAREN, A., AND CLINE, J. 1951The mechanism of virus attachment to hostcells. I. The role of ions in the primary reac-tion. J. Exptl. Med., 93, 65-88.

258 [VOL. 70

on June 6, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

CLASSIFICATION OF BACTERIAL VIRUSES

RIJEGAMER, W. 1954 The role of cations in the

reversible inactivation of coliphage in mediaof low ionic strength. I. The competitiveaction between H+ and other cations. Arch.Biochem. and Biophys., 50, 269-284.

SWANSTROM, M., AND ADAMS, M. H. 1951 Agar

layer method for production of high titerphage stocks. Proc. Soc. Exptl. Biol. Med.,78, 372-375.

WATSON, J. D. 1950 The properties of X-rayinactivated bacteriophages. I. Inactivationby direct effect. J. Bacteriol., 60, 697-718.

WELSH, J. N., AND ADAMS, M. H. 1954 Photo-dynamic inactivation of bacteriophage. J.Bacteriol., 68, 122-127.

WILLIAMS, R. C., AND FRASER, D. 1953 Mor-phology of the seven T-bacteriophages.J. Bacteriol., 66, 458-464.

1955] 259

on June 6, 2018 by guesthttp://jb.asm

.org/D

ownloaded from