Monday, February 18, 2013

"Death Rode Ruthless..."

 A modern reanalysis of official reports reveals that some 1,000 lives were lost in ‘Our National Calamity’

Death rode ruthless on the waters that night while hundreds saved their lives by what seemed miracles of chance,” wrote Ernest P. Bicknell, the National Director of the American Red Cross, long after the 1913 floodwaters had receded. Initial newspaper headlines wildly exaggerated estimated fatalities—but official reports significantly underreported fatalities. 

Banner headline of the Cleveland Plain Dealer for Wednesday, March 26, 1913, shows how newspapers greatly overestimated initial numbers. Such initial overestimation still happens based on best possible information in the first moments of crisis: recall that right after Sept. 11, 2001, first estimates of possible lives lost were hazarded to be as high as 50,000, the number of people who worked daily in the World Trade Towers, assuming all were present and none escaped. [Photo credit: Trudy E. Bell]

Just how many people did die as a result of the March 1913 natural disaster? And why are official reports so inconsistent?
First, to count a death, you have to know about it. Drowning in the flood or being crushed under tornado wreckage is obviously a death directly attributable to the disaster. So clearly is dying of a heart attack from terror, or succumbing to exposure after 48 hours in subfreezing temperatures on a roof or in a tree. But other disaster-related deaths—such as dying weeks later from a severe injury that developed tetanus or gangrene, or months later from pneumonia, typhoid fever, or smallpox contracted in a refugee camp—often occurred after official reports were published. While hospitals may have thought to report such a death to an agency collecting statistics, families of individuals who died at home likely did not—especially if they were immigrants not fluent in English or were among the uneducated poor (and make no mistake, the  poor were disproportionately hit because their homes were often on undesirable low land or even flood plain close to factories). Nor did authors of official reports have the resources to contact every county coroner across a dozen states for official causes of death. So in 1913 (and likely also in natural disasters today), official death counts are virtually certain to be below actual fatalities.
One of the sad and gruesome necessities in the 1913 flood was taking care of the bodies of those who died from drowning, injury, or pestilence. Here bodies are shown in the cavernous National Cash Register garage that was converted to an emergency morgue; coffins are stacked behind them, and mud-stained bundles of clothing lie at the feet of each shrouded body. Note the guard standing watch. [Credit: Miami Conservancy District]
Second, in principle you find all the bodies, line them up, and count them. But the 1913 tornadoes and floods were so violent that people simply vanished without a trace. For weeks afterwards, unidentified bodies were recovered down the Ohio River or even down the Mississippi. Absent DNA testing, when a half-decayed corpse was found, bloated and mangled beyond recognition, there was no definitive way to tie it back to a specific person missing in some state upriver. Thus, body counts are only partial counts. How partial? A year after the disaster in May 1914, George Burba, Secretary to Ohio Governor James M. Cox, declared: “After the waters had subsided, 428 bodies were recovered, and probably half as many more were never found.” Moreover, undercounting was worst precisely in those places where the calamity was most violent—where tornados whirled heavy objects (including bodies) for miles or where walls of water swept through a city, carrying away trees and houses (and people). Disappearance compounded the tragedy: In at least one instance, absent a body, a woman was barred from collecting life insurance to support her children when her husband was swept away.
Third, fatalities should be counted over the full multistate geographical area afflicted from Nebraska to New York and down the Mississippi. In 1913, however, not even the Federal government was able to achieve geographical completeness. Not for lack of trying. The U.S. Geological Survey and the Weather Bureau sought definitive statistics. But their sources of information were incomplete. The USGS, for example, sent out questionnaires to 200 cities over 5,000 population, but only 120 returned answers; moreover, cities under 5,000—of which there were (and still are) many in the Midwest—did not get polled. The American Red Cross, the U.S.’s official disaster relief agency since 1905, devised a stellar registration system for recording the plights of tornado and flood sufferers; but in 1913 the Red Cross was still sufficiently small that its few dozen agents and 236 nurses could not be everyplace—so they focused efforts primarily in the 112 hardest-hit communities of Ohio (including Dayton), primarily along Ohio’s five major rivers. Fully half the counties in Indiana decided to care for their own victims, so the Red Cross had even less of a presence. Outside of Omaha, Nebraska, and Lower Peach Tree, Alabama, its presence in other regions was sparse or nonexistent. No Red Cross, no records.
Some counts were done by organizations whose concerns were confined to smaller territories (states or even just individual river valleys) or specific areas of interest (e.g., flooding alone or tornadoes alone). Most notable was the work of the Ohio State Board of Health headed by Secretary and Executive Officer Eugene F. McCampbell. Physicians and sanitation engineers spread out over the state from flood week through the end of April, disseminating disinfectants and instructions, seeing that animal carcasses and supplies contaminated food were destroyed (yes, unscrupulous merchants tried to sell meat that had been submerged), and reporting on houses destroyed, people injured or left homeless, and deaths. The teams clearly note they did not visit every community nor examine all situations with their own eyes; they also noted that in several towns officials were antagonistic to their information-gathering or sanitation help.
