Skip to main content
Back to Steel City Stories
The Day Donora’s Air Killed the Town — Steel City Stories
HistoryChapter 23 of 4118 min read

The Day Donora’s Air Killed the Town

Steel City Stories

Preparing the first passage.

Steel City Stories · Chapter 23

The receipt is not a headline. It is a man walking in front of an ambulance with a flashlight, because the street in Donora had disappeared. In the last days of October 1948 the Monongahela Valley town, and the neighboring community of Webster across the river, filled with a smog thick enough that emergency vehicles could not see the curb. A 2018 reconstruction in the American Journal of Public Health, drawing on the contemporary record, describes Donora’s first responders, an ambulance and a firetruck, being guided through that darkness, and then firemen going door to door with oxygen when even the walking guide stopped working. Doctors told people with chronic illness to leave. The roads closed anyway. The first death in that account is placed at about two o’clock on the morning of Saturday, October 30. Within twelve hours, seventeen people in Donora and Webster were dead. By the time the episode was counted, the figure that entered the federal investigation was twenty.

That sequence matters because Donora did not experience the smog as a surprise from a clear sky. Residents of a town that lived on two industrial plants already knew heavy smoke. The American Journal of Public Health account names those plants as the American Steel and Wire works and the Donora Zinc Works, miles of riverfront that treated a dirty sky as the ordinary price of a paycheck. Smoke, in the Monongahela towns of that era, could be read as proof that the mills were running. A clear day could mean layoffs. The disaster of late October was exceptional even by that standard, and it was exceptional in a particular way. It was weather acting on an industrial system that had already been allowed to load the valley with sulfur compounds, metal fumes, coal smoke, and the exhaust of river traffic. The fog did not invent the pollution. It held it in place.

A town built into a bend

Donora sits in Washington County, about thirty miles south of Pittsburgh, inside a horseshoe bend of the Monongahela. The same 2018 medical history describes hills and cliffs rising at least four hundred feet around the borough, and an inversion layer that held effluent below about one hundred fifty feet, an altitude at which people actually lived. Geography was not a poetic backdrop. It was a container. When the wind dropped, the valley had nowhere to send the plume. American Steel and Wire, part of the United States Steel orbit, operated the steel and wire production that gave the town its industrial name, and the Zinc Works that later sediment studies would tie to a sharp rise in cadmium, lead, and zinc after the works opened in 1915. Later writers sometimes speak of Donora Steel as the steel side of that complex. The corporate reality on the ground was American Steel and Wire’s mill and the Zinc Works that belonged to the same industrial system. Treating them as one anonymous “factory” hides the point. This was a company landscape, with steel, wire, and zinc sharing a riverbank and a workforce.

The Zinc Works had a particular reputation, and later investigators had reasons to look there. Zinc smelting throws off sulfur compounds, carbon monoxide, nitrogen dioxide, fluorides, and fine particles carrying heavy metals. The steel and wire side burned coal, ran furnaces, and added its own stack emissions. Neither plant was a secret. They were the reason the town existed in its twentieth-century form. Houses, stores, a high school football team, and a Halloween parade all sat downwind of a payroll. When the air turned lethal, the people breathing it were not visitors. They were the town.

The episode did not announce itself as a new kind of emergency. Elizabeth Jacobs and other later chroniclers, as used by the 2018 journal article, record that the annual Halloween parade still went down Main Street on October 29. The next day the Donora Dragons played a football game in visibility so poor that the field itself was hard to see. Those details are not color. They show how completely industrial smoke had been normalized. A community that will march children through a street it cannot see has already decided, for years, that the air is a condition of life rather than a violation of it. The telephone calls to the town’s doctors came anyway. Respiratory distress spread through houses. Chronic patients were told to get out, and then the roads refused them.

