divisions of pulmonary suppuration, i.e., endobronchial and

THE EXPERIMENTAL PRODUCTION OF ABSCESSOF THE LUNG *

BY ELLIOTT C. CUTLER, M.D.AND

S. A. SCHLUETER, M.D.OF CLEVELAND, OHIO

FROM THE LABORATORY OF SURGICAL RESEARCH, THE LAKESIDE HOSPITAL AND THE WESTERN RESERVE

UNIVERSITY SCHOOL OF MEDICINE

ABSCESS of the, lung is a general term applied to a wide variety of sup-.purative conditions occurring within the lung. These conditions differ greatlyin their mode of production, their morbid pathology and their response to

r~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4~

Hz'',e!fi..'.. "iK§;;,D.,./.

FIG. I.-Segment of femoral vein tied at base and held open for the insertion of a piece of lead and thebacterial emulsion.

therapy. They enter into a common category merely because they occupy thesame organ, and they differ as widely as do comparable conditions within theliver, such as amoebic abscess and the abscess of an ascending pylephlebitis.Certainly the post-pneumonic abscess, the bronchiectatic abscess, and the post-operative abscess of the lung have a quite dissimilar etiology. All of thesevarieties undoubtedly belong in one or the other of the two great pathologicaldivisions of pulmonary suppuration, i.e., endobronchial and parenchymatoussuppuration. We feel, however, that further study of this subject from the

* Read before the American Surgical Association, May 26, I926.256

EXPERIMENTAL ABSCESS OF THE LUNG

point of view of morbid pathology is less likely to give a full appreciationof the condition than an attempt at experimental reproduction.

The studies reported here concern solely post-operative abscess of thelung. They form part of a general study of post-operative pulmonary com-plications and were undertaken in the hope that further proof might befound for the concept that a large proportion of such complications are dueto embolism from the operative wound. We have 1, 2 for ten years acceptedthe suggestion made in I900 by Mikulicz 3 that embolism might be the causeof certain so-called post-operativepneumonias. He was led to this asser-tion by the occurrence of such com-plications following operations undercocaine amesthesia. It is now gener-ally accepted that, in addition to mas-sive pulmonary embolism, certain ofthe other pulmonary complications, T

such as pleurisy and pneumonia, mavhave a similar etiology. Could weprove that post-operative abscess ofthe lung resulted from the same.mechanism, considerable weight wouldbe added to the explanation of thesemost serious operative sequelae. More-over, such a simplification, by bring-ing many complications within thelimits of a single mechanism wouldforcibly indicate the lines along which FIG. 2.-Segment of vein in glass cannula filledwith salt solution about to be forced into tbethese undesirable sequelae might be jugular vein.avoided, since the blame for the complications would then rest squarely uponsurgical technic and operative skill.

We have been able to assemble from the literature i908 cases of abscessof the lung. Of these 29.6 per cent. are post-operative, and 14.6 per cent.,or approximately one-half, follow the operation of tonsillectomy. This fre-quency following tonsillectomy has resulted in the conception that abscess ofthe lung is a complication peculiar to this type of operation. Investigationdoes not substantiate this. The fact is that tonsillectomy is a very commonoperation and that abscess of the lung is a relatively infrequent post-operativecomplication. It follows the operation of tonsillectomy no more frequentlythan operations in any other sept;c field.

The apparent frequency of abscess of the lung following tonsillectomywas thought to be due to the special liability in this operation to aspirationof infected buccal content. It has long been known that during every generalanaesthetic mouth contents are aspirated into the lung,4''- 6 and in the case oftonsillectomy, with the operative field close to the respiratory orifice, the dan-

17 257

CUTLER AND SCHLUETER

:0

0420

'4

0

..t~~~~~.

