In the winter of 1811–1812, a series of very large earthquakes shook much of North America, rolling out from the middle Mississippi Valley to be felt strongly all along the Ohio River, in many Atlantic coastal regions, and as far north as Upper Canada. These temblors soon became known as the New Madrid earthquakes, after the small Mississippi River port near the areas of greatest disturbance. Understanding of these earthquakes has shifted radically over the last two hundred years with the successive remembering and forgetting of first-hand descriptions of the events. Accounts of the quakes originally generated to share and attempt to understand widespread natural upheaval were collected and employed as scientific evidence in the early nineteenth century, ignored when scientific instruments promised more quantitative and exact knowledge, then returned to years later as the earthquakes themselves became both more important and less understood because of changes in scientific models.

When the New Madrid earthquakes initially struck, people described them in letters and more formal reports throughout newspapers and journals. Witnesses couched their perceptions in the form of narratives, describing what they had felt, heard, smelled, and done during the shocks. These accounts were a form of conversation: they contributed each writer's particular perspective on events that were generally agreed to be intelligible only when many individual accounts from many different places could be collected and compared.

By the turn of the twentieth century, dramatic changes in seismological work had rendered such narrative accounts obsolete. Seismographs became the way to understand the Earth's movement – their written record gave exactitude that promised global reach, unlimited by the parochial perspective of breathless individual stories. On the centennial of the quakes, geologist Myron L. Fuller compiled and studied the original nineteenth-century accounts, but many decades passed before other researchers extended his research. The radical explosion of plate tectonic theory in the 1960s confirmed researchers in their lack of interest in events in the Mississippi Valley. Instrumental data from plate boundary regions such as Japan and the American West Coast gave researchers compelling and abundant material with which to work, even as plate tectonics suggested research questions far removed from narratives of the American heartland. Written accounts of the New Madrid earthquakes faded into obscurity as frontier tall-tales, and the earthquakes themselves became regarded as long-ago curiosities – indeed, barely regarded at all.

Yet gradually, in the 1980s and 90s, a few seismologists began to insist that the events at New Madrid were significant and worthy of study – and that the aged written record might pose interesting questions for the new plate tectonic model. Earthquakes powerful enough to be felt across the country? Earthquakes that had opened enormous sandblows? Earthquakes of such apparent power, hundreds and hundreds of kilometers from a plate boundary, in what was thought to be a seismically stable continental interior? The written record of these otherwise obscure nineteenth-century accounts insisted upon this strange environmental truth. In 1973, Saint Louis University seismologist Otto Nuttli pushed mainstream seismology to re-examine the environmental and textual basis of Myron Fuller's New Madrid earthquake history. With increasing intensity over the last two decades, researchers of mid-continent seismicity have responded, giving credence to New Madrid accounts and sifting through old archives and newspapers to find more.

Today, so-called intraplate or stable continental region earthquakes pose a significant problem in seismology. What, after all, causes seismic upheaval in what is otherwise a quiet continental interior? Are powerful quakes likely to recur? The New Madrid events offer one of the most significant examples upon which researchers can draw. Two-hundred-year-old written accounts by those who experienced them firsthand are a main piece of evidence for this pre-instrumental series of earthquakes, and they are today both prized and dissected, as researchers take them apart in search for answers about extent, size, and meaning of these quakes. This renewed focus on the textual reports of the New Madrid earthquakes bears important parallels with early-nineteenth-century investigation of the quakes: researchers are, after all, focused on exactly the same pieces of paper (or their electronic reproductions). Yet the meaning of these accounts has subtly but significantly changed: no longer narrative contributions to a sprawling, dispersed conversation about causes, questions, and puzzles, first-hand New Madrid accounts appear instead as textual data divorced from the context of exchange and mutual knowledge-building. They are mined as texts but no longer read as narratives, their communicative content no longer valuable to the experts whose conclusions now build our science of mid-plate earthquakes.

Following bits of paper through time in this search for meaning in the early American “Far West” reveals how narrative accounts come in and out of visibility. In very recent research, earth scientists become historians, looking for historical evidence with historians’ tools. Yet there is change even in seeming continuity: researchers have recently turned back to long-neglected narrative sources, but with an entirely different perspective on how to read and use them. No longer part of a conversation, accounts of the New Madrid quakes become dry pieces of evidence. In tracing how these puzzling North American earthquakes were experienced, remembered, forgotten and then rediscovered, writing the history of texts – the original accounts and the landmark papers that call them into visibility at different moments – becomes a way to write a history of earthquake science over two centuries.

Earthquakes of the Mississippi Valley

The facts of the New Madrid earthquakes remain an open question. Fierce debate rages in earth sciences literature about the most basic parameters of the quakes: How many main shocks were there? How big were they? What fault were they based on? To say “what really happened” is a literal as well as philosophical impossibility.

All observers agree on at least some basic outlines. Beginning in the middle of the night of 16 December 1811, the middle Mississippi valley was shaken by three (or perhaps four) major tremors, along with swarms of mid-sized and smaller ones.Footnote ¹ Main shocks occurred on 16 December, 23 January, and 7 February. Significant felt tremors continued in dense clusters for years after the 1811–12 events, tapering off only in the early 1820s. Seismic shaking continues to be felt by animals, people, and instruments today. Because the quakes were felt alarmingly as far away as the east coast, and because reports at the time and lasting environmental evidence all agree on certain spectacular disturbances around the epicenters, modern researchers concur that the quakes were very large. How large? Estimates of moment magnitude by credible authorities have ranged from 7.0 to 8.7, though most current ranges are closer to 7 than 8 (Hough Reference Hough, Fréchet, Meghraoui and Stucchi2008, 355; Hough and Bilham Reference Hough and Bilham2006, 79). Since earthquake magnitude is a logarithmic scale, this is a huge range indeed.

Accounts of the quakes immediately began to fill private correspondence and public media. People recorded and exchanged impressions, memories, and information all through that icy winter and the next spring – and indeed for a number of years after. People in the middle Mississippi region of present-day southeast Missouri and northeast Arkansas generally felt the most drastic shaking, yet even people far distant, in major cities along the east coast and communities in what was then Upper Canada, sometimes felt or heard significant indications of the “hard shocks.” These quakes were by no means limited to the Mississippi Valley, and curiosity about them was widespread in North America.

The tremors had effects on the lands of the Mississippi Valley and on the river itself. A portion of western Kentucky lurched upward, raising itself up into an enormous dome, still visible today as a seemingly gentle bulge in flat floodplain. The sudden uplift turned the continent's central river back on itself: the mighty Mississippi ran briefly but dramatically backward before the current wore down the new-made falls. Huge tremors during all three major shocks created enormous waves that tore into boats and ate away huge swaths of riverbank. At least one group of people, Indians traveling along the river, was killed by these vast underminings of riverbank. Numerous other river travelers perished on the river itself. People on flatboats saw wreckage and empty craft floating downriver long after the hardest of the shocks. On the Mississippi River, long-buried trees rose again to the surface along with new snags and deadly sawyers (uprooted trees whose heavy roots sank to the river's bed but whose branches bobbed dangerously at the surface). Several whole islands disappeared, confounding navigation. The quakes created new lowlands in northeast Arkansas, raised up portions of southeast Missouri, and leveled the southeast Missouri village of Little Prairie, which was abandoned. The uplift of western Tennessee dammed a smallish creek, creating the substantial and still existing Reelfoot Lake (Fuller [1912] Reference Fuller1995; Hough and Bilham Reference Hough and Bilham2006, chap. 4; Smith Reference Smith1812; Flint [1826] Reference Flint and Grattan1932, 219; New Madrid Compendium n.d.; and Owen Reference Owen1856, 117–119 and Plate 6).

