STUDIES ON PRIMARYATYPICAL PNEUMONIA.

1I. OBSERVATIONSCONCERNINGTHE RELATIONSHIP OF ANON-HEMOLYTICSTREPTOCOCCUSTO THE DISEASE L2

By LEWIS THOMAS,3 GEORGES. MIRICK,J EDWARDC. CURNEN,3JAMESE. ZIEGLER, JR.,' AND FRANKL. HORSFALL, JR.3

(From the United States Navy Research Unit at the Hospital of The Rockefeller Institute for MedicalResearch, New York City)

(Received for publication August 14, 1944)

In a preliminary communication from thislaboratory (1), the isolation of a non-hemolyticstreptococcus from the lungs of 2 fatal cases ofprimary atypical pneumonia was reported. Itwas shown that convalescent sera of numerouspatients with primary atypical pneumonia werecapable of agglutinating this microorganism andsome sera yielded precipitates with appropriateextracts of it. It was also shown that, exceptin rare instances, similar reactions were notdemonstrable with acute-phase sera from thesame patients, sera from patients with otheracute infectious diseases, or sera from normalindividuals.

Since the publication of these preliminary ob-servations, additional strains of this non-hemo-lytic streptococcus have been isolated from thelungs and sputa of other patients with primaryatypical pneumonia. A comprehensive study ofthe biological and immunological characteristicsof this microorganism has been undertaken, andthe results are described in detail in separatecommunications (2, 3). The serological reac-tions obtained with this streptococcus have beeninvestigated, with additional specimens of serumfrom patients with primary atypical pneumonia,patients with other diseases, and normal persons,by means of a number of different immunologicalprocedures. Of the patients with primary atyp-ical pneumonia, 106 were studied in the Hospital

1 The Bureau of Medicine and Surgery of the UnitedStates Navy does not necessarily undertake to endorseviews or opinions which are expressed in this paper.

2The work described in this paper was done under acontract, recommended by the Committee on MedicalResearch, between the Office of Scientific Research andDevelopment and The Rockefeller Institute for MedicalResearch.

'Lieutenant Commander (MC), U.S.N.R.4Lieutenant (MC), U.S.N.R.

of The Rockefeller Institute, and an analysis ofthe clinical, roentgenological, and laboratoryfindings in these patients is presented in theaccompanying paper (4).

It is the purpose of the present paper to sum-marize the information which has been obtainedconcerning certain properties of this strepto-coccus, the methods by which it was isolated fromlung tissue, sputum, and throat swabs, and theserological reactions which were demonstrable withthe convalescent serum of patients with primaryatypical pneumonia. Evidence has been ob-tained which indicates that the strains of thismicroorganism which have been studied consti-tute a homogeneous group and belong to a singleserological type, and may readily be differen-tiated from other varieties of non-hemolyticstreptococci. It will be shown that the positiveserological reactions obtained with this strep-tococcus and convalescent serum from patientswith primary atypical pneumonia were due tospecific antibodies which developed during the*course of the illness. Moreover, evidence willbe presented which indicates that these anti-bodies were independent of the properties of non-specific complement fixation (5) and cold hemag-glutination (6, 7), which were also exhibited bycertain of the convalescent sera studied.

For convenience, this streptococcus will bereferred to in this report as "streptococcus MG."The problem of classification of streptococcusMGis discussed in a separate paper (2).

MATERIALS AND METHODS

SpecimensSpecimens of sputum from patients with primary

atypical pneumonia were frozen rapidly and stored at- 700 C. Sputa from patients with other acute infectionsof the respiratory tract were treated in an identical manner.

227

THOMAS, MIRICK, CURNEN, ZIEGLER, JR., AND HORSFALL, JR.

Lung tissues from fatal cases of primary atypical pneu-monia and from patients who died of other conditions werealso stored at -70° C." These various specimens werestored for periods ranging from a few days to 2 years beforethey were cultured. In addition to the frozen materials,fresh specimens of sputum and throat swabs from patientsand normal persons were also cultured.

Culture mediaFor the isolation of streptococcus MGfrom specimens of

lung tissue, Brewer's thioglycollate broth was employed(8). This medium was unsatisfactory for the culture ofsputum specimens or throat swabs because of the rapidovergrowth of streptococcus MGby other microorganisms.For the isolation of streptococcus MGfrom such speci-mens, a semiselective medium was employed which per-mitted the growth of this microorganism and other varietiesof non-hemolytic streptococci but inhibited the growthof most other bacterial species. This medium, which wassimilar in some respects to that recommended (9) for theselective cultivation of Streptococcus salivarius, was com-posed of the following constituents:

Proteose peptone DifcoYeast extract BactoBeef extract BactoGlucose C. P.Gentian violetSodium azideSulfapyridineDistilled water

5 grams

5 grams

3 grams

10 grams

2 mgm.

200 mgm.

500 mgm.

1000 ml.

The pH of the completed medium was adjusted to 7.2with 1N NaOHand the medium was then sterilized in theautoclave at 15 pounds pressure for 15 minutes.

For the m-aintenance and continued subculture ofstreptococcus MGin the laboratory, Todd-Hewitt brothwas employed. When cultures were stored at 40 C. forperiods of longer than 1 week, defibrinated rabbit bloodwas added to the broth in a concentration of 2 per cent.Blood broth cultures of streptococcus MGremained viablefor as long as 2 months at 40 C.

Nutrient agar containing 5 per cent sucrose was em-

ployed for the differentiation of streptococcus MGfromStr. salivarius. The latter organism has been shownto produce very large, mucoid colonies on agar containing5 per cent sucrose (10). In contrast, streptococcus MGgrew on this medium in the form of small fluorescentcolonies.

Rabbit antiseraAntiserum was prepared by the intravenous immuniza-

tion of rabbits with saline suspensions of heat-killed strep-tococcus MG. The details of the procedure used for

'The authors are grateful to Dr. A. A. Applebaum,Toledo, Ohio, Major J. H. Dingle, M.C., A.U.S., FortBragg, N. C., Dr. W. S. Tillett, New York, N. Y., Com-mander J. A. de Veer, M.C., U.S.N.R., New York, N. Y.,and Dr. W. C. Von Glahn, New York, N. Y., who gen-

erously supplied specimens of lung tissue.

immunization are described elsewhere (3). The immunesera thus prepared usually possessed agglutination titers of1: 2500 or higher, and also caused definite capsular swellingof streptococcus MG.

Procedures for isolation and identificationA loopful of sputum was inoculated into approximately

4 ml. of the semiselective medium described above, andincubated at 370 C. for 24 to 48 hours. Throat swabs wereplaced directly in 4 ml. of this medium and incubated ina similar manner. Lung tissue suspensions were culturedin Brewer's thioglycollate medium.