4)4

4-4)

0c

0 0

.50 ~~~~~~~~~~~4) 0

~44

004

0 4)

_ £~~~~~~~~~~~~~~~~~~~~~~~~~~a'258


ger seemed too obvious. Moreover, the reports of such complications fromoperators working upon patients in the upright position 7 seemed to furtherjustify this feeling. As a result of such reasoning, considerable investigativework has been done attempting to reproduce a similar pulmonary suppurationin animals,8 9, 10 Many kinds of bacteria, plugs of meat and foreign materialshave been introduced into the lung by insufflation, by the bronchoscope andby various ingenious methods calculated to lead the infected material into thefiner ramifications of the bronchial tree. So far as we can determine, no onehas been able to repro-duce in animals typicalabscess of the lung bysuch methods.

There is, moreover,another side to the prob-lem. In the first place,post-operative pulmonarysuppuration is not pecu-liar to the operation oftonsillectomy. Again thisdire sequela f o I I o w supon tonsillectomy whenthe operation is per-formed under local anes- lthesia."' 12 Further, theclinical history is not thatof an immediate post-o p e r a t i v e pulmonaryupset. In fact, t h e r e FIG. 5.-Rontgenogram of Dog Y 38-Experiment I, five days afterfrequently o c c u r s a embolism. A well-defined abscess cavity is present.period of normal convalescence until, at a period seven to fourteen days post-operative, preceded by or synchronous with pleuritic pains, the symptoms ofpulmonary disease commence and gradually are intensified. In addition thereis the evidence that the endobronchial apparatus is well adapted to a defenseagainst infection. Inspired foreign bodies rarely result in true pulmonarysuppuration, though there may occur endobronchial irritation, infection andsubsequent dilatation of that part of the air passages lodging the foreignbody.'3 As further evidence of the great defensive mechanism within thebronchi, may we cite the following case:

E. T., forty-three, complained of dysphagia. The diagnosis of carcinoma of theoesophagus was made by direct visualization December 4, I924. He was treated by theimplantation of radium seeds. At the patient's request, gastrostomy was performedSeptember 30, I925. March i8, 1926, the patient commenced regurgitating by mouthfood given via the stomach catheter. This was accompanied by paroxysms of coughing.This condition continued until his death April 29, 1926. Autopsy showed complete stenosisof the cesophagus and a fistula betwee4 the cesophagus just below this point and' the left

259


primary bronchus; the left lung was clear; there was pneumonia in the right lowerand middle lobes.

It seems established in this case that for at least five weeks the patienthad a communication between his stomach and his left lung, and yet the leftlung was unaffected.

We had long felt that evidence of this nature was sufficient to justifygrave suspicion of the conception that aspiration was the cause of pulmonarysuppuration. Knowing of the failure of the experimental work in which

attempts have been madeto reproduce abscess ofthe lung by the instilla-tion of infected materialsinto the air passages, itoccurred to us that post-operative abscess of thelung might well be of)emolic origin as are

many other post-opera-tive complications.

Before attempting toproduce abscess of thelung by the use of in-fected emboli and inorder to free ourselvesof criticism by those whostill hold to the theorythat aspiration gives rise

FIG. 6.-Rontgenogram of Dog Y 38-Experiment I, eight days to this condition, we per-after embolism. The abscess in the left lower lobe is distinct. Thelead fragment lies in a clear central zone. formed a cons i d e r able

number of experiments in which we placed infected materials with a broncho-scope into the finer ramifications of the bronchial tree. We used pieces ofinfected meat, infected tonsil, peanuts, etc. In no one of the seventeenattempts could we reproduce typical abscess of the lung. This was notsurprising to us because other investigators had failed previously in simi-lar attempts.