What the Public Health Service actually counted

The investigation that followed is the reason Donora is a chapter in public health rather than a local ghost story. Borough leaders, the United Steelworkers, the state of Pennsylvania, and American Steel and Wire itself all pressed for an outside inquiry, because nobody on the ground trusted a study run only by an interested party. The United States Public Health Service sent a team under H. H. Schrenk, chief of its environmental investigations branch, a man who had spent twenty years with the Bureau of Mines in Pittsburgh. Harry Wexler of the Weather Bureau joined the authorship. In 1949 the Federal Security Agency published Public Health Bulletin No. 306, Air Pollution in Donora, Pa.: Epidemiology of the Unusual Smog Episode of October 1948. It was labeled a preliminary report. A final report was never published.

The bulletin’s findings, as abstracted by the Environmental Protection Agency’s HERO record of the study and by the Department of Energy’s bibliographic summary, are specific enough to quote as numbers and still too careful to flatten into a slogan. An episode of heavy smog from October 27 to October 31 caused twenty deaths. A household canvass put questionnaires in front of thousands of residents. The study concluded that 5,910 people, 42.7 percent of all persons living in the area, were affected. Of those, 2,148 were mildly affected, 2,322 moderately affected, and 1,440 severely affected. Illness was essentially an irritation of the respiratory tract and exposed mucous membranes. Cough was the most prominent single symptom. Incidence and severity rose with age. More than 60 percent of people sixty-five or older were affected. The ages of the dead, for the eighteen fatalities on whom the investigators had information, ran from fifty-two to eighty-four. Sixteen of the fatalities had preexisting cardiorespiratory disease. Autopsies on fatal cases found capillary dilatation, hemorrhage, edema, and bronchiolitis. Some domestic animals, including dogs, became ill and died. Sex, race, occupational status, length of residence, and physical activity did not, in the investigators’ reading, explain who got sick.

An Environmental Protection Agency citizen handbook later compressed the same disaster into a sentence that is close to the chapter’s familiar formulation and not quite identical to the bulletin’s percentage. It said that before wind and rain cleared the air, twenty people were dead, and more than 5,900 of the town’s total population of approximately 14,000 became sick. A 2001 note in the American Journal of Public Health used a wider illness range, an estimated 5,000 to 7,000 persons in a town of 14,000, with some 400 hospitalized, and twenty dead before rain on October 30 and 31. The 2018 revisit said twenty dead and an additional 5,900, described as 43 percent of the population of Donora, and then broke the nonfatal toll into 1,440 seriously ill and 4,470 mild or moderate. Those are not contradictions invented for drama. They are what happens when a preliminary federal count, a percentage of “the area,” and later retellings of “the town” are asked to be the same sentence. The defensible core is the bulletin’s: twenty deaths in the episode, and 5,910 people, 42.7 percent of persons living in the studied area, made ill. Roughly 14,000 is a fair description of the community’s scale. It is not a license to pretend every later percentage is the same denominator.

The chemistry that would not settle

Atmospheric sampling looked for particulates, trace metals, sulfur dioxide, total sulfur, carbon dioxide, carbon monoxide, and nitrogen oxides at the zinc and steel plants. The HERO abstract of the bulletin is explicit about the conclusion that popular memory keeps trying to simplify. No single substance was responsible. The illnesses could have been produced by a combination of sulfur dioxide, its oxidation products, and particulates. The concentration of contaminants tracked atmospheric stability. A bibliographic summary of the same report puts the probable mechanism as sulfur dioxide acting together with particulates, and notes that zinc and other metals were high in the air near the industrialized area. That is a synergy finding. It is not a verdict that one stack, one metal, or one molecule has been convicted while the others are excused.

The 2018 journal article is equally plain that the Public Health Service positively identified neither a single contaminant nor a precise source. It describes three contributing factors. The first was pollution from the American Steel and Wire plant and the Donora Zinc Works, with the Zinc Works identified in particular because of hydrogen fluoride, carbon monoxide, nitrogen dioxide, sulfur compounds, and heavy metals in fine particles. The report also counted river traffic and household coal furnaces. The second factor was the weather. The smog settled on October 26 and 27 with a temperature inversion. A cold front arrived, the wind failed, and colder air sat under warmer air like a lid. The third factor was the horseshoe bend and the hills. Five days of effluent stayed in the bowl. Recommendations followed: reduce plant effluent, and build a warning system that would tell the valley when the weather was about to trap the next load. Those recommendations already concede the real subject. The danger was repeatable. It was a system of emissions plus stagnation plus a town placed in a river bend, not a freak fog that arrived once and left no policy behind.

