Mapping Kentucky's frontier
trails through Geographical information and cartographic applications.(Report)
The Geographical Review - July 1, 2010
Karl B. Raitz
Word count: 9125.
The Road is that physical sign, or symbol, by which you will best understand any age or people[,] ... for the Road is a creation of [people] and a type of civilized society.
--Horace Bushnell, 1864
During the United States' prerevolutionary period the trans-Appalachian West, though largely terra incognita to people living on the Eastern Seaboard and occupied by significant numbers of native peoples, lay open to initial forays by colonial hunters, explorers, surveyors, and settlers. Between 1750 and 1780, white hunters and explorers established some of the earliest overland travel routes to traverse the portion of western Virginia that became the Commonwealth of Kentucky. First-generation traces and trails across open, unsettled country tended to be transductive; that is, routine use of routes encouraged settlers to establish farms and businesses such as taverns and blacksmith shops, thereby reinforcing the alignment of subsequent roads along the same general corridor (Figure 1). Therefore, many of these initial routes became long-standing avenues of economic, urban, and social development (Newton 1970,134; Newton and Raphael 1971, 250).
Four early routes entered Kentucky at or near the southeastern gateway where concurrent gaps truncate Cumberland and Pine mountains. The Indians' Warriors' Path, partially adopted by white pioneers, led north to the Indian Old Fields in eastern Clark County and Indian settlements at Lower Shawneetown on the Ohio River. Daniel Boone's Trace linked to Fort Boonesborough. The Wilderness Road and Skagg's Trace led to Fort Harrod at Harrodsburg and the Falls of the Ohio at Louisville. These routes enabled an influx of migrants into Kentucky, increasing the state's population to nearly 221,000 by 1800 (Speed 1886 [1971], 7).
PREMISE AND PROBLEM
Our principal objective was to map at the largest practicable scale the Kentucky sections of the four primary eighteenth-century frontier trails: the Warriors' Path, Skagg's Trace, Boone's Trace, and the Wilderness Road--and the sites associated with pioneer settlement linked by these trails. Eighteenth-century explorers' maps and reports, surveyors' notes, travelers' diaries, and other information sources about the western frontier were often locationally incomplete, imprecise, or distorted by present standards. Information from such sources may be adequate to suggest highly generalized route alignments but are of limited value for drafting large-scale trail maps. Generalized small-scale route maps illustrate only elementary connections between origins and termini and perhaps imply linkages to medial points, although they paradoxically imply that the landscape is bereft of distinguishing features and detail. Small-scale maps may, for example, encourage the reader to envision complicated topography as a simplistic isotropic surface or interpret a tortuous track as a passable road with elementary attributes: straight and level, hard and smooth (Lardner 1851, 315-316; Upton 1997, 175). Such a road has long been the highway engineer's ideal. But frontier travelers moved on neither isotropic surfaces nor ideal roads. Without accurate detail, small-scale route maps invite the reader to ignore experiential questions that relate to human perceptions, culture-based priorities, environmental context and influence, and contingency and decision making.
While acknowledging the idiosyncrasies inherent in historical records, we sought to glean from them precise points and routes and, where warranted and possible, to compile those locations into an accurate route map. We wished to do this consistently and unambiguously, because subtle distinctions in the exact location of historical routes and related settlement sites will influence the nature of the questions one is prompted to pose and the reliability of the conclusions that one can draw about pioneer knowledge, preferences, and behaviors as they relate to way finding and route and settlement site choices (Elliott and Talbert 2002). Detailed, large-scale maps invite interpretative questions about the quotidian lives of individuals and families: What were their destinations and route choices, their travel modes and exigencies, and their movement strategies when they encountered a varied physical environment? Knowing the exact alignment of a historical route alignment encourages map readers and landscape observers to, for example, consider the pioneer travelers' rationale for establishing long-distance travel routes through diverse, unfamiliar territory replete with potential hazards. Did first-generation migrants simply adopt extant buffalo and Indian trails as their own frontier routes? Or did they establish routes according to tacit or natural knowledge and pragmatic concerns that, whatever their destination, the most desirable trails--"best-choice routes"--should necessarily cover the shortest distance, maintain a desired direction, and engage the lowest possible gradient (Thackray 1974, 674)? How did originary trails compare regarding these primary locational criteria? If practical eighteenth-century migrants knew or intuited that travel time and energy expended correlated with route distance, direction, and gradient, did they also deem these considerations important in their choice of routes?