It should be noted that streptococcus MGgrew verypoorly, or not at all, when suspensions of lung tissue con-taining it were streaked on the surface of blood agarplates, either under aerobic or anaerobic conditions. Thereason for this is not clear, since broth cultures of strepto-coccus MGgrew readily when inoculated on blood agar,and formed small, grey, coniform colonies. The failureto obtain growth with direct cultures of lung tissue onblood agar was apparently not due merely to the presenceof small numbers of streptococci in the lung tissue sus-pension, since, as will be shown below, these suspensionseven when diluted to 10-5 yielded growth in thioglycollatemedium.

Whenvisible growth in broth had occurred, a loopful ofthe culture was mixed with a loopful of rabbit anti-streptococcus MGserum together with a loopful of 1 percent methylene blue, and was then examined for capsularswelling. A positive quellung reaction was taken toindicate the presence of either streptococcus MGor Str.salivarius, type I. Evidence is presented elsewhere (3)which shows that the latter microorganism possesses cer-tain antigenic components also present in streptococcusMG.

In order to differentiate streptococcus MGfrom Sir.salivarius, type I, the cultures were then inoculated on thesurface of 5 per cent sucrose agar plates. Single samplesof the small, fluorescent colonies formed on this mediumby streptococcus MGwere inoculated in beef-infusionbroth. Following growth in broth, the cultures wereagain tested for quellung in rabbit anti-streptococcus MGserum.

ISOLATION OF STREPTOCOCCUSMG

FROMLUNG TISSUE

Lung tissues from 8 fatal cases of primaryatypical pneumonia, and from 6 patients whodied of other causes were cultured by themethods described above. The results areshown in Table I. Streptococcus MGwas iso-lated from 6 of the 8 specimens of lung tissuefrom patients with primary atypical pneumonia.In contrast, this microorganism was not isolatedfrom any of the 6 lung specimens from patientswith other diseases, including 1 case of psittaco-sis, 1 of traumatic death, and 4 of pneumonia due

228

STREPTOCOCCUSMGANDATYPICAL PNEUMONIA

TABLE I

Isol4tion of streptococcus MGfrom the respiratorytract of human beings

Positive culturesSpecimen Diagnosis Numberntpatients

Number Per cent

Primary atypical 8 6 75Lung pneumonia

Other diseases* 6 0 0

Primary atypical 97 53 55pneumonia

Acute respiratory 39 13 33infection withoutpneumonia

Sputum or Pneumococcal 19 4 21throat swab pneumonia

Other bacterial 15 3 20pneumonias

Influenza A 9 0 0

Normal persons 57 7 12.5

* 1 psittacosis, 1 trauma, 4 bacterial pneumonia.

to other varieties of bacteria. The 6 patientswith primary atypical pneumonia whose lungtissues yielded positive cultures died between the8th and 22nd day after the onset of the disease,while the 2 negative specimens were from pa-tients who died on the 25th and 34th days,respectively, after onset.

Str. salivarius was not encountered in any ofthe cultures of lung tissue from patients with pri-mary atypical pneumonia. Four of the speci-mens contained small numbers of other bacteriawhich could be eliminated from the cultures bypreparing higher dilutions of the suspensions oflung tissue before culturing in thioglycollatebroth. It is noteworthy that pure cultures ofstreptococcus MGwere obtained from 4 of thelung specimens when dilutions as high as 10-5were cultured.

Occasional colonies of the strains of non-hemolytic streptococci isolated from lung tissuewere encountered which did not exhibit quellungin the presence of rabbit anti-streptococcus MGserum. Whengrown in broth, subcultures fromthese particular colonies produced a flocculenttype of growth, rather than the diffuse type ofgrowth characteristic of streptococcus MG.Further comments on these strains will be found

below under the heading: Dissociation of strep-tococcus MG.

ISOLATION OF STREPTOCOCCUSMGFROMSPUTUM

Specimens of sputum or throat swabs frompatients with primary atypical pneumonia, pa-tients with other infectious diseases, and normalpersons were cultured in the semiselective me-dium by the methods described above. Theresults are included in Table I. StreptococcusMGwas isolated from 53 of a total of 97 speci-mens from patients with primary atypicalpneumonia. This microorganism was also re-covered from 13 of 39 specimens from patientswith acute upper respiratory infections withoutpneumonia, or so-called "catarrhal fever," from4 of 19 specimens from patients with pneumo-coccal pneumonia, and from 3 of 15 specimensfrom patients with other types of bacterialpneumonia, but was not obtained from thesputum of 9 patients with influenza A. Strep-tococcus MGwas isolated from the throat cul-tures of 7 of a total of 57 normal persons.

BIOLOGICAL CHARACTERISTICS OF

STREPTOCOCCUSMG

A total of 59 strains of streptococcus MGhavebeen subjected to a detailed investigation. Eachof these strains was selected on the basis of thefollowing criteria: (1) quellung in rabbit anti-streptococcus MGserum, and (2) the formationof small fluorescent colonies on 5 per cent sucroseagar. Thirty-three of these strains were alsotested by the agglutination technique with acute-phase and convalescent serum from selected pa-tients with primary atypical pneumonia. Eachof these strains was agglutinated by convalescentserum but not by acute-phase serum. Thecharacteristics of the agglutination reactionswith these strains were identical to those notedwith the first strains of streptococcus MGiso-lated (1).

None of the 59 strains produced greening orhemolysis of rabbit blood agar when incubatedfor 24 hours. However, longer periods of in-cubation resulted in the production of distinctgreening by some strains. Several strains pro-duced more definite greening when grown onhorse or sheep blood agar.

A detailed account of the biological properties

229

THOMAS, MIRICK, CURNEN, ZIEGLER, JR., AND HORSFALL, JR.

of streptococcus MGis given in a separate paper(2). It was found that there was a strikinghomogeneity in the properties of different strainsof streptococcus MG, and that these propertiesmade it possible to differentiate this microorgan-ism from other non-hemolytic streptococci. Itwas also found that streptococcus MGwas notpathogenic for any of the common laboratoryanimals. It was highly resistant to bile, and tothe bacteriostatic action of sulfapyridine, astested by the method of MacLeod and Mirick(11). It was however, susceptible to the effect ofpenicillin.

Effect of storage at - 70° C.Most of the specimens of lung tissue and spu-

tum from patients with primary atypical pneu-monia had been stored at - 70° C. for periodsvarying from a few days to two years. It wastherefore of interest to determine the effect ofstorage under these conditions upon freshlyisolated cultures of stre'ptococcus MG. Youngbroth cultures of 19 strains of this bacteriumwere stored at -70° C. for 3 months. At theend of this time, subcultures from these tubeswere made in beef-infusion broth and in gentianviolet-sodium azide-sulfapyridine broth. It wasfound that 3 of the strains failed to grow ineither medium, and that 7 others grew in beef-infusion broth but failed to grow in the semi-selective medium. With 5 of the broth sub-cultures, capsular swelling in the presence ofrabbit antiserum could no longer be demon-strated. The adverse effect of storage of strep-tococcus MGunder these conditions suggeststhat this microorganism may be more readilyisolated from freshly obtained specimens ofsputum or lung tissue than from specimens whichhave undergone prolonged storage at - 70° C.

