Skip to main content
Back to Steel City Stories
The Night the Three Rivers Took Pittsburgh — Steel City Stories
HistoryChapter 28 of 4117 min read

The Night the Three Rivers Took Pittsburgh

Steel City Stories

Preparing the first passage.

Steel City Stories · Chapter 28

The receipt is a gauge, and the gauge embarrassed the forecast. Dollar Bank’s institutional history of the St. Patrick’s Day flood, written from the records of a Fourth Avenue bank that stayed in the path of the water, puts the numbers in clock time. Flood stage at the Point was twenty-five feet. The rivers passed that stage between eight and nine on the morning of March 17, 1936. The prediction in circulation was a crest of thirty-three feet. At nine o’clock on the evening of Wednesday, March 18, the high water stood at forty-six feet. Twenty-one feet above the stage the city called a flood, and thirteen feet above the number downtown had been told to expect. The Heinz History Center’s account of the same days uses the same crest, forty-six feet at the Point, and describes it as more than twenty feet over flood stage. Two independent Pittsburgh institutions, a bank and a history museum, agree on the height. That agreement is what allows this chapter to say forty-six. A round memory of “about forty-six” is no longer necessary.

The water that made the number had a weather history the History Center states without mystery. Nearly two inches of rain fell on March 16, on top of a winter the same account measures at sixty-three inches of snow. Warmth melted the snowpack. Creeks in the upper Allegheny and the upper Monongahela rose first. On St. Patrick’s Day the rise reached the North Side and then the streets of downtown, and businesses that had survived ordinary spring floods went under in hours. Dollar Bank adds the mechanism at watershed scale: melting snow and forty-eight hours of rain sent the Allegheny drainage over its banks. Pittsburgh’s mistake, if the word mistake can be used for a city built inside a confluence, was not that it failed to notice rain. It was that its forecast, its utilities, and its streets were calibrated for a thirty-three-foot river in a week when the river had forty-six feet available.

What forty-six feet covered

The Golden Triangle did not flood at the edges and spare the business district. Dollar Bank’s topography is a set of street names. Water reached the pavement of Tenth Street on the Allegheny side and Wood Street on the Monongahela side. Trolley cars on Penn Avenue went under to their roofs. The Jenkins Arcade, Horne’s department store, and the Roosevelt Hotel took water to their second stories. By eleven on the morning of March 18, Market Street was waist deep. The stock exchange, then at 229–233 Fourth Avenue beside the Benedum-Trees Building, suspended trading at 10:40 a.m. Banks from Wood Street toward the Point closed. Dollar Bank names First National and Farmers Deposit at Fifth and Wood, and the People’s-Pittsburgh Trust at Fourth and Wood. In the afternoon the river was still rising, at about a foot an hour. Elevators failed, then heat, then electricity, then the telephones. Thousands of workers tried to leave. Thousands more were upstairs in buildings that had become islands.

The History Center’s photograph of Fifth Avenue and Market Square on March 18 is the public version of the same hour. More than half of downtown businesses were underwater. The famous pictures are downtown because downtown is where the cameras and the banks were. The same essay is careful about the rest of the map. Johnstown, Sharpsburg, and Etna were in the water. So were communities across eleven states in the Northeast and the mid-Atlantic, and the District of Columbia. A Pittsburgh chapter that stays inside the Golden Triangle will miss the watershed that produced the gauge reading. A national chapter that never stands on Wood Street will miss what forty-six feet did to a concentrated downtown. Both scales are true. The Point is where the Allegheny and the Monongahela become the Ohio. Everything upstream that melted arrived there with a single number attached.

