U.S. Army Medical Department, Office of Medical History
Skip Navigation, go to content







AMEDD MEDAL OF HONOR RECIPIENTS External Link, Opens in New Window






Chapter VII







During the existence of the Army Neurosurgical Laboratory the original plan of investigation was developed and extended. In some lines of activity progress was achieved, while in other phases no definite advances were made. The work of the laboratory is summarized in this chapter.


One of the most frequent of the sequelae of meningitis (particularly of the chronic type) is the development of an internal hydrocephalus. The experimental production of the condition was found to be possible in both adult cats and kittens. Previous investigators had been able to reproduce the condition by blocking the intraventricular passages of the cerebrospinal fluid; in this laboratory the experimental obstruction to the normal escape of the fluid was caused by a lesion in the meninges and not by blockage in the ventricular system. The method used for this purpose was the injection into the subarachnoid space of a 5 to 10 percent solution of lampblack in Ringer's solution. The carbon particles thus introduced either mechanically blocked the meningeal passages of the cerebrospinal fluid or occasioned a sterile meningitis which, in turn, accomplished a similar obstruction. 2

Such subarachnoid injections of suitable amounts of lampblack in adult animals caused an almost immediate lethargy and sleepiness; the more acute cases remained in this condition for several days. In the milder cases the animals were a little more quiet than normal, with the progressive development of a lethargy. The lesion at autopsy in such adult cats was a typical hydrocephalic enlargement of the lateral ventricles.

The younger animals (kittens), in which the ossification of the skull was not yet complete, showed even more striking abnormalities after such injections of lampblack. Not only did the ventricles enlarge enormously at the expense of cerebral cortex, but the whole head became bulging and relatively enormous. The fontanelles opened widely; in some animals in which the cranial bones were already united by bony union, the fontanelles were reformed. The clinical and pathological pictures were typically those of an infant with an internal hydrocephalus.

As soon as it was found possible to reproduce this pathological condition invariably, therapeutic measures aiming at its relief were undertaken. A certain degree of amelioration followed the creation of an artificial connection between the subarachnoid space and the superior sagittal sinus.

a Report of Investigations conducted at the Army Neurosurgical Laboratory, Baltimore, Md. By Capt. Lewis H.Weed, M. C.


Difficulty in obtaining cerebrospinal fluid by lumbar puncture in laboratory animals led to the adoption of puncture through the occipitoatlantoid ligament as a convenient method (Dixon and Halliburton). The technique for performing this puncture proved to be quite simple, and with a little practice the needle may be inserted into the fluid reservoir, lying just beneath the ligament, without injury to the structures of' the medulla. The use of this operation greatly facilitates the obtaining of spinal fluid from the cadaver. It should prove of value in man when there is a block in the spinal subarachnoid space due to meningitis adhesions; it should be serviceable likewise in irrigations of the spinal subarachnoid space for the removal of the pus and débris due to infection. With the increase in experience in the use of this puncture the route may prove of value for the introduction of serum in early cases of epidemic meningitis. It was found that a much better spread in the basilar cisterns and over the convexity of the brain was obtained when a suspension of india ink was injected by the occipitoatlantoid route than when it was introduced in the lumbar region.

Research in the Army Neurosurgical Laboratory demonstrated that the intravenous injection of hypertonic solutions exerted a marked effect on the pressure of the cerebrospinal fiuid. Hypotonic solutions, when injected intravenously into an etherized animal, caused an enduring rise in the pressure of the fluid as determined in a manometer connected with the subarachnoid space. On the other hand, the intravenous injection of a strongly hypertonic solution brought about an initial rise in the pressure of the cerebrospinal fluid followed quickly by a marked fall, often to negative values. The employment of a solution isotonic with the blood occasioned no lasting change in the pressure of the cerebrospinal fluid. The logical explanation of the alterations in the pressure of the cerebrospinal fluid following the intravenous injection of hypotonic and hypertonic solutions, seems naturally to be related to the experimental change in the osmotic value of the blood. Data regarding the absolute or relative osmotic values of the body fluids were not obtained, but it appeared that the alteration in the salt-content of the blood produced experimentally could only be compensated by fluid readjustments within the tissues.


Investigations on cats showed that the intravenous injection of a strongly hypertonic solution (30 percent NaCl or saturated NaHCO3) was followed by a marked decrease in the size of the brain. This change in cerebral volume occurred in the unopened skull; but if the skull was trephined and the dura opened, the brain after such injection could be seen to fall away several millimeters from the inner surface of the skull. Intravenous injection of a hypotonic solution (water) caused a marked swelling of the brain,. Such increase in brain bulk was noted in the intact cranium; if the skull was opened, tense herniae of the cerebral substance, protruding several millimeters, invariably

resulted from the experimental procedure. Animals subjected to these experimental procedures promptly became normal on recovery from the anesthetic.

These changes in brain bulk were independent of the volume of the fluid injected, as was demonstrated by control injections of Ringer's solution, which did not alter the size of the brain. The age of the animal was found to playa noticeable part in this phenomenon, the brains of old cats failing to respond readily to such intravenous injections, especially of hypotonic solutions. No histological changes were demonstrated in the brains of animals subjected to experimental alteration of brain bulk but with opened skulls (removal of restrictions to change in volume). In those animals which were not trephined, internal changes, recognizable microscopically, were found quite constantly. The clinical application of this phenomenon of volume-change of the brain should be of value in cases of increased intracranial tension, cerebral herniation, cerebral edema in acute infections or injuries.