Although distance, direction, and gradient initially seem to define static qualities of a topographic surface, complicated terrain actually presented the historic traveler with a large number of route choices that likely required weighing alternatives. One route might be shorter but much steeper than another. An alternate route might have low gradients but follow muddy creek bottoms or be oblique to the traveler's preferred direction. Varied terrain, therefore, presented travelers with options, although the more difficult the topography the narrower the field of practical choices.
We explore a method that permits an objective assessment of historical best-choice routes in the context of desirable route attributes. We reconcile best-choice routes with the confirmed locations of road-related eighteenth-century sites, such as road fragments, stations (fortified houses), and forts. A strong relationship between best-choice routes and road-related sites implies that frontier travel was often a carefully considered endeavor that demonstrated topographic awareness and a vernacular sense of mechanical efficiency.
KENTUCKY AS A FRONTIER DESTINATION
Many eighteenth-century emigrants from the mid-Atlantic states--Pennsylvania south to the Carolinas--held as their foremost objective land acquisition in central Kentucky's fertile Bluegrass region, a frontier area that was, by the 1770s, already reputed to be "The Eden of the West" (F. Walker [1775] 1901, 164). But reaching Kentucky via overland routes required that migrants deal with contentious topography, including the wall-like Blue Ridge, the complexly interdigitated Ridge-and-Valley landscape of western Virginia, and the formidable Cumberland Mountain, which shunted frontier advancement toward the southwest. To reach the coveted Bluegrass country emigrants first had to best the 1,500-foot-high Cumberland Mountain that, for some 120 miles, demarcates the eastern edge of the Appalachian Plateau and presents a continuous topographic barrier to cross-country transportation routes (Brown 1948, 184-186) (see Figure 1). The mountain's unbroken eastern face was commonly known as "White Rocks" (Lower Pennsylvanian quartz sandstones of the Breathitt Group). Cumberland Gap, a 600-foot-deep air gap in Cumberland Mountain, offered the only practicable passage across the great barrier.
A second linear ridge, Pine Mountain, parallels Cumberland Mountain some 10 miles to the northwest and mirrors Cumberland's geological and topographic profile. The crest of Pine Mountain rises about 1,200 feet above adjacent valley floors. North of Cumberland Gap the Cumberland River cross-cut a water gap through Pine Mountain, permitting comparatively easy passage through a narrows at river level. Because the felicitously positioned Pine Mountain Gap obviated a topographic dead end, Cumberland Gap became the primary entry point into the mountainous Appalachian Plateau for the Warriors' Path, Boone's Trace, and other pioneer roads.
Beyond the twinned Cumberland and Pine Mountain ridges, emigrants moving on horseback or on foot with packhorses encountered the labyrinthine mountain expanse of Kentucky's Appalachian Plateau. To reach the Bluegrass country travelers had to move north and northwest and traverse some 70 miles of rugged uplands with local relief ranging from 500 to 1,200 feet. Trails were necessarily contingent upon the motive power of human and horse to negotiate gradient and distance and thus were sensitive to topographic advantages (Schivelbusch 1986, 9). First-generation routes, therefore, were "natural" tracks, not constructed roads, which sometimes followed old buffalo traces or Indian trails through the region's dendritic maze. Whichever route the emigrants chose, they had to traverse steep mountains with pinched, stony ridges and slopes heavily wooded with thicket-like understories of laurel (Rhododendron maximum). Creek bottoms could be narrow and rock-strewn or swampy and choked with trees and cane (Arundinaria gigantea) (Gist 1898, 154; F. Walker [1775] 1901, 163; compare Harlan and Denny 2003, 57, 124-129).
The northern boundary of the Appalachian Plateau is demarcated by the conglomerate-capped Cumberland Escarpment, which delineates the Pennyroyal country to the northwest and the Bluegrass plain to the north. At the escarpment, rugged hills abruptly transition to rolling topography more amenable to movement. Crossing this prominent topographic break, Boone's Trace and the Wilderness Road descended from the plateau through steep creek valleys and knobby hills onto the Bluegrass plain, where route choices expanded dramatically. Partly forest, partly parkland savannah, the low-amplitude surface of the limestone plain has a local relief of about 120 feet. Travelers here could favor routes that prioritized desired direction and shortest distance, thereby allowing them to directly approach favored points such as perennial springs, salt licks, and stations.