Soluble specific substance of streptococcus MGWhen sterile filtrates of broth cultures of

streptococcus MGwere mixed with homologousimmune rabbit serum, definite precipitation oc-curred. Precipitation did not occur when suchfiltrates were mixed with normal rabbit serumor with the serum of rabbits immunized withpneumococci or beta-hemolytic streptococci.This finding together with the positive quellungreactions obtained with homologous immune

rabbit serum suggested that streptococcus MGmight possess a capsular polysaccharide analog-ous to those of type-specific pneumococci.Evidence has been obtained which indicates thatstreptococcus MGdoes possess a capsular poly-saccharide and that the type-specific immuno-logical reactions obtained with this microorgan-ism are dependent upon the presence of thissubstance (3). The results of precipitation testswith the capsular polysaccharide and serumfrom patients with primary atypical pneumonia,intradermal tests with the substance in patientswith this illness, and tests for the antigenicityof the polysaccharide in normal human beingswill be given below.

Dissociation of streptococcus MGBy means of prolonged cultivation in broth

containing homologous immune rabbit serum,dissociation of streptococcus MGwas induced.The so-called R variants obtained by this pro-cedure no longer were capable of synthesizingdemonstrable capsular polysaccharide, and failedto give a positive quellung reaction with rabbitanti-streptococcus MGserum (3). The serumof rabbits immunized with R variants was cap-able in high dilution of agglutinating the Rvariants but either failed to agglutinate encap-sulated strains of streptococcus MGor agglu-tinated them only at very low serum dilutions.In their biological and immunological properties,these R variants closely resembled certain singlecolony strains described above, which were ob-tained from the lung tissues of fatal cases ofprimary atypical pneumonia and which werenegative in the quellung test. The availableevidence suggests the possibility that these latterstrains may represent similar non-encapsulatedvariants of streptococcus MG.

Antigenic relationship of streptococcus MG toStreptococcus salivarius, type I

The results obtained in studies on the bio-logical properties of streptococcus MG(2), indi-cated clearly the marked differences between thecharacteristics of this microorganism and thoseof the Str. salivarius group. One striking andeasily recognized differential property is theircolonial morphology on nutrient agar containing5 per cent sucrose. As was stated above, strep-

230

STREPTOCOCCUSMGAND ATYPICAL PNEUMONIA

tococcus MGproduced only small fluorescentcolonies on this medium, whereas Sir. salivariusproduced very large mucoid colonies. In con-trast with the significant differences in the bio-logical properties of these microorganisms, evi-dence also has been obtained which indicatesthat streptococcus MGpossesses certain anti-genic components in commonwith Str. salivarius,type I (3). Various serological cross-tests re-vealed antigenic similarities between the capsularpolysaccharides of these 2 different species ofnon-hemolytic streptococci. However, cross ab-sorption tests with immune rabbit sera showedthat although streptococcus MGand Str. sali-varius, type I, were related antigenically theywere not antigenically identical and could bedistinguished from each other in immunologicaltests with appropriately absorbed antisera.It should be pointed out that streptococcus MGwas not immunologically related to Str. sali-varius, type 11 (3).

SEROLOGICALREACTIONS WITH

STREPTOCOCCUSMG

Agglutination tests with streptococcus MGwere carried out with all available sera from 193patients with primary atypical pneumonia. Ofthese patients, 106 were studied in the Hospitalof The Rockefeller Institute, and the remaining87 patients were studied in other hospitals.6 Ascontrols, similar tests were performed with serafrom 321 normal persons and 120 patients withother diseases, including psittacosis, pneumo-coccal pneumonia, acute respiratory infectionswithout pneumonia ("catarrhal fever"), scar-let fever, influenza A, rheumatic fever, andother streptococcal infections. Agglutinationtests were also carried out with sera from 9 nor-mal human beings who were vaccinated intra-dermally with the capsular polysaccharide ofstreptococcus MG, and from 2 normal humanbeing who were vaccinated subcutaneously witha saline suspension of heat-killed streptococcusMG.

Selected specimens of serum were also testedby precipitation and complement fixation tech-

6 The authors are grateful to Dr. G. P. Berry, Rochester,N. Y., and Dr. W. T. Longcope, Baltimore, Md., whogenerously supplied specimens of sera from their patientswith primary atypical pneumonia.

niques against the capsular polysaccharide ofstreptococcus MG, and by the quellung tech-nique with the intact microorganism. Skintests with this polysaccharide were performedin 22 patients with primary atypical pneumonia,14 patients with other diseases, and 14 normalpersons.

The methods employed in these studies, andthe results obtained are presented in the follow-ing section.

AGGLUTINATION TESTS

Materials and methodsBacterial suspensions for agglutination tests were pre-

pared from 18-hour broth cultures of streptococcus MG.The bacterial cells were washed 3 times in 0.85 per centNaCl and suspended in sufficient saline to give a turbidityapproximating No. 5 in the McFarland scale. The strep-tococci were killed by heating at 650 C. for 1 hour. Mer-thiolate, in a final concentration of 1: 10,000, was addedas a preservative.

Serial 2-fold dilutions of unheated serum were made insaline. Each serum dilution was then mixed with an equalvolume of streptococcal suspension. The final dilutions ofserum tested ranged from 1: 10 to 1: 320. Dilutions lowerthan 1:10 were not used in routine tests because of thefrequent occurrence of non-specific agglutination of strep-tococcus MGin 1:2 or 1:4 dilutions of either normalhuman or rabbit serum. The tubes containing the serumand the streptococcal suspension were placed in a water-bath at 370 C. for 2 hours, followed by 18 hours in the ice-box at 40 C. They were then again placed in the water-bath at 37° C. for 2 hours, after which the tubes wereshaken and readings of the degree of agglutination weremade. The final period of 2 hours at 370 C. was found tobe of importance since non-specific agglutination of strep-tococcus MGoccasionally occurred at icebox temperatures.Such agglutination disappeared when the mixtures werebrought to 370 C.

Sera were not heated at 56° C., because it was found thatthe agglutination titer of convalescent serum was some-times reduced by heating at this temperature.

In estimating the degree of agglutination, the followingstandards were adopted. A designation of 4+ was as-signed to tubes in which agglutination was complete, witha solid plaque or disc of bacteria and a clear supernatantfluid. Agglutination with large clumps and clear super-natant fluid, but without complete settling of the bacteriato the base of the tube, was designated as 3 +. Agglutina-tion with incomplete clearing of the supernatant fluid wasdesignated as 2+. Agglutination with turbid fluid, butwith particles visible to the unaided eye, was designated as1+. Agglutination which required the use of a hand lensfor visualization was designated as A. The agglutinationtiter was taken as the highest dilution of serum in whichreactions of 1+ or more were observed.

231

THOMAS, MIRICK, CURNEN, ZIEGLER, JR., AND HORSFALL, JR.

