The Contradiction of Doctor Lewis

The system she served was temporary, improvised, incomplete. It was also, for this moment, the only system that functioned. On the morning of October 21, 1918, a nurse at the Philadelphia Naval Hospital drew fluid from the lung of a dying sailor with a glass pipette, sealed it in a sterile tube, and labeled it with a number. The sample would travel six miles across the city to the Henry Phipps Institute, arriving before noon. The journey required no special vehicle, no escort, no urgency in the streets. The city had grown accustomed to movement between the wards and the laboratories, between those who tended the dying and those who sought to name what killed them.

At the Henry Phipps Institute, the sample passed through the hands of a laboratory assistant and came to rest in a wooden rack designed to hold such tubes in vertical rows. The rack sat on a bench beneath a north-facing window, where the light remained steady through the afternoon.

Dr. Paul A. Lewis received the tube with a nod. He was forty-one years old, the director of the institute, a man whose reputation for precision had followed him from the Rockefeller Institute to his appointment at the University of Pennsylvania in 1917. His laboratory occupied the second floor of a brick building at 36th and Spruce Streets, a location chosen for its proximity to the university hospital and its distance from the industrial smoke of the river wards.

The building had been constructed in 1905 with funds from Henry Phipps, the steel magnate, who had specified that the institute should study tuberculosis and the larger problems of preventive medicine. Lewis had accepted the directorship with the understanding that his research would not be confined to any single pathogen.

The tube contained sputum, cloudy and viscous, drawn from a lung that had filled with fluid in the course of forty-eight hours. The sailor had been healthy when he arrived at the Navy Yard in early October. He had developed fever on October 18, cough on October 19, and by the morning of October 21 his breathing had grown so labored that the medical officers had classified him as moribund. The classification was not a diagnosis. It was a preparation. Lewis understood the distinction. His task was to move backward from the preparation to the cause, to find in the cloudy fluid the organism responsible for the cascade of symptoms that had overtaken the city.

He prepared a slide. The method was one he had refined over fifteen years of laboratory work: a thin film of the sample, fixed by heat, stained with gentian violet and then with Gram’s iodine, washed with alcohol, counterstained with safranin. The Gram stain separated bacteria into two categories, positive and negative, based on their retention of the primary dye. Lewis had used this method to identify the pneumococcus in cases of lobar pneumonia, the meningococcus in epidemic cerebrospinal fever, the diphtheria bacillus in throat cultures. Each organism had its characteristic shape and staining pattern, its predictable behavior under the microscope. The method worked because bacteria were consistent, because their forms followed rules that could be learned and applied.

He placed the slide on the stage of his microscope and adjusted the light. The objective was a Zeiss oil-immersion lens, 1/12 inch focal length, capable of resolving structures to 0.2 microns. He had learned to use such instruments during his years at the Rockefeller Institute, where he had worked with the director on poliomyelitis and with Hideyo Noguchi on syphilis. The Rockefeller had been the most advanced research institution in the country, perhaps in the world. Lewis had left it for Philadelphia with some reluctance, persuaded by the opportunity to build a department and by the university’s promise of independence. The Henry Phipps Institute reported to no hospital board, answered to no health department. Its work was pure research, the search for knowledge without immediate application.

What he saw through the eyepiece was familiar and strange. The field was crowded with cells, debris, the detritus of inflammation. Among them moved small rod-shaped organisms, stained red by the safranin counterstain, indicating that they were Gram-negative. Their form was consistent: short bacilli, sometimes occurring in pairs, occasionally in longer chains. The morphology matched his expectations. He had seen such organisms before, in previous epidemics, in the literature that had accumulated since Richard Pfeiffer’s first description in 1892.

The Pfeiffer bacillus, Haemophilus influenzae, had been identified as the causative agent of influenza in the pandemic of 1889-1890. The identification had been disputed, then accepted, then disputed again as subsequent epidemics failed to yield the organism in all cases. But the consensus among bacteriologists in 1918 leaned toward Pfeiffer. The bacillus was difficult to culture, fastidious in its requirements for growth, easily missed in inadequate preparations. Its absence from some cases did not disprove its presence in others. Lewis had accepted this reasoning. It was the reasoning of his profession.