The methods by which brain injuries and local infections were produced in experimental animals varied much in detail, but the investigation was primarily related to the study of the general principles underlying the spread of infection and the possibility of the experimental control of the process. In every instance the dura was perforated; in some of the experiments, parts of the cortex were removed; in others, perforations communicating with the lateral ventricles were made, while in others only the most superficial layers of the brain were injured. Foreign bodies of bone or metal were sometimes introduced into the wound; while at other times the skull was fractured, with puncture of the central nervous system. In a word, attempts were directed toward reproduction of cerebral wounds similar to those occurring at the battlefront. It was early discovered that an abscess in the brain of a cat developed only with great uncertainty unless a massive dose of bacteria was introduced into the wound. Simple puncture wounds were found to heal per primam although no efforts were made to sterilize the instruments or tract. The animal's resistance to infection even after puncture of the dura seemed quite remarkable. Gross examination of such infected brains showed principally great swelling, with necrosis of the tissue and protrusions through the wound opening. In about one-half of the cases recorded a generalized meningitis was present at death, either from direct or indirect infection of the lateral ventricles in the region of the wound and subsequent spread to the meninges through the foramina of Magendie and Luschka. A complete restoration of the destroyed nervous tissue was never attained, although function might have been taken over by other cells; and in some instances a normal individual, to all intents and purposes, was preserved. An attempt was made by the tissue to combat the process of destruction instituted by the infecting organism; this defense. as well as the removal of the debris by large phagocytic cells, could be seen microscopically. The tendency of the infective process to invade the subarachnoid space from the point of injury was not marked, but in a third of the animals the infection entered the subdural space, forming there a subdural abscess.

These acute traumatic lesions were very different from the more slowly growing abscesses, extending from the cranial air sinuses. The latter, occurring frequently in man, may be differentiated by the relatively slight swelling and dislocation of the cerebral substance, and by the development of a definite connective tissue capsule between the infective focus and the sound parenchyma. The traumatic abscess in the experimental animal extended rapidly along the fiber tracts. No encapsulation could be demonstrated in any of the observations. Healing took place by the ingrowth of connective tissue.

The toxicity of certain antiseptics upon the central nervous system was tested by direct injections into, or irrigation of, the subarachnoid space. Practically all of the chemical bodies employed possessed definite intraspinous toxicity, so that, unless given in suitable dilution and amount, immediate death of the animal ensued. With chlorainine and flavine on subarachnoid injection, in addition to the initial toxicity, death in five to ten days was brought about in consequence of the direct injury to the central nervous system. With injection of small amounts of a suitable dilution the animals remained apparently normal but all showed at autopsy pathological changes in the meninges. The lesions consisted of a more or less complete obliteration of the meningeal (subdural and subarachnoid) spaces with serofibrinous exudate; in the more severe cases the nervous system became involved in a process of destruction by direct continuity from the meninges. The blocking-off of the subarachnoid space by this exudate was complete in one case, as demonstrated by the subsequent injection of india ink; it was not, however, sufficient for the localization of an infection. The subarachnoid injection of lysol and potassium permanganate, in the presence of an otherwise fatal meningeal infection, did not prolong the life of the animal.

In the earlier experiments the irrigation was limited to the spinal canal, and for this operation the first puncture needle was inserted into the subarachnoid space through the occipitoatlantoid ligament, and the second in the lumbar region. Between these needles fluid could be passed through the spinal subarachnoid space in either direction, although the descending route (from cervical to lumbar) was, as a rule, selected. Later, in order to include the cerebral meninges in the irrigation, needles were introduced into the subarachnoid space in the vertex area; from there the flow could be conducted either to an occipitoatlantoid or to a lumbar needle.

Irrigations of the spinal and cerebral subarachnoid spaces were well tolerated by cats if the irrigating fluid was composed of sodium chloride, potassium chloride, and calcium chloride in proper proportions (modified Ringer's solution). If, however, the irrigations were made with isotonic solutions of sodium chloride alone, various toxic effects became very apparent. Many of these animals died during or immediately after the irrigation; if this immediate toxicity was survived, convulsive seizures and acute mania were almost invariable. Recovery from such attacks was frequent. Single irrigation of infected

meningeal spaces with modified Ringer's solution prolonged the life of the animals as compared with the controls. The period of survival in many cases was doubled as a result of this washing out of the infected meninges. Multiple irrigations were not attempted.

The method employed to produce compression of the cord was the injection of paraffin into the epidural space; subsequently the spinal fluid above and below the area of compression was examined. As a result of such epidural compression of the spinal cord a partial transverse myelitis, manifested chiefly as incomplete paraplegia, resulted. In these cats the spinal fluid obtained from below the area of compression usually differed greatly from that above; the former showed almost constantly a greater protein-content, was usually scanty in amount. and frequently of a yellow color, and at times clotted rapidly and completely. The fluid taken from above was uniformly normal. As in man, the fluids which clotted and contained the greatest amount of protein were found in the animals showing greatest symptoms of pressure upon the spinal cord. The protein associated with the mild aseptic meningitis present in these cases was relatively insignificant as demonstrated by the fact that the fluid, obtained from the cisterna magna, often contained white cells equal in number to those from the lumbar region, but showed only slight increase in protein. In some cases a well-marked vascular engorgement of the pial vessels below the area of compression was demonstrated as was also an abnormal amount of serum in the subarachnoid space at the level of compression. Transudation into the lumbar sac was apparently the pathological process operative in the formation of these fluids so rich in protein.


Experiments were undertaken to determine the relative value of various kinds of bone for bridging experimental defects in the skull, the protection of the underlying structures, and the return of the contour of the head to its normal convexity. The problems of cranioplasty could not be met entirely by the experience obtained from osteoplastic work on long bones. The logical material for use in crainioplasty consisted of plates of cranial bones as the requirements of protection for the brain and restoration of the shape of the head could be immediately accomplished and the formation of exostoses was avoided. Animal experimentation (on cats) indicted that either living or dead grafts could he used effectively in the head. In the case of dead grafts, the bone might be removed during routine autopsies, sterilized by boiling or in the autoclave, and kept until needed for the operation. In man living grafts were recommended: but if they were not available, plates of sterilized cranial bone were preferred to any other tissue.