The failure was a systems failure, which is why a city of enormous industrial plant could still be stopped by water. Mills do not run without power, telephones, dry motors, and men who can reach the gate. Dollar Bank records that more than sixty thousand steelworkers were idle until flooded plants could be put back. That figure is a bank history’s regional labor claim, not a union ledger re-added here, and it is large enough to change the meaning of the flood. The disaster was not only a wet Triangle. It was a stoppage of the wage system that surrounded the Triangle. Police and the Pennsylvania National Guard threw a picket around the lower streets. Dollar Bank says nobody got below Wood Street without a pass. The pass was an admission that the ordinary municipal map had been replaced by a shoreline. Looting was the fear. Isolation was the fact. A fire, a heart attack, or a short circuit on the wrong side of the line was now a problem for boats and for men wading a street that had been a trolley route the day before.

Paper, vaults, and the drying line

The most Pittsburgh document of the flood may be a wet stock certificate. Downtown banks had trusted basements. The water proved the concrete porous. Vaults that had been sealed still took the river in, and safe-deposit boxes gave up paper. Dollar Bank describes certificates hung on lines like laundry and ironed, by the customers themselves, back into something a market would accept. The work ran into April. One customer salvaged a letter from Theodore Roosevelt by that domestic process. Electricity was out for days. Some banking halls could operate by daylight. Union Trust worked by candlelight. Pitt National Bank, at Fifth and Liberty, moved temporarily up the Triangle. Dollar Bank opened two windows so Union National’s customers, displaced from Fourth and Wood, could cash checks, deposit money, and pay mortgages and rent. The officers of Union National answered with a letter and a newspaper notice. Neighborliness in this episode was a teller window, not a sentiment.

The vaults that were rebuilt afterward are the flood translated into capital expenditure. Farmers Deposit spent three hundred thousand dollars in the summer of 1936 on a new safe-deposit vault fourteen feet above the main floor, specified to stay dry through a sixty-foot stage. Pitt National waterproofed a vault on its main floor by December. Colonial Trust waterproofed as part of a 1937 renovation. Union National built a wall around its vault. Keystone National added bulkheads rated to fifty-four feet. Dollar Bank’s directors, in July 1937, approved a Mosler vault, fifteen by eighteen feet, copper plate between steel, built against acetylene torches as well as against rivers, at a cost of one hundred thousand dollars. It was installed by W. F. Trimble and Sons and, the bank notes, is still in service on Fourth Avenue. A city that has just watched forty-six feet does not buy a sixty-foot specification for decoration. The bankers believed the gauge could be beaten again, and they moved the paper up.

Property-loss figures in the bank history and in later Corps of Engineers summaries converge on about two hundred fifty million dollars and then diverge about what that money covers. Dollar Bank says two hundred fifty million in property loss throughout the Pittsburgh district, tens of thousands driven from their homes, and fifty thousand people in Red Cross shelters in libraries and school gymnasiums. A Corps of Engineers historical vignette on the Pittsburgh District, as published by the agency, describes the spring 1936 flood in western Pennsylvania as inundating sixty percent of downtown Pittsburgh, damaging one hundred thousand buildings, killing more than two hundred people, and approaching two hundred fifty million dollars in regional damage. The History Center, focused on the regional human toll of this flood, says sixty-two dead, more than five hundred injured, and one hundred thirty-five thousand homeless, with millions of dollars in damage to homes, businesses, and industry. Sixty-two and more than two hundred cannot be the same body count unless the geographic frame has changed. This chapter will not average them. The Pittsburgh-centered museum essay and the bank history support a regional disaster whose local death toll is on the order of sixty-two, with a much larger homeless figure. The Corps vignette supports a wider spring flood, measured across western Pennsylvania, with a death toll above two hundred and a building count of one hundred thousand. The two hundred fifty million dollars appears in both the bank’s district property figure and the Corps’ regional damage figure. Money traveled better than bodies did, which is itself a warning about how disasters get totaled.