MAPPING HISTORICAL ROUTES
From the 1830s to the 1920s, state-mandated road construction practices gradually relocated and straightened roads away from historical tracks, for modern equipment permitted engineers to reduce gradients and shorten distances. Although current roads and historical tracks utilize some of the same topographic advantages, such as the Cumberland and Pine Mountain gaps, conflating current roads and historical tracks for the purposes of large-scale mapping is problematic at best. Instead, reconstruction of frontier travel routes can draw historical information from four types of primary sources: narrative reports by explorers, surveyors, and travelers, and interviews such as those collected by the historian Lyman Draper; (1) historical maps; public documents, including property deeds and county road survey records; and historical archaeology reports. Archival sources must be used with care to assure that key sites can be accurately transposed to the map. Some first-generation explorers and surveyors provided reports or diaries that described drainage and topography--the environmental context in which they found themselves. These descriptions often contained significant Ideational errors or provided stream and place-names that were either duplicated elsewhere or subsequently changed (Filson 1784a; T. Walker 1888). On the other hand, many eighteenth-century residents, court officials, and others wrote detailed reports, diaries, and records that described route and topographic landmark locations in considerable detail. This information eventually permitted scholars to retrace original route alignments and note with some precision the location of road-related sites (Herold 1936; Cotterill 1946; Hammon 1970, n.d.; Conforti 2004).
General histories may be useful resources for plotting first-generation routes, provided that they are based on information that is sufficiently specific to allow locating key places, such as passes, springs, stream junctions, and station sites (Speed [1886] 1971; Kincaid 1947). Nevertheless, braided routes or multiple tracks seem to confuse some authors, who mistake one route for another or combine several routes into one so that individual or braided trails are not discernable in the text or on maps (Speed [1885] 1971, 1; Hulbert 1902, 73, 90-91). Other authors seem flummoxed by complex topography and misplaced stream valleys or conflate such prominent landmarks as Cumberland Gap and Pine Mountain Gap (Eckert 1995, xxxiii-xxxiv).
HISTORICAL ROUTE-CHOICE PRIORITIES I: DISTANCE AND DIRECTION
One might conjecture that historical maps drawn by surveyors or drafted from their notes by competent cartographers should provide useful information about the location of historical route alignments and landmark sites. Unfortunately, observational inaccuracy or absence of reliable instrumentation often led eighteenth-century cartographers to place key landmarks, such as falls, stream junctions, ridge gaps, and forts or other settlement features, well away from their actual locations (Lewis 1987, 561; Rumsey and Williams 2002, 4-8). Displacement of key locations, in turn, mitigates attempts to pose experiential questions about frontier travel and practical long-distance navigation through complex topography. Thomas Walker, a representative of the Loyal Land Company, kept a diary of his exploratory trip into Kentucky that provides an example. He and five companions passed through the Cumberland and Pine Mountain gaps into the Kentucky mountains in the spring of 1750. Few of the place-names he cited remain current, for other explorers and settlers renamed many of them. In general, Walker's stream toponyms and notations on topographic features were of marginal value to eighteenth-century cartographers who wished to establish or confirm landmark locations (T. Walker 1888, 55-56; 1898, 77-84; Dobbs 2006, 110-111).
Based on information provided by frontier explorers, traders, and surveyors, in 1755 Lewis Evans published an important map of the northeastern quadrant of what is now the United States, extending from the Atlantic Coast to the Falls of the Ohio (see Reed 1915, 221-222; Raisz 1937, 377; Klinefelter 1971, 44-50). Even though the map was widely regarded as a milestone in colonial graphic arts and the best map of its day, its portrayal of the country west of the Appalachians was based largely on description rather than competent survey (Klinefelter 1971, 47). By current standards Evans's map was replete with distance and direction distortions and site displacements. Other publishers extensively copied and republished the information on Evans's map, including many of its errors (Pownall 1776). In 1782 John Filson traveled from Philadelphia to Kentucky to gather information for a new map. Advantaged by a gradual escalation in exploration and surveying in the trans-Appalachian West after 1755, and using interviews and surveying trips with Daniel Boone and other frontiersmen, Filson published an updated Kentucky map two years later (Filson 1784b; Slotkin 1973, 270-275; King 1996, 115) (Figure 2). Although Filson believed that his map was credible, it abounds with site displacements and generalizations (Wright 1942; Walton 1956, 34; Nobles 1993, 11; Edney 1996, 188). Elihu Barker and other mapmakers subsequently corrected errors, refined site locations, and formalized place-names (Barker 1795), but technical limitations and a paucity of survey reports often meant that the early Kentucky map was an apocrypha of distorted information.