Throughout southeast Missouri, northeast Arkansas, and western Tennessee, the earth's movement produced “earthquake cracks,” some more than a mile long and apparently meters deep (most began eroding immediately so their depth and extent is the subject of yet more debate). In southeast Missouri and eastern Arkansas, pressure from the earth's movements geysered fountains of liquefied sand up into the air – some of them twenty or thirty feet up, or even higher – to create immense, round “sand blows.” The quakes thus re-made the local landscape in dramatic and visible ways, rending the surface of the earth, spewing forth material from underground, and scarring the landscape. Yet such changes to soft-soiled floodplain were vulnerable to swift erosion. Unlike fault lines that tear rocky outcrops, these quakes left comparatively few unambiguous environmental traces. The narratives of people at the time remain a main form of evidence recoverable at ever-increasing historical distance.

Earthquake Accounts of the Early Nineteenth Century

When touched by these bewildering events, people all over eastern North America wanted to understand what they had experienced. Where had these earthquakes been felt? Did people in different locations feel them at the same time? Were they based in a particular geographic region? (Slow communication and uncertain mails made these hard questions to answer in 1811 and 1812.) What causes such a mighty “commotion”? In early-nineteenth-century discussion of quakes, these various concerns related and intertwined.

Accounts of the New Madrid earthquakes took their form from familiar ways of recounting important events and processes. Earthquake reports of the early nineteenth century were generally written as letters. Like reports of new settlement sites, relations with Native groups, or recent sickness, many accounts of the quakes began as letters from one individual to another person or family, but were then sent to a local paper. Particularly vivid recountings were reprinted several times over, appearing often in newspapers successively further East. Some accounts form barely a paragraph; others might run for pages of studied comparison and commentary. They generally all partook, however, in the sense that they were one person telling, narrating, experience to others. “[T] the most extraordinary effect I saw,” wrote a man named James Fletcher from Little Prairie, in a letter later reprinted in a Chillicothe, Ohio, newspaper, “was a small lake. . .the bottom of which is blown up higher than any of the adjoining country, and instead of water is filled with a beautiful white sand” (“The Earthquakes” 1812). Accounts like Fletcher's were a personal telling of experience which could then be shared (for an accessible compilation of these accounts, see New Madrid Compendium n.d.).

Common elements of the accounts included a narrative of the author's experience during the worst of the tremors and a description of the major environmental effects. From the beginning, writers emphasized how hard it was to understand what was going on, or why such upheaval was occurring. People often thought themselves under sudden attack – the huge booming noises of the largest shocks frequently sounded to frightened people like artillery (Broadhead Reference Broadhead1902, 82; Smith Reference Smith1812, 32; Sargent Reference Sargent1815, 354). In many places, witnesses saw flashes of light, or a dull gleaming glow (Smith Reference Smith1812, 21). Those further away noted changes to water sources that often involved befouling stench (Montgomery Reference Montgomery1812). Sulfurous stench might be a clue to the volcanic nature of earthquakes (a common theory) or a sign of the nearness of eternal damnation. Or, perhaps, both: God was never far from such conversations, even as observers struggled to find naturalistic explanations for seismic tumult.

In this profoundly agricultural world, observers frequently sought out correlations with weather. Former Mississippi Territorial Governor Winthrop Sargent experienced cloudy weather and rain along the Mississippi River before many of the larger shocks, but then also recorded early on the morning of Tuesday, 20 December, “a shock of greater severity and duration than any of the preceding – the sun rose brilliant with a clear sky” (Sargent Reference Sargent1815, 357). Sargent's apparent surprise at the calm clarity of morning weather in the face of a great quake indicates the way in which seemingly flat observations carried implicit arguments – here, about the failure of expected correlations between disruption below the earth and in the sky.

Most nineteenth-century accounts contain little in the way of explicit theoretical declarations, but that does not mean theorizing was absent. Instead, as with Sargent's weather observations, people often tried to work through what must have happened by figuring out what they felt and what their sensations meant. Reports frequently included extensive discussion of the apparent directionality of the quakes and their associated phenomena. Cherokee leader Charles Hicks wrote to missionary friends about the “great perplexity” of his community in the Cherokee nation after the first of the hard shocks: “Our living quarters were moving so much that they seemed close to falling down. Just before this a loud noise was heard from the north-northwest and some people saw a flash of light from right where the noise in the air had begun” (letter of 17 Dec 1811, quoted in McClinton, Gambold, and Gambold Reference McClinton2007, 461). Such reasoning led many to conclude that the quakes came from much farther away, in the west – perhaps the site of a yet unknown volcano. While often seemingly a-theoretical, earthquake accounts carried musings or arguments about causes and effects, evidence and truth, in their structure and emphasis: it is interesting to note where noises or tremors come from, since that might reveal what caused them; smells of sulfur might be a useful clue to volcanic origin; careful observation of weather surrounding the tremors might provide the key to unlock the complex relationship between climate and earthquakes.Footnote ²

Familiar narratives of health and illness helped give form to scientific accounts. The heaving earth was often figured as a living body. William Leigh Pierce wrote of his experience on the Mississippi River near the epicenter of the first December shocks that “For the first two days, the veins appeared to run a due course from W. to E., afterwards they became more variable, and generally took a N.W. direction” (An account of the great earthquakes . . . 1812, 11). Just as those by a bedside would write letters to those further off explaining the coming on or drawing off of fevers, the gradual increase or sudden explosion of boils or inflammation, the alternately restive or calm manner of someone agitated by high temperature, those experiencing convulsions of the earth told their stories with an attention to detail and to the changing balance of forces of disturbance.

Early-nineteenth-century earthquake reports were narratives: they tell a story, recounting the unfolding of a process. Those who experienced earthquakes noted the weather preceding quakes, the behavior of animals, unusual breezes, the sounds of the tremors, the way in which air seemed to change or stay still during them, and the ways in which movement or noise or smell or lights or a combination of all those would come on, grow in strength, and then fall off. Their language and their concerns were those of people waiting by a bedside for the consequential shift that would tell whether a grave illness would lift or intensify.

What Was the Extent of the Disturbance?

A main interest for many people who composed the accounts of New Madrid disturbance was to figure out how widely shared their experiences were (see “A Subscriber” 1811). Whenever possible, writers compared reports thus far to estimate the extent of disturbance (“Several shocks by earthquakes. . .” 1812). “We have not heard how far the shocks have extended into the country,” concluded one early report of the first tremor from Charleston, “except that they were felt at Rantowless [a nearby community]” (Smith Reference Smith1812, 8). William Leigh Pierce, writing 10 Jan 1812, described the range of felt effects of quakes in some detail: the earthquake was “by no means confined to the Mississippi. It was felt in some degree throughout the Indiana Territory and the states of Ohio, Kentucky and Tennessee. . . .From thence it ranged the Ohio River, encreasing in force until it entered the Mississippi, and extending down that river to Natchez, and probably a little lower. Beyond that it was not perceived” (An account of the great earthquakes . . . 1812, 13). In such narratives, readers mapped for themselves the connection of felt effects that bound together portions of their nation both old and new.

Newspapers that carried news of the New Madrid quakes frequently also carried snippets of reports about earthquakes elsewhere (see “Severe shocks of an earthquake were felt in France on the 12th and 18th of December 1812”). This in part reflects the extent of reporting about New Madrid: the quakes continued for many months, and reports about them came out over a long period, so a number of earthquakes took place elsewhere during that time. But such coverage speaks to a deeper facet of early-nineteenth-century scientific conceptions: a sense of global interconnection, in which experiences in one part of the world may indicate a more profound disturbance elsewhere.