The law that followed the water

The political response could not be a Pittsburgh ordinance. Rain that falls on the upper Allegheny is not inside the city limits when it falls, and it is inside the city when it arrives. The History Center dates the legislative answer to June 1936. The Flood Control Act authorized a system of dams and reservoirs to be built by the Army Corps of Engineers and by affected communities. The sequence matters. The rivers crested on March 18. The statute is a June law. The flood did not implement the act. It supplied the exhibit. Flood-control bills had been argued before 1936. WESA’s later interview with History Center curator Lauren Uhl records the local memory that earlier proposals had stalled, and that the northeastern and mid-Atlantic disasters of that spring, including Pittsburgh’s, were what Congress looked at when it reconvened. A gauge reading of forty-six feet is not a statute. It is the kind of number that makes a statute harder to postpone.

The Corps vignette is specific about what the 1936 act did for this watershed. It assigned the Corps a national flood-damage mission, and it specifically authorized nine reservoirs in the Allegheny basin above Pittsburgh. The first two dams built under that impulse, at Crooked Creek and Tionesta Creek, were completed in 1940, ahead of schedule. The Pittsburgh District eventually built sixteen flood-control reservoirs and forty-two local protection projects. Those works are not the March flood. They are the institutional sediment of the March flood, and they took years, a war, and later authorizations to become the system people now point to when the rivers rise and stop at a lower stage. Reservoirs hold water that has not yet become a downtown street. They are upstream because the problem was upstream. Pittsburgh could sandbag the Point. It could not freeze the snowpack in New York and the northern Allegheny counties. The act conceded that fact in statutory form.

Local protection and upstream storage answer different failures. A wall at a plant gate assumes the river will be high and tries to keep a single property dry. A reservoir assumes the river will be high and tries to keep the peak from forming. The 1936 flood defeated both the informal version of the wall, which was a second-story office and a basement vault, and the absence of storage. Bankers responded with the wall, moved upward and waterproofed, rated to fifty-four or sixty feet because they had seen forty-six. The federal government responded with storage because a waterproof vault does not help a house in Millvale, a mill in the valley, or a worker who cannot reach either. The History Center’s note that high-water marks remain on downtown buildings is the right monument. The mark is a measurement, still visible, of a night when the three rivers occupied the street grid. The reservoirs are the argument about whether that measurement has to be repeated.

Upstream of the Point

A flood at Pittsburgh is a vote of the whole watershed, counted at one gauge. The Allegheny reaches the Point from the north and the east, draining country far outside Allegheny County. The Monongahela reaches it from the south, through the industrial valleys that had spent a century putting mills, rail yards, and houses on bottom land because bottom land was where the coal boats and the works had to be. The Ohio leaves the Point already carrying both. When both upper rivers rise together, the city at the junction has no third direction to send the water except through its own streets. March 1936 was that coincidence. Snow measured by the History Center at sixty-three inches, plus nearly two inches of rain on March 16, plus Dollar Bank’s forty-eight hours of rain on a thaw, is a recipe that does not require a broken dam. It requires a landscape that was already full. The thirty-three-foot forecast was a statement about how much of that landscape the weather office thought would arrive. Forty-six feet was the amount that arrived. The error is thirteen feet. Thirteen feet, on a street grid, is the difference between a wet first floor and a second story, which is exactly where Dollar Bank places the water in the Jenkins Arcade, Horne’s, and the Roosevelt Hotel.

Ordinary floods had taught the wrong lesson. The History Center’s account, and the later public explanation of the high-water marks, treat a rise to flood stage as a seasonal nuisance Pittsburgh knew how to endure. Endurance is a rational policy when the river stops near the prediction. It is a trap when the river does not. Businesses that had survived twenty-five-foot and thirty-foot years had located their records, their inventory, and their customers accordingly. Banks had located their vaults in basements because basements had been, in the living memory of the officers, dry enough. The porous concrete and the laundry lines of wet bonds are what a rational basement looks like after an irrational river. The pass system below Wood Street is what a police department looks like when the nuisance has become a shoreline. None of these were improvisations of a city that had never seen high water. They were improvisations of a city that had seen high water and had budgeted for the usual amount.