Clarence Mills, a professor of experimental medicine at the University of Cincinnati, studied the episode on his own and argued that the federal investigation had not faced the possibility of long-term harm. Editorials in the American Journal of Public Health and the New England Journal of Medicine in 1950 called the report a missed chance to look harder at both chronic pollution and acute disasters. The 1950 public-health editorial, as summarized in the later revisit, also noted findings that still complicate any single-pollutant story, including a higher death rate among nonwhite residents than among white residents and the role of poor housing in how badly people were hurt. Those are disparities inside the event. They are not a chemical name.

Two University of Pittsburgh biostatisticians, Antonio Ciocco and Donovan Thompson, later did the work the preliminary bulletin had not done. Their studies of Donora deaths from 1948 to 1957 found more cardiovascular deaths than expected among people who had reported heart or respiratory disease at the time of the smog. In one comparison, seventy cardiovascular deaths were observed in that previously ill group, against thirty-nine expected. A companion look at cancer, set against the Arsenal district near Pittsburgh, has been reanalyzed by later authors as a higher proportion of cancer deaths in Donora. A lake-sediment study about six miles northeast of town found cadmium, lead, and zinc rising sharply after the Zinc Works opened in 1915, and still elevated decades after the works stopped. The smog weekend was the body count that made the valley famous. The metals in the mud say the exposure was a long industrial residency, not a four-day accident.

Why the plants kept running until they could not

During the episode the zinc works was shut down at the height of the incident. That fact, recorded in the 2001 American Journal of Public Health note, is often told as if the shutdown were a confession. It was an emergency measure inside a disaster, and the investigation that followed refused to convert it into a simple assignment of blame. The same note records the finding that then-recommended workplace levels of airborne contaminants had not been exceeded, and that the inquiry fixed no firm blame on any party. Workplace standards and outdoor air in a trapped valley are not the same measurement. A standard written for a worker on a shift does not describe a town of children, old people, and heart patients breathing the same valley for five days with the wind gone. The failure to name a single defendant was partly a limit of the science in 1949 and partly a political fact. American Steel and Wire had asked for the study. The Steelworkers had asked for it. The borough had asked for it. A preliminary report that said “no single substance” could be used by every party. It could also be used, more honestly, as a description of combined exposure.

The town’s doctors, by the journal’s count of the crisis, were eight. They were not an industrial-hygiene service. They were local physicians whose telephones became the surveillance system. The fire department became an oxygen-delivery service. Funeral capacity, hospital beds down the valley, and the question of whether a patient could be driven out of the bowl were the actual public-health infrastructure on the night of October 30. None of it had been built for a valley-wide inhalation injury. That is the institutional fact underneath the flashlight. Donora had mills capable of loading the air and a medical system scaled to ordinary sickness. When the inversion arrived, the mismatch was the disaster.

From a valley to a national authority

The Meuse Valley in Belgium, in December 1930, had already shown the pattern. Over roughly twelve miles of river industry, a temperature inversion and a narrow valley coincided with sudden respiratory collapse. More than sixty people died in about a day, and observers reached for mustard gas and Sahara dust before the medical conclusion settled on factory pollution trapped by weather and terrain. Donora was the American repetition, closer to Pittsburgh’s own industrial geography, with a federal epidemiology team rather than a rumor. London’s killer fogs would make the same point again on a larger stage. The lesson that survives comparison is not that fog is occasionally poisonous. It is that concentrated combustion, metals, and sulfur become a mass-casualty event when a landscape and a weather system stop dispersing them, and when nobody has the authority to slow the source before the deaths begin.