Many of those writing about the tremors of 1812 linked them to upheaval around the globe. The first volume of the Transactions of the Literary and Philosophical Society of New York was largely given over to a series of reports by noted physician, scientific thinker, and Congressional representative Samuel Mitchill on the environmental disturbances of 1811–12, from earthquakes in North and South America to volcanoes in St. Vincents, explosions and convulsions in the Azores, even a “terrible snow storm” that had raged along the Atlantic seaboard from Cape Hatteras to Massachusetts Bay in late December (Mitchill Reference Mitchill1815b; Mitchill Reference Mitchill1815e; Mitchill Reference Mitchill1815a; Mitchill Reference Mitchill1815d; Mitchill Reference Mitchill1815c). Not mere curiosities, these were terrestrial symptoms of deeper and powerful disturbance. An article in August of 1811 in a Vermont paper reprinted news from London, reporting “dreadful forms and convulsions of nature in different parts of the world” – floods of lakes and oceans and “strange fluxes and refluxes of the sea.” These were possible clues to a larger disturbance: “Those who recollect that similar events took place before the great earthquake at Lisbon are prepared to expect, that a similar convulsion has taken place in some distant part of the globe” (“Foreign News, London” 1811). Tremors and noises functioned as a barometer to the internal state of a dangerously disrupted earth.

Powerful bodily metaphors shaped interpretation and narrative. Earthquakes were a symptom of the earth that could indicate far-off danger, just as inflammation or disruption in one particular part of the body could, in the nineteenth-century view of bodily systems, indicate an ailment or imbalance elsewhere. Frameworks of health and well-being helped shape conversation about terrestrial as well as somatic processes (Valencius Reference Valencius2002).

A good deal of theorizing focused on the curious sequence of many large earthquakes.Footnote ³ In the fall of 1755, within less than a month, large earthquakes leveled Lisbon, Portugal, and shook New England. Then in the spring of 1812, barely weeks after the huge tremors that shook the Mississippi and Ohio Valleys, another enormous quake in Venezuela devastated Caracas and helped bring about the end of the short-lived Venezuelan revolutionary government (McCook Reference McCook, Buchenau and Johnson2010). Like Samuel Mitchill, a number of thinkers posited connections between these seemingly paired catastrophes (for example, Duden [1829] Reference Duden1980, 159). Missionary, travel writer, and man-about-the-Mississippi-Valley Timothy Flint wrote of the season of New Madrid quakes that on one night in particular people observed “continued glare of vivid flashes of lightning, and repeated peals of subterranean thunder, seeming to proceed. . .from below the horizon.” These signs were significant not simply for that particular local environment, but also for telling the truth of events far distant: “the night, so conspicuous for subterranean thunder, was the same period in which the fatal earthquakes at Carraccas occurred, and they seem to suppose these flashes and that event parts of the same scene” (Flint [1826] Reference Flint1968, 217, reflecting common nineteenth-century spelling of the city). Such accounts look within the earth for explanation but are also concerned with knitting together events across the globe into one coherent story of environmental disruption, with perhaps a central set of causes. They bear witness to the search for a global science, even along the American frontier.

Letters and Their Collation as a Form of Scientific Evidence

In this period, letters were a crucial form of scientific conversation. Interpersonal correspondence was, after all, the only way of gathering information in an emerging scientific context that lacked centralized system (Goldstein Reference Goldstein1994; cf. Fan this issue). Although by the 1820s European researchers were beginning to collect historical earthquake accounts to look for seasonal, meteorological, or geographical patterns, little parallel effort had yet been mounted in North America (Oldroyd et al. Reference Oldroyd, Amodor, Kozák, Carneiro and Pinto2007, 343–4). Collations of many individuals’ accounts served as the beginnings of such seismic catalogues. And collations and collections were a crucial way in which knowledge of the quakes dispersed and ramified.

Newspaper reports in the weeks and months after the quakes collated witnesses’ accounts and offered brief commentary and comparison, often with reference to similar historical earthquake accounts. Philadelphia printer Robert Smith and a press in Newburyport, Massachusetts, quickly bundled these into longer collections of the same essential framework of first-person accounts accompanied by commentary (Smith Reference Smith1812; An account of the great earthquakes . . . 1812). Samuel Latham Mitchill's 1815 report compiled first-hand accounts and added extensive appendices of observations and theories about other earthquakes worldwide (Mitchill Reference Mitchill1815b). These compilations were the milestones in the early-nineteenth-century era of understanding of New Madrid, in which individuals’ felt experiences were part of a conversation about the causes of natural phenomena. By the turn of the century, changes in technology and scientific models turned attention within the earth sciences elsewhere.

Text and Context in the Age of Instrumentation

Two developments usurped the primacy of written narratives for understanding earthquakes: seismometers and plate tectonics. Both of these revolutions – the instrumentalization of seismic observation in the decades surrounding the turn of the twentieth century and the reconceptualization of the earth's composition and movement in a stunningly swift period of the early 1960s to mid 70s – created new views of earthquakes and seismic action. Both offered new insights. Both transformed and pushed forward earth science. Both led to a near total erasure of the first-hand New Madrid accounts, as evidence not accessible to instrumental observation and not well explained by the plate tectonic model.

Seismology, Jesuits, and Instruments in St. Louis

Over the first two-thirds of the twentieth century, what study there was of New Madrid seismicity took place because of the instruments and institutional focus of the earth and atmospheric sciences program at Saint Louis University, a Jesuit institution and bedrock of St. Louis intellectual culture. Saint Louis University formed a mid-American node in the sustained, long-distance networks of careful and educated Jesuit observers that were integral to the broad-based sciences of earth and sky (Geschwind Reference Geschwind1998; Udías Reference Udías and Kölbl-Ebert2009; Udías and Stauder Reference Udías and Stauder1991; Udías and Stauder Reference Udías and Stauder1996). SLUFootnote ⁵ was an early center of weather study in the continental interior: two Jesuit professors from the Theology and Philosophy program began sending meteorological observations to federal government observers in 1860. Founded in 1818 as the first university west of the Mississippi, Saint Louis University was proud to host the first department of geophysics in not just the trans-Mississippi West, but the western hemisphere (it was created in 1925). In 1909, in the wave of American earthquake investigation that followed the 1906 San Francisco earthquake, the program installed a Weichert inverted pendulum seismograph in the basement of a university building (Saint Louis University n.d.; Saint Louis University Department of Earth and Atmospheric Sciences n.d.-a). SLU's instruments soon anchored the Jesuit Seismological Service, a network of seismographs established after 1909 at Jesuit schools across the U.S. and Canada, at the time the only continental network employing uniform instrumentation. Yet cooperation proved difficult to sustain and the network was shortlived, ceasing in 1922 (Udías Reference Udías and Kölbl-Ebert2009, 140; Udías and Stauder Reference Udías and Stauder1991 and Reference Udías and Stauder1996; Saint Louis University Department of Earth and Atmospheric Sciences n.d.-a). It would be revived three years later as part of continued development of instrumental seismology.

