Friday, May 1, 2015

Terror in Terre Haute

The violent tornado that ripped through southern Terre Haute, Indiana, on Easter night, March 23,1913, may have been more than one twister, and its full path of destruction extended over 25 miles

Lightning crashes repeatedly, luridly lighting the parlor where John Hanley and his family were trying their best to ignore the violent thunderstorm and enjoy being together the rainy night of Easter Sunday 1913. Then around 9:45 PM, over booming thunder and howling winds and drumming rain, Hanley hears a growing roar of what 
Oil painting, possibly of the Terre Haute tornado, was featured as the cover of a leaflet by the New York Underwriters Agency advertising tornado insurance. The unidentified location may have been of a rural area southwest or northeast of Terre Haute itself. If so, artistic license is liberal. The actual tornado struck not in sunlight but well after dark—nearly 10 PM Easter night—in the midst of horrific lightning and torrential downpour, and very likely people were not running across farm fields so near it. Credit: Ray Thomas collection of postcards on the 1913 flood 
sounds like a fast-approaching express train. He opens the front door—and beholds a towering tornado just blocks away, bearing down in his direction and sweeping up whole houses in its fury.

No time to run for the storm cellar—. Yelling he knows not what, Hanley gathers his family around him in the small hall to huddle behind the strong front door and its protective outer storm door. Seconds later, heavy timbers fly through the parlor window and across the room in a cascade of shattering glass. In moments, the beautiful home is wrecked, along with Hanley’s three-story warehouse of awnings and construction materials behind it. Had the family remained seated in the parlor, all five would have been killed.

The destroyed Hanley house likely looked something like the Dix house, shown here, the morning after the tornado roared through Terre Haute. Credit: Terre Haute’s Tornado and Flood Disaster, Wabash Valley Visions and Voices
In less than two minutes, the tornado roars across southern Terre Haute, Indiana, destroying some 250 homes, ruining businesses and livelihoods, and killing a score of friends and neighbors and maiming hundreds more. Along the path of destruction—which varies from 100 to 500 yards wide—fires ignite as crossed electrical wires short, gas lines burst, and glowing coals from overturned kitchen and bedroom stoves that cold  night scatter onto carpets, bedding, and curtains. Through the torrential downpour falling on the ruins of human lives rise wails of agony and mourning. 

Wide-angle view of a few blocks of destruction a day or so after the Terre Haute tornado. Note that many people had umbrellas, as heavy rains were continuing, and in the next day or two flooding was widespread. Credit: New York Underwriters Agency advertising leaflet in Ray Thomas collection of postcards on the 1913 flood 
Reconstructing Terre Haute’s disaster
One long-standing mystery to me has been the fact that today the Terre Haute tornado is remembered just for striking one portion of one city, as if it touched down there and nowhere else. Moreover, text references to it both then and now as “the Terre Haute tornado” imply the assumption that it acted completely alone. Yet, violent tornadoes are more typically part of a larger rotating regional-scale supercell thunderstorm system that tends to generate multiple tornadoes—as indeed happened four hours earlier that same night Easter Sunday, 1913 in Omaha, Council Bluffs, and elsewhere across Nebraska, Iowa, and Missouri (see “‘My Conception of Hell). And as strong vortices, they also tend to persist along paths miles long. 

The Terre Haute tornado destroyed the factory buildings of the Root Glass Works, but did not destroy the company itself, which two years later (1915) went on to design and patent the iconic Coca Cola bottle, this year celebrating its centennial. Credit: Engineering News
So for years, my big questions were: did the Terre Haute tornado indeed act alone? And what was the full extent of its path of destruction? To research those questions, a year ago (April 2014), I photocopied articles on the Terre Haute tornado from microfilmed pages of 10 local 1913 newspapers in Vigo and surrounding counties housed at the Indiana State Library in Indianapolis.

The path of the Terre Haute tornado never was mapped either at the time or later—or if it was, such a map seems never to have been published in local newspapers or in Monthly Weather Review, the official journal of the U.S. Weather Bureau. But the commemorative booklet Terre Haute’s Tornado and Flood Disaster, March twenty-three to thirtieth, nineteen hundred and thirteen issued by the Terre Haute Publishing Co. and heavily relying on newspaper accounts and photographs, compiled many individual stories—many of which include names and street addresses of victims and of buildings destroyed.

Cover of the commemorative booklet Terre Haute’s Tornado and Flood Disaster, March twenty-three to thirtieth, nineteen hundred and thirteen issued by the Terre Haute Publishing Co. Street addresses in this booklet allowed me to plot the destruction of the tornado through Terre Haute. Credit: Wabash Valley Visions and Voices

So with the aid of Google Maps, I spent an entire day plotting scores of 1913 addresses on a modern map of Terre Haute to see what emerged.