A Method for the Production of Abscess of the Lung.14-We then at-tempted the production of abscess of the lung by the instillation of septicemboli. Dogs were used in these experiments, and once the procedure becamestandardized we were able to produce abscesses in IOO per cent. of our ani-mals. During the early experiments we iniserted into the jugular vein of theanimals pieces of infected tonsil, infected meat, etc., and though we achievedsome abscesses, it was frequently the result that an extensive pneumonitis ofthe lobe in which the embolism lodged was produced. This rapidly brokedown and often perforated into the pleural cavity resulting in death. We

260


felt that, though it was necessary to have the correct number and type oforganisms present, it was equally necessary that some sort of local immunitybe produced in the pul-monary tissue before themajority of the organ-isms were set free. Itseemed to us simpler toset up a temporary arti-ficial barrier about theinfected embolus than tor a i s e beforehand thedefense reaction in thepulmonary field. T h eprocedure, which finallybecame standardized asthe most satisfactory forthe production of abscessof the lung, consisted inconstructing an embolusof a segment of vein FIG. 7.-Left lower lobe removed from Dog Y 38-Experiment I,sixteen days after embolism. A small abscess is still present.(Fig. i ) which was filledwith a culture of the desired organisms. We excised a small segment of thefemoral vein, ligated one end, filled this capsule with the organisms and added

a piece of lead filingcoated with paraffin torender it inert.. This bitof metal enabled us torecognize in immediaterontgenograms the si tewhere the embolus lodged.After tying off the otllerend of this small capsuleit was set free in the

~li~ jugular or femoral vein(Fig. 2). In over sixtvper cent. of our experi-ments the embolus lodgedin the left lower lobe,due, we believe, to the

w s i ; - straighter course of thevessel and the greater vol-ume of blood going to this

FIG. 8.-Lungs removed at necropsy from Dog I (normal control) lobe This course wasthirty-six hours after the intravenous injection of a clot infectedwith virulent B. coli organisms. The left lower lobe is quite taken, therefore, for thedeinqplv t-innRlitiAtp2. Xuen:seiy c;omsoituat2u.

261


same reason that directs the lodgement of emboli in human cases, where themajority go to the right lower lobe.

The following experiment is typical of the series, and serves to illustratethe simplicity of the procedure:

Protocol.-Experiment I.-Dog Y 38, weight I2.I kg., November II, 1925 was givenmorphin, gr. Y4. Under ether anaesthesia, a segment was removed from the right femoral

vein, inoculated with cul-tures of staphylococcusaureus, B. coli and pneumo-coccus, type II, and intro-duced into the left jugularvein. A r6ntgenogramlocalized the foreign bodyin the left lower lobe, Fig. 3.

November I3, I925.-According to the r6ntgeno-gram there was beginninginfiltration about the for-eign body, Fig. 4. Theanimal ate his food well.

November i6, 1925.-Arontgenogram of the chestshowed a definite abscesscavity in the left lower lobew here a clear zone contain-ing the bit of lead appearedin the centre of the infil-trated area, Fig. 5. The

- _ animal ate only a sm a ll

FIG. 9.-R6ntgenogram of Dog II, fifty-four hours after the intra- amu of oodvenous injection of a clot infected with virulent B. coli organisms. November I9, 1925.-There is a beginning consolidation of the right lower lobe. The abscess cavit in the

left lower lobe was larger. The surrounding infiltration of the lung remained aboutthe same, Fig. 6.

November 27, I925.-Subsequent r6ntgenograms of the chest had shown a subsidenceof the infiltration and the abscess cavity had decreased considerably in size. The leftlower lobe was removed by operation. In the removed lobe an area of induration couldbe felt near the tip of the lobe. The overlying visceral pleura was considerably thickened.On section, thick indurated tissue was encountered within which a small abscess cavitywas still present, Fig. 7. Microscopically, the lining of the cavity was composed chieflyof fibrous tissue. Considerable cell exudation was still present.

Discussion.-By this method of freeing infected emboli into the venouscircuit, we have been able to produce true abscess of the lung. Such anabscess starts as a parenchymatous lesion. Its establishment depends un-doubtedly upon many factors. Under the circumstances imposed by us inour experiments, the type of organism present and the physical property ofthe covering of the embolus seem to be of some importance. We found thatfreeing a simple infected and uncovered clot into the venous circuit usuallyresulted in a general pneumonitis rather than a walled-off abscess. It wouldseem as if the covering of the venous segment permitted the establishment

262


of a walling-off process before the full effect of all the organisms in the cap-sule was produced. Or the influence of the venous segment as an organicforeign body may be of impoftance. Possibly this means that the actualphysical make-up of any single infected embolus may play a dominant r6leas to whether abscess of the lung is or is not to result. It is conceivable thatclots of variable structure and infectivity may be set free from the operativefield. Those clots, in which the majority of the organisms are centrallylocated, and in which theclot has an unusuallvt o u g h peripheral coat,should be more likely tocause an abscess.