The Board’s 145-page “Special Report on the Flood of March, 1913,” issued phenomenally fast in May, reported that in Ohio “approximately 430 lives were lost”—434 to be precise (totaling them on an Excel spreadsheet). However, for the Board’s year-end annual report in December—which is almost never cited in later histories— McCampbell adjusted some death counts in such a way as to call his first official numbers into question. He raised some numbers (e.g., those for Tiffin, from 19 to 30) while he reduced others (e.g., those for Hamilton, from 85 to 72, based strictly on body count). The December report also notes that in the city of Delaware, 18 bodies were found but another 21 persons were still missing. Depending on which numbers seem most acceptable, the resulting official death toll for Ohio is left somewhere between 422 to 470. In December, McCampbell doesn’t even hazard a final total.
No other state board of health or other agency appears to have mobilized a similar effort. One Federal report by the U.S. Public Health Service did visit five Kentucky cities to provide sanitation services and collect information, but did not record death counts or injuries(!).In Indiana, one official report documents only property damage along the lower White River in seven counties, but does not address deaths. The only other “official” records about the 1913 flood in Indiana are two brief typed manuscripts compiled in 1935 and 1937 in the Indiana State Archives, one of which consists of a compilation of statistics from newspaper reports. Figures in this latter, expressed in ranges, suggests that the death toll in Indiana was 100 to 200 lives—which may be fair given that Ohio’s topped 400 and that Indiana was as struck by surprise as Ohio.
The axiom (sometimes attributed to Carl Sagan) that “absence of evidence is not evidence of absence” comes to mind. In the absence of official numbers from other states, one unfortunate result has been that McCampbell’s numbers for Ohio published in May 1913 have been taken by subsequent historians to represent all fatalities in the entire natural disaster—with the possible implication that Ohio was the only state affected. That is a serious error. It might be one key as to the obscurity of the monumental scale of the storm system that stretched from Nebraska to New York (I remain baffled how something so enormous could be forgotten), including dissociating the flooding from the devastating tornadoes.
The Weather Bureau reported statistics about the entire geographical scale of the flooding—excluding the tornadoes—in its special Bulletin Z titled The Floods of 1913 in the Rivers of the Ohio and Lower Mississippi Valleys. It contains the only attempt at collecting definitive information on property damage from states bordering the Mississippi after the flood crest had exited the Ohio River. Ohio is given a death toll of 467 (tallied by county, not city, on pages 67-69); Bulletin Z also quotes an official investigation at Hamilton that put the body count at 92 but the “loss of life, probably” of 150—more than 50 percent higher (page 55). Death tolls are not provided for other states.
The U.S. Geological Survey in its Water-Supply Paper The Ohio Valley Flood of March–April, 1913, Including Comparisons with Some Earlier Floods drops the Ohio death count by an even 100 to 367—making one suspect a copying error from Bulletin Z. The Water-Supply paper, however, is one of the few reports with statistics about flood deaths outside of Ohio: It cites a death count of 39 for Indiana, 4 for West Virginia, 2 for each of Illinois and Pennsylvania, and 1 for Kentucky for a total of 48 flood deaths in those five states. Why should those numbers be so much lower than Ohio? Keep in mind, the flood was a rolling disaster. The flood crests of the upper Ohio River didn’t reach the Mississippi River until early April, and took another three weeks to roar down the Mississippi—bursting levees en route—until exiting out the Atchafalaya mouth of the Mississippi around May 1. Citizens of Paducah, Kentucky and other cities down the Ohio and Mississippi Rivers had days of warning, enough actually to build substantial wooden flood-refugee housing on higher ground in advance of the approaching flood crests. The death toll in Ohio—and also in Indiana—was an order of magnitude greater because the disaster struck with no warning. That being said, 39 flood deaths in Indiana seems so low as to be suspect as an undercount: after all, in Peru, Logansport, and elsewhere along the Wabash—whose headwaters are in Ohio along the same continental divide over which the heaviest rainfall concentrated—the flood came as unexpectedly as it did in Ohio.
Meantime, the Red Cross remained in Ohio doing relief work through August 1913. Possibly because of spending five months on the ground and being able to count additional deaths that were not immediate, the Red Cross cited higher numbers that kept edging upward. For example, although the Ohio State Board of Health ascribes 98 deaths to Dayton, but 1914 the Red Cross had upped Dayton's death toll to 116. By October 1913, Bicknell stated in an official report on donations and outlays for the disaster: “About 600 persons were drowned in the entire flood area.” Unfortunately, he did not specify what state(s) comprised that area. In 1914, Red Cross worker Winthrop D. Lane stated, however, that the flood “killed 625 people in two states alone,” clearly meaning Ohio and Indiana. It is not stated whether these figures are based on body count, but Bicknell clearly states “drowned.”