The initial investigation of an outbreak of lethargic encephalitis in Camp Lee, Va., was intrusted to the staff of this laboratory. Nine cases of this disease were examined with complete pathological studies of four. The onset of

symptoms was always insidious--headache, malaise, weakness, and vertigo being commonly complained of. Early symptoms of probably greater significance were sore throat, diplopia, and invariable fever. It was unusual to find signs of organic nerve disease in the first week of illness, hut by the second week--sometimes later still--a widespread organic neurological disorder became evident, when cerebral symptoms appeared. Drowsiness occurred in almost every case, frequently developing into coma, and at times alternating with at state of irritability or anxiety. In spite, however, of an apparently clouded mental condition, orientation and cerebration were usually unaffected until just before death. Long projection fiber tracts to arms and legs showed pro-found disturbance in seven cases as indicated by ataxia, spasticity, and clonus. The only symptoms and signs of a focal character were referable to the brainstem, and these were present in all. Diplopia was complained of in seven of the nine cases, although oculomotor palsy was seldom seen, doubtless because of its transitory nature. The second most frequent disorder was weakness of the facial muscles, usually one-sided, a condition seen in five cases.

Macroscopically, all the brains examined appeared alike. A great degree ofen gorgement of all vessels was conspicuous, and free blood was present in the meninges as evidenced by a pink tint to the pia. The chief seats of the lesions were the brain stem and basal ganglia. The essential pathologic processes were found to be a perivascular exudation and a diffuse infiltration of parenchyma. While both types of lesion varied greatly in intensity, extent, and symmetry, they occurred especially in the gray matter about the canal, fourth ventricle, and aqueduct, though deeper tissues were also affected and the white matter was not spared. The cells concerned in both types of lesion were all mononuclear; a small mononuclear cell and at large mononuclear cell, frequently phagocytic, together with the lymphocytes and plasma cells, were recognized. Polymorphonuclear leucocytes were absent even in the cases of short duration. That the diffuse infiltrating exudate was not necessarily related to a destructive process was borne out by the normal or only slightly changed appearance of nerve cells in its midst: however, when the exudate was excessive, marked nerve cell changes, including neuronophagia, resulted. Hemorrhages were few in nunber and very small. Blood vessel changes were of two types. There was almost constant evidence of proliferation of the intima in vessels in areas of exudation, those in unaffected territory usually showing no abnormality. The second type of lesion noted was infiltration of the vessel walls (especially intra-adventitial), with mononuclear cells, chiefly lymphocytes and plasma cells. The cord and organs in cases examined appeared essentially normal. Lesions in the cerebral cortex were in all cases either nonexistent or negligible. No organisms were seen and cultures from the cerebrospinal fluid and from the nervous system post mortem were negative.

One of the chief problems presented for investigation in the Army neurosurgical laboratory was that of meningitis. In general it was not proposed to deal with the meningococcic infections but rather to ascertain whether the pyogenic, nonmeningococci forms of meningitis could be treated successfully

by measures other than serum therapy. The first need in such an investigation was that of standardization of the infection; i. e., the experimental production of a uniformly fatal meningitis. It was hoped that an organism capable of bringing about meningeal infections in the laboratory animals in numbers analogous to those causing infection in man could be found. Previous work on meningitis in animals had, almost without exception, dealt with the sub-arachnoid injection of very large doses of organisms (one-half to four agar slants). The infections under such circumstances were hardly uniform, even in the monkey.

The search in this laboratory for an organism highly virulent within the meninges of experimental animals was successful. The investigation necessitated the testing of many strains of bacteria within the meninges and the later accentuation of the pathogenicity of those which possessed "natural virulence." Later the major portion of the work became centered about the study of the factors which favored the invasion of the meninges from the blood stream; the production of such hematogenous meningitides was investigated from many angles. The results of the subarachnoid injection of the various organisms will be detailed first and the study of infections of the meninges by organisms circulating within the blood stream will be given later.


Practically all of the experiments dealing with the direct subarachnoid injection of organisms were done on cats, the inoculations being made into the meninges through the occipitoatlantoid and lumbosacral ligaments. Organisms, possessing no natural virulence within the meninges of the cat, were discarded after a few trials. If the organism seemed to possess some pathogenicity, but too large a number, on initial injection, was required to produce a fatal meningitis, the virulence was raised by passage through the meninges of a series of animals.

The reaction caused by the growth of an organism in the meninges was determined by the clinical manifestations of the animal, characteristic changes in the cerebrospinal fluid, and the pathological lesions of the central nervous system at necropsy. Although the clinical manifestations varied somewhat, the more acute cases were characterized by general weakness, convulsions, extensor rigidities of the muscles, accompanied by frequent spontaneous out-cries. These signs of cortical irritation appeared in paroxysms, so that, were the animal observed during a quiescent period, little could be noted unless there actions were elicited by appropriate stimuli. Chronic meningitis exhibited signs of neuromuscular disturbance--ataxia, variations in gait (spasticity),weakness and paralysis of certain muscle groups-which were probably the result of permanent destructive lesions in the central nervous system.

The microorganisms which generally cause a fatal meningitis in man (meningococcus, pneumococcus, streptococcus, staphylococcus, influenza, etc.) were found to possess but slight natural pathogenicity for the meninges of a cat. This animal also proved to be comparatively insusceptible to these organisms on intravenous inoculation, especially if they were of human origin. Another group of organisms (miscellaneous bacilli) was found to be capable

of the production of a fatal meningitis upon the injection of massive doses into the subarachnoid space, but their virulence could not be raised markedly. A third group of organisms was found to possess great virulence within the meninges of the laboratory animals used. Of this group (B. pyocyaneus, B. coli, B. paratyphosus-B., and the mucosus capsulatus group) a strain of B. lactis aerogenes possessed the greatest natural virulence for the meninges. The chief disadvantages of this organism were related to its extreme virulence, the difficulty experienced in producing an immune serum, and its relative infrequency as a meningeal invader in man; these characteristics made it difficult of use in the investigation of certain other phases of the study of experimental meningitis, although for the purpose of this laboratory it was invaluable. This strain of B. lactis aerogenes was virulent within the meninges of all the common laboratory mammals; in cats, the subarachnoid injection of as few as 20 organisms (as determinel by plating) produced a meningitis causing death within 24 hours. The organism was originally obtained at autopsy from the heart's blood and lungs of a man dying of bronchopneumonia.