It would seem, how-ever that such an ex-

planation could apply toonly a limited number ofactual cases. This neces-sitated further studv con-cerning the mechanismby which l o c al tissueimmunity might be pro-duced. It was apparentthat there must be pres-ent some factor whichtended to keep the proc-ess well localized. Unless

a clt ws scontruted FIG. Io.-Rontgenogram of Dog II, seventy-two hours after theaclotwas so constructed intravenous injection of a clot infected with virulent B. coli organisms.

that it had' a tough outer The right lower lobe is densely consolidated.coat, that factor must deal with local tissue immunity. We felt that, if wecould raise the local immunity of an animal by vaccination with an organ-ism to be used later in an infected thrombus, we might well establish a highlocal immunity which would restrain spreading of organisms from the localfield, bring about an intense local resistance, and thus create an abscess.Moreover, as post-operative abscess of the lung occurs in cases already infectedand thus already immunized, it would appear that such experiments wouldmore nearly reduplicate what actually occurs in human cases. The followingprotocols of experimental and control animals demonstrate the proof ofthese contentions.

Experimental Stud(ies of the RO'le of Local I unmtitity in the Produtctiont ofAbscess of the Lung.-Four animals were comprised in each experiment.One animal was used as a normal control, one was immunized by injecting anavirulent strain of B. coli intravenously, one animal had small sterile emboli(starch granules) set free in the jugular vein in an.attempt to see if non-infectious material would also create high local resistance, and the fourth

263


animal was submitted to an abdominal operation (lateral intestinal anasto-mosis) in an effort to see whether such a procedure would bring about adefinite general immune reaction sufficient to create local pulmonary resist-ance and thus wall off the subsequent pulmonary infection. We felt thatwe covered the major possibilities for the establishment of a high local resist-ance in the lung with the normal control, the pre-operative aseptic emboliccontrol and the simple operative control animals. When these animals had

1) e e n sufficiently pre-pared thev were sub-mitted to embolism.

The embolus was

plrepared as follows: Avirulent strain t of B.coli was obtained, grownon an agar slant, andthe culture washed offwith salt solution andthoroughly s h a k e n;blood was drawn underaseptic precautions, the

bid E -| suspension of B. coli atonce added and the tulbethoroughly shaken to en-sure equal mixing befor2clotting. The whole was

then allowed to clot ande i g h t e e n hours later

FIG. II.-Lungs removed at necropsy from Dog II, seventy-two t ohours after the intravenous injection of a clot infected with virulent B. Lhe clot was carefullycoli organisms. The right lower lobe is densely consolidated. divided into four equal

parts and the clot slipped into the jugular vein of all four animals. Eachfragment of clot was roughly IO x 4 mm. in size.

The experimental work dealing with this phase of the question is notsufficiently complete to permit us at thi.s tim-e to give final reports and fullproof of our ideas regarding the role of local tissue immunity and the methodof its production. We have, however, sufficient experimental evidence tostrongly indicate that the establishment of such a local immunity plays adominant part in the production of abscess of the lung using the methoddescribed above.

Before reporting such experiments may we repeat the following facts:(1 ) simple infected clots set free in the. jugular vein of dogs usually give riseto a diffuse pneumonitis which will result in fatality or recovery, according

t Secured from Dr. B. Steinberg of the Department of Pathology, Western ReserveUniversity. This organism was of sufficient virulence to kill a dog in six hours by theintraperitoneal injection of one washed agar slant mixed with gum tragacanth.