Such instrumentation and institutional cooperation held powerful consequences not only for how earthquakes could be observed, but who observed them. By the first few decades of the twentieth century, many researchers were skeptical (or stoically resigned) about the role of public observation. Rather than observation by happenstance, dependent upon individual initiative, earthquake science required coordinated effort by large groups working under central direction (Davison Reference Davison1978, 184; Coen article, this issue). To maintain, much less read and compare, often delicate instrumentation required teamwork over long periods, routinization of methods as well as machines, increasingly specialized spaces to anchor and operate seismological sensing equipment, and increasingly specialized staff to tend to and interpret it, from secretaries to lab technicians to Ph.D. students (for one example, see Hough Reference Hough2007, chap. 6).

From the early 1900s through the mid-1970s, Saint Louis University was home to one such node of instruments and personnel central to studying seismicity in the mid-continent. At SLU, the Rev. James B. Macelwane, S.J., was for much of the early to mid twentieth century a leader in seismology, researching and publishing on topics from microseismicity to the operation and design of seismometers to the seismic effects of local quarrying. As researcher, educator, and tinkerer, “Father Mac” introduced students to the work of making and reading seismograms and caring for seismic measuring equipment (Blum Reference Blum1956; Byerly and Stauder Reference Byerly and Stauder1958; Heinrich Reference Heinrich1956, 14; Howell Reference Howell1990, 131–2). During his tenure, the SLU department installed Wood-Anderson seismographs at their own array in 1927, and began operating Wood-Anderson seismographs in the 1930s in Little Rock, Arkansas, and Cape Girardeau, Missouri, to study New Madrid seismicity (Blum Reference Blum1956, 10; Byerly and Stauder Reference Byerly and Stauder1958; Saint Louis University Department of Earth and Atmospheric Sciences n.d.-a).

At the urging of fellow scientists at the National Research Council and the Carnegie Institution of Washington, as well as international Jesuit colleagues, Macelwane succeeded in reorganizing the defunct Jesuit network in 1925 into the Jesuit Seismological Association (Blum Reference Blum1956; Byerly and Stauder Reference Byerly and Stauder1958; Hodgson Reference Hodgson1956; Saint Louis University Department of Earth and Atmospheric Sciences; Udías Reference Udías and Kölbl-Ebert2009, 140; Udías and Stauder Reference Udías and Stauder1991, 189). The SLU Central Station continued to amass and publish data and locate earthquake epicenters until the early 1960s, complementing reports by the U.S. Coast and Geodetic Survey, the International Summary in London, and the Bulletin of the Central Seismological Institute in Strasbourg (Udías Reference Udías and Kölbl-Ebert2009, 140). As more powerful computing eliminated the need for prior methods of data compilation and for duplicate efforts in the location of earthquakes, many stations – Jesuit and secular – ceased observation, but eleven Jesuit stations formed part of the 125-station-strong World Wide Standard Seismological Network established and maintained by the U.S. government beginning in 1962. Later, in 1974, SLU installed the first stations of the Upper Mississippi Valley Seismic Network and continued to be a center of seismological research (ibid., 138, 141; Saint Louis University Department of Earth and Atmospheric Sciences n.d.-a).

Gradually, instruments operated by Saint Louis University were joined by those from other universities and governmental institutions in adjacent states. Over the mid-twentieth century, greater sensitivity of instruments and a slow increase in the number of observation stations led to the detection of more frequent and smaller seismicity. Earthquake listings for Mississippi Valley tremors that had been purely textual became more quantitative over the course of the twentieth century, as instruments not only detected smaller and smaller tremors, but (by the 1930s) timed them to the hundredth of a second (Jackson Reference Jackson1979, 49, 61, and Table 2; on the significance of such changes, see Canales Reference Canales2009). Historic narratives of a frontier-era earthquake ceased to be central to understanding the middle Mississippi environment. Instead, SLU's seismometers, along with those in a few other regional stations, became the primary testimony to seismicity in the American heartland.

In mid-America as throughout the globe, the study of earthquakes after the turn of the twentieth century was done through instruments and through visual representation. The jagged peaks and valleys of seismograms substituted for the word-pictures of heaving earth or waving trees as researchers sought the origins of particular quakes and of earthquake phenomena generally. Earth science did not, of course, transform into a visual science with the advent of seismograms: geology had a profoundly visual basis from its origins as a modern science (Rudwick Reference Rudwick1976; Rudwick Reference Rudwick2005). Nonetheless, the fundamental evidentiary base of seismology in particular did shift, in broad terms, from words to numbers and graphic traces in the early twentieth century. More subtly, the subject matter of seismologic reports narrowed. Human tragedy did not disappear from seismology, but God certainly did: expert language no longer included the prayers to divine Providence and pleas for God's mercy that were a taken-for-granted part of writing about disaster in the early nineteenth century.

Further, from the late-nineteenth through the mid-twentieth century the primary earth science investigations of the mid-American continent were taken up with the search for coal, oil, and gas reserves (Lucier Reference Lucier2008). In the context of rapidly-developing insights yielded by fossil fuel exploration and by earthquake study centered on the west coast, the Mississippi Valley seemed neither seismically interesting nor economically rewarding. Researchers looked forward to the increasing precision and abundance of instrumental data, and outward to the West Coast's series of demonstrably and recently serious earthquake regions. Almost no one looked back at the outdated and seemingly outlandish reports of cataclysm in mid-continent.

In this era of instrumental seismology, one serious investigation of the historical quakes of the mid Mississippi Valley emerged.Footnote ⁶ This 1912 report received little notice at the time but has had great impact over the last three decades. On the centennial of the tremors, one geologist went out into the woods of north Arkansas and the sloughs of southeast Missouri to document evidence of major earthquakes, based on the first-hand accounts he collected from a century before. His work is now cited throughout contemporary science as the beginning of serious seismology of the quakes, but at the time he was quite literally a voice in the wilderness.

Fuller's Investigations

Myron Fuller graduated from MIT in 1896 and then worked for several years as an instructor in the Department of Geology. He began work on fossils, but he published more than eighty articles and studies on topics including glacial geology, various formations, and earthquakes in Jamaica and continental North America. In his later career, he became a private “oilogist,” exploring sites in North America and northern China (Schrock Reference Schrock1982, 39, and M-7; Myron L. Fuller correspondence 1917). But he became best known for his publicly-funded studies of underground oil and water and investigations of ground-water contamination.

In his work on underground fluids as in his research on mid-continent earthquakes, Fuller was a pioneer. He was one of the first researchers in the newly-formed US Geological Survey to study ground-water, and he is responsible for a brief but significant period of attention to ground-water pollution within USGS (and within American science broadly), from 1905 to 1911 (Davis Reference Davis2004). His work on the New Madrid quakes came about, in fact, through a colleague's interest in the effects of earthquakes on underground sources of water (Fuller [1912] Reference Fuller1995; Shepard Reference Shepard1905).

Myron Fuller's research reflects the way in which geology was done from the eighteenth century until the last few decades: Fuller tromped around. In all, he made three autumn trips to the region where Tennessee, Arkansas, and Missouri come together (the Missouri bootheel, the northwest part of Tennessee, and northeast Arkansas) where the quakes’ effects had been most intense. On his first trip, in the fall of 1904, Fuller traveled with E. M. Shepard by dugout canoe and horseback “up the old De Soto trail” (Fuller [1912] Reference Fuller1995). The fact that later researchers on Hernando De Soto debate the Soto expedition's precise route does not change the historical importance of Fuller's description: he and his colleague were traveling areas that had first been traversed by Europeans in the mid-fifteenth century and about which formal European and American knowledge had in many ways not significantly deepened (Hudson Reference Hudson and Galloway1997). The second year, Fuller went out on horseback with a Mr. C. B. Baily, the city engineer of the small town of Wynne, Arkansas, who like many in northeast Arkansas in the late nineteenth and early twentieth century was involved in supplying lumber to the railroads hungry for it. From “timber explorations,” Fuller explained, Baily “had become familiar with the earthquakes features in the still almost untouched forests north of the St. Louis, Iron Mountain, & Southern Railway between Memphis and Wynne” (Fuller [1912] Reference Fuller1995). As happened throughout knowledge of New Madrid – and happens still – a few locals knew about features of a landscape that was still almost as unknown to outsiders as it had been to a would-be conquistador.