He made a note in his laboratory book. The entry gave the date and the source of the sample. It specified the staining method. It described what the microscope revealed. He did not record his conclusion, not yet. Conclusions required confirmation, and confirmation required additional samples, cultures, animal inoculations. The methodical progression was ingrained in his training. He had learned it from Theobald Smith, from the Rockefeller director, from the German bacteriologists whose textbooks he had studied in translation. Science advanced through the accumulation of evidence, the elimination of alternative explanations, the replication of results. Haste led to error. Error led to false confidence, to therapeutic misdirection, to the discrediting of the scientific enterprise itself.

But haste was upon him. Outside the laboratory windows, the city was drowning. The Bureau of Health had recorded 428 deaths from influenza and pneumonia on October 19, 517 on October 20, 554 on October 21. The mortuaries had exhausted their supply of coffins. The potter’s field at League Island had received more than two hundred bodies in three days, buried in trenches dug by prisoners from the county jail.

The volunteer nurses, the seminarians, the women of the Emergency Aid Society, had constructed a parallel system of care that operated alongside and in spite of the official machinery. Wilmer Krusen, the director of the Bureau of Health, had issued closure orders on October 3, then expanded them, then retreated into administrative silence as the death toll mounted. The city government was not designed for catastrophe. It was designed for the management of routine, the processing of permits, the collection of taxes. The epidemic had exposed the design. Lewis, in his laboratory, worked in a space that remained insulated from the chaos, but the insulation was temporary, psychological, perhaps illusory. The samples arrived with increasing frequency, each one carrying the same urgent question.

He had begun his systematic study on October 15, when the magnitude of the epidemic had become undeniable. His methodology was comprehensive: samples from Navy Yard personnel, from civilian hospital patients, from autopsy material supplied by the city coroner’s office. He cultured the organisms on blood agar, chocolate agar, the specialized media required for fastidious bacteria. He inoculated rabbits, mice, monkeys, observing the course of disease in animals that could not describe their symptoms. He stained and restained, comparing the morphology of organisms from different sources, looking for the consistent agent that would explain the inconsistent clinical picture. The work required twelve to sixteen hours each day. He had suspended his teaching duties, his administrative responsibilities, his participation in the university’s war-related committees. The epidemic was his single occupation.

The Pfeiffer bacillus appeared in the majority of his preparations. This was the finding he reported to his colleagues, first in informal consultation, then in a formal communication to the Philadelphia Medical Journal. The organism was present in the sputum of acute cases, in the blood of fatal cases, in the lung tissue obtained at autopsy. Its absence from some samples did not trouble him. The bacillus was known to be difficult to isolate, easily overgrown by contaminating flora, destroyed by improper handling. The negative cases were technical failures, not evidence against the hypothesis. He had seen this pattern before, in his work on poliomyelitis, where the causative virus could be recovered from only a fraction of cases. The absence of proof was not proof of absence. The logic was sound, the methodology rigorous, the conclusion consistent with the prevailing scientific consensus.

He prepared additional slides from the October 21 sample, varying the staining method to confirm the morphology. The organisms remained consistent: small Gram-negative bacilli, sometimes pleomorphic, occasionally showing bipolar staining that gave them the appearance of safety pins. The description matched Pfeiffer’s original account. The match was satisfying. It suggested that the current epidemic was continuous with previous influenza pandemics, that the accumulated knowledge of three decades could be brought to bear on the present crisis. Therapeutic recommendations could follow: serum therapy, developed for other bacterial infections, might be adapted to target the Pfeiffer bacillus. Public health measures could be refined, directed at the known mode of transmission. The uncertainty that paralyzed the city’s response could be replaced by scientific certainty, by the confidence that came from understanding the enemy.

But the satisfaction was incomplete. Lewis was too experienced to ignore the anomalies that accumulated at the margins of his data. Some cases showed no bacilli, despite careful technique. Some cultures grew other organisms, streptococci and staphylococci, that could also produce pneumonia. The clinical course of the disease was more rapid, more violent, than the textbooks described for influenza. Patients died within hours of symptom onset, their lungs hemorrhaging, their faces turning the dusky cyanosis that signaled terminal hypoxia. The Pfeiffer bacillus was not known to produce such effects. It was a secondary invader, a complication of viral respiratory infection, not a primary killer.