264


to the virulence of the organism; (2) an infected embolus enclosed in a cap-sule (segment of vein) will result in an abscess of the lung.

Protocol-Experiment A.-Dog I (normal control), weight I6.9 kg., May I5, 1926,was given morphin, gr. Y4. Under novocain anaesthesia the left jugular vein was exposed,opened and the eighteen hour clot infected with the virulent B. coli organism wasintroduced.

May i6, I926.-Twenty-four hours after the injection of the clot the dog appearedill and would not eat. The respiratory rate was so rapid that an X-ray plate of the chestcould not be taken. It wasnoticed that at times the dogwould cough and bring upa frothy, bloody sputum.The rectal temperature was41.3' C.

May 17, 1926.-The dogdied thirty-six hours afterthe injection of the infected.clot.

Necropsy.-There. w a sconsiderable cloudy hemor-rhagic fluid in the left pleuralcavity. The lungs showedsome congestion at the rightlower base; otherwise thelobes on the right side wereair-containing and apparentlynormal. The left upper lobeshowed some congestion. Theleft lower lobe was quitedensely consolidated, Fig. 8.On section the cut surface ofthis lobe was of a deep red-dish color (red hepatizationstage of pneumonia).

Dog II (bacterial immu-nized control), weight 8 kg.,May Ij, I926, was given FIG. 12.-Lungs removed at necropsy from Dog III, eighteenMay15,1926, was given hours after the intravenous injection of a clot infected with virulent Bmorphin, gr. Y/4. An attempt coli organisms. The right lower lobe is consolidated. The right upper

hadbeenmade to immunize and middle lobes and the left lower lobe show some congestion.had been made to immunizethis animal by injecting intravenously a platinum loopful of avirulent B. coli organismssuspended in IO c.c. of salt solution. Three injections at intervals of two days had beencarried out ten days previously. The left jugular vein was exposed under novocain anas-thesia, opened, and the eighteen hour clot infected with the virulent B. coli organismwas introduced.

May i6, I926.-The dog ate most of his food and did not appear ill although his tem-perature was 39.70 C. A r6ntgenogram of the chest showed normal lung fields.

May I7, 1926.-The dog ate his food. The rectal temperature was 39.9g C. A r6ntgeno-gram of the chest was taken in the morning. The lung fields appeared quite normal.Another r6ntgenogram of the chest was taken in the late afternoon and showed a begin-ning area of cloudiness in the right lower lobe, Fig. 9.

May i8, 1926.-When seen about 9 A.M. the dog was in extremis. The respiratory265


rate was very rapid and the rectal temperature was 400 C. A r6ntgenogram of the chestjust preceding death showed a dense cloudiness of the entire right lower lung, Fig. IO.

Necropsy.-There was a slight amount of purulent hemorrhagic fluid in the rightpleural cavity. The left lung appeared normal. The right upper and middle lobes showedsome congestion but were air-containing. The right lower lobe was densely consolidatedand cut with resistance, Fig. ii. The cut surface of this lobe presented a deep reddishcolor. There was no localized abscess present. The animal had apparently establishedsome degree of immunity which was, however, not sufficient to prevent the diffuse lowerlobe infection. It is interesting to note, however, that this animal, a smaller animal than

the normal control, was ableto withstand the infectionfor a greater period of time.

Dog III (pre-operativeaseptic e m b o I i c control),weight io kg., AMay 15, 1926.was given morphin, gr. .This animal had had threeintravenous inj ections at in-tervals of two days of IO c.c.of starch solution ten. dayspreviouslv. Under novocainanlesthesia the left jugularvein was exposed, opened,and the eighteen hour clotinfected with the virulent B.coli organism was introduced.

May i6, i926.-The ani-mal died about eighteenhours following the injection.