The routine culturing at necropsy of the heart 's blood of animals dying of experimental meningitis revealed certain interesting facts in regard to the transfer of infection between the meninges and the blood stream. In all cases of meningitis caused by the injection of organisms into the subarachnoid space, cultures of heart 's blood showed the presence of the same bacteria with which the animal had been inoculated intraspinously.

The study of the meningeal reaction produced by subarachnoid inoculation of a large number of organisms, was made largely upon formalin-hardened material. The meningitis resulting from such subarachnoid inoculation could be grouped into three pathological types. The first, a focal subacute meningitis, showed small accumulations of exudate in isolated foci, especially in the deeper layers of the pia. Organisms were usually absent from the meninges and the blood, both culturally and on microscopic examination. The second, an acute, low-grade, exudative meningitis, was characterized by a scanty or a considerable exudate, of polymorphonuclear or lymphocytic cells. Organisms, if present, were but few in number and were considered to possess only mild subarachnoid virulence. The third group consisted of those cases of massive acute meningitis, in which there was evidence of extreme virulence and proliferation of the organism. In this type alone was death considered to be due primarily to the meningitis.

In the cases of acute fulminating meningitis the exudate soon passed beyond the subarachnoid space into the ventricles, and a little more tardily invaded the substance of the brain and spinal cord. In 24 hours approximately one-half of the specimens showed such involvement of the central nervous system; this invasion usually occurred by direct extension from the ventricles and canal. The exudate also spread outward with the nerve roots, and a patchy or diffuse epidural infection then resulted. The dura itself became infiltrated at areas of root "perforation." The blood stream was early infected in such acute

meningitis, as seen on section and on culture, and it is likely that this septicemia played an important role in the death of the animal. In the less acute forms of meningitis, the arachnoid membrane more or less effectively limited the spread of infection from without and from within.


In the course of the study of the processes involved in the localization of an infection within a focus in the nervous system, certain physiological reactions of the cells lining the subarachnoid space were noted. When active or inert particles of matter were injected into the subarachnoid cavity of a living animal, the cells lining the space hypertrophied, lost their normal attachments, and engaged in removing the debris. The importance of such a formation of free cells from fixed elements in any process involving destruction and repair in the meninges (infection, hemorrhage, etc.), seemed great. The reaction of the cellular membrane to such particulate matter was a slow one and appeared to be well under way only after the first 24 hours; dead bacteria might betaken up and removed by the leucocytes before the arachnoid cells showed any signs of activity. The most striking results occurred after subarachnoid injection of laked blood, due probably to the fact that it had no toxic effect on the cells and could be utilized by them.

This reaction of the cells lining the subarachnoid space apparently plays a part in the defensive process against infection. Many of the earliest cases of meningitis (six hours) showed a marked proliferation of these cells: the exudate frequently was largely mononuclear. In other cases roughly half of the cells were mononuclear while the remainder were polymorphonuclear. On section the cell borders of the arachnoidea (particularly the trabeculae) seemed to be high and proliferating; the origin of at least a considerable portion of the mononuclear elements in the exudate seemed certain. The proper control of such a defensive cellular reaction will ultimately accomplish much in the therapy of infectious processes.

In meningeal infections there is always a cellular and protein exudate in the cerebrospinal fluid. With different bacteria and etiological factors, such pathological conditions within the meninges cause ialso variations ill the concentrations of glucose, of sodium chloride, and of urea, and perhaps other changes both in organic and inorganic constituents. The increased cellular and protein constituents of the cerebrospinal fluid must, however, be considered as giving greatest indication of a reaction caused by the injection of a known microorganism into the subarachnoid space.

The number of cellular elements in the normal cerebrospinal fluid of the experimental animals was found to be rather variable; in a series of routine examinations the counts of the white cells ranged from 0 to 10 per c. mm. No erythrocytes were demonstrated in the cerebrospinal fluid from normal cats in a large number of examinations. The protein content varied from 0.1 to 0.5 grams per liter, 0.25 grams being a low normal and 0.5 grams a high. bordering

on a pathological fluid. The colloidal gold reaction of a number of normal fluids gave a similar curve to that obtained by normal fluids from man. In acute meningitis the cerebrospinal fluid of the cat was found to contain from200 to 22,400 white blood cells, 2 to 17 grains of protein to a liter, and to give are action in any of the three zones in the gold sol test. The cerebrospinal fluid. of chronic meningitis in a cat proved to be practically normal, with exception of the gold sol reaction, in which a change to the paretic or luetic zones was recorded. In some of the chronic cases there was a slight increase in the number of white blood corpuscles and in the content of protein. The colloidal gold reaction was of service in demonstrating a pathological cerebrospinal fluid but showed no specific zone reaction, except to a slight extent in the cerebrospinal fluids from animals with chronic meningitis. There was indeed a great similarity in the Lange test as applied to the cerebrospinal fluids from cases of experimental chronic meningitis and those of general paresis in human patients. It was suggested that reaction in the paretic or even the luetic zone denotes the presence of a small amount of protein, as in a chronic lesion of the central nervous system.