[FIGURE 2 OMITTED]
One can assess distortions in distance and direction on historical maps by comparing the latitude and longitude values assigned to selected sites (Harley 1968, 65-66). Thomas Pownall's 1776 version of the Evans map has two Kentucky locations that one can use to establish latitude and longitude coordinates: Big Bone Lick and the Falls of the Ohio (see Figure 1). Pownall's coordinates placed the lick 94 miles away from its actual location and moved the falls 122 miles. The 1784 Filson map and the 1795 Barker map record five common landmarks that invite comparison: Cumberland Gap, Fort Boonesborough, Saint Asaph's, Fort Harrod, and the Falls of the Ohio. Each map has substantial coordinate errors. The farther west from a prime meridian a given site is positioned the greater the displacement error becomes, and all of the locational displacements on eighteenth-century map noted here were westerly (Walton 1956, 62). Cumberland Gap, the easternmost point, is displaced 17 miles by Filson and 8 miles by Barker. The Falls of the Ohio, the westernmost point, suffers a 66-mile displacement on the Filson map, whereas Barker managed to reduce the error to 35 miles. If locational precision is a primary concern--as it must be if one's purpose is to begin to comprehend and appreciate the priorities and decisions of frontier trail users--information drawn from period maps must be employed with considerable skepticism when mapping historical roads according to a datum.
Locational control--scale, direction, distance, and accurate site placement--was one of several issues that eighteenth-century cartographers confronted. Another was the problem of illustrating three-dimensional topography--landform elevation and slope shapes and gradients--on two-dimensional paper. Many American mapmakers employed the ancient "molehill" technique to suggest hills and mountains (see Figure 2). The molehill was an iconic topographic symbol drawn from a high oblique perspective that provides only a general impression that an area might be hilly. Although the molehill technique provides no height, slope, or surface shape information, the Pownall, Filson, and Barker maps employed them. As an information source for accurate route reconstruction on contemporary maps, this type of topographic representation has little value.
HISTORICAL ROUTE-CHOICE PRIORITIES II: GRADIENT
Replacing a highly generalized, small-scale, regional migration route map, whether historical or current, with a modern, comprehensive, large-scale map requires the cartographer to reconcile patchy historical information (time) and localized topographic detail (space) (Cappon 1974,507). Although historical maps may position a station, salt lick, or stream confluence some distance from its actual location, cartographers can establish objective ground truth for sites if actual locations can be confirmed by grid coordinates, by archival and archaeological research, or by references to known locations (Knoerl 1991, 106). Extended route segments confirmed by ground truth can be readily incorporated into a large-scale map (Conforti 2004). Other segments may contain only trail fragments and scattered settlement sites that can be confirmed. Mapping a route that connects disparate points and alignments requires interpreting the topographic surface to establish an optimal alignment; that is, a best-choice route that recognizes the pioneers' intuitive knowledge and preference for a way that followed the lowest local and aggregate gradient change, encountered the fewest obstacles, was aligned to the desired direction, and traversed the shortest distance. We recognize, of course, that even a high density of confirmed route-related sites cannot affirm with absolute certainty that, when mapped at a large scale, a given alignment is correct in every detail, even when it is the best choice (Plewe 2002, 441-443). Nevertheless, mapping both confirmed and best-choice routes has the advantage of directing field and archival research to specific places, and route location can then be corroborated through land surveys, property deeds, and other documents.
Identifying frontier migrants' way-finding priorities facilitates mapping historical routes. Eighteenth-century explorers, surveyors, and settlers were probably rational decision makers who pragmatically wished to minimize the effort their travel required (Bond 1923, 323). Contingencies presented by potential impediments may have forced travelers to detour from desired or established routes, thereby contributing to braided trails. Flooded rivers, for example, offered a type of periodic obstruction that induced travelers to alter their travel timing. By 1795 folk understanding of Kentucky's seasons had improved, so most migrants traveled the Wilderness Road during the dry fall season. Of the 9,010 people who arrived in Kentucky via the Wilderness Road in 1795, 59 percent traveled during the three driest months: September, October, and November (KG 1796; Cappon 1974, 508).
A brief review of state road law suggests a measure of historical continuity in pragmatic route choice. Kentucky's 1797 road law directed county courts to appoint three or more "suitable persons" to "view" a proposed route and report on its comparative "advantages and disadvantages." The law further directed that roads should follow the "nearest and best way" (GACK 1795, 54; Crump 1895, 7; Raitz and O'Malley 2004, 424-425). Viewers' interpretations of advantages and disadvantages usually resulted in recommendations that roads adopt a straightforward priority of shortest distance, consistent directional bearing, lowest gradient, and absence of obstacles (Lardner 1851, 314; Schivelbusch 1986, 21). Although the formal legal conventions associated with route choice after 1797 have no bearing on the Warriors' Path or the course of Skagg's Trace, Boone's Trace, or the Wilderness Road, all of which preceded the state's road legislation, it is apparent that early legislators framed the state's road law using the same pragmatic logic that pioneer migrants had found practicable.