Nec-opsy.-There was aA _ | slight a m o u n t of cloudy

hemorrhagic fluid in the rightFIG. 13.-Rontgenogram of Dog IV, forty-eight hours after the P 1 e u r a 1 cavity. The left

intravenous injection of a clot infected with virulent B. coli organisms. lower lobe showed some con-The left lower lobe shows a diffuse cloudiness. gestion but otlierwise the leftlung was air-containing. The right upper and middle lobes showed a moderate degree ofcongestion. The right lower lobe was densely consolidated and there was a slight amountof exudate on the visceral pleura, Fig. I2. On cut section the lobe showed numerousgrayish patches, but a reddish color was predominant.

Dog IV (simple operative control), weight I6.9 kg., May I5, 1926, was givenmorphin, gr. ¼4. This animal had been submitted to an abdominal operation (lateralintestinal anastomosis) two days previously.. Under novocain anesthesia the left jugularvein was exposed, opened, and the eighteen hour clot infected with the virulent B. coliorganism was introduced.

May 16, 1926.-The dog appeared somewhat ill but ate part of his food. The rectaltemperature was 40.5° C. A r6ntgenogram of the chest showed an indefinite area ofcloudiness in the left lower lobe.

May 17, I926.-The animal ate a part of his food. The respiratory rate was definitelyincreased and it was noticed that the dog coughed occasionally. The rectal temperaturewas 40.7° C. A rontgenogram of the chest showed a definite area of mottled cloudinessin the left lower lobe, Fig. 13. A r6ntgenogram taken late in the afternoon showed aneven greater area of cloudiness in the left lower lobe.

May I8, I926.-The animal continued ill but ate some food. The rectal temperature266


(DnQ-o

0

pWo

00

Ocn0P

=0

00m'

IQo

09''

-o<

_-N"

o oa

OQ

- , 'r.A2.;,.s

CD( F-o,,,.

0 C+z .

, * o

0 rA

$CDUc+o

o.o oq

267


was 38.90 C. A r6ntgenogram of the chest showed an even greater area of consolidationof the left lung, Fig. I4.

May I9, I926.-The animal's general condition seemed improved. The rectal tem-perature was 400 C. A r6ntgenogram of the chest showed the area of cloudiness in theleft lower lung to be less dense but there were numerous areas of lessened density sug-gestive of abscess formation, Fig. I5. It was decided to remove the left lower lobeby operation. At operation numerous friable adhesions plastered the left lower lobe tothe lateral chest wall and to the adjoining upper lobe. These were easily broken downand the lobe was removed. The removed lobe was consolidated throughout except for

a small area near the tipw h i c h was air-containing.The lobe was of a dark redcolor and presented numer-ous areas of grayish exudateon the visceral pleura beneathwhich areas of softeningcould be felt. Near the hilusof the lobe there were sev-eral areas of broken downlung tissue, Fig. i6. Thelobe was sectioned and sev-

eral small abscesses wereencountered containing thickwhitish pus. Near the hilusof the lobe a fairly largeabscess cavity was encoun-tered approximately i cm. by3 cm., Fig. 17.

Discussion.-T heabove experiment demon-strates f a c t s that havebeen corroborated repeat-edly in this laboratory.

FIG. I6.-Left lower lobe removed at operation from Dog IV,ninety-six hours after the intravenous injection of a clot infected with In the first place, viru-virulent B. coli organisms. The lobe is quite solid and numerous let infeted 1 o tareas of exudate present on the visceral pleura beneath which are ent n ecte c o s seareas of softening. areas of softening. free in the venous circuitresult in fatal pneumonitis. Histological examination shows a process identi-cal with true lobar pneumonia. In the second place, immunization with aviru-lent organisms or by previously operating upon the animal in a field whereorganisms are already present (the intestines) yields a varying degree ofprotection. Both Dog II (bacterial immunized control) and Dog IV (opera-tive control) outlived Dog I (the normal control animal). Aseptic emboliseem to instigate insufficient immunity to affect the outcome (Dog III). Inthis particular experiment, the normal control animal (Dog I) survived eigh-teen hours, the bacterial immunized control animal (Dog II) survivedseventy-two hours and the operative control animal (Dog IV) was apparentlyrecovering when lobectomy was performed. Abscess of the lung resulted onlyin Dog IV. We feel that this indicates that abscess results when immunity

268


has been sufficiently stimulated. The fact that Dog IV outlived Dog IIindicates that immunity reactions were more highly stimulated.

