Case Study · Tornado · 1925
March 18, 1925. The deadliest tornado in United States history tore 219 miles across Missouri, Illinois, and Indiana and killed 695 people in 3.5 hours. Not one of them received a tornado warning. The government had made that impossible — the U.S. Weather Bureau had officially banned the word "tornado" from all forecasts since the 1880s, to prevent public panic. The Tri-State Tornado's legacy is every warning system that exists today.
Missouri, Illinois & Indiana · March 18, 1925
At approximately 1:00 PM on Wednesday, March 18, 1925, a tornado materialized near Ellington, in southeastern Missouri. The weather forecast for the day had been normal. There had been no severe weather watches, no storm warnings, no alerts of any kind — not because forecasters lacked concern, but because a standing policy of the U.S. Weather Bureau explicitly prohibited the use of the word "tornado" in any official forecast. The policy had been in place since the late 19th century. Its logic: the public would panic, and the resulting chaos would cause more harm than the tornado itself. The Tri-State Tornado would become the most complete refutation of that logic in the history of American meteorology.
The tornado moved northeast at an average speed of 62 miles per hour — faster than most cars could travel on the roads of 1925. It crossed into southern Illinois, where it destroyed towns in rapid sequence: Gorham, De Soto, Murphysboro, West Frankfort. In just 50 minutes after crossing the Mississippi River, it had killed 553 people. In Murphysboro, the largest city in its path, it killed more than 230 people and destroyed more than 80 percent of the city. In West Frankfort, it killed 149 people in one pass. The tornado continued into Indiana, where it destroyed the town of Griffin and killed 76 more. It finally dissipated at approximately 4:30 PM near Pike County, Indiana, having traveled 219 miles. The Evansville Museum's centennial documentation of the event records the total: 695 deaths, more than 2,000 injuries, tens of thousands left homeless — in a period before the word "tornado" had been spoken in any official government communication about weather.
Mar 18, 1925
Date
695
Deaths
219 miles
Path Length
62 mph
Avg Forward Speed
F5 Equivalent
Intensity
The NWS Heritage documentation of the event records the formal policy clearly: "The Weather Bureau had a policy not to even use the word 'tornado' in any forecast, to avoid inciting panic." This policy, documented in Britannica's account as having been in force since the late 19th century, meant that the residents of Murphysboro, West Frankfort, De Soto, and every other town in the Tri-State Tornado's path woke up on March 18, 1925 with a normal forecast and no reason to believe the day would be anything other than ordinary. 553 of them would be dead by 2 PM. In the context of the modern warning system — where Doppler radar can identify tornado-producing rotation 10–15 minutes before touchdown, and where Wireless Emergency Alerts reach every phone in a warning area automatically — the policy that killed 695 people on March 18, 1925 is not an abstraction. It is the baseline from which every advance in tornado forecasting measures its distance.
The Science
Think of a tornado's warning window as the time between when you know it's coming and when it arrives. In a modern tornado event with a 10-minute warning lead time, a tornado moving at 30 mph gives you about 5 miles of cushion — enough time to reach a basement or interior room. The Tri-State Tornado moved at an average of 62 mph and peaked at 73 mph. At 62 mph, a 10-minute warning gives you only about 10 miles — and the tornado's width, up to 1.7 miles at its peak, means the actual dangerous zone is enormous. With no warning at all, as was the case on March 18, 1925, the survival calculus depended entirely on whether you happened to already be in a protected location when the tornado arrived. In the poorly constructed frame houses of 1920s southern Illinois — most without basements — almost nothing could protect a resident who did not know to seek shelter.
The Wikipedia documentation of the Tri-State Tornado records its records precisely: longest path of any U.S. tornado (219 miles), longest time on the ground (3 hours 45 minutes), most deaths from a single tornado in U.S. history (695). The 3News meteorological history analysis puts the 50-minute Illinois segment in context: "To put this number in context, the number of people killed in tornadoes from 2013–2023 was 527 people." The Tri-State Tornado killed more people in one 50-minute stretch than the entire United States tornado death toll over a decade, one century later. The difference is almost entirely the warning system — Doppler radar, NWS watches and warnings, and Wireless Emergency Alerts that the Tri-State's victims never had.