The travelers' desire to avoid steep slopes has an ironic corollary when it is combined with the concern to minimize route distance. A low-gradient valley is the product of an actively eroding stream that produces a floodplain across which it channels broad meander loops. The lower the stream gradient--and the less effort required by people or horses to move along its floodplain--the more sinuous the stream becomes. A stream with a high sinuosity index--channel length divided by valley length along a given reach or stream segment--inevitably has an elongated channel and therefore does not necessarily provide a shortest-distance route and may significantly depart from a straight line or constant direction (Mueller 1968). Following a high-sinuosity stream channel meant that a traveler's journey would be substantially longer than would an "as-the-crow-flies" route. Such a route might also compel the traveler to make repeated stream crossings (Kilpatrick 1921, 368). Eighteenth-century maps, based on fragmentary information, often generalized rivers and creeks into directional trend lines. Filson's 1784 map, for example, includes the Ohio River's denotative great north bend between the Great and Little Miami rivers near Cincinnati, but it also portrays most streams, especially the Ohio's tributaries, with an affected sinuosity, thereby implying that they are more or less alike in gradient and have prominent floodplains and uniform land surfaces (Walton 1956, 62-72).
IDENTIFYING BEST-CHOICE ROUTES
Kentucky's frontier travel accounts suggest that some early explorers followed buffalo traces and Indian paths that, in turn, followed valley bottoms or ridge crests. In lieu of established routes, explorers and surveyors marked out their own paths. Given that both informal (folk navigation strategies) and formal (legally prescribed) travel preferences directed migrants and road builders to seek the shortest distance and lowest-gradient routes, we devised a method that allows us to interpret likely best-choice routes across the areas where precise route locational information was absent from the historical record. The method employs GIS software to combine U.S. Geological Survey (USGS) large-scale topographic maps and USGS hill-shaded base maps for Kentucky. The hill-shaded maps assist visualization of slope, and the topographic maps provide contour lines that permit assessment of gradients. The software also permits measurement and comparison of route distances (Cappon 1974, 512; compare Wentworth 1930; Robinson 1946; Calef and Newcomb 1953).
The highest elevation in Kentucky is the 4,139-foot-high Black Mountain, which lies between Cumberland and Pine mountains on the state's southeastern border. The elevation at the Falls of the Ohio (Louisville), an important terminus for the Wilderness Road, is 380 feet. The aggregate drop over the 160-mile distance from the vicinity of Cumberland Gap to the falls therefore exceeds 3,700 feet. After it passes the narrows at Pine Mountain Gap the prevailing direction of flow of the Cumberland River, including its primary tributaries, the Laurel and Rockcastle rivers, is west-southwest. Travelers moving northwestward through this section were therefore obliged to cross repeatedly from one drainage basin to the next. North of the Rockcastle drainage basin the prevailing stream flow direction changes to northwest as streams move down gradient toward the Ohio River by way of the Kentucky River, the trunk stream for the state's southeastern and central sections. Here, trails could more readily follow ridgelines or valleys for extended distances, thereby maintaining a desirable direction of movement.
Frontier migrants encountered two types of slope profiles in crossing Kentucky's mountainous terrain. Ridge spurs, or noses, were often convex in linear cross-section, whereas tributary valley slopes were often concave (Figure 3). Being convex, the steepest segment of a low-angle ridge spur was closest to the primary stream's valley, and the ease or difficulty of climb might have been evident from a trail following the valley bottom. Tributary streams frequently created concave slopes that, when approached from the trail in the primary valley, were initially low angle, given their close proximity to the floodplain of the trunk stream, but steepened progressively upstream, perhaps to abruptly terminate in a "box" at the head of a valley surrounded by rocky slopes too steep to climb (Gist 1898,154; Dalrymple, Blong, and Conacher 1968). Buffalo roads often ascended convex ridge spurs to gain a ridgetop; some frontier trail sections followed spurs to ridge crests where their directional alignments were favorable.