These experiments will be elaborated upon and repeated, but even thesingle group reported would seem to indicate that in the establishment of post-operative abscess of the lung the two factors of embolismn front the w'ouiidand the local imimune reactions in the lung play dominant r6les. The typeof organism present, the physical character of the clot, immunitv establishedby previous infectionwith similar organisms,the presence of organicmatter other than simpleclot in the embolus, andthe number and virulence /of both pathogenic andsaprophytic o rga n i s m swithin the bronchial pas-sages themselves mayplay a variable part.

Why such lesions indogs tend to heal withintwo weeks unless theykill in the first few daysseems to be explained(i) by the horizontal FIG. 17.-Appearance of the lobe removed from Dog IV, after sec-

plane of the dog's bron- tioning. One fairly large abscess can be seen near the hilus.chial tree enabling freer drainage than in man and (2) by the fact thatanimals have, because of more frequent exPosure, a far higher resistanceto all infection.

REFERENCES'Cutler, E. C., and Hunt, A. M.: Post-operative Pulmonary Complications. Arch. Int.

Med., vol. xxix, p. 449, 1922.2 Cutler, E. C., and Hunt, A. M.: Post-operative Pulmonary Complications. Arch. Surg.,

vol. i, p. I14, 1920.9Mikulicz, J.: Die Methoden der Schmerzbetaubung nod ihre gegenseitige Abgrenzung.

Verhandl. d. deutsche. Gesellsch. f. Chir., vol. xxx, p. 560, I90I.'Hoelscher, R.: Experimentelle Untersuchungen uber die Entstehung der Erkrankungen

der Luftwege nach A_:thernarkose. Arch. f. klin. Chir., vol. Ivii, p. I75, I898.Kelly, R. E.: AnTesthesia by the Intratracheal Insufflation of Ether. Brit. M. J., vol. ii,

p. II2, July 20, 1912; The Intratracheal Insufflation of Ether. Brit. M. J., vol. ii,p. 617, September 14, 1912.

6 Lemon, W. S.: Aspiration. Experimental Study. Arch. Surg., vol. xii, p. I87, 1926.'Manges, Ml.: The Occurrence of Abscess of the Lung after Tonsillectomy. Am. J. Surg.,

vol. xxx, p. 78, I9 I 6.8Aschner, P. W.: The Pathology of Lung Suppuration. ANNALS OF SURGERY, VOl. lXXV,

P. 321, I922.

269


Lambert, A. V. S., and Miller, J. A.: Abscess of the Lung. Arch. Surg., vol. viii, p. 446,I924; Miller, J. A., and Lambert, A. V. S.: The Treatment of Abscess of the Lung.Am. J. M. Sc., vol. clxxi, p. 8i, I926.

Schlueter, S. A., and Weidlein, I. F.: Post-operative Lung Abscess. An ExperimentalStudy. Arch. Surg., in press.

Simpson, J. R., and Noah, H. G.: Report of Two Cases of Lung Abscess Following,Tonsillectomy under Local Anaesthesia in Tubercular Subjects. Penn. M. J., vol.xxiii, p. 332, I920.

Porter, W. B.: Pulmonary Abscess Following Tonsillectomy under Local Anaesthesia.Virginia M. Month., vol. xlvii, p. 6o6, I92I.

"3Jackson, C.: Suppurative Diseases of the Lung Due to Inspirated Foreign BodyContrasted with Those of Other Etiology. Surg., Gynec. and Obst., vol. xlii,P. 305, I926.

"Holman, E., Weidlein, I. F., and Schlueter, S. A.: A Method for the ExperimentalProduction of Lung Abscess. Proc. Soc. Exper. Biol. and Med., vol. xxiii, p. 266.I926.

270

divisions of pulmonary suppuration, i.e., endobronchial and

Documents