A Centennial History of New Madrid

In the plainspoken prose of a government agent, Myron Fuller wrote a surprising and new centennial history of earthquakes based on entirely old-fashioned sources: aged writings and half-eroded environmental traces. The only scientific publications on the New Madrid quakes at all contemporary to his, a pair of articles from 1902, were simply compilations of extended quotations from historical narratives (Broadhead Reference Broadhead1902; McGee Reference McGee1902). Such work was fully in the tradition of late-nineteenth-century seismology, but Fuller attempted an in-depth integration of historical accounts with reconstruction and analysis of environmental damage. Fuller's interest in earthquakes was also part of the rapid expansion of American seismology following the devastating 1906 San Francisco earthquake: his illustrations explicitly compare hundred-year-old sand blows in the New Madrid region with much more recent smaller sand blows produced by the San Francisco quake (Fuller [1912] Reference Fuller1995, Plate VIII). Yet Fuller's report reflects very little of the instrumental focus of seismology taking hold in the first decade of the century: instead, he documented what must have happened by comparing what earthquake narratives said had happened with the evidence that he could decipher from the heavily wooded swamplands of southeast Missouri and northeast Arkansas. Pulling those together, he documented that the 1811–1812 quakes had been felt at very large distances, their local effects were extremely powerful, and they were far from unique in the environmental history of the region.

Fuller framed his book as a synthesis of observation and textual investigation, made necessary by the relative absence of knowledge of the quakes and by their potential significance as evidence of “the geologic effects of great disturbances upon unconsolidated deposits,” that is, the loose soils of the Mississippi embayment. Although scientific literature had long listed the New Madrid events as significant tremors, he observed in his introduction, “the memory of it has lapsed from the public mind.” As a result, he explained, “The story of the earthquake is told in two ways – in the quaint, picturesque, and graphic accounts of contemporaries, and in the equally striking geographic and geologic records” (ibid., 7, 9). Throughout his slim volume, he integrated these two kinds of sources.

Like early-nineteenth-century researchers, Fuller collected and collated what historians would term primary sources of the quakes, and he used them to great effect. He opened with a summary of the events of the quakes drawn from historical narratives, and then used historical accounts at every section to buttress, detail, and ground his subsequent environmental investigations. In an extensive investigation, Fuller documented that much of the swampy area of northeast Arkansas and southeast Missouri was not, as those in the region assumed (and often still assume) the primeval condition of unchanged land, but was instead “sunk land” created by the earthquakes. His evidence was both observational – deadened trees still erect in standing water – but also textual. In one passage that combines his own investigations by dug-out canoe with references to several narratives of the earthquakes’ effects, Fuller quoted a hunter from the late nineteenth century on the recent earthquake origin of the “sunken land”: “In some of the lakes,” wrote the old trapper, “I have seen cypresses so far beneath the surface that with a canoe I have paddled among the branches” (ibid., 70). Fuller himself had canoed many of the same bayous: he and this grizzled trapper had the same experience, both brought forward as legitimate evidence for seismological history.

Many of Fuller's avenues of research not only called upon nineteenth-century sources but hearkened back to nineteenth-century forms of exploration and potential explanation. In an attempt to investigate causes and correlations of the tremors, Fuller created a chart and diagram listing the relation of the major shocks to phases of the moon and time of day, as well as to weather conditions. He attempted to use reports of felt sensation to figure out the directionality of the earth's vibrations. Fuller's report thus collected older forms of knowledge and investigation even while transmitting knowledge and analysis of the quakes into a new century (ibid., 33–42).

In a portion of his account that was completely ignored until very recent investigations by teams in the new subfield of “paleoseismology,” Fuller insisted that the 1811–12 events were only the latest in a series of very large temblors to hit the region. As evidence, he pointed to trees of several centuries’ growth emerging from fault scarps along the St. Francis river in northeast Arkansas, to subtle topographic evidence of former uplift prior to 1811, and, significantly, to accounts of quakes from the 1770s and ‘90s recorded by Daniel Drake after the 1811–12 sequence, as well as to accounts of Indian oral histories preserved by early investigators at New Madrid, including Charles Lyell (ibid., 11–13, 109–110, and Plate II).

In a bare hundred and few pages, Myron Fuller demonstrated the serious extent and environmental consequences of the historic earthquakes, documented prior seismicity in the region, and warned of the effects of future quakes (ibid., 109–110). Later, during the gradual resurgence of interest in historical New Madrid seismicity that began in the late 1970s, Fuller's book was the main source of information (Jackson Reference Jackson1979, 42–46). Yet for more than half a century, Fuller's report lay apparently unknown.

Myron Fuller's New Madrid Quake History

Myron Leslie Fuller was a government man. His history of the New Madrid earthquakes – like many of his publications – was published by a government agency, in this case the Department of the Interior. The publishing history of Fuller's report mirrors the conflicts and contours of study of New Madrid: his account came out in 1912 as a staid government publication, was reprinted in the 1950s by a private firm which promoted history of the southeast Missouri region by reprinting historical documents, was then again reprinted several times in small editions by David Stewart's Center for Earthquake Studies beginning in 1989, and then came out from Stewart's essentially private press in 1995, in a paperback edition with a lurid cover (Fuller Reference Fuller1958; Fuller [1912] Reference Fuller1995, Foreword). More recently, it has been republished on the US Geological Survey website (Fuller Reference Fuller1912 n.d.).

The 1995 edition of Myron Fuller's The New Madrid Earthquake is a complicated, multilayered document. Though the bulk of the text is simply a facsimile reprint of Fuller's 1912 report, the book was published by and has a foreword by David Stewart, a contested figure whom many in the scientific community have disavowed (ibid.). Once a leading figure – indeed, director of the Central U.S. Earthquake Consortium and founding director of the Center for Earthquake Studies at Southeast Missouri State University – Stewart had been a major force supporting Iben Browning, a one-time meteorologist with a sideline in earthquake prediction.

In 1990, Browning predicted a repeat of the New Madrid quakes for a specific day that December, and Stewart urged officials and residents of the region to heed this dire warning. They did, and a flurry of fear and preparation swept central and southeast Missouri.Footnote ⁷ The St. Louis newspaper ran ads for “earthquake kits” in preparation. Media vans crowded the several blocks of the New Madrid downtown, jostling for something to photograph. Then, on the appointed day, in as dramatic a way as nothing can possibly happen. . .nothing happened (Spence, Herrmann, Johnston, and Reagor Reference Spence, Herrmann, Johnston and Reagor1993).

As a result of David Stewart's approach to seismology of the region, and in particular of his support for Browning, he and his work became anathema to most of the geological and seismological community of the mid-continent. His press (the somewhat archly named Gutenberg-Richter Publications, based in a small town in Missouri) existed as a form of protest against scientific researchers who would have nothing to do with Stewart or his view of New Madrid. Yet Fuller's original text is cited and referred to as a touchstone by all who speak of New Madrid – and Stewart's edition of the book, lurid cover drawing and all, represents the main way most interested readers could find a handy copy before the recent on-line republication by USGS. Ironically, his view of the public risk from future earthquakes in the New Madrid region (if not his methods of understanding or predicting such risk) is now much closer to the officially promulgated view than it was in the late ‘80s. Expert panels now convene public events in order to raise public awareness and urge earthquake preparation (one example: Protecting America 2006; much more is planned for the New Madrid Bicentennial: “New Madrid Bicentennial website”).