What changed after Donora was not the Monongahela’s shape. It was the claim that air could be governed. In 1955 Congress passed the Air Pollution Control Act, which gave the Public Health Service money to research the causes and control of air pollution. Later Clean Air Acts, in 1963, 1970, and 1990, added standards and enforcement that the 1949 bulletin did not have the power to impose. The 2001 public-health note is careful: Donora and other environmental calamities bolstered support for that legislation. They did not, by themselves, write the statutes. The path from a preliminary bulletin to a national enforcement regime ran through more deaths, more science, and more politics. Remembering Donora as the place where “clean air started” is a local museum’s compressed truth, and the journal article that opens on that sign does not treat the sign as hype. The more precise historical statement is that Donora made air pollution legible as an epidemiological event in the United States, and that legibility was a precondition for treating the air as a public-health jurisdiction rather than as an industrial courtesy.

The mills did not vanish in November 1948. People went back to work in the same bend of the river. The preliminary report asked for less effluent and for weather warnings. It did not close the Zinc Works. The long argument over fluorides, sulfur, particles, and metals continued in technical papers because the bulletin had refused a single villain, and because a single villain would have been better politics than the record allowed. Residents lived with the memory of neighbors who had died in their houses, and with the everyday smoke that had not been declared illegal. That combination, grief beside a still-operating stack, is why a freak-fog story is not just incomplete. It is a way of returning the blame to the sky.

Pittsburgh’s air, measured in a smaller town

Donora was not a rural accident on the edge of the steel district. It was one of the Monongahela mill towns that fed the same industrial system Pittsburgh claimed as its identity. Wire, zinc, and steel from that bend moved into a regional economy of bridges, nails, galvanized goods, and the furnaces farther up and down the river. The city that sold itself as the workshop of the world had already spent decades treating smoke as the visible form of payrolls. Pittsburgh’s own ordinances against dense smoke were older than 1948, and they had never converted the Mon Valley into a clean-air jurisdiction. Donora made the regional bargain impossible to file as a nuisance. When a town of roughly fourteen thousand can count its sick in the thousands over a single stagnation, the smoke is no longer a picture of prosperity. It is an exposure with a denominator.

The investigation’s method is part of the historical event, not a footnote. House-to-house questionnaires, medical interviews, three autopsies, veterinary reports, and air-sampling stations were an attempt to build a population picture instead of a mill-gate rumor. Schrenk’s team had to separate the people who coughed because the valley always coughed from the people whose illness belonged to those days. The bulletin’s answer was statistical and clinical at once. The smog illness rose with age, clustered in people whose hearts and lungs were already damaged, and presented as irritated airways. That pattern is how an investigator shows a shared inhalation injury without pretending every symptom in a smoky town began on October 27. It is also how a careful report refuses the comfort of a murder weapon with one name. Sulfur dioxide was in the samples. Particulates were in the samples. Carbon monoxide was in the samples. Metals were high near the plants. The authors’ sentence, preserved in the federal abstract, is that the illnesses could have been caused by sulfur dioxide and its oxidation products acting with those particles. “Could have been” is the language of a preliminary report that knows the valley was a mixture.

The requests for the study are themselves a map of power. A company that asks the Public Health Service to investigate is trying to get ahead of a worse story. A union that asks is trying to put member deaths on a federal page. A borough council that asks is trying to borrow an authority it does not have. The state that asks is admitting that a Washington County emergency has become a Pennsylvania problem. The fact that all of them converged on the same agency does not make the agency neutral in result. It means that in 1948 the United States still had no standing air-pollution police, so a disaster had to recruit a research team. Twenty-five investigators arrived after the rain. They could count the dead. They could not reopen the week and turn the stacks down on October 26, which is when the inversion, in the later reconstruction of the bulletin, first settled on the town.