The Corps of Engineers’ later system is the correction aimed at the watershed rather than at the basement. Nine reservoirs authorized in the Allegheny basin above Pittsburgh are a geographic admission. The water that hurt the Golden Triangle had been a tributary yesterday. Hold it on Crooked Creek, hold it on Tionesta, and the crest at the Point changes without anyone in downtown lifting a sandbag. Those two dams, finished in 1940, are the vignette’s evidence that the 1936 act was not only a press release. The larger system the same vignette describes, sixteen flood-control reservoirs and forty-two local protection projects, is the work of decades, and this chapter does not pretend every dam was funded in the June statute. Local protection, the walls and the bank works, defends a place. Storage defends a hydrograph. Pittsburgh needed both because it had concentrated so much of its taxable property inside a junction, and because the junction’s inputs were in other counties and other states. A city ordinance cannot zone the snowpack. A federal act can build the dam that waits for it.

The human figures refuse to be tidied, and the refusal is part of the history. Sixty-two dead, in the History Center’s regional sentence, is a toll large enough to end any story in which the flood was only an expensive inconvenience to banks. More than two hundred dead, in the Corps’ spring-flood sentence, is a toll that includes a wider western Pennsylvania disaster than the Triangle photographs show. One hundred thirty-five thousand homeless, in the museum account, is a population larger than many of the mill towns in this book, living in other buildings while the water went down. Fifty thousand in Red Cross shelters, in the bank account, is a subset of care, not a contradiction of the homeless figure, unless someone forces the two numbers to describe the same night in the same boundary. This chapter leaves them in their sources. What they share is scale. The flood moved more people than it killed, by an order of magnitude, and it killed enough that a single agreed integer is less important than the fact that the sources will not yield one. Property at two hundred fifty million dollars is the figure both the bank and the Corps are willing to approach. Dollars are easier to estimate than bodies when the boundary of “the region” moves.

The confluence as a machine

Pittsburgh exists where it exists because the rivers were a machine for moving coal, ore, glass, and steel. The machine’s price is that the same confluence is a drain. In ordinary years the drain misbehaved and the city mopped. The History Center’s curator has described flooding as an annual event, which is why March 17 at twenty-five feet was not, by itself, a civic emergency. The emergency was the second day, when the annual event refused the forecast and kept the appointment through Wednesday night. Forty-six feet converted a familiar nuisance into a shutdown of telephones, lights, mills, and the paper inside the vaults. A city can be brilliant at production and still be one utility failure away from paralysis when those utilities share a floodplain. That is the night the rivers took Pittsburgh, in the only sense the records support. They did not abolish the city. They occupied its lower streets, stopped its exchange, idled a regional workforce, and left a gauge number that two local institutions still print as forty-six.

The aftermath splits. Within months, banks bought height and copper plate. Within the year, Congress bought a national flood program and named Allegheny reservoirs. Within four years, Crooked Creek and Tionesta were dams rather than drawings. The dead, whether the frame is the History Center’s sixty-two or the Corps’ wider two hundred, were not refunded by either purchase. Homelessness on the order of one hundred thirty-five thousand, in the museum’s regional figure, describes a population living in other people’s buildings while the laundry lines of wet bonds were still up in the banks. The flood’s lesson is not that the three rivers are malicious. It is that a confluence city which stores its wealth, its records, and its workers inside the reach of a thirty-three-foot expectation will eventually meet a forty-six-foot river. The public system built afterward, reservoirs upstream and a Corps mission written into the Flood Control Act of 1936, is the admission that the expectation had been a private gamble with a watershed the city did not own.