[FIGURE 3 OMITTED]
One can assess how gradient affects movement by employing common conventions, in part derived from engineering premises that equate grade with effort and energy expended. For example, a gradient of 1 in 160 or a l-foot rise over a 160-foot horizontal distance (0.6 percent, or 0.36[degrees]) is a standard maximum for railroad grades, whereas a i-in-60 gradient (1.7 percent, or 0.95[degrees]) is unusually steep for railroad grades and is used only in special circumstances and for limited distances. The nineteenth-century English road builder Thomas Telford adopted a maximum road gradient of 1 in 30 (3.3 percent, or 1.9[degrees]), whereas French engineers adopted 1 in 20 (5 percent, or 2.9[degrees]) as a maximum gradient for their roads. Kentucky's 1853 road law mandated 1 in 8 (12.5 percent, or 7[degrees]) as a maximum gradient for state roads, a grade that road engineers considered the steepest permissible for horse-drawn freight traffic. A gradient of 1 in 2 (50 percent, or 26.6[degrees]) is considered too steep for walking (Frost 1910, 48-49).
Slope exacts a pragmatic limit on the load that a person on foot, a packhorse, or a horse and wagon can move uphill or downhill. One can estimate grade resistance per ton of load as the constant of 2,240 (a long ton) divided by the rate of gradient. A rise of 88 feet per mile equals a grade of 1 in 60 and creates a resistance of 38 pounds per ton of load. Therefore, 10,000 pounds (5 standard tons) of freight loaded onto a large wagon encounters 560 pounds of resistance on a grade of 1 in 20 (a rise of 264 feet per mile), and 3,300 pounds of resistance on a grade of 1 in 8 (660 feet per mile) (Frost 1910, 14-17). One can read grade on large-scale topographic contour maps by measuring or interpreting contour lines. USGS topographic maps record landscape information by naming or labeling, by recording measurable elevations, and by representing reality through iconic or color symbols (Tufte 1990, 81). Typically, 1:24,000 (7.5') topographic quadrangle maps for Kentucky use intervals of 10 or 20 vertical feet between contour lines. A model contour map of a creek flowing through varied terrain illustrates the relationships between contour line spacing, contour interval, and slope gradient (Figure 4). Widely spaced contours equate to low slope angles and are typically associated with stream floodplains or ridge lines in otherwise rugged topography. Intermediate slopes may be found in tributary valleys or on the spur or nose of a ridge. The exemplar Pioneer Creek floodplain in Figure 4 provides a possible low-gradient route that, given the transect from A to A', might be 1 in 160 or less. Climbing the ridge spur by way of B to B' would engage a grade of 1 in 30; C to C' would be a grade of 1 in 8. Slopes of 1 in 2, as illustrated by D to D', are impassible for normal foot traffic.
MEASUREMENT AND REPRESENTATION OF ROUTES
Mapping historical road alignments at a large scale using information that can be confirmed by legal documents, cadastral coordinates, or ground truth is a straightforward undertaking. For poorly documented alignments, connecting known road sections and sites to others along a best-choice, albeit conjectural route, can be accomplished by interpreting the land surface. With practice, one can gauge slope variations along a series of transects and estimate best-choice routes along low-gradient valleys or ridge crests. One can therefore arrive at an optimal route that will connect frontier stations, forts, road fragments, or other confirmed road-related landmark locations.
Route reconstruction that permits consideration of route choices and decision making necessarily requires details about topography and cultural features, which can be obtained if one can map at a very large scale, say 1:12,000. Using GIS software permits mapping on a topographic surface at any scale but entails dealing with the problem of adapting historical place and distance information to route-reconstruction mapping, an issue that relates, in part, to the density of confirmed locational points or lines. Publication format limitations and a low frequency of confirmed data points for some sections of Kentucky's frontier trails have probably prompted mapmakers to largely ignore the problem of establishing detailed route alignments between known points and to simply generalize entire routes (compare Mason 1951, map insert, 140-141; Billington and Ridge 1982, 169). Along the transect from Cumberland Gap to the Falls of the Ohio some extended route sections have been thoroughly researched using property deeds and other historical documentation and were field inspected for ground truth (Conforti 2004; Hammon 1970). Other sections may include no more than one or two confirmed information points (named stream junctions, stations, springs, or salt licks, for example) or lines (creek alignments, known road segments) for 10 miles or more. Some sections, such as the Laurel "levels" near London, may have five or more confirmed points over a comparable distance.
Trail mapping through an area rich in confirmed data points might begin at a spring or the site of a fort and follow a stream valley to other confirmed points. Mapping a trail across a landscape spare in data points requires "stitching" together two or more short links to form an intermediate-length route segment (1-5 miles) which includes a known landmark that serves as a directional target. Intermediate-length road sections may possess a comparatively large number of possible route choices, especially where tributaries enter a trunk stream valley. Intermediate links sparsely marked by confirmed points may be treated as a series of short route segments on a heading toward a distant landmark that are compared and evaluated for low-gradient passages. Long-distance road links (5 miles or longer) may include several known landmarks. By working from either end of a route segment containing known landmarks one may join short-distance links into intermediate links to establish a general alignment that permits forward and backward sighting to confirm that the route complies with the migrant's posited priorities while also aligning with landmark targets.