The Britannica account of the Tri-State Tornado notes that it was unusually wide and low to the ground, appearing more like a moving dark wall of debris than the recognizable funnel shape most people associate with tornadoes. Survivors described not recognizing it as a tornado until it was nearly upon them. In the absence of any warning system, the visual recognition of the storm was the only signal residents had — and the storm's unusual appearance, combined with the speed of its movement, meant that by the time anyone understood what they were seeing, there was often no time to act.
Timeline
01
~1:00 PM: Tornado forms near Ellington in Reynolds County, Missouri. Weather forecast that day: normal. The Weather Bureau has officially banned the word "tornado" from all forecasts. The tornado moves northeast at 62+ mph through Annapolis, Biehle, and Frohna, killing 11. It crosses the Mississippi River in less than 30 minutes.
02
~1:30–2:20 PM: Tornado enters Illinois and destroys Gorham, De Soto, Murphysboro (230+ deaths, 80% of city destroyed), West Frankfort (149 deaths, including 33 children in a schoolhouse), Parrish (22 deaths, entire town leveled), and Hamilton County farms. 553 lives taken in 50 minutes. The Weather Bureau does not issue a tornado warning — it cannot.
03
~4:00 PM: Tornado crosses the Wabash River into Indiana. Destroys Griffin (entire town leveled), damages the Owensville area and Princeton. 76 killed in Indiana. Red Cross and Indiana National Guard arrive within hours. 4:30 PM: Tornado dissipates in Pike County, Indiana. 3 hours 45 minutes on the ground. Total: 695 dead, 2,027+ injured.
04
1925–1952: The Tri-State Tornado begins to change the Weather Bureau's position on tornado warnings over the following decades. 1948: First experimental U.S. tornado forecast issued by Air Force meteorologists. 1952: Weather Bureau establishes Severe Local Storms unit (SELS) — predecessor to the Storm Prediction Center. The ban on tornado forecasting, in practice, ends. The modern warning system's origin story begins on March 18, 1925.
Human Decisions
What worked — in the aftermath
The Indiana Historical Bureau's account of the Tri-State Tornado documents that help arrived in Griffin within hours of the tornado's passage — Red Cross teams, the Indiana National Guard, and assistance from neighboring towns. The scale of the disaster mobilized relief networks that were, for their era, genuinely impressive in their speed and organization. Recovery in Griffin was measurable within weeks: a schoolhouse, a church, and a grain elevator rebuilt within a year. The community response to the storm was not the failure. The absence of any warning system was.
The Tempe Today analysis of the Tri-State Tornado's centennial notes that the event "shaped tornado research and safety protocols in the decades since." The enormous death toll — still unequalled by any single tornado a century later — made the ban on tornado forecasting politically and morally untenable over time. The Air Force's first experimental tornado forecast in 1948 and the Weather Bureau's establishment of the SELS unit in 1952 trace their lineage to the pressure to prevent a recurrence of what happened on March 18, 1925.
The decision that caused it
Britannica's account of the Tri-State Tornado is direct: the Weather Bureau had a policy prohibiting the word "tornado" from any forecast "to avoid inciting panic." The policy reflected a paternalistic calculation: officials believed the public could not be trusted to receive warning information about tornadoes without causing harmful overreaction. The Tri-State Tornado is the cost of that calculation: 695 people who could not protect themselves because the system that was supposed to protect them had decided that warning them would cause more harm than the tornado itself. The policy was wrong. The death toll is the proof.
The frame houses of 1920s southern Illinois provided almost no resistance to F5 winds. Most homes in Murphysboro and West Frankfort lacked basements — the single structural feature that provides meaningful protection against extreme tornado winds. The combination of no warning and no safe refuge in the path created a near-total vulnerability: residents had neither the time to flee nor a place to shelter if they had tried. The modern tornado shelter program — FEMA safe rooms, school storm shelters, community shelters in mobile home parks — addresses the physical gap that 1925's housing stock could not.
Despite the Tri-State Tornado's catastrophic demonstration of what happened without tornado warnings, the Weather Bureau did not immediately reverse its policy. The NWS Heritage documentation confirms that the formal establishment of tornado forecasting capability did not occur until 1952 — 27 years after the Tri-State event. In those 27 years, tornadoes continued to kill Americans without warnings. The institutional inertia that kept a demonstrated-to-be-fatal policy in place for nearly three decades is a case study in how slowly life-safety policy can change, even in the presence of overwhelming evidence that change is necessary.