In determining the natural virulence of 102 strains of 24 groups of microorganisms, it was found that certain bacteria or groups of bacteria were more capable than others of producing a fatal infection when injected directly into the meningeal spaces. The term "natural” virulence has been used to denote the pathogenicity of recently isolated strains of microorganisms, before increase in virulence by animal passage. It was found necessary to enhance the natural intrameningeal virulence of strains of B. mucosus capsulatus, hemolytic streptococcus, meningococcus and B. paratyphosus, as after being kept for some time on the ordinary laboratory media their virulence so far decreased as to render them of little value for experimental work. Failure to increase the intrameningeal virulence (though the intravenous pathogenicity was raised considerably) by successive intravenous injections of the culture in cats caused the trial of "animal passage " by means of inoculation directly into the meninges. It, was found possible by this method of direct subarachnoid inoculation to increase the virulence of four strains of microorganisms, representing as many groups, to the degree indicated: B. lactis aeroqenes, 0.000,000,000,01 c. c. of a 24-hour broth culture killing in 24 hours (cats); B. paratyphlosus-B., 0.0001 c. c. (cats):hemolytic streptococcus (cats), 0.001 c. c., and (rabbits), 0.0005 e. c.; meningococus (rabbits), 0.001 c. c. By such passage, the intravenous virulence was hardly increased; the intrameningeal and intraperitoneal virulence increased to the same degree. By combined intraperitoneal and intrameningeal methods, approximately the same degree of pathogenicity was developed with streptococcus in rabbits. The intrameningeal virulence became at least 500 times greater than the intravenous. The ratio of the intrameningeal and intravenous pathogenicity of B. lactis aerogenes, of B. paratyphosus-B., of streptococcus on cats and on rabbits, and of meningococcus became, respectively, 10,000,000 to 1: 10,000 to 1; 1,000 to 1; 500 to 1, when the intrameningeal method of animal passage was employed.


The marked virulence of the B. lactis aerogenes within the meninges of the cat led to its use in other laboratory mammals. For all of these (guinea pig, white rat, rabbit, monkey) the same intraspinal pathogenicity held. Later, experiments leading to another end were undertaken. These concerned the intravenous injection of this organism in doses of from 0.5 to 1 c. c. of a 24-hour broth culture. None of these cats developed meningitis, but remained normal throughout the period of observation. In one case, shortly after such experimental injection into the blood stream, cerebrospinal fluid was removed by puncture. The next day the animal showed signs of meningeal irritation, and a second puncture, with withdrawal of cerebrospinal fluid, was made. This fluid was definitely turbid, contained 5,800 white blood cells, and yielded a positive culture of B. lactis aerogenes. In the film preparation from this fluid many bacilli were present. The animal died in 28 hours, and at necropsy atypical exudative meningitis was found.

Experiments on cats were immediately devised to test out the possible relationship of this withdrawal of cerebrospinal fluid during an artificial bacteremia to the later production of a meningitis.18 The series were so arranged that the control animals were given double the intravenous dose of B. lactis aerogenes (usually 0.5 c. c. of a 24-hour meat infusion broth culture). These control cats remained normal and showed no signs of meningeal infection, and were usually killed at the end of a month for histologic control. The other cats in the series were given the unit dose of organisms intravenously (usually 0.25 c. c. of the same culture), and two minutes afterward, cerebrospinal fluid was withdrawn. In the routine experiment, from 1 to 2 c. c. of fluid was permitted to escape, and the animal then allowed to recover from the anesthetic. As contrasted the next (lay with the control cat, which, though receiving double the intravenous dose, was normal and active, the punctured cat would exhibit signs of meningeal infection. Customarily within 24 hours, the typical signs of such meningeal involvement were present; the animal was somewhat hyper-sensitive, ill, and cautious, and moved only on urgent necessity. The tendency toward convulsions became more outspoken as time elapsed, and spontaneous seizures were noted. During such a spasm death often occurred, or the animal went into a coma and died without further signs of meningeal irritation. Pathologically, an acute exudative leptomeningitis was invariably found.

The experimental procedure outlined above was repeated in scores of animals; the production of meningitis by intravenous inoculation followed by release of cerebrospinal fluid was so certain and so regular that it became the chief method for the experimental production of the infection. The work was controlled in many ways, in addition to the routine method of giving one animal in the series an intravenous injection of double the amount of the culture but without puncture.

The release of cerebrospinal fluid was brought about by either lumbar oroccipitoatlantoid puncture. Approximately the same amount of fluid was withdrawn by either method; the end result was identical. Release of fluid by both procedures, with proper dosage of the organism within the blood stream, resulted invariably in infection of the meninges.

That this phenomenon of an acute meningitis following intravenous inoculation with release of cerebrospinal fluid was not peculiar to eats was demonstrated by a series of experiments on rabbits, guinea pigs, white rats, and monkeys. In each of these species, the control was given the same or double the intravenous dose as the animal from which the cerebrospinal fluid was removed at the height of the artificial bacteremia. In every case, the control remained well and normal until killed, while the animals from which spinal fluid was removed promptly developed a typical meningitis with death in 96 hours or less. In the two monkeys at our disposal, the same procedure was carried out; the control (receiving only the intravenous injection) remained normal for seven months, while the monkey receiving the intravenous injection, followed by release of cerebrospinal fluid, died in 54 hours with typical signs of meningitis. The two monkeys were subjected to cistern puncture 48 hours after the initial injection; in the control, the cerebrospinal fluid contained no white cells and the culture was negative, while the fluid of the other had 14,000 white blood cells and gave a positive culture.

The time relations between the withdrawal of cerebrospinal fluid and the intravenous injection of organisms were found to be of importance. In no case did meningitis develop when the cerebrospinal fluid was released 30 or more minutes before the intravenous inoculation. If the puncture were done, however, only a few minutes before the inoculation into the blood stream, infection of the meninges occurred as in other observations in which routine withdrawal of fluid was accomplished immediately after the intravenous injections of organisms.