Four men carry a flood victim past the Dayton Courthouse in Ohio. Walking behind them are a man and a woman, likely family of the deceased. [Credit: Dayton Metro Library]
Given the Red Cross’s meticulous record-keeping for purposes of granting relief aid, let’s take those numbers at their word. So the total for flood deaths in Ohio and Indiana was likely about 625, likely including at least some who died much later from infection or other storm-related complications. Let’s also add the additional 11 deaths the USGS counted outside of Indiana, plus some reasonable allowance for people who were swept away but their bodies not recovered. In round numbers, that would bring the total to the neighborhood of 650 who died as a result of angry waters.
Let us not forget the Good Friday windstorm and the Good Friday and Easter tornadoes. For tornadoes, a widely recognized definitive secondary source is the two-volume reference Significant Tornadoes by Thomas P. Grazulis, who scoured newspapers on microfilm and other references in all 50 states to compile records of all tornadoes of F-2 and greater from 1880 to 1989. According to Grazulis’s research, all the Good Friday tornadoes in Alabama, Georgia, and Mississippi claimed 50 lives. All the Easter Sunday tornadoes in Nebraska, Iowa, Indiana, plus Louisiana and Missouri killed 192 people. Another seven lost their lives to tornadoes on Easter Monday. All those tornado fatalities add 249 deaths to the about 650 likely for the flood zone, bringing the death toll to about 900. 

Right-click on the image of the table to download the JPEG and view it larger on your own computer. Table ©2013 Trudy E. Bell
The real eye-opener is the death toll from the mammoth Good Friday windstorm that swept the eastern half of the U.S. from the Great Lakes to the Gulf of Mexico, not counting deaths from tornadoes. There are no official reports for these onesies and twosies, so numbers must be gleaned from local newspapers—but deaths at a few at a time are what local dailies cover best, including enough details about individual victims (name, town, profession) that duplication can be eliminated in AP wire stories run in several papers. Deaths caused by the straight winds from chimneys falling, worker blown from scaffolds, vehicles overturned, etc. total at least 66 across 11 states (Arkansas, Illinois, Louisiana, Michigan, Mississippi, Missouri, New York, Ohio, Pennsylvania, Tennessee,  and West Virginia)—a number tallied from just three Dayton newspapers in stories published on March 21 and 22, and thus not at all comprehensive. Yes, the Good Friday windstorm alone was so vast and powerful that hurricane-force straight winds killed more people across 11 states than did the Good Friday tornadoes that devastated Lower Peach Tree and elsewhere.
In short, the total storm-related death toll from Good Friday, March 21 (the Good Friday windstorm and tornadoes) through the following Friday, March 28 (the flood peak in the Hudson River) is conservatively greater than 900, and very likely over 1,000.
The sad truth is we may never know with exactness.
But two important things are clear. First, a realistic death toll for the full multistate extent of “Our National Calamity” was far greater than the 400s often cited for Ohio alone—in fact, evidence is strong that anywhere in the 400s is likely an undercount even for Ohio. Moreover, official death counts—in a natural disaster so violent that bodies were swept away or destroyed, and people died later of disaster-related injury or disease—should be regarded cautiously only as minimums.

©2012–2013 Trudy E. Bell. For permission to reprint or use, contact Trudy E. Bell at

Next time: So Many Ways to Die

Selected References
Bicknell, Ernest P., “The Ohio Flood of 1913—Our First Great Relief Task,” The Red Cross Courier, September 1914. 
Burba, George F., “The State’s Part in the Emergency,” [one article of a special section by American Red Cross authors, "When Disaster Comes"] The Survey 32 (5): 113-153, May 2, 1914.

Bybee, Hal P., and Clyde A. Malott, “The Flood of 1913 in the Lower White River Region of Indiana,” Indiana University Studies II (22): 105–223, October 1914.

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.

Henry, Alfred J., The Floods of 1913 in the Rivers of the Ohio and Lower Mississippi Valleys. Bulletin Z. U.S. Department of Agriculture. Weather Bureau. Washington, Government Printing Office, 1913.

“Historical Study of Floods in Indiana,” compiled in 1935. Plus O’Harrow, Dennis, “Indiana Flood Damage,” State Planning Board of Indiana, February 1937. Many thanks to Nancy Germano for providing me PDFs of these two typed manuscripts in the Indiana State Archives.

Horton, A. H. and H. J. Jackson, The Ohio Valley Flood of March–April, 1913, Including Comparisons with Some Earlier Floods, (Department of the Interior, United States Geological Survey, Water-Supply Paper 334, Government Printing Office, Washington, D.C., 1913).

Lumsden, L.L., Sanitation of Flood-Stricken Towns and Cities, with Special Reference to Conditions Observed in River Towns and Cities of Kentucky, U.S. Public Health Service, Public Health Reports, Reprint No. 131, June 13, 1913; Washington, Government Printing Office, 1913.

[McCampbell, E. F.] Twenty-Eighth Annual Report of the State Board of Health of Ohio. For the Year Ending December 31, 1913. (Columbus, Ohio: The F. J. Herr Printing Co., 1914.

McCampbell, E. F., “Special Report on the Flood of March, 1913,” reprinted from Monthly Bulletin Ohio State Board of Health, May 1913; pp 299–445.

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 shipping, complete with inscription of your choice; for details, e-mail me at )


No comments:

Post a Comment