Map of the southern half of today’s city of Terre Haute, plotting the location of damage by address given in the 1913 commemorative booklet Terre Haute’s Tornado and Flood Disaster. Credit: base map Google Maps; 1913 tornado damage plot Trudy E. Bell
Several revelations emerged. First, the city of Terre Haute in 1910 was Boomtown, USA. It had almost the same population as it does today: over 58,000 compared to 61,000, making it then one of the nation’s top 100 populous cities. It was also growing fast, Even so, its city limits were smaller and surrounded by fields and farmland instead of urban sprawl and suburbs (today Terre Haute’s entire statistical metropolitan area encompasses over 170,000 people). 

Second, street numbering and names today must differ on some streets. Google Maps could not plot any of the addresses in the booklet given for Lockport Road, so those data are missing from my map. Neighborhoods must have also changed names. For example, the booklet states that tornado damage was particularly bad in Krumbhaar Place, “the new sub-division recently opened on the south side of the city”; I could find no subdivision with that name today, just a single Krumbhaar Street in what might be the approximate area. Another hard-hit area I could not find was Gardentown (also spelled Garden Town), apparently an unincorporated community six or eight miles south of the city just north of Prairieton and largely devoted to truck farming for fruit and vegetables and greenhouses for florists. Appeal to readers: If you know more about the historical geography of Terre Haute, please contact me.

Another general view of tornado destruction in the Terre Haute. Credit: Terre Haute’s Tornado and Flood Disaster, Wabash Valley Visions and Voices

Third, it is clear from the booklet’s text that several newspaper reporters or other authors sought to be as thorough as possible, clearly visiting hospitals and walking along ruined streets. But the accounts are jumpy in geography and some of the anonymous writers were more complete than others in specifying locations.

Nonetheless, the map I was able to construct of the tornado’s path of destruction through Terre Haute reveals tantalizing structure. Are the variations in width due to actual variation in width of the tornado’s funnel of destruction, or merely incompleteness or limitations in data gathered or published? Do separations in areas of destruction reveal that the tornado hopped along its path, or did it just pass through what were open fields in 1913 until encountering another group of buildings?

And could it have been a multiple-vortex tornado with several small, short-lived smaller subvortices that orbit around the main funnel: subvortices that actually deal some of the worst death and destruction? 

Multiple-vortex tornado with half a dozen small, short-lived, but exceptionally violent subvortices, photographed near Altus, OK on May 11, 1982. Credit: U.S. National Oceanic and Atmospheric Administration (NOAA)
A 2013 article by Mike McCormick for the Terre Haute Tribune-Star written for the centennial of the Terre Haute tornado describes it as a “multi-funneled tornado” shortly before 11 PM. But the article cites no reference for either the time—which is clearly documented as 9:45 PM in Monthly Weather Review and elsewhere—or the assertion about multiple funnels. 

McCormick also writes that “According to several witnesses, the storm cloud divided. One section, containing two shafts [sic], headed east between Hulman and Washington streets. The other went northeast, killing Dr. Mahlon Moore at his office at 629 College Ave.” I plotted McCormick’s locations on my map in blue (to distinguish them from the locations cited in the booklet). However, I have not yet found any 1913 primary source for his statements, although the commemorative booklet does show a photograph of Moore’s devastated office (with no address). 

Ruins of office of Dr. Mahlon Moore; if the address given by Mike McCormick is correct, might Moore have been killed by a subvortex? Credit: Credit: Terre Haute’s Tornado and Flood Disaster, Wabash Valley Visions and Voices 
But if McCormick is right about the location of Moore’s office, that would be strong evidence for another vortex some distance from the main funnel. Moreover, I wonder whether the patches isolated damage removed from the main path of my map suggests the possibility of damage from subvortices, which last only a few seconds but are exceptionally destructive.

The map of the destruction I compiled from the booklet, plus the booklet’s stated variations in the width of destruction, is tantalizingly suggestive of the cycloidal marks carved into farm fields from multiple-vortex tornadoes.  

Cycloidal marks in farm fields left by a multiple-vortex tornado. Credit: U.S. National Oceanic and Atmospheric Administration (NOAA)
25+mile path
Both the booklet and local newspapers published immediately afterwards in and around Vigo County and in Indianapolis clearly describe additional tornado destruction in and north of Prairieton—a town of about 700 population 8 to 10 miles southwest of Terre Haute. The Brazil Daily Times and The Crawfordsville Journal also detailed damage in Perth, an even smaller town (400 population) about 20 miles northeast of Terre Haute, as well as in Glenn and East Glenn, the western part of Seelyville, and Ehrmandale in between. The northeasternmost report of damage was a mile and a half west of Carbon. 