During the height of a suitable artificial bacteremia, the release of spinal fluid invariably caused a meningitis. In one series of experiments, the punctures were delayed for various periods after the intravenous injection. Animals from which cerebrospinal fluid wats removed within three hours after the intravenous injection developed meningitis. But also animals receiving somewhat larger intravenous doses could not be punctured five hours afterwards without developing infection of the meninges. Hence a striking time relation between the degree of the artificial bacteremia and the withdrawal of cerebrospinal fluid seemed established. It must be assumed that following the intravenous injections of B. lactis aerogenes the number of bacteria in the circulating blood was constantly diminishing so that in practically all cases of simple intravenous injection the blood was sterile in 24 hours. Consequently delay in removing cerebrospinal fluid became comparable to the initial administration of a smaller intravenous dose of organisms. Apparently the number of organisms circulating in the blood stream at the time of puncture is one of the crucial factors in determining the infection of the meninges.

Other observations were made to determine how soon the infection of the meninges occurred after the release of cerebrospinal fluid during the height of the bacteremia with B. lactis aerogenes. In these typical experiments the animals were killed in one hour, two hours, four hours, and six hours after the injection and puncture. The membranes of the central nervous system were then examined in the fresh by means of smears taken from the subarachnoid space, and the findings later controlled by histologic sections. In the one-hour

case, after repeated search one or two large bacilli, morphologically identical with B. lactis aerogenes, were found in the cerebral leptomeninges but not elsewhere within the meninges. The two-hour specimen yielded bacteria only in the cranial portion of the pia-arachnoid, but in considerable numbers. Practically no cellular exudate was found. Many polymorphonuclear and mono-nuclear cells were present in the four and six hour animals; the infection seemed to be largely within the cerebral meninges, but apparently to a lesser degree the spinal portion of the subarachnoid space was involved. These findings indicated that the infection of the meninges occurred almost immediately after the release of cerebrospinal fluid during the bacteremia.

Pathologically, the meningitis produced by this means was comparable to infection of the meninges in man. In a great majority of cases (particularly those under 48 hours' duration) the gross distribution of the exudate was almost entirely cerebral; but in a small percentage of the early cases, the exudate was wholly confined to the spinal meninges. In the more prolonged infections, the involvement of the subarachnoid space was customarily universal. Microscopically the exudate consisted of polymorphonuclear leucocytes, large mono-nuclear cells, a few phagocytes, and the infecting organisms.

It was essential to determine whether the meningitis produced by intravenous inoculation and release of cerebrospinal fluid was the result of infection due to a possible leakage of blood along the track of the needle into the subarachnoid space. The evidence was strongly in favor of the idea that the determining factor in the infection was the reduction of the pressure of the cerebrospinal fluid, with associated vascular changes.

With such results following the intravenous injection of B. lactis aerogenes with release of cerebrospinal fluid, experiments were undertaken to ascertain if the same procedure, but with other organisms, would produce meningitis. The difficulty here was that with the ordinary cultures in a laboratory, the intravenous toxicity was high in comparison to the intraspinal. However, it was possible to repeat these experiments and confirm the finding on the cat with two other organisms, B. pyocyaneus and B. paratyphosus-B. On rabbits a similar result has been obtained with strains of meningococci and of streptococci, procured from an Army camp.

The facilitation of infection of the meninges from. the blood stream by the release of cerebrospinal fluid seemed established as a biological factor by the production, in this laboratory, of a typical meningitis with five different organisms. The conditions for the successful production of such an experimental meningitis concerned two factors, both apparently of determining importance. In the first place, it was demonstrated that the organism used must possess relatively great virulence within the meninges and be capable of multiplication there, even when in small numbers. The strain of B. lactis aerogenes best fulfilled this requirement when introduced into the subarachnoid space of the common laboratory mammals; its virulence in these animals was comparable to that of organisms causing meningitis in man. The other important condition dealt with the number of organisms circulating in the blood stream at the time of release of the cerebrospinal fluid; if this was not great enough, no infection took place.

It must not be assumed from the statements made that a meningitis could not be produced by B. lactis aerogenes after simple intravenous injection. Such a meningitis was caused by the introduction within the blood stream of massive doses of the organism. The amount necessary to produce such a meningitis was many times the customary intravenous injection; such animals were killed usually by the septicemia and not by the meningitis. Many of these animals died from the overwhelming intravenous injection without development, as shown by later necropsy, of any meningeal infection at all.

Practically all of the data indicated that the infection of the meninges by organisms circulating in the blood stream, following removal of cerebrospinal fluid, was due to alteration in the pressure of the cerebrospinal fluid and the associated vascular changes within the cranium. It was assumed that the withdrawal of the spinal fluid was partially compensated by an immediate vascular dilatation, particularly on the venous side. This vascular readjustment really involved a slowing of the blood flow through the cerebral vessels; it was thought possible that this slowing of the flow might facilitate in growth of organisms from the blood stream into the meninges.

Experiments to test this hypothesis were carried out.19 In the first group, the retardation of cerebral blood flow was brought about by digital compression of the jugular veins and adjacent tissues of the neck for two minutes; in the second series, the heart was completely stopped for 30 seconds by excessive administration of ether. Both series were subjected to the necessary controls; every animal was given a suitable intravenous injection of organisms before the secondary facilitating procedure was carried out. In the two series, a fatal meningitis occurred in 50 per cent of the animals; the clinical manifestations of the disease were typical.

These experiments seemed to indicate strongly that the infection of the meninges from the blood was closely associated with cerebral vascular changes. That only one-half of the animals developed a fatal meningitis could well be accounted for by the necessary variability of the experimental procedures employed. It did appear, however, that the removal of the cerebrospinal fluid was more certain than these other procedures as a facilitating factor. At this time, experiments dealing with the effect of intravenous injections of solutions of different concentrations upon the pressure of the cerebrospinalfluid were being conducted in this laboratory. It was found, as recorded in a foregoing paragraph, that the intravenous injection of a strongly hypertonic solution markedly lowered the pressure of the cerebrospinal fluid, often to negative values. Such intravenous injections were immediately combined with suitable intravenous injections of organisms, virulent within the meninges; a fatal meningitis invariably occurred as in the experiments in which the pressure of the cerebrospinal fluid was reduced by withdrawal of fluid. The relation of the low pressure of the cerebrospinal fluid and its associated vascular changes was thereby demonstrated.