In a further irony, the “fault finders guides” published in the 1990s by Stewart and his few remaining colleagues are an accessible and engaging way for most laypeople to understand the impact of the earthquakes on the terrain surrounding the epicenters. Containing photographs of sand blows and other still-visible earthquake features, the guides are far more immediate and comprehensive than anything available today from official earthquake-related agencies (originally several volumes, they are now published together: Knox and Stewart Reference Knox and Stewart1995). Yet their provenance makes them unpalatable, so these earthquake guides are not available in any credible or government-related agency studying New Madrid (they are only available on-line or at hybrid institutions like the tiny New Madrid Historical Museum just off the levy on the small main street of present-day New Madrid).

As Fuller tromped and wrote, Saint Louis University had just put in place its initial array of seismometers. Instrumental readings were becoming the way to understand ground movement. Knowledge on the ground – knowledge hard-earned by climbing up and around tree limbs and wading through morasses alive with cottonmouths and copperheads – was no longer the only way to comprehend local territory.Footnote ⁸ Seismographs gave researchers access to tremors far from them, even on the other side of the globe, even as seismograms made precise information portable, a medium of exchange.

Seismograms of very small seismic movement became the way to understand the potential of the region – not wild tales of huge upheaval that seemed just a part of regional tall-tales. Perhaps the Mississippi flowed backward, just as the legendary keelboatman Mike Fink wrestled alligators bare-handed: the river and those who worked it were always good for colorful exaggeration, but that was a far cry from sober evidence. As researchers Arch Johnston and Eugene Schweig observed in the mid-1990s, accounts of the quake elicited “widespread modern disbelief” (Johnston and Schweig Reference Johnston and Schweig1996). Over the first two-thirds of the twentieth century, Myron Fuller's report faded into the obscurity of a little-read, earnest government report.Footnote ⁹ The New Madrid quakes became something of a crackpot science. Moreover, in the late 1960s seismology convulsed with the second of its major upheavals – this one just as profound.

Change to Earthquake Models

In the 1960s, plate tectonics abruptly broke apart the world of earth science and put it back together in a different way. By the late 1960s, the discipline of seismology was presented with three key elements that would effect the change to a model of plate tectonics: the World Wide Standard Seismograph Network (referred to, in this acronymic era of American Big Science, as WWSSN); the increasing availability of high-speed computing; and the ability, through WWSSN seismograms, to study P (longitudinal) waves. Through a series of dramatic publications, debates, and encounters, conventional wisdom decisively shifted by about 1973 (Oreskes and Le Grand Reference Oreskes2001, esp. 150–151). Suddenly and abruptly, scientists understood what caused most earthquakes in most parts of the world: vast tectonic plates, grinding, colliding, and bumping up over each other. Eureka! Yet the exciting, rapid development of the plate tectonic model held little interest in seismicity that was located far from a tectonic plate boundary. The Mississippi Valley is in the middle of a tectonic plate. The theory of plate tectonics does nothing to explain why earthquakes happen there. Attention, funding, and research turned to all that could be explained and revealed with the revolutionary new model of tectonic plates. The swamps of the middle Mississippi remained a scientific wasteland.

For most of the twentieth century, the New Madrid quakes existed as little more than a historical footnote. Old texts of an old quake were simply not at the center of scientific interest or research. Only a few outliers – creative researchers often with minimal funding – began to investigate the New Madrid events in the late 1970s and early ’80s. Spurred by the anomaly of quakes unexplained by the plate tectonic model, geologists and seismologists in the region looked up old earthquake accounts as sources of information, rather than exaggerated tall tales. Cautiously, researchers began to ponder the New Madrid quakes as an event of the past and a possibility of the future.

Otto Nuttli and the Rediscovery of Narrative in Late Twentieth-Century Seismology

Only slowly has awareness of the New Madrid earthquakes re-emerged in scientific circles, impelled in large part by renewed awareness of the usefulness of original earthquake accounts. Scientists studying the New Madrid tremors are forced to rely on colorful frontier stories full of barking dogs and terrified hunters and sulfurous stench, not the cool, abstract certainty of instrumental measurements or recorded jags on a roll of seismograph paper.

Those who today work on mid-American earthquakes generally cite a 1973 article by St. Louis University seismologist Otto Nuttli as a landmark in the rediscovery of historical narratives and thus of New Madrid seismicity. Nuttli focused attention on New Madrid-area seismic history in a 1973 paper published in the Bulletin of the Seismological Society of America, a leading journal in the field (Nuttli Reference Nuttli1973). In this piece, Nuttli resurrected and resumed Fuller's practice of combining textual research on the 1811–1812 events with other forms of seismological evidence. Moreover, he attempted to translate narrative evidence into quantified forms such as magnitude estimates that would be accessible with the techniques and conventions of instrumentally-based seismology.

Otto Nuttli was a researcher thoroughly of the mid-Mississippi. Reared in St. Louis and trained and then employed by Saint Louis University, Nuttli was steeped in SLU's focus on seismology and rose to become a main figure in the field. Nuttli contributed to studies of the core-mantle boundary, shear waves near the Earth's surface, and comparisons between plate-boundary and what are now termed intraplate quakes. Throughout, local context shaped his intellectual legacy: Nuttli was positioned to investigate and insist upon the historical reality and scientific validity of seismicity in the mid-continent. Throughout his research, he observed and insisted upon different models for the west coast than for central and eastern North America.Footnote ¹⁰ He developed a magnitude scale that involved using phases from small events at short to moderate distances, one that would come to be called a Nuttli magnitude (Mitchell Reference Mitchell1988; Howell Reference Howell1990, 127, 103–4, 127).

In his 1973 BSSA article, he argued for a reassessment of the New Madrid earthquakes. Because there was no instrumental record, he argued, researchers have to consider “indirect sources of information, such as published accounts of the effect of the earthquakes upon the landform and upon people” (Nuttli Reference Nuttli1973). This hardly seems a ringing endorsement of historical method: to term first-hand accounts of people's own experiences “indirect sources of information” is bruising to most historians’ sensibilities. And yet Nuttli's essay was a substantial redirection of seismological research. While remaining firmly committed to deciphering the earthquake through evidence of physical disruption, Nuttli argued that written accounts of the quake could indeed help reveal those environmental truths.

Setting himself the task of recovering physical events from written sources, Nuttli practiced a novel kind of seismology. “In an attempt to obtain as complete a set of information as practical,” he explained, “and to avoid possible errors resulting from using secondary or tertiary sources, the present author has gone back to the original newspaper accounts wherever possible.” Newspaper accounts were the heart of Nuttli's reassessment of New Madrid – and his paper literally carried them. The issue of the BSSA that published Nuttli's essay included a microfiche card tucked into a pocket in the back cover, containing a “a retyped copy of the newspaper accounts of the earthquakes” (ibid., 228). Present-day online archives of the BSSA include the microfiche typescript in PDF form alongside Nuttli's paper itself, the strokes of a manual typewriter now digitized but still carrying his commitment to uncovering and sharing the most direct evidence of the quakes.