He considered these objections and set them aside. The objections were theoretical, based on prior assumptions that the current epidemic might violate. Science progressed through the revision of assumptions, not their rigid application. The evidence of his microscope was concrete, immediate, subject to verification by others. He photographed his preparations, preserving the visual record for publication. He prepared cultures for shipment to the Hygienic Laboratory in Washington, where colleagues could replicate his findings. The scientific method required this openness, this willingness to submit private observation to public scrutiny. He would not be accused of haste, of premature conclusion, of the kind of speculation that had discredited previous influenza research.

On October 22, he received a visit from two medical officers from the Navy Yard. They came in uniform, their faces showing the fatigue that Lewis had learned to recognize in all who worked through the epidemic. They brought additional samples, drawn from sailors who had died in the previous twenty-four hours, and they brought questions. The Navy’s medical service was overwhelmed, its facilities inadequate to the demand. They needed guidance on treatment, on isolation procedures, on the expected duration of the outbreak. Lewis received them in his office, a small room adjacent to the laboratory, furnished with a desk, two chairs, a bookshelf containing the standard texts of bacteriology and pathology.

He explained his findings. The Pfeiffer bacillus, he said, was the causative agent. The disease was influenza, modified in severity by circumstances that remained to be determined—virulence of the strain, susceptibility of the population, environmental factors. The therapeutic implication was clear: treatment should target the bacillus, through serum therapy where available, through supportive care where serum was lacking. Isolation of cases was essential to prevent transmission. The duration of the outbreak could not be predicted with precision, but historical precedent suggested a curve of six to eight weeks, with decline following the exhaustion of susceptible individuals.

The officers took notes. They asked about the negative cases, the patients from whom no bacilli could be isolated. Lewis acknowledged these exceptions and minimized their significance. The bacillus was difficult to culture, he repeated. Technical failures were inevitable in epidemic conditions. The positive findings were the significant ones, establishing the etiology of the disease and the direction of therapeutic effort. He spoke with the confidence of a man who had spent his career establishing such facts, whose reputation rested on the reliability of his laboratory work. The officers left with their notes and their samples, and Lewis returned to his microscope.

The encounter troubled him, though he did not acknowledge the trouble directly. The officers had not challenged his conclusions, but their questions had exposed the gap between laboratory certainty and clinical uncertainty. The sailors continued to die despite every intervention. The serum therapy, prepared from the blood of recovered patients and administered to the acutely ill, showed no clear benefit. The isolation procedures, rigorously applied at the Navy Yard, had failed to contain the spread. The epidemic was not behaving like previous influenza outbreaks, not following the expected curve of rise and fall. Something was different, something that his identification of the Pfeiffer bacillus did not explain.

He worked through the evening of October 22, examining additional samples, preparing cultures for the morning’s inoculations. The laboratory was quiet, the building nearly empty. The university had suspended most of its operations, the students dispersed to their homes, the faculty engaged in war service or epidemic duty. Lewis worked alone, as he preferred. Collaboration was necessary for verification, but discovery required solitude, the sustained concentration that excluded distraction. He had learned this habit in his years at the Rockefeller, where the competition for priority was fierce and the hours were long. The habit had served him well. His bibliography included more than fifty papers, his appointments had come with increasing prestige, his current position was among the most secure in American medical science.

At 10 p.m., he made his final entry for the day. The laboratory book recorded the examination of twelve additional samples. Eight yielded the Pfeiffer bacillus. Four showed only secondary organisms. He noted the morphology, the staining characteristics, the growth patterns on artificial media. He did not note his growing sense that the pattern was incomplete, that the organism he was finding did not account for the disease he was seeing. Such doubts were not appropriate for the laboratory record. They belonged to the private realm of hypothesis, of intuition, of the hunches that preceded formal investigation. He would pursue them, but not yet, not until the primary work of identification was complete.

He walked home through streets that had grown strange in the weeks of the epidemic. The normal sounds of the university district—students returning from theaters, the clatter of restaurants, the occasional automobile—had diminished to near silence. The streetcars still ran, but with reduced frequency, their passengers sparse and masked. The restaurants were closed by order of the health department. The theaters, the dance halls, the gathering places of ordinary life, stood dark. Lewis walked past them without attention, his mind occupied with the problem of the bacillus. The city around him had become a backdrop, a setting for the intellectual drama that consumed his waking hours.