The rivers did not take the hills. They took the working floor of the city, which is why the photographs that survived are of downtown streets rather than of a vanished Pittsburgh. Fifth Avenue at Market Square, the North Side approaches, Millvale and the valley towns in the History Center’s wider frame, and the mill gates where more than sixty thousand steelworkers waited on dry motors are the same event at different elevations. A bank on Fourth Avenue could reopen by daylight and candle because its banking hall sat above the water that had ruined its vault. A family in a bottom-land house did not have a banking hall. The Red Cross shelters in libraries and gymnasiums, fifty thousand people in the bank history’s figure, are the institution that appeared in the gap. Shelters are not reservoirs. They are what a city uses in the weeks before a law, and in the years before a dam. March had the shelters. June had the statute. 1940 had the first two dams. The people who slept in the gymnasiums lived in that interval, after the gauge had already been read and before either dam existed. A city can survive that gap. It should not be asked to call the gap a plan.

Stand at the Point with the two local numbers that match. Twenty-five feet is flood stage. Forty-six feet is where the water stood at nine on the night of March 18, thirteen feet past the forecast of thirty-three. Wood Street and Tenth Street are the lateral limits Dollar Bank gives the inundation. The second stories of Horne’s, the Jenkins Arcade, and the Roosevelt Hotel are the vertical limit. The stock exchange stopping at 10:40 in the morning, the vaults weeping through concrete, the steelworkers idle, the National Guard pass below Wood Street, and the certificates ironed like shirts are the civic systems that failed together. None of them failed because Pittsburgh lacked industry. They failed because the industry, the paper, and the housing had been placed inside a drain, and the drain exceeded the number the city had agreed to respect.

The remedy was not a better attitude toward the rivers. It was jurisdiction. The Flood Control Act of June 1936 gave the Corps a national assignment and, in the agency’s own vignette, nine reservoirs above Pittsburgh on the Allegheny. Crooked Creek and Tionesta, done by 1940, are the first physical proof. The later district system of sixteen reservoirs and dozens of local protection works is the long proof. High-water marks still painted on downtown walls are the proof that the city chose to remember the gauge even after the dams existed. Remembering the gauge without the act turns the flood back into weather. Remembering the act without the gauge turns it into a legislative accident. The night the three rivers took the lower city is the night a forecast broke, a vault leaked, and a country that had treated Pittsburgh’s floods as local bad luck finally wrote upstream storage into federal law. The dead remain in dispute as to number and boundary. The crest does not.

Evidence Ledger

What remains disputed

The Point crest is not the disputed item. Dollar Bank and the History Center both say forty-six feet, and Dollar Bank supplies the clock time, 9 p.m. on March 18, against a twenty-five-foot flood stage and a thirty-three-foot forecast. The death toll is disputed because the frames are different and the numbers do not meet. The History Center says sixty-two dead in the region. The Corps vignette says more than two hundred for the spring 1936 flood in western Pennsylvania, in a paragraph that also covers damage well beyond the Golden Triangle. This chapter does not blend those counts into a midpoint. Homelessness is given by the museum as one hundred thirty-five thousand and is not re-derived here. The two hundred fifty million dollars appears as district property loss in the bank history and as regional damage in the Corps account. Downtown’s share is “more than half of businesses” at the History Center and “sixty percent of downtown” at the Corps. Those are compatible descriptions, not independent surveys. How many of the nine authorized Allegheny reservoirs were built exactly as the 1936 act drew them, and how much of the later sixteen-reservoir system depends on later laws, is a Corps institutional history this chapter summarizes rather than re-maps dam by dam. The flood did not, by itself, appropriate the money or pour the concrete. It made the appropriation difficult to refuse.

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.

1936 Pittsburgh flood0007

1936 Pittsburgh flood0007

Wikimedia Commons · 1936 Pittsburgh flood0007.jpg · Mother of "upheld" on Flickr [2] · CC BY 3.0

(River: View of the Monongahela River from Mt. Washington)

(River: View of the Monongahela River from Mt. Washington)

Wikimedia Commons · River View of the Monongahela River from Mt. Washington (85.4.40).jpg · American · Public domain

The Point Pittsburgh 1910 1920

The Point Pittsburgh 1910 1920

Wikimedia Commons · The Point Pittsburgh 1910 1920.jpg · Part of Detroit Publishing Company Photograph Collection · Public domain

Related Stories