Nuttli read newspaper accounts carefully. He dismissed and did not reprint in his typescript a series of letters about an event later proven to be a hoax (though he does not give these the dignity of explanation, they were likely a series of reports about the explosion of an eastern volcano). He acknowledged that not all that occurs appears in newspapers: “Although aftershocks were felt at least through 1813,” he observed, “they apparently were not considered newsworthy, especially when compared with such events as the War of 1812 and Napoleon's adventures in Europe” (ibid., 228). Nuttli's analysis reflects a careful reading of newspaper accounts against one another, checking for internal consistencies.Footnote ¹¹ Though his conclusions would take the next several decades to influence the field, Nuttli's essay re-introduced to North American earthquake study the whole notion of examination of historical record as a legitimate and worthy technique of research into this earthquake sequence.

His conclusions were surprising to his many readers who would regard the middle of North America seismically quiet. First, he used newspaper accounts of the earthquakes to construct an isoseismal map of the first main shock, a map “characterized by an unusually large felt area, with MM [Modified Mercalli] intensities of V [on a scale of I-X] as far away as the southeast Atlantic coastal area.” Nuttli thus reinforced Myron Fuller's claim about the extent of the earthquakes and their status as an actual, historical event worth investigation. Nuttli then offered estimates of the magnitude of the shocks, estimating the December 16 shock as a body-wave magnitude of 7.2, Jan 23 as 7.1, and Feb 7 as 7.4 (ibid., abstract). These estimates would ultimately draw other researchers into the question of New Madrid seismicity and form one important element of a fierce and wide-ranging debate about the size of the quakes.

Historicism in Memphis – CERI

Just as Saint Louis University was central to the instrumental understanding of mid-continent seismicity in the first two-thirds of the twentieth century, the Center for Earthquake Research and Information has been central to renewed attention and debate regarding the New Madrid quakes over the last three decades. Founded in 1977, now with more than 55 staff and students, CERI is an unusual conglomerated entity affiliated with the University of Memphis but also housing the local field office of the US Geological Survey (Center for Earthquake Research and Information 2009; Dr. Chuck Langston Takes Helm at U of M's Earthquake Center 2008). From the outset of the program, historical narratives have been central to many CERI researchers’ work.

As with most developments having to do with seismology, the key moment for the creation of CERI was an earthquake: a magnitude 5 event in September 1976, centered at Marked Tree, Arkansas, that brought down telephone lines, damaged roofs, broke small objects, blew out a few windows, and in the words of founding director Arch Johnston, “scared the hell out of people.” The tremor also frightened state legislators into action. The state of Tennessee created what was then the Tennessee Earthquake Information Center (later renamed the Center for Earthquake Research and Information). A state legislator who was a recent graduate got it centered at what was then Memphis State University (now the University of Memphis), the school closest to the tremors (Jackson Reference Jackson1979, 69; Johnston Reference Johnston2009a; Johnston Reference Johnston2009b).

Arch Johnston reflects conventional thinking in his own approach to earthquakes in the 1970s. A young geophysicist just finishing a Ph.D. in Colorado after a five-year term in the Air Force during the American war in Vietnam, Johnston got the job as director of TEIC without thinking much about historical quakes or lower-level seismicity. As he explained, as a geophysicist, he felt that on a global scale, anything under about an M5 “does not register” – and the seemingly casual term is telling: certainly seismicity that did not register on an instrument was not useful data for Johnston or most of his contemporaries. Like many people who find themselves in a new job after getting out of school, he found that his interests were about to be radically reoriented because of financial exigency and institutional priority: “There I was, didn't really care about small magnitude earthquakes, much less historical earthquakes” – and then got funding for a seismograph network to study small magnitude earthquakes, in a zone where there had been large but pre-instrumental quakes. He read Otto Nuttli's 1973 piece and contacted him for advice and critiques. After turning only “grudgingly” to narrative accounts of quakes, Johnston recounts, “I was pretty quickly forced to realize that we had to use historical accounts – we had to see just how quantitatively we could use qualitative accounts.” Slowly, and with a certain degree of reluctance, Johnston found himself turning to what he began to see as “a very worthy scientific problem, and one that was not appreciated in the scientific community”: the nature, extent, and recurrence of mid-continent seismicity (Johnston Reference Johnston2009a; Johnston Reference Johnston2009b).

In 1982, Johnston published his first major piece on New Madrid-area seismicity, an article in Scientific American (Johnston Reference Johnston1982). At around the same time, other professions were beginning to examine long-buried evidence of the 1811–1812 quakes. In 1976, historian James Lal Penick of Loyola University in Chicago, a Jesuit school, published a well-respected account of the impact of the quakes, especially on the growth of American society in the region (subsequently revised as Penick Reference Penick1981).

Yet despite modest professional interest, larger currents within seismology ran against any consideration of historical narratives as anything but sideline curiosity. In 1988, Arch Johnston enquired with the Saint Louis University earth sciences program about where Otto Nuttli's work had been archived. It had not. After Nuttli's death, the department chair at SLU had gone in to clean out his office and had thrown everything away. As Johnston concluded, in a wry mid-south drawl, “That gives you an idea of the value placed by a pretty high-level seismologist on historical work” (Johnston Reference Johnston2009b; confirmed by Harper 2010).

At CERI, however, one of Nuttli's essential projects – the list of primary source accounts that had appeared as supplementary material for the BSSA issue in which he had published his 1973 piece – has been preserved and expanded. The current website for CERI features the New Madrid Compendium, a collection of first-person nineteenth-century accounts of the quakes (New Madrid Compendium n.d.). Some are retyped, recent entries are available as PDFs, and all are indexable by specific effects as well as source and date. The Compendium is essentially a higher-tech Web-based version of Otto Nuttli's 1973 database. It began when Arch Johnston, in the process of his own work, took the annotated bibliography from James Penick's book on the New Madrid earthquakes and tracked down all those sources. By the mid-‘90s, Johnston had “about ten big notebooks” of New Madrid accounts, and in 1999 he hired Nathan K. (Kent) Moran, who was just completing his dissertation in the University of Memphis history department, to continue the project and move it to the Web (Johnston Reference Johnston2009b; Moran Reference Moran2009a; Moran Reference Moran2009b).

For more than ten years, in a small room of the CERI offices in Memphis, Kent Moran marshaled a small team of undergraduate research assistants and balky microfilm machines in his mission for archival completeness. He is currently finishing his years-long project of reading the entire run of every American newspaper he can locate for the few months surrounding the main 1811–1812 quakes (Moran Reference Moran2003; Moran Reference Moran2009a). All of these sources are available through the CERI website, alongside the scientific context such as an earthquake catalog of known mid-continent quakes. Present-day disaster planners, seismologists, and historians turn to the CERI website to find instrumental data on the past hundred years of seismicity in the region, as well as research particular aspects of the experience of the New Madrid quakes (Hough and Bilham Reference Hough and Bilham2006; Langston Reference Langston2010).

The New Madrid Compendium is a project of straightforward historicism, yet it is at once consistent with and distinct from older historical approaches. The goals of the project are not far from Mitchill's or those of other early-nineteenth-century compilers: comparison of myriad accounts can help reveal useful connections and information. The project's overall parameters, however, are far distant from Samuel Mitchill's world. The original New Madrid accounts were deeply embedded in a context of newspaper reporting and narrative accounting for changing processes – whether processes of body or of land. Little of that context appears today in the CERI Compendium. Kent Moran has a finely nuanced sense of the politics and focus, whether local or national, of the papers he and his team have worked through, and he is committed to reading entire issues of the newspapers he locates, so that he can get a feel for how and why the different newspapers present news the way they do. Yet little of this larger framework appears on the website, where the various entries appear either as disembodied, retyped text, or more recently as scanned-in snippets separate from their commercial or political context. Moran works almost entirely through microfilmed newspapers, not the increasingly sophisticated and plentiful on-line databases of period publications (rightly pointing out that text-recognition is not always consistent or successful).