TABLE II

Results of agglutination tests against streptococcus MGwithacute-phase and convalescent serum in

primary atypical pneumonia

Dilution of serumCase Day of

number disease1:10 1:20 1:40 1:80 1:160 1:320

284 6 0* 0 0 0 0 08 0 0 0 0 0 0

14 4 4 4 4 4 320 4 4 4 4 4 329 4 4 4 4 3 057 4 3 1 0 0 0

286 8 0 0 0 0 0 010 0 0 0 0 0 016 4 1 0 0 0 032 4 4 4 4 2 042 4 4 4 2 0 054 4 2 0 0 0 0

* Figures refer to degree of agglutination4 = complete agglutination with disc formation1 = slight agglutination0 = no agglutination

The results of agglutination tests againststreptococcus MGwith serum specimens ob-tained from 2 selected patients at various periodsduring the course of primary atypical pneumoniaare shown in Table II. It will be noted that, inboth instances, agglutination did not occur withsera obtained early in the course of the illness,whereas marked agglutination took place withsera obtained later. With sera from one patient,the agglutination titer was found to be 1: 320 onboth the 14th day and 20th day after onset,

1: 160 on the 29th day, and 1:40 on the 57th day.With sera from the other patient, the agglutina-tion titer was 1:20 on the 16th day, 1:160 on

the 32nd day, 1:80 on the 42nd day, and 1:20on the 54th day after onset. As is indicated bythe results shown in Table II, the agglutinationproduced by the convalescent serum of thesepatients consisted of large plaque-like discs ofbacteria with a clear supernatant fluid. Vigor-ous shaking of the tubes caused some breakingup of the larger clumps but obvious agglutina-tion persisted even after prolonged shaking.The gross appearance of the agglutinated bac-teria was indistinguishable from that which was

observed with this microorganism in the presence

of anti-streptococcus MGrabbit serum.The results of agglutination tests against

streptococcus MGwith acute-phase and con-

valescent sera from 193 patients with primaryatypical pneumonia are summarized in Table I II.It will be seen that of 158 acute-phase sera whichwere obtained during the first week of the dis-ease, only 5 were capable of agglutinating thismicroorganism. In contrast, 130 (67.4 per cent)of a total of 193 convalescent sera agglutinatedstreptococcus MG in serum dilutions rangingfrom 1:10 to 1:320. It will be noted that theconvalescent sera from 54 patients (28.4 per cent)of the entire group, possessed an agglutinationtiter of only 1: 10. Each of these sera was re-

tested, with similar results. The significance ofagglutination titers of 1:10 is obviously open to

TABLE III

Results of agglutination tests with streptococcus MGand the serum of human beings

Number Agglutination titer Total Percent-Diagnosis Serum of number age of

patients <1:10 1:10 1:20 1:40 1:80 1:160 1:320 positive positive

Primary atypical pneumonia Acute 158 153 3 1 1 0 0 0 5 3.2Conv. 193 63 54 35 19 16 4 2 130 67.4

Acute respiratory infection Acute 36 36 0 0 0 0 0 0 0 0without pneumonia Conv. 36 31 5 0 0 0 0 0 5 13.9

Other acute infectious diseases Acute 83 83 0 0 0 0 0 0 0 0Conv. 84 82 0 0 1 1 0 0 2 2.4

Normal persons 75 75 0 0 0 0 0 0 0 0

Normal persons 246 210 29 7 0 0 0 0 36 14.6(Training Station)

Vaccinated persons Before 11 11 0 0 0 0 0 0 0 0(Strep. MG) After 11 0 0 3 3 2 3 0 11 100.0

232

STREPTOCOCCUSMGAND ATYPICAL PNEUMONIA

question, particularly since serum dilutions lowerthan 1:10 were not employed as a routine.Convalescent sera from 76 patients (39 per cent)possessed agglutination titers of 1:20 or more,and convalescent sera from 41 patients (21 percent) had agglutination titers of 1:40 or more.As is shown in the accompanying paper (4),there appeared to be a positive correlation be-tween the severity of the illness and the agglu-tination titer of the convalescent serum forstreptococcus MG.

In Table III, the results of similar agglutina-tion tests with sera from patients with otherdiseases are also shown. It will be seen thatthe convalescent sera of 5 patients with acuterespiratory infections without pneumonia, so-called "catarrhal fever," agglutinated strepto-coccus MGat serum dilutions of 1:10. Noneof the sera from 31 other patients with a similarillness agglutinated this microorganism. Nega-tive results were obtained with both acute-phaseand convalescent sera from 82 patients withother diseases, including psittacosis (11 pa-tients), pneumococcal pneumonia (30 patients),influenza A (21 patients), rheumatic fever (6patients), early pulmonary tuberculosis (4 pa-tients), and scarlet fever (10 patients). Theserum of one patient with subacute bacterialendocarditis and streptococcus viridans bac-teremia possessed an agglutination titer of 1:80and the serum of another patient with empyemadue to a Group F beta-hemolytic streptococcushad an agglutination titer of 1:40. None of thesera of 75 normal persons, who had no history ofrecent respiratory infection, were capable ofagglutinating streptococcus MG. In agglutina-tion tests with sera from 246 persons living atclose quarters in a naval training station, 36 sera(14.6 per cent) showed agglutination titers of1:20 or less. The sera from this latter groupwere obtained at a time when an outbreak ofacute upper respiratory infections was in prog-ress, and it is probable that many persons in thegroup had either been exposed or were recoveringfrom such infections.

The results of agglutination tests with serafrom 11 human beings who were given injectionsof either the capsular polysaccharide or heat-killed suspensions of streptococcus MGare alsoshown in Table III. Agglutinins were not

demonstrable in the serum of these 11 personsbefore vaccination. Following vaccination witha suspension of heat-killed streptococci, 2 per-sons developed agglutinins in titers of 1: 20 and1: 40, respectively. Nine persons, who wereinjected with the capsular polysaccharide, de-veloped agglutinins in titers ranging from 1:20to 1:160. The agglutinins were demonstrablein serum obtained 8 days after vaccination, andwere still present in undiminished titer 7 monthslater. It will be seen that the agglutinationtiters observed with sera from vaccinated personswere comparable to those obtained with theconvalescent sera of patients with primary atyp-ical pneumonia. The type of agglutination pro-duced by the former sera was indistinguishablefrom that produced by the latter sera.

Agglutinins for streptococcus MGin the seraof patients with primary atypical pneumoniausually appeared between the second and fourthweeks of the disease and reached maximum titersduring this period. In most cases, the agglutina-tion titer began to diminish during the fifth andsixth weeks, and in some instances agglutininswere no longer detectable at this time. How-ever, the sera of certain patients still possessedsignificant titers as late as nine weeks after onset.The variability in the time of appearance anddisappearance of agglutinins for streptococcusMGwas such that it was found advisable to ob-tain specimens of serum at weekly intervals fromeach patient, whenever this was possible.