Further observations were made to determine the effect of preliminary subarachnoid injections of protein (autologous, homologous, and heterologous)upon the intrameningeal lodgment of organisms circulating within the bloodstream. Flexner and Amoss had earlier demonstrated that poliomyelitis

could be experimentally produced by intravenous injection of the virus, provided that the permeability of the meninges had previously been altered by intraspinous injection of serum, salt solution, etc. Subsequently Austrian applied this same procedure to experimental meningococcic meningitis in rabbits. Austrian recorded the production of a fatal meningitis in three out of twenty rabbits given intravenous injections of meningococci after preliminary intraspinous injections of serum. In this laboratory preliminary injections of serum into the subarachnoid space were given and later suitable intravenous injection of B. lactis aerogenes was made. In 6 out of 39 cats, a typical fatal meningitis was produced; the other animals remained normal in every way. As a facilitating mechanism the preliminary subarachnoid injection of serum has been found to be by no means as effective as the reduction in the pressure of the cerebrospinal fluid and the associated vascular changes.

The interpretation of these many and varied experiments is necessarily related to the mechanism of facilitation of the infection of the meninges from the blood stream. The intravenous injection of suitable dosage of an organism virulent within the meninges did not of itself produce meningitis; such an injection had to be combined with an experimental procedure which facilitated invasion of the subarachnoid space by bacteria. Of these various procedures, withdrawal of cerebrospinal fluid by puncture or reduction of its pressure by intravenous injection of strongly hypertonic solutions was most efficacious. Measures slowing the intracranial blood flow (cerebral venous congestion or stoppage of the heart) caused infection of the meninges in only half of the cases. Preliminary subarachnoid injections of serum resulted in infection of the meninges in but 6 out of 39 experiments.

In this acute hematogenous meningitis, there occurred a more or less wide-spread distribution of purulent matter throughout the subarachnoid space, obliterating all characteristic markings and obscuring contours. With the discoloration due to pus and hemorrhage, there was associated a fairly constant swelling of the nervous system itself, rendering tense the dura of both brain and spinal cord. In a great majority of cases (particularly those under 48 hours' duration) the gross distribution of the exudate was almost entirely cerebral; but in a small percentage of the early cases, the exudate was wholly confined to the spinal meninges. Evidence favored the view that the cortical meninges were the site of earliest infection, with rapid spread to other portions of the subarachnoid space and to the cerebral ventricles. The invasion of the ventricles occurred early in all fatal cases. Infection of the ventricles and canal led to involvement of the substance of the brain and cord. The exudate which accompanied all the meningeal infections was polymorphonuclear or mononuclear in character and its distribution followed closely that of the organisms. In the earlier cases, the exudate was often slight, but in those of longer duration it became massive, imparting to the entire meninges a discolored appearance. Extension outward from the subarachnoid space occurred in many cases and two methods of the accomplishment of this process have been observed. One was through apparent adhesions of arachnoid and dura with infiltration by

leucocytes and swelling at this point. A more common site of extension was in the area where the dura is "pierced" by the nerve roots, the exudate frequently accompanied both the anterior and posterior roots outward for a short distance, invading the dura and sometimes passing through it to its external surface. With the strains of B. lactis aerogenes, meningococcus, and streptococcus employed, there was produced an acute hemorrhagic-purulent meningitis, in which the bacteria appeared to be rapidly increasing in number. With B. pyocyaneus and B. paratyphosus-B., a mild acute meningitis resulted, with very few organisms to be seen. Control animals receiving inoculations alone showed no pathological lesion of the central nervous system or merely a mild "febrile reaction," recognizable only microscopically.

The microorganism, B. lactis aerogenes, which proved so extremely virulent in the central nervous system, exhibited a similar pathogenicity for the eye. The release of fluid from the anterior chamber of the eye, and the congestion of the cerebral circulation during an experimental bacteremia, resulted in the production of a purulent panophthalmia. It is interesting, however, that in the course of several hundred experiments on meningitis in which this bacterium was inoculated into the blood stream of cats, an ophthalmia was produced only once. When procedures analogous to the withdrawal of spinal fluid were carried out on the eye, the infection was limited to the one organ operated upon, the opposite eye and the central nervous system being unaffected. The close correspondence between the anatomical and physiological processes in the eye and central nervous system was emphasized by this production of panophthalmia by intravenous inoculation followed by facilitating measures.

A series of cases was observed in the base hospital at Camp Jackson by members of the staff of this laboratory for the purpose of determining the relationship, if any, of diagnostic lumbar puncture, in the presence of a septicemia, to the subsequent occurrence of meningitis. To this end, blood cultures were taken at the time of the lumbar puncture and the cases were followed subsequently with reference to the development of meningitis. In a number of cases in which a septicemia was present, the first diagnostic lumbar puncture yielded a clear and normal cerebrospinal fluid in which no organisms could be demonstrated; later fluids from these cases were turbid and contained the organism isolated from the blood culture. Two cases of pneumococcus septicemia developed a meningitis, subsequent to diagnostic lumbar punctures yielding normal fluids. In one of meningococcic septicemia, a negative spinal fluid at the time when the blood culture was positive, was obtained; the presence of an early meningitis was observed at autopsy. The case suggested a possible relationship between the release of cerebrospinal fluid during the septicemia and the subsequent meningitis. Three other patients with meningococcic septicemia were observed; lumbar puncture in all three yielded initial negative

fluids. Within 48 hours two of the three had developed a definite meningitis. Interpreted from the standpoint of the experimental work in this laboratory, the relationship of the withdrawal of cerebrospinal fluid during a septicemia to the development of a meningitis was indicated.