This fracturing of narratives into snippets of text – which is after all perhaps necessary in a searchable online database – does, though, seem related to the lack of connection between the Compendium and broader public conversation about the quakes. Thus far there is little sense of whether the Compendium has seen an uptick of interest with the bicentennial of the New Madrid quakes, or how the Compendium might be promoted as an educational resource, perhaps for school syllabuses that might use the rich and detailed source material. Certainly the New Madrid Compendium reflects the most broadly accessible set of the quake accounts created over the past two hundred years, yet it does not seem to be supported by anything like the public interest that apparently drove the myriad reprintings and newspaper stories about the quakes in the comparatively much more limited media of the early nineteenth century.

Narratives and the Debate about Magnitude

At the same time, New Madrid accounts like those collected in the CERI Compendium are assuming an ever-greater role in contemporary scientific exchange, especially the heated disputes over the magnitude of the 1811–1812 events. Magnitude is for experts a complicated concept, with a lot of variations, but at core the question “how big?” has a certain intuitive urgency for scientists and laypeople alike. Certainly for disaster planning and forecasting the sense of how large tremors were in the past is extremely significant. And on this crucial question of analysis there is a high degree of disagreement, conducted at high volume (summarized efficiently as well as civilly by Hough Reference Hough, Fréchet, Meghraoui and Stucchi2008, 352).

In 1996, Johnston and USGS colleague Eugene S. Schweig, also based at CERI, opened the debate. In an article that summarized work done since Nuttli's 1973 article, they challenged seismologists to unravel the “enigma” of the quakes, and in particular they emphasized the sheer size of the tremors. Insisting that geophysical and geological work had largely verified historical accounts of tumultuous upheaval, they asserted that “The sequence included at least six (possibly nine) events of estimated moment magnitude M.7 and two of M ≈ 8” (Johnston and Schweig Reference Johnston and Schweig1996).

In seismological terms, them's fightin’ words. Since then, seismologist Seth Stein at Northwestern University has been a leader in assailing the magnitude assessment, causal mechanisms, and unnerving future implications of those who are, roughly speaking, advocates of high New Madrid magnitudes.Footnote ¹² Significant elements of this dispute have played out in the public sphere, in debates about public regulatory priorities, particularly with respect to building codes in Memphis, Tennessee (for a recent chapter, see Charlier Reference Charlier2009). In the seismological literature, debate has ranged over almost all aspects of the tremors, including whether GPS systems show movement along a fault, whether models applied to other quake systems can be trusted to apply to intraplate quakes, what counts as a main shock versus an aftershock, and what recurrence interval paleoseismology reveals. But much of the debate over the New Madrid quakes, and particularly about their magnitude, rages about how to read old texts.

Debate over the magnitude of the 1811–12 New Madrid quakes hinges to a large measure on interpretation of historical accounts. In a broader sense, these debates are part of a growing reassessment of techniques and insights to be gained through historical seismology (see Fréchet, Meghraoui and Stucchi Reference Fréchet, Meghraoui and Stucchi2008). In 2000, Hough and three colleagues – including a political scientist – offered a landmark revision of original magnitude assessments hinging on careful reading of historical accounts as well as a reassessment of settlement patterns (Hough et al. Reference Hough, Armbruster, Seeber and Hough2000; further developed in Hough Reference Hough, Fréchet, Meghraoui and Stucchi2008, 361–362). Hough and her co-authors built on a feature of “felt reports” that had long been recognized: the quakes were experienced much more in certain places far away, such as lowlands in North Carolina, and had damaged buildings in Cincinnati, but had not awakened sleepers much closer to the epicenters in Kentucky. Researchers had long recognized that the loose soils of coastal plains and bottomlands along rivers shake much more than, say, the central hills of Kentucky, where a thin layer of soil covers well-anchored bedrock (see Jackson Reference Jackson1979, 10–11). Hough and her co-authors were the first to argue that patterns of settlement on the frontier dictated that very few people would be present to report lack of damage in most places, while many settlers would be concentrated in bottomland areas of atypically high “site response.” But the content as well as placement of historical narratives was also crucial to Hough and colleagues’ downgrading of New Madrid intensity. In their 2000 piece, the co-authors reexamined felt reports and translated their reports of damage differently. Further, they argue that Nuttli simply misread his own primary source in assigning one important moment magnitude reading – and point out a likely transcription error in the date for another shock (Hough et al. Reference Hough, Armbruster, Seeber and Hough2000, 23, 845). Poor historical analysis, in other words, led to poor seismology.

Hough has further developed her fine-toothed approach to historical documentation in a series of publications with several collaborators. In a 2004 summary article, she argued that “Overall, many of the accounts do not appear to support values as high as those originally assigned.” To make this argument, she carefully dissects previous analysis:

Here, to understand the past environment of the Mississippi valley and therefore the future seismic risk of a large American city, seismologists argue about phrasing and language in an early-nineteenth-century newspaper.

In such debates researchers are, moreover, returning to earlier, pre-instrumental forms for evaluating ground shaking. Prior to instrumental records of quakes, seismologists used damage to buildings as a way to describe and convey the characteristics of shaking. Early seismological researcher Robert Mallet, for instance, advocated the careful, comparative use of wall cracks and earth fissures to study earth motion during the 1850s (Dewey and Byerly Reference Dewey and Byerly1969, 190). Distinctions in kind and extent of building damage are still part of the diagnostic criteria for evaluating the intensity of ground motion, though today they are generally used in conjunction with other, instrumental forms of evidence as well. So in this passage, as in other contemporary discussions of New Madrid, modern researchers dissect two-hundred-year-old texts to make in essence a pre-instrumental assessment of ground motion.

In a recent summary, Susan Hough in some sense comes full historical circle in seismological understanding of the quakes. They were once overlooked and seen as unimportant, then for a brief period regarded as the largest earthquakes to rock the continental United States – and now, in her arguments, are downgraded in size to roughly equivalent to the 1886 Charleston quakes (Hough Reference Hough, Fréchet, Meghraoui and Stucchi2008, 351, 365). Unlike past dismissal of the New Madrid quakes, however, Hough bases her assessments in part on careful analysis of archival records. She has identified several new sources that help demonstrate the comparative lack of catastrophic effects in relatively nearby settlements like the river village of Sainte Genevieve (Hough et al. Reference Hough, Armbruster, Seeber and Hough2000). Moreover, Hough's analysis is full of insights about the very process by which records came to be examined: she notes, for instance, the “biases that can be associated with early archival records,” in particular “the fact that dramatic effects are more likely to be documented than less dramatic effects” (ibid., 363). In the work of Susan Hough and her collaborators, the very texture, siting, and characteristics of nineteenth-century settlement and observational writing are at issue in reconsidering the New Madrid quakes.

Such recent scientific work possesses a very fine-grained sense of textual explication, in which scientists bring keen tools to the meanings of words, phrases, locations, or building damage. At the same time, recent researchers have left unseen in these accounts the thick character of discursive exchange and essential qualities of narrative – the ways in which early earthquake accounts took the forms of letters, and of reports on illness, telling a story about processes that were interconnected with a vast web of often ill-understood correspondences and states. In this move from narrative conversation to extractable text – from readers contemplating interconnection to readers poring over and astutely extracting snippets of usable, comparable, and at times even quantifiable information – the seemingly identical documents appear in fundamentally altered form.