The Warriors' Path was long established by the time colonial-era explorers arrived and noted the route in their reports (Gist 1898,156; Kincaid 1947). As the primary trade route linking the Cherokee and Catawba in East Tennessee and the Carolinas and the Shawnee in the mid-Ohio Valley, the path connected Cumberland and Pine Mountain Gaps with Lower Shawnee Town at the mouth of the Scioto River west of present-day Portsmouth, Ohio (Henderson, Jobe, and Turnbow 1986, 81; compare Dobbs 2009, 332). We mapped the path as an exercise in best-choice route interpretation and soon realized that historical maps were of little value in plotting the route. Both Pownall and Filson suggested the path's presence on their maps, but they placed it--represented as a straight line--well away from its actual corridor (Pownall 1776; Filson 1784b; Pratt 1942, 237-240).
Utilizing confirmed landmarks along the Warriors' Path, including gaps and stream junctions, we employed GIS software to visualize the topographic surface and selected a route that connected the known points while also prioritizing low gradient, desired direction, and shortest distance. Beginning at Cumberland Gap, we plotted the Warriors' Path along a broad parabolic curve, trending northwest to intersect the Indian Old Fields. From there the path recurved northeast toward Lower Shawnee Town at the mouth of Ohio's Scioto River (Figure 5). On the Bluegrass plain, south of the Indian Old Fields at Eskippakithiki, we could map the path's route with a relatively high level of confidence. The trail's northern portion has less authoritative documentation, and long segments required interpretation. The entire route largely followed stream valleys, and although it passed through mountainous topography it crossed only three primary drainage divides, thereby limiting the number of ridge crossings required. Toponyms still identify many features associated with the Warriors' Path, among them "War Gap," "War Fork," "War Creek," and "Station Camp Creek."
COMPOSITION AND COMPARISON OF ROUTES
From Cumberland Gap to the Falls of the Ohio, frontier folk established routes other than the Warriors' Path through both the mountain and Bluegrass sections. Mapping these routes utilizing GIS software invites comparison and assessment of how they represented a response to concerns about distance, direction, and gradient. Some 53 miles north of Cumberland Gap, at the Rockcastle River west of Hazel Patch, one frontier route bifurcated into Skagg's Trace and the Wilderness Road. For a straight-line distance of 22 miles these routes ran in parallel, crossing at Brodhead, before rejoining into a single track at Crab Orchard. Though roughly corresponding in direction, these tracks differed in character and locational rationale, so they provoke questions about their creation and use (Figure 6).
[FIGURE 6 OMITTED]
Skagg's Trace was an early hunter's trail probably named for three brothers--Henry, Charles, and Richard Skaggs--who hunted in the Kentucky mountains for many years prior to permanent settlement. From a low-lying, six-stream junction near Hazel Patch the route pointed northwest toward the Dick's (later renamed "Dix") River drainage and the region's first settlement at Harrodsburg, and beyond that to the Falls of the Ohio. Skagg's Trace followed the Rockcastle River downstream to the East Fork of Skegg Creek (likely a later misspelling of "Skagg's"), then northwestward to the trace head near the present site of Mount Vernon. West of Mount Vernon the route dropped over the Cumberland Escarpment and angled northwest to Brodhead and Crab Orchard. The segment of Skagg's Trace from its junction with the Wilderness Road near the Rockcastle River to Crab Orchard is 31 miles. Although it adopted a westerly course along the Rockcastle River bottom, Skagg's Trace realigned northwestward at the East Fork of Skegg Creek and maintained that trajectory to Crab Orchard (Kincaid 1947, 113-114; Hammon 1970, 118-123). The route's low creek-bottom sections necessitated multiple creek crossings--one traveler reported crossing Skegg Creek forty times--and the rich bottomland soils produced dense stands of cane that forced riders and their packhorses into a narrow path (Kincaid 1947, 141).