To prevent the possible accidental production of a meningitis as the result of diagnostic lumbar puncture, it was recommended that careful consideration be given the bacteriological study of the blood before such punctures were attempted; and that in acute diseases, in the absence of definite signs of irritation of the central nervous system, lumbar puncture should be avoided unless it was first conclusively shown that the blood stream was free of infection.


(1) Weed, Lewis H., Capt., M. C.: The Experimental Production of an Internal Hydrocephalus. Carnegie Institution of Washington (Publication No. 272), 1920, 425.
(2) Wegeforth, P., Ayer, J. B., and Essick, C. R., Captains, M. C.: The Method of Obtaining Cerebrospinal Fluid by Puncture of the Cisterna Magna (Cistern Puncture). The American Journal of the Medical Sciences, Philadelphia, 1919, clvii, No. 6, 789.
(3) Weed, L. H., Capt., M. C., and McKibben, P. S., 1st Lieut., S. C.: The Effect of Intravenous Injections of Various Concentrations upon the Central Nervous System. The Anatomical Record, Philadelphia, 1919, xvi, No. 3, 167.
(4) Weed, Lewis H., Capt., M. C., and McKibben, P. S., 1st Lieut., S. C.: Pressure Changes in the Cerebrospinal Fluid following Intravenous Injection of Solutions of Various Concentrations. The American Journal of Physiology, Baltimore, Md., 1919, xlviii, No. 4, 512.
(5) Weed, Lewis H., Capt., M. C., and McKibben, P. S., 1st Lieut., S. C.: Experimental Alteration of Brain Bulk. The American Journal of Physiology, Baltimore, Md., 1919, xlviii, No. 4, 531.
(6) Essick, C. R., Capt., M. C.: Pathology of Experimental Traumatic Abscess of the Brain. Archives of Neurology and Psychiatry, Chicago, 1919, i, No. 6, 673.
(7) Wegeforth, P., and Essick, C. R., Captains, M. C.: The Effect of Subarachnoid Injections of Antiseptics upon the Central Nervous System. The Journal of Pharmacology and Experimental Therapeutics, Baltimore, Md., 1919, xiii, No. 4, 335.
(8) Weed, L. H., and Wegeforth, P., Capts., M. C.: Experimental Irrigation of the Subarachnoid Space. The Journal of Pharmacology and Experimental Therapeutics, Baltimore, Md., 1919, xiii, No. 4, 317.
(9) Ayer, J. B., Capt., M. C.: Cerebrospinal Fluid in Experimental Compression of the Spinal Cord. Archives of Neurology and Psychiatry, Chicago, 1919, ii, No. 2. 158.
(10) Wegeforth, P., Capt., M. C.: Note on Experimental Cranioplasty. Annals of Surgery, Philadelphia, 1919, lxix, No. 4, 384.
(11) Wegeforth, P., and Ayer, J. B., Capts., M. C.: Encephalitis Lethargica. The Journal of the American Medical Association, Chicago, July 5, 1919, lxxiii, 5.
(12) Felton, L. D., Contract Surgeon, and Wegeforth, P., Capt., M. C.: The Production of Experimental Meningitis by Direct Inoculation into the Subarachnoid Space. Monographs Rockefeller Institute Medical Research, New York, 1920, No. 12, 5.
13) Ayer, J. B., Capt., M. C.: A Pathological Study of Experimental Meningitis from Subarachnoid Inosulation. Monographs Rockefeller Institute Medical Research, New York, 1920, No. 12, 26.
(14) Essick, C. R., Capt., M. C.: The Formation of Macrophages by the Cells Lining the Subarachnoid Cavity in Response to the Stimulus of Particulate Matter. Carnegie Institution of Washington (Publication No. 272), 1920, 377.
(15) Felton, L. D., Contract Surgeon: Analyses of Cerebrospinal Fluid of Cats with Meningeal Infections. Johns Hopkins Hospital Bulletin, Baltimore, Md., 1919, xxx, 242. 
(16) Felton, L. D., Contract Surgeon: The Intraineningeal Virulence of Microorganisms. Monographs Rockefeller Institute Medical Research, New York, 1920, No. 12, 45.
(17) Weed, L. H., Wegeforth, P., Ayer, J. B., Capts., M. C. and Felton, L. D., Contract Surgeon: The Production of Meningitis by Release of Cerebrospinal Fluid during an Experimental Septicemia. The Journal of the American Medical Association, Chicago, 1919, lxxii, No. 3, 190.
(18) Weed, L. H., Wegeforth, P., Ayer, J. B., Capts., M. C. and Felton, L. D., Contract Surgeon: The Influence of Certain Experimental Procedures upon the Production of Experimental Meningitis by Intravenous Inoculation. Monographs Rockefeller Institute Medical Research, New York, 1920, No. 12, 57.
(19) Weed, L. H., Capt., M. C.: Sur l'infection expérimentale des méninges par des germes, contenuts dans le sang circulant. Archives médicales belges, Bruxelles, 1920, lxiii, No. 1, 1.
(20) Ayer, J. B., Capt., 'M. C.: Experimental Acute Hematogenoous Meningitis--A Pathological Study. Monographs Rockefeller Institute Medical Research, New York, 1920, No. 12, 113.
(21) Wegeforth, P., Capt., M. C.: Experimental Production of Panophthalmia by Infection from the Blood Stream. Archives of Ophthalmology, New York, 1919, xlviii, No. 3, 276.
(22) Wegeforth, P., Capt., M. C., and Latham, Jos. R., 1st Lieut., M. C.: Lumbar Puncture as a Factor in the Causation of Meningitis. American Journal of the Medical Sciences, Philadelphia, 1919, clviii, No. 2, 183.