The cascade lesson
The Tri-State Tornado is not primarily a story about meteorology. It is a story about what happens when a warning system is deliberately suppressed to prevent the public from having information that might change their behavior. The Weather Bureau's reasoning was not malicious — it reflected a genuine belief that panic was a greater danger than the tornado itself. That belief was tested on March 18, 1925, and it failed catastrophically. The modern tornado warning system — watches, warnings, Doppler radar, Wireless Emergency Alerts — exists because the Tri-State Tornado made the cost of no warning system impossible to minimize. For the households who live in it today, the lesson is simple: the warning system was built from 695 deaths. Use it.
What Changed
The Tri-State Tornado's legacy in warning system development was gradual but direct. The event demonstrated, at the cost of 695 lives, that the policy of not warning the public about tornadoes was indefensible. The Weather Bureau's ban remained formally in place for years, but the event created pressure that eventually produced formal tornado forecasting capability. The Air Force meteorologists F.S. Fawbush and R.C. Miller issued the first experimental tornado forecast in 1948. In 1952, the Weather Bureau established the Severe Local Storms unit (SELS) in Kansas City — the direct predecessor to the Storm Prediction Center in Norman, Oklahoma, which issues tornado watches for the entire country today.
The NSSL May 3, 1999 documentation notes that significantly fewer losses in that event occurred "because of applied knowledge and new technologies developed through" meteorological research — a direct contrast to the zero-technology environment of 1925. The deployment of the NEXRAD Doppler radar network in the 1990s, which can detect the rotating winds inside a thunderstorm that produce tornadoes before the tornado touches the ground, is the single largest advance in tornado warning lead time since the Tri-State event. Average tornado warning lead time in 1925: 0 minutes. Average tornado warning lead time today: approximately 13 minutes. The distance between those two numbers is measured in 695 deaths.
What You Can Do Now
The Tri-State Tornado's lesson is not to fear what you can't control. It's to use the warning system that was built from 695 deaths — because no one in 1925 had access to what you have today.
Wireless Emergency Alerts automatically reach every cell phone in a tornado warning area. They don't require an app, a subscription, or pre-registration. Go to your phone's settings, confirm WEA is enabled, and verify the sound is not muted. In 1925, 695 people had no warning system. You have one that will wake you up at 3 AM if a tornado is approaching your area. Use it.
Tornado preparedness guideA tornado warning gives you an average of 13 minutes. That is not enough time to decide where to shelter, find a flashlight, locate your family members, and move to safety. Know your shelter location now: lowest floor, interior room, away from windows. If you have a basement, that's the answer. If not, identify the most interior room on the lowest level of your home. Walk to it once so the path is automatic. The Tri-State Tornado's victims had no warning. You'll have 13 minutes. Use them efficiently.
Tornado shelter guideMobile homes provide almost no protection against tornado winds. The Tri-State Tornado's death toll included a disproportionate number of people in structures that could not shelter them. Modern manufactured housing is more robust than 1925 frame construction, but it is not a tornado shelter. If you live in a mobile home park, know the location of the nearest community storm shelter and your route to reach it. A tornado watch — which means conditions are favorable for tornadoes — is your signal to identify your shelter route. Don't wait for the warning.
Mobile home tornado safetyWireless Emergency Alerts work when you're awake with your phone nearby. A NOAA Weather Radio, kept on the bedside table, provides a loud alarm even when your phone is silenced, your cell service is degraded, or the power is out. Tornadoes occur at night, and nighttime tornadoes kill more people per event because people are asleep and don't receive the warning. The Tri-State Tornado struck in afternoon daylight. Many of the deadliest tornado events in the past two decades — including Joplin 2011 — struck after dark in some areas. A weather radio is the insurance policy for the warning that might otherwise not wake you.
Emergency kit essentialsA tornado watch means conditions are favorable for tornado development in a broad area — prepare and monitor. A tornado warning means a tornado has been detected by radar or confirmed by spotters in a specific area — take shelter immediately. The distinction is the most important two-word vocabulary lesson in tornado preparedness. Many people treat a tornado watch with the urgency of a tornado warning, and many treat a tornado warning with the complacency of a tornado watch. The correct response to a warning in your area is to be in your shelter before the warning is issued.
Tornado warning guideTornado case study series
Joplin 2011 covers warning fatigue. Tuscaloosa 2011 covers the limits of visible warnings. Moore 2013 covers school shelter failures. Bridge Creek 1999 covers the limits of even the best warning system. Together, they cover every documented failure mode in tornado preparedness.
Full tornado case study seriesSources