Rain on the last days of the month broke the stagnation. The 2001 public-health note places the dispersal on October 30 and 31. The EPA handbook describes wind and rain ending a four-day grip that it dates from October 27. The difference of a day in the onset is real in the secondary literature and small beside the structure of the event. A cold-air lid lasted long enough for twenty deaths and thousands of illnesses, and then the atmosphere did what the institutions had not done. It moved. People who treat that rain as the hero of the story have missed the investigation’s own recommendation. The Public Health Service wanted less effluent and a weather warning, which is an admission that the next inversion was a scheduling problem, not a miracle to be awaited. A town that needs rain in order to breathe is a town whose production plan includes the hope that the wind will return before the funerals do.

What the flashlight was pointing at

Stand again on that street. The ambulance is real in the reconstruction because the town’s emergency service was real, small, and blind. The man with the flashlight is there because the industrial atmosphere had become optically solid. Behind him are zinc and steel works that had been loading the valley for years, a temperature inversion documented by the Weather Bureau scientist whose name is on the federal bulletin, and a population the Public Health Service counted house by house. Twenty people died in the episode the bulletin studied. Thousands were made ill, on the order of 5,910 people and 42.7 percent of the area, a figure later retold as about 5,900 of roughly 14,000. The illnesses looked like irritated lungs and damaged hearts, concentrated among the old and among people already sick. No investigator the federal team trusted was willing to say that one chemical, alone, had done it.

The public-health system that came out of that week was not a moral. It was an institution that had not existed in usable form on October 26. Before the inversion, smoke was a local nuisance and a sign of wages, policed, if at all, as a workplace exposure or a soot complaint. After the bulletin, the United States had a documented case in which ordinary industrial emissions, ordinary coal heat, ordinary river traffic, and a stagnant atmosphere had killed and sickened a whole community. Monitoring, research appropriations, warning logic, and eventually enforceable air standards are the slow machinery built so that the next valley would not have to discover the same fact with a flashlight. Donora’s air killed people because a production system had been allowed to use a river town as its vent, and because no public authority was yet organized to treat that use as a predictable threat to life. The fog lifted. The jurisdiction had to be invented.

Evidence Ledger

What remains disputed

The immediate death toll of twenty is the Public Health Service episode count and the figure used by later medical reviews. It is not a claim about every excess death in the following months or years. Ciocco and Thompson’s work argues for elevated cardiovascular mortality in the next decade among people who were already ill in 1948. That is a different question from the twenty deaths inside the smog week, and it remains an epidemiological inference rather than a list of named victims. The illness percentage is also unstable at the margins. The bulletin’s 5,910 and 42.7 percent apply to persons living in the studied area, which included Webster as well as Donora. Later summaries round to about 5,900 and sometimes call the denominator the town of 14,000, or 43 percent of Donora. The 2001 journal note’s range of 5,000 to 7,000 ill is wider still. Housing quality, racial differences in death rates, and the share of blame belonging to household coal, river traffic, coke ovens, the steel works, and the Zinc Works were all acknowledged and none was settled as a sole cause. Fluoride, sulfur dioxide, and metal particles each have advocates in the later literature. The preliminary report’s own sentence still governs this chapter: no single substance was responsible. A final Public Health Service report was never issued, which means the 1949 bulletin remains the controlling federal epidemiology and also an unfinished document.

This is an original public expansion of the short website chapter. It does not replace the Google Drive FULL MASTER.

Archival photographs relevant to this chapter. Sourced from Wikimedia Commons with attribution — not modern stock art.

Webster-Donora Bridge

Webster-Donora Bridge

Wikimedia Commons · Webster-Donora Bridge.jpg · PennDOT · Public domain

Donora Pennsylvania Street Scene mid 1970s

Donora Pennsylvania Street Scene mid 1970s

Wikimedia Commons · Donora Pennsylvania Street Scene mid 1970s.jpg · Paul Gipe · CC BY-SA 4.0

Donora, Washington County, Pennsylvania, 1901 LOC 2002620680

Donora, Washington County, Pennsylvania, 1901 LOC 2002620680

Wikimedia Commons · Donora, Washington County, Pennsylvania, 1901 LOC 2002620680.jpg · Fowler, T. M. (Thaddeus Mortimer); Moyer, James · Public domain

Related Stories