The specificity of the positive agglutinationreactions observed with convalescent serum wasinvestigated by testing selected samples ofstrongly positive serum against suspensionsprepared from cultures of other varieties ofbacteria. These included Hemophilus influenzae,Slaphylococcus aureus, pneumococci, and GroupA hemolytic streptococci. In addition, 18strains of other non-hemolytic streptococciwhich had been shown to be negative in thequellung test with anti-streptococcus MGrabbitserum were used. None of these various bac-terial species was agglutinated by selected con-valescent sera from patients with primaryatypical pneumonia.

It has already been mentioned that an anti-genic relationship exists between streptococcusMGand Str. salivarius, type I, which is ap-

233

THOMAS, MIRICK, CURNEN, ZIEGLER, JR., AND HORSFALL, JR.

parently due to antigenic similarities in thecapsular polysaccharides of these two micro-organisms (3). It was therefore of interest todetermine whether suspensions of Str. salivarius,type I, were also agglutinated by convalescentsera. Acute-phase and convalescent sera from99 patients with primary atypical pneumonia,each of whomwas found to have developed ag-glutinins for streptococcus MG, were testedagainst suspensions prepared from cultures ofStr. salivarius, type I, and Str. salivarius, typeII.7 It was found that the convalescent sera of6 of these 99 patients also were capable of ag-glutinating Str. salivarius, type I, and that theconvalescent sera of 3 other patients agglu-tinated Str. salivarius, type II. With sera fromthe remaining 90 patients, agglutination did notoccur in tests against either type I or type IIStr. salivarius. Acute-phase and convalescentsera from 45 patients with primary atypicalpneumonia, no one of whom developed demon-strable agglutinins against streptococcus MG,were also tested against suspensions of Str.salivarius, type I, and Str. salivarius, type II.It was found that none of the sera from these 45patients was capable of agglutinating either typeof this microorganism.

An analysis of the increases in titer observed inagglutination tests against streptococcus MGwith sera from patients with primary atypicalpneumonia is shown in Table IV. In a groupof 99 patients whose sera yielded positive results,and from whom2 or more samples of serum were

TABLE IV

Comparison of agglutination titers of acute-phase andconvaksscent sera of patients with

primary atypical pneumonia

Increase in titer ofStreptococcal Number convalescent serum

suspension patients _ _

OX 2X 4X 8X 16X32X64X

Streptococcus MG 99 3* 47 25 13 8 2 1

StrcPtococcus salivarius, 99 93 3 1 1 0 1 0type I

Streptococcus salivarius, 99 96 1 1 1 0 0 0type II

* Figures refer to number of patients who showed desig-nated increase in titer.

I The strains of Str. salivarius, type I and type II, wereobtained through the courtesy of Dr. J. M. Sherman ofCornell University School of Agriculture, Ithaca, N. Y.

obtained during the course of the disease, a2-fold increase in the titer of the convalescentserum was demonstrated in 47 patients, a 4-foldincrease in 25, an 8-fold increase in 13, a 16-foldincrease in 8, a 32-fold increase in 2, and a 64-foldincrease in 1. In 3 patients, agglutinins werepresent in the acute-phase serum and no increasein the titer of the convalescent serum wasdemonstrable.

Analyses of the increases in titer observed inagglutination tests against either Sir.' salivarius,type I or type II, with sera from this same groupof 99 patients, are also shown in Table IV. Itwill be seen that in tests with Sir. salivarius,type I, no increase in the titer of the convalescentsera of 93 patients was observed, while a 2-foldincrease in titer was demonstrated in 3 patients,a 4-fold increase in 1, an 8-fold increase in 1,and a 32-fold increase in 1. It will also benoted that in tests with Str. salivarius, type II,no increase in the titer of the convalescent seraof 96 patients was observed, while 2-fold, 4-fold,and 8-fold increases, respectively, were demon-strated in 3 patients.

It has already been mentioned that non-encapsulated, so-called R variants of strepto-coccus MGwere obtained by inducing dissocia-tion of this microorganism. These R variantspossessed antigens which were immunologicallydistinct from the capsular polysaccharide ofstreptococcus MG(3). When acute-phase andconvalescent sera from selected patients withprimary atypical pneumonia were tested, by theagglutination technique, against such Rvariants,it was found that the convalescent sera aggluti-nated the R variants whereas the acute-phasesera did not. The agglutination of R variantsproduced by convalescent sera was distinctlydifferent from the agglutination of encapsulatedstreptococcus MGproduced by the same sera.Agglutination of R variants was characterizedby the formation of very small bacterial clumpswhich were easily dispersed by gentle shaking.It is noteworthy that the agglutination of Rvariants produced by the serum of rabbits,immunized with either streptococcus MGor itsR variant, was similar in all respects to thatproduced by the serum of patients convalescentfrom primary atypical pneumonia. These find-ings suggested that convalescent sera possessed

234

STREPTOCOCCUSMGAND ATYPICAL PNEUMONIA

TABLE V

Results of agglutination tests with streptococcus MGand its R variant

Serum dilution (reciprocal)StreptococcalSerum Immunized with suspension

10 20 40 80 160 320 640 1280 2560 5000

Normal MG 0 0 0 0 0 0 - - -

R 0 0 0 0 0 0 -7 - - -

Rabbit Streptococcus MG MG 4 4 4 4 4 4 4 3 1 OR + + + + + + 0 0 0

R variant of strepto- MG + 0 0 0 0 0 0 0 0 0coccus MG R + + + + + + + + 0

Normal MG 0 0 0 0 0 -_ _ _R 0 0 0 0 0 - - - - -

Human Streptococcus MG MG 4 4 1 4 0 - _ _ _R + + + 0 - _

Capsular polysac- MG 4 4 4 4 2 0 - _ _charide of strepto- R 0 0 0 0 0 0 - - _coccus MG

Patient with Acute phase serum MG 0 0 0 0 0 0 - _ _primary R 0 0 0 0 0 0 - - - -atypicalpneumonia Convalescent phase MG 4 4 4 4 2 -

serum R + + + + + - _ - _

MG= encapsulated streptococcus MGR = R variant of streptococcus MG

antibodies specifically directed against at least2 separate and distinct antigenic components ofstreptococcus MG. Further evidence support-ing this interpretation was obtained from im-munization experiments with streptococcus MG,its capsular polysaccharide, and an induced Rvariant in rabbits and in human beings. Theresults of certain experiments bearing upon thispoint are summarized in Table V. It will beseen tl,at when either rabbits or human beingswere immunized with encapsulated streptococcusMG, they developed agglutinins for both the Rvariant and the encapsulated form. Whenrabbits were immunized with the R variant,they developed agglutinins against the R variantbut not against the encapsulated streptococcus.When human beings were immunized with thecapsular polysaccharide, they developed agglu-tinins against the encapsulated streptococcus,but not against the R variant. It would there-fore appear that the serum of patients con-valescent from primary atypical pneumonia re-acted in agglutination tests against streptococcusMGand its R variant in a manner analogous to

the serum of rabbits or human beings immunizedwith encapsulated streptococcus MG. It ap-pears evident that such convalescent sera reacteddifferently in these tests from the sera of humanbeings, immunized only with the capsular poly-saccharide, or the serum of rabbits, immunizedonly with the R variant.