On the morning of October 23, he received a telegram from Surgeon General Rupert Blue. The message was brief, formal, requesting a summary of Lewis’s findings for distribution to Public Health Service officers nationwide. The request was not unusual. The Surgeon General’s office had established itself as the clearinghouse for epidemic information, coordinating the response of local health departments through bulletins and directives. Lewis had contributed to such bulletins before, in previous epidemics of meningitis and poliomyelitis. His reputation gave weight to his conclusions, made them suitable for official transmission.

He composed his reply with care. The telegram allowed no elaboration, no qualification, no acknowledgment of uncertainty. He stated his finding: the Pfeiffer bacillus was the causative agent of the current epidemic. He stated his method: examination of sputum, blood, and tissue from Navy and civilian cases, with consistent isolation of the organism. He stated the implication: therapeutic and preventive measures should be directed accordingly. The message was sent at 11 a.m., and by evening it had been relayed to health officers in every state.

The transmission marked a transition. Lewis’s finding, still preliminary in his own estimation, became fixed in the official record. The Surgeon General’s office had no mechanism for subsequent revision, no procedure for incorporating new evidence that might contradict the initial conclusion. The bulletins accumulated, each one building on the previous, creating a structure of authority that resisted modification. Lewis understood this dynamic. He had participated in it from both sides, as contributor and recipient. But understanding did not prevent participation. The urgency of the epidemic demanded action, and action required the confidence that only authority could provide.

He continued his laboratory work through October 23 and 24, accumulating additional samples, refining his techniques, seeking the confirmation that would transform hypothesis into established fact. The Pfeiffer bacillus remained present in the majority of his preparations. The exceptions remained exceptions, noted but not emphasized. He prepared a formal paper for the Journal of the American Medical Association, summarizing his findings and their implications. The paper would appear in the November 2 issue, reaching thousands of physicians who looked to scientific authority for guidance in treating their patients.

The publication would cement his conclusion in the medical literature. It would also expose him to criticism, to the replication studies that would test his findings in other laboratories, other cities, other epidemiological circumstances. Lewis welcomed this exposure. Science advanced through criticism, through the clash of competing interpretations, through the eventual triumph of evidence over assumption. He was confident in his methods, secure in his reputation, prepared to defend his conclusions against all challengers.

But the confidence was not absolute. In the late hours of October 24, after the laboratory assistants had departed and the building had settled into its nocturnal silence, he examined a final set of slides from autopsy material obtained that afternoon. The tissue was from a sailor who had died within twelve hours of symptom onset, whose clinical course had been too rapid for secondary bacterial infection. The slides showed the same inflammatory changes, the same cellular debris, the same Gram-negative bacilli that he had come to expect. But they showed something else as well: areas of lung tissue that appeared normal under low magnification, that contained no organisms at all, that had been destroyed by a process that left no visible trace.

He studied these areas for an hour, changing objectives, varying the light, searching for the agent of destruction. He found nothing. The tissue was simply dead, killed by something that had passed through without remaining to be identified. The finding contradicted his established conclusion. It suggested that the Pfeiffer bacillus, even if present, was not sufficient to explain the disease. It opened the possibility that another agent, invisible to his methods, was the true cause.

He recorded the observation in his laboratory book, marking it with a query that indicated its problematic status. He did not revise his telegram to the Surgeon General, did not withdraw his forthcoming publication, did not alert his colleagues to the emergence of doubt. The observation was single, uncorroborated, possibly artifactual. Science required the accumulation of evidence, not the pursuit of every anomaly. He would investigate further, would seek additional cases that showed the same pattern, would determine whether the finding represented a genuine phenomenon or a technical error.

But the investigation would take time, and time was not available. The epidemic continued its course, the deaths mounted, the demand for answers overwhelmed the capacity for careful study. Lewis’s finding, transmitted through official channels, published in the leading medical journal, became the basis for therapeutic decisions across the country. Physicians administered antipneumococcic serum, developed for other bacterial infections, in the hope that it would affect the Pfeiffer bacillus. Public health officers directed their efforts at the control of bacterial transmission, confident that they understood the enemy they faced. The resources of the medical profession were mobilized around a hypothesis that was plausible, authoritative, and wrong.

In the Hygienic Laboratory in Washington, the cultures Lewis had shipped sat in incubators, replicating his results for distribution to other research centers. In Philadelphia, the Bureau of Health recorded another 489 deaths on October 24. In the laboratory at 36th and Spruce, Lewis closed his book and extinguished the light, leaving the query mark in the margin where no one but he would see it.