Plotting those areas on a broader-area map suggests that the path of the Terre Haute tornado could have been 25 to 30 miles long, as the paths line up nicely. There is also the possibility of the three areas of destruction being wreaked by different twisters, but almost no newspaper accounts indicate the time locations were hit, which would be essential in sorting out the truth. 

Map of tornado damage in various locations reported in half a dozen local newspapers reveals that the path of the Terre Haute tornado was at least 25 miles long. Credit: base map Google Maps; 1913 tornado damage plot Trudy E. Bell
Plotting specific locations on the streets of Prairieton, Seelyville, and Perth was almost impossible: in such small communities, clearly everyone knew everyone else and local landmarks, so destruction is described only by giving the owners’ names without street addresses or just the names of local parks long gone. That makes it almost impossible for someone a century later without detailed knowledge of local history or access to public records of property ownership to map the extent of damage. Again, I welcome contact from any reader who can help.

Why was more information not preserved about the path and timing of the Terre Haute tornado? Reporters in Omaha and Council Bluffs and elsewhere did history a huge service in preserving a very detailed and thorough record of the family of 10+ tornadoes that struck Easter night 1913. Why is the record sketchier in Indiana?

Ruins of Olson house. Note umbrellas, as it was raining hard and flooding followed a couple of days later. Credit: Credit: Terre Haute’s Tornado and Flood Disaster, Wabash Valley Visions and Voices

The answer dawned when I was photocopying the newspapers on microfilm in the Indiana State Library: tragically, the city of Terre Haute was unique in suffering both violent tornado damage (like Nebraska, Iowa, and Missouri) and record flooding (like Ohio and other states) in the Great Easter 1913 storm system. Indeed, Indiana, like Ohio, was at the epicenter of the 1913 flood. On Easter Sunday, rain in Terre Haute was already heavy, and floodwaters began overflowing river banks the next day. Not only did record-high floodwaters confront Terre Haute residents with more urgent worries than tracing a tornado’s path through the open countryside, but also nature itself was immediately obliterating that very evidence. 

Death undercount
Published death counts for the Terre Haute tornado range from 17 to 21. Seventeen—the number given in the booklet—is a clear fact-checking error and significant undercount: simply cross-checking the names of fatalities described in the booklet’s text with the names given in “Toll of the Tornado” reveals the omission of at least three people whose bodies were discovered: Mrs. Moses Carter and Mrs. Leonard Sloan and her day-old infant. Also, The Crawfordsville Journal reported “one or more” people killed in Prairieton. So the verified minimum is no fewer than 20 killed, and perhaps closer to 23.

And of course, as discussed already in a detailed analysis of fatalities during the Great Easter 1913 storm system and flood (see “‘Death Rode Ruthless…’” ), people injured by the Terre Haute tornado could have died weeks or even months later of complications and thus not have been counted as tornado deaths at the time the booklet was published.
© 2015 Trudy E. Bell

Next time: Never Before Seen

Selected references
Special thanks go to Ray Thomas for high-resolution scans of the New York Underwriters Agency leaflet and permission to use images from his amazing website of postcards from the 1913 flood

In addition to the sources already cited in the text, these also proved especially useful:

“Big Storm Passes West of Brazil” and “Damage Near Carbon,” both in The Brazil Daily Times, March 25, 1913, p. 1.
Edwards, Roger, “The Online Tornado FAQ,” U.S. National Oceanic and Atmospheric Administration 
Grazulis, Thomas P., Significant Tornadoes, 1880-1989. St. Johnsbury, VT: Environmental Films, 1991. Classic and fascinating two-volume reference detailing virtually every U.S. tornado F2 and greater for more than a century. Grazulis now runs The Tornado Project.

“Perth in Path of Disastrous Storm” The Brazil Daily Times, March 24, 1913, p. 1.

Shannon, Charles W., “Soil Survey of Clay, Knox, Sullivan and Vigo Counties, Indiana,” Thirty-Sixth Annual Report of Department of Geology and Natural Resources, Indiana 1911, Indianapolis, 1912, pp. 137–280. Brief description of Garden Town is on page 275.

 “Tornado and Flood Damage at Terre Haute, Ind.,” Engineering News 68(15): 738–739, April 10, 1913.

“Tornado at Terre Haute, Ind., March 23, 1913,” Monthly Weather Review 41(3): 483–484, March 1913.

Bell, Trudy E., The Great Dayton Flood of 1913, Arcadia Publishing, 2008. Picture book of nearly 200 images of the flood in Dayton, rescue efforts, recovery, and the construction of the Miami Conservancy District dry dams for flood control, including several pictures of Cox. (Author’s shameless marketing plug: Copies are available directly from me for the cover price of $21.99 plus $4.00 shipping, complete with inscription of your choice; for details, e-mail me), or order from the publisher.