PRECIPITATION TESTS

The capsular polysaccharide of streptococcusMGwas employed in precipitation tests, usingthe capillary tube method (12), with the acute-phase and convalescent sera of 16 patients withprimary atypical pneumonia. Similar tests werealso made with the sera of 10 normal persons,before and after the intradermal injection of thispolysaccharide.

The methods used for the extraction and purifi-cation of the capsular polysaccharide of strepto-coccus MGhave been described in detail else-where (3). When the serum of rabbits im-munized with streptococcus MGwas testedagainst the capsular polysaccharide, visible pre-cipitation occurred with the polysaccharide in

235

THOMAS, MIRICK, CURNEN, ZIEGLER, JR., AND HORSFALL, JR.

dilutions as high as 1:1,000,000. For tests withhuman sera, a 1: 10,000 dilution of the polysac-charide was found to be satisfactory.

Convalescent sera from 11 of the 16 patientswith primary atypical pneumonia producedvisible precipitation with the capsular poly-saccharide. None of the acute-phase sera fromthese patients yielded precipitates under iden-tical conditions. Each of the sera from 1Dnormal persons was negative. Four weeks afterthe injection of the capsular polysaccharide,however, specific precipitins were demonstrablein the sera of 9 of these 10 persons.

Absorption tests were carried out with selectedsamples of convalescent serum, in order to de-termine whether the agglutination and precipita-tion reactions were produced by the same anti-body. It was found that absorption of suchserum with either a suspension of streptococcusMGor with its capsular polysaccharide resultedin the removal of both the agglutinins for thestreptococcus and the precipitins reactive withthe polysaccharide.

The precipitation test was found to be oflimited usefulness with sera from patients withprimary atypical pneumonia, since positive reac-tions were only obtained with sera which pos-sessed relatively high agglutination titers. More-over, positive precipitation tests were usually notobtained for as long a period during convales-cence as were positive agglutination tests.

QUELLUNGTEST

Convalescent sera from 5 selected patientswith primary atypical pneumonia, each of whichhad a high agglutination titer for streptococcusMG, were tested for their capacity to producecapsular swelling of this microorganism. A loop-ful of undiluted serum was mixed with a loopfulof an 18-hour broth culture of streptococcus MG,and a loopful of 1 per cent methylene blue. Themixtures were then examined under the micro-scope for the presence of a positive quellungreaction. Definite quellung was produced by 2different samples of convalescent serum, ob-tained from 1 patient during the second andthird weeks of the disease. Each of these serapossessed an agglutination titer of 1: 320. Withthe convalescent sera from 4 other patients,

which had lower agglutination titers, capsularswelling was not observed.

COMPLEMENTFIXATION TESTS

Complement fixation tests were performedwith the acute-phase and convalescent sera of 12selected patients with primary atypical pneu-monia, using the capsular polysaccharide ofstreptococcus MGas an antigen. In other testswith this antigen, complement fixation had beenshown to occur in high dilutions of anti-strepto-coccus MG rabbit serum. The method em-ployed in these tests was the same as that usedin tests with rabbit antiserum, and is describedelsewhere (3).

It was found that the convalescent serum of 1patient in a serum dilution of 1:32 producedcomplement fixation with the capsular poly-saccharide of streptococcus MG. The convales-cent sera of each of the other 11 patients werenegative in this test, although some of thesesera possessed relatively high agglutination titersand were positive in precipitation tests.

SKIN TESTSWITH THECAPSULARPOLYSACCHARIDE

OF STREPTOCOCCUSMG

The capsular polysaccharide of streptococcusMGwas injected intradermally in 22 patientswith primary atypical pneumonia, 14 patientswith other acute infectious diseases, and 14normal persons. Similar tests were also per-formed in 5 persons who had received an intra-dermal injection of the polysaccharide 2 weekspreviously.

The skin test consisted of the intraderjnal in-jection in the anterior surface of the forearm, of0.1 ml. of sterile saline containing 1O jug. of poly-saccharide. A control injection of saline wasmade in a corresponding area on the other arm.

In 7 of the 22 patients with primary atypicalpneumonia who were given an intradermal injec-tion of polysaccharide during the second or thirdweek after onset of the disease, a definite re-sponse occurred at the site of inoculation. Thisreaction, usually consisted of an itching whealsurrounded by a zone of erythema. It appearedwithin 2 hours after the injection and persistedfor 24 hours or longer. In 15 of these 22 pa-tients, the reaction was negative. Negative

236

STREPTOCOCCUSMGAND ATYPICAL PNEUMONIA

results were also obtained in 14 patients withother infectious diseases and in 14 normalpersons.

In each of the 5 normal persons who had pre-viously received an injection of the polysac-charide, positive skin reactions occurred whichwere similar in all respects to those observed inthe patients convalescent from primary atypicalpneumonia.

There appeared to be no direct correlationbetween the presence of circulating antibodiesand positive skin reactions in the patients withprimary atypical pneumonia, and it was notpossible by means of serological tests to deter-mine at what period during convalescence thereaction might become positive. Repeated skintests in the same individual during the course ofthe disease produced results which could not beinterpreted, since a single injection of poly-saccharide stimulated the production of circu-lating antibodies against streptococcus MGandproduced sensitization of the skin to a secondinjection 1 week later. The skin test was con-sidered to be of limited practical usefulness,although the results obtained with it providedfurther evidence for the presence of specificantibodies against streptococcus MGin patientsconvalescent from primary atypical pneumonia.

PROPERTIES OF THE AGGLUTININ IN

CONVALESCENTSERUM

Certain properties of the agglutinin for strep-tococcus MG in the convalescent sera of 2patients with primary atypical pneumonia wereinvestigated. The agglutination titer of thesesera was not affected by alterations in pH over arange extending from pH 4.5 to 9.0. It wasfound by fractional precipitation with ammoniumsulfate that the agglutinin was present solely inthe water-insoluble portion of the globulin frac-tion of the serum. These properties are amongthose which characterize specific antibodies (13).

The agglutinin was found to be relativelythermolabile, as compared with certain otherantibodies. Four-fold or greater reductions inthe agglutination titers of certain sera wereproduced by heating the undiluted serum for30 minutes at 560 C. and the agglutinin wasalmost entirely destroyed by heating at 600 C.for a similar period. It is of interest that the

same degree of thermolability was encounteredwith the agglutinins which developed in theserum of monkeys which had been inoculatedintratracheally with streptococcus MG. More-over, heating for 30 minutes at 560 C. reducedthe agglutination titers of the sera of humanbeings who had received intradermal injectionsof the capsular polysaccharide of this micro-organism.

It has been shown that the acute-phase seraof patients with primary atypical pneumoniacontain an abnormal protein which reacts inprecipitation tests with the C-polysaccharide ofpneumococcus (4). This C-reactive protein isalso known to be present in serum during theearly stages of many other varieties of acuteinfectious disease (14). It was of importanceto determine whether any relationship existedbetween this protein and the agglutinin forstreptococcus MGin the serum of patients withprimary atypical pneumonia. In a series of 18patients, C-reactive protein was found to bepresent almost invariably in acute-phase serum,while agglutinins for streptococcus MGwere notdemonstrable in any of these sera. On the otherhand, C-reactive protein was not present in anyof the convalescent sera at the time when ag-glutinins for streptococcus MGhad reachedtheir maximum titers. These findings indicatethat the agglutinin for streptococcus MGwasdistinct, from the C-reactive protein.

It has also been shown that the convalescentsera of certain patients with primary atypicalpneumonia are positive in the cold-hemagglu-tination test (6, 7), and in complement fixationtests with a variety of unrelated tissue antigens(5). Each of these non-specific serologicalphenomena was encountered with the sera ofcertain patients in this series (4). However,many convalescent sera were found to possessagglutinins for streptococcus MGin high titerwhile cold-hemagglutinins were not demon-strable. Furthermore, the sera of a few patientswere capable of causing cold-hemagglutination inhigh titer, although they did not possess agglu-tinins for streptococcus MG. Similarly, a num-ber of sera which possessed agglutinins forstreptococcus MGfailed to exhibit the propertyof non-specific complement fixation. These ob-servations suggested that the agglutinins for

237

THOMAS, MIRICK, CURNEN, ZIEGLER, JR., AND HORSFALL, JR.

TABLE VIResults of cross-absorption tests with the convalescent serum of a patient with primary atypical pneumonia

Serum dilutionSerum Test -___ ___-___

1:10 1:20 1:40 1:80 1:160

Unabsorbed 1) Agglutination of streptococcus MG 4 4 4 4 42) Cold hemagglutination 4 4 2 2 03) Complement fixation with mouse lung 4 4 1 0 0

Absorbed with strep. MG 1) Agglutination of streptococcus MG 0 0 0 0 02) Cold hemagglutination 4 4 2 2 03) Complement fixation with mouse lung 4 4 0 0 0

Absorbed with human RBC 1) Agglutination of streptococcus MG 4 4 4 4 32) Cold hemagglutination 0 0 0 0 03) Complement fixation with mouse lung 4 4 1 0 0

Absorbed with mouse lung 1) Agglutination of streptococcus MG 4 4 4 4 42) Cold hemagglutination 0 0 0 0 03) Complement fixation with mouse lung 0 0 0 0 0

streptococcus MG were independent of theother 2 serological phenomena. In order to de-termine whether any direct relationship existed,cross-absorption tests were carried out withseveral samples of convalescent serum which werepositive in each of these 3 different tests. Theresults of a typical cross-absorption experimentare illustrated in Table VI. It will be seen that,before absorption, this serum produced agglu-tination of streptococcus MGin a titer of 1:160,cold-agglutination of human group 0 erythro-cytes in a titer of 1: 80, and complement fixationwith normal mouse-lung antigen in a titer of1:40. Following absorption of this serum withstreptococcus MG, the agglutinins for this micro-organism were completely removed while thetiter of both cold-hemagglutination and non-specific complement fixation remained almostunchanged. Similarly, when the serum wasabsorbed with human group 0 erythrocytes, orwith a suspension of mouse-lung tissue, the cold-hemagglutinins or the property of non-specificcomplement fixation were removed, while theagglutinin titer for streptococcus MGremainedthe same as before absorption. These resultsindicated that the agglutinin for streptococcusMGin convalescent serum was independent ofthe 2 non-specific serological phenomena men-tioned above.

DISCUSSION

Non-hemolytic streptococci are known to bepresent commonly in the upper respiratory

tract of human beings, and are generally re-garded as non-pathogenic saprophytes. Micro-organisms belonging to this diverse group areencountered in cultures of the throats of manynormal persons. The classification of non-hemolytic streptococci has not yet been com-pleted although many workers have entered uponthe problem. One distinct species of non-hemolytic streptococci, Str. salvarius, has beenwell characterized and can be distinguishedreadily from other non-hemolytic streptococci,both by biological and immunological tests (10).The so-called Str. mitis group appears to possessmuch less distinctive properties and probablycannot be considered to be a homogeneousspecies (10). The other varieties of non-hemo-lytic streptococci commonly present in the upperrespiratory tract have not so far been classifiedsatisfactorily.

The microorganism which has been referredto in this paper as streptococcus MGwas readilydifferentiated from either Str. sakivarius or theSir. mitis group. The strains of streptococcusMGwhich have been studied showed a suffi-ciently striking homogeneity as regards boththeir biological and immunological character-istics to warrant the conclusion that they con-stituted a distinct species and belonged to asingle serological type.

Streptococcus MG was isolated from' thelung tissues of 6 of a total of 8 patients who diedof primary atypical pneumonia, and was notobtained from the lung tissues of 6 patients who

238

STREPTOCOCCUSMGAND ATYPICAL PNEUMONIA

died of other causes. It was isolated from thesputum or throat swabs of 53 of a total of 97patients with primary atypical pneumonia, and20 of a total of 82 patients with other acuteinfectious diseases, as well as 7 of a total of 57normal persons.

Evidence was obtained which indicated thatin 4 of the lung tissues from fatal cases of pri-mary atypical pneumonia, streptococcus MGwas present in numbers of the order of 100,000or more per gram of lung tissue.

It was found that 67 per cent of 193 patientswith primary atypical pneumonia developed,during convalescence, antibodies directed againststreptococcus MG. The available evidence in-dicates that the serological reactions obtainedwith this non-hemolytic streptococcus and theconvalescent sera of patients with primaryatypical pneumonia were due to the presence ofspecific antibodies and not to some non-specificproperty of these sera analogous to that re-

sponsible for cold-hemagglutination (6, 7) or

complement fixation with various tissue antigens(5). With but relatively few exceptions, suchantibodies were demonstrable only in the con-

valescent sera of these patients and not in theserum of patients with other infectious diseases.Positive reactions did not occur when convales-cent sera possessing high agglutination titers forstreptococcus MGwere tested against a numberof other bacterial species. Str. salivarius, type I,which has been shown to possess certain antigensin common with streptococcus MG, was ag-

glutinated only by some 6 per cent of convales-cent sera which were capable of agglutinatingthe latter microorganism. Moreover, it hasbeen shown that the antibodies in the convales-cent sera of patients with primary atypicalpneumonia reacted with at least 2 separate anddistinct antigenic components of streptococcusMG, namely, its capsular polysaccharide andone or more somatic antigens.

The significance of this non-hemolytic strep-tococcus in relation to primary atypical pneu-

monia is not yet clear. There appear to be a

number of possible explanations which must beconsidered separately. First, it seems possiblethat the observed serological reactions might bethe result of a coincidental antigenic relation-ship between this non-hemolytic streptococcus

and some other agent, perhaps a virus, which isitself the causative agent in primary atypicalpneumonia. This possibility may be consideredas analogous to the positive serological reactionsobtained with members of the B. proteus groupin various rickettsial diseases. There are, how-ever, certain reasons for doubting the validityof this hypothesis. It would hardly accountfor the presence of this streptococcus in con-siderable numbers in the lungs of fatal cases ofprimary atypical pneumonia. Furthermore, itwould be surprising if there were an antigenicrelationship between 2 separate and antigeni-cally distinct components of this bacterium andsome other agent. In view of the presence inconvalescent serum of antibodies directed againstboth the capsular polysaccharide and somaticantigens of streptococcus MG, such an assump-tion would be necessary.

Secondly, it seems possible that this non-hemolytic streptococcus might occupy the r6leof secondary invader in primary atypical pneu-monia. Inthis regard, it is of interest that otherinvestigators have called attention to the fre-quent occurrence of non-hemolytic streptococciin primary atypical pneumonia (15, 16). Thepossibility that these microorganisms may repre-sent secondary invaders has been suggested(17, 18). An association between non-hemo-lytic streptococci and other clinical varieties ofpneumonia has also been reported by a number ofinvestigators (19 to 22). The possibility thatstreptococcus MGmay be involved in primaryatypical pneumonia as a secondary invadercannot be excluded. However, this term usuallyimplies an accidental or random associationbetween an infectious agent and a disease,rather than an association with such a relativelyhigh proportion of cases as is indicated by theresults of this study. It seems rather unlikelythat an accidental association should occur withmicroorganisms which, in each instance, weremembers not only of a single species but alsoof one serological type.

Thirdly, it seems possible that this non-hemolytic streptococcus, either alone or inconcert with some other infectious agent, mightbe primarily involved in the pathogenesis ofprimary atypical pneumonia. The availableevidence is not sufficient to warrant the accep-

239

THOMAS, MIRICK, CURNEN, ZIEGLER, JR., AND HORSFALL, JR.

tance of this hypothesis at the present time,although the results of this study suggest pathsfor further exploration.

SUMMARY

Non-hemolytic streptococci, comprising a ho-mogeneous group and belonging to a singleimmunological type, have been isolated from thelungs and sputa of patients with primary atyp-ical pneumonia. Specific antibodies directedagainst this species of streptococcus have beendemonstrated in the convalescent serum ofpatients with primary atypical pneumonia bymeans of a number of serological techniques.With few exceptions, such antibodies were notdemonstrable in the serum of these patientsduring the acute phase of the disease, in theserum of patients with other diseases, nor in theserum of normal persons.

The significance of these results is not yetclear, but they suggest the possibility that thisnon-hemolytic streptococcus may, in some man-ner, be implicated in the pathogenesis of primaryatypical pneumonia.

BIBLIOGRAPHY

1. Thomas, L., Mirick, G. S., Curnen, E. C., Ziegler, J.E., Jr., and Horsfall, F. L., Jr., Serological reac-tions with an indifferent streptococcus in primaryatypical pneumonia. Science, 1943, 98, 566.

2. Mirick, G. S., Thomas, L., Curnen, E. C., and Horsfall,F. L., Jr., Studies on a nonhemolytic streptococcusisolated from the respiratory tract of human beings.I. Biological characteristics of streptococcus MG.J. Exper. Med., 1944, 80, 391.

3. Mirick, G. S., Thomas, L., Curnen, E. C., and Hors-fall, F. L., Jr., Studies on a nonhemolytic strepto-coccus isolated from the respiratory tract of humanbeings. II. Immunological characteristics of strep-tococcus MG. J. Exper. Med., 1944, 80, 407.HII. Immunological relationship of streptococcusMGto Streptococcus salivarius, Type I. J. Exper.Med., 1944, 80, 431.

4. Curnen, E. C., Mirick, G. S., Ziegler, J. E., Jr.,Thomas, L., and Horsfall, F. L., Jr., Studies onprimary atypical pneumonia. I. Clinical featuresand results of laboratory investigations. J. Clin.Invest., 1945, 24, 209.

5. Thomas, L., Curnen, E. C., Mirick, G. S., Ziegler,J. E., Jr., and Horsfall, F. L., Jr., Complementfixation with dissimilar antigens in primary atypical

pneumonia. Proc. Soc. Exp. Biol. and Med., 1943,52, 121.

6. Peterson, 0. L., Ham, T. H., and Finland, M., Coldagglutinins (autohemagglutinins) in primary atyp-ical pneumonias. Science, 1943, 97, 167.

7. Turner, J. C., Development of cold agglutinins inatypical pneumonia. Nature, 1943, 151, 419.

8. Brewer, J. H., Clear liquid mediums for the "aerobic"cultivation of anaerobes. J.A.M.A., 1940, 115, 598.

9. Rose, K. D., and Georgi, C. E., Improved technic fordemonstrating the presence of Streptococcus sali-varius on eating utensils. Proc. Soc. Exper. Biol.and Med., 1943, 53, 25.

10. Sherman, J. M., Niven, C. F., Jr., and Smiley, K. L.,Streptococcus salivarius and other nonhemolyticstreptococci of the human throat. J. Bact., 1943,45, 249.

11. MacLeod, C. M., and Mirick, G. S., Quantitative de-termination of the bacteriostatic effect of the sul-fonamide drugs on pneumococci. J. Bact., 1942,44, 277.

12. Swift, H. F., Wilson, A. T., and Lancefield, R. C.,Typing group A hemolytic streptococci by Mprecipitin reactions in capillary pipettes. J. Exper.Med., 1943, 78, 127.

13. Marrack, J. R. The Chemistry of Antigens and Anti-bodies. Medical Research Council, special reportseries, No. 230. His Majesty's Stationery Office,London, 1938.

14. Tillett, W. S., and Francis, T., Jr. Serological Reac-tions in Pneumonia with a Non-protein SomaticFraction of Pneumococcus. J. Exper. Med., 1930,52, 561.

15. Kneeland, Y., Jr., and Smetana, H. F., Currentbronchopneumonia of unusual character and un-determined etiology. Bull Johns Hopkins Hosp.,1940, 67, 229.

16. Peterson, 0. L., and Finland, M., Modern treatmentof pneumonia. M. Clin. North America, 1943, 27,1291.

17. Thomas, H. M., Jr., The role of alpha hemolyticstreptococcus in pneumonia. Bull. Johns HopkinsHosp., 1943, 72, 218.

18. Rhoads, P. S., The probable incidence and clinicalfeatures of "virus" pneumonia. Radiology, 1943,40, 327.

19. Gerhartz, H., Akute Enterokokkenpneumonie unterdem Bilde des Typhus. Deutsche med. Wchnschr.,1934, 60, 206.

20. Senerchia, F. F., Jr., and Livengood, H. R., Strepto-coccus viridans pneumonia. Report of EightCases. New York State J. Med., 1941, 41, 143.

21. Solomon, S., and Kalkstein, M., Pneumonia due tothe Streptococcus viridans. Am. J. M. Sc., 1943,205, 765.

22. Annals of Pickett-Thomson Research Laboratory,Monograph 9, 521-569, 1933.

240