In 1795 the Kentucky Legislature passed an act advocated by Governor Isaac Shelby to relocate and shorten the road from Crab Orchard to Cumberland Gap. Shelby wished to enhance businesses in the state by upgrading the road to permit passage of wagons capable of hauling one ton of weight when pulled by four horses. Large freight wagons of the day were capable of hauling five times this weight across good roads, so the modesty of Shelby's objective speaks to his sober acknowledgment that a mountain road had to overcome significant topographic obstacles. The governor appointed three road commissioners whose responsibilities included locating and surveying the new route and employing laborers to clear a 30-foot-wide track through the forest (Littell 1809, 3: 275-277). From Crab Orchard to Hazel Patch the surveyor's new road lay east of the old Skagg's Trace route. When completed, the new route was officially designated the "Wilderness Road" for the first time. People traveling the new road found it a difficult and uncomfortable undertaking. The early industrialist Moses Austin followed the new road into central Kentucky in 1796 shortly after it was opened, for example, and when he finally reached Crab Orchard he observed that "much work with many bridges may make a good road, [it is] opend for wagons and much work don on it [but] much more must be don to make it tolerable" (Garrison 1900, 525).
Combining GIS analysis and cartographic visualization, we created a profile map juxtaposing Skagg's Trace and Shelby's Wilderness Road (Figure 7). On a horizontal plane the two routes are only about 6 miles apart at their widest points of divergence: where the Wilderness Road crosses Wildcat Mountain and Skagg's Trace followed the East Fork of Skegg Creek. In cross-section the two routes are substantially different. The Wilderness Road segment was primarily a ridge route, whereas Skagg's Trace followed river and creek bottoms for a considerable distance (Figure 8). Following James Vance (1961, 375-376), we created a total relief index for each route by using GIS to count the number of ascending and descending contour lines that each road crossed, dividing that value into the elevation change from lowest to highest points, and multiplying that number by 100. According to the resulting relief index, higher values indicate a more efficient, lower total gradient change along a given route. A relief index value of too, for example, would suggest a level route.
[FIGURE 7 OMITTED]
The path followed by Skagg's Trace started at about 875 feet elevation near Hazel Patch and rose to 1,225 feet over the 26-mile distance to Brodhead. It dropped below 1,000 feet before reaching Brodhead, where it entered the Dick's River watershed. It then crossed a series of low hills on its approach to English's Station. The Skagg's Trace relief index is 25.3. The trace was somewhat more sinuous than the Wilderness Road: Its route covered 31 miles from its origin to English's Station, whereas the Wilderness Road track between the same two points was 26 miles. A sinuosity index can be calculated by measuring the actual road distance between beginning and end points and dividing that number by the straight-line distance between the same two points, or 21 miles for each route. Skagg's Trace has a sinuosity index of 1.48, meaning that the route required traveling nearly one-half mile extra for every mile one would have traveled had the route followed a straight line.
The Wilderness Road, by comparison, climbed from 875 feet at Hazel Patch to the crest of Wildcat Mountain at 1,265 feet before plunging into the Rockcastle River valley, where it forded the stream near the 845-foot level. The road then climbed to the 1,440-foot crest of Chestnut Ridge before descending the Cumberland Escarpment northwest of Mount Vernon. It crossed one more significant ridge before dropping into the Dick's River valley near English's Station. The relief index for this Wilderness Road segment is 15.5, suggesting that the route had a cumulative gradient change some 39 percent greater than Skagg's Trace, which was primarily a valley route. The trace crossed only one primary drainage divide between the Rockcastle and Dick's River watersheds. The creek bottoms along Skagg's Trace produced luxuriant vegetation, especially cane. Although cane proved high-quality fodder for the pioneers' horses, the dense canebrakes made passage exceptionally difficult (Hammon 1973, 262-264). The unfortunate combination of thick canebrakes and difficult stream crossings created formidable obstacles to wagon travel. These conditions were probably a primary reason why Isaac Shelby strongly advocated that the state legislature sanction surveying and clearing a new route that would ostensibly permit wagon traffic.
Over the mountainous section from near Hazel Patch to English's Station, Shelby's Wilderness Road was the more direct route: It was about 5 miles shorter than Skagg's Trace and has a lower sinuosity index of 1.24. Although its aggregate gradient is substantially greater, the Wilderness Road segment was primarily a ridge route, a choice that would, once cleared of trees and laurel, provide a more direct passage. Road surveyors may have presumed that, because wagons were pulled by draft animals, their route's steep and rocky slopes would have been less troubling than would the problem of negotiating thick, rapidly growing canebrakes and making multiple crossings of unbridged streams. Merely clearing a route through mountainous terrain did not assure ease of travel by wagon, however. During the Civil War, nearly seven decades after the Wilderness Road was initially surveyed and cleared, both Union and Confederate armies found the route logistically intractable. Although officers on both sides assigned troops to road building and repair duty, the steep and rocky Wilderness Road was littered with broken army wagons and dead draft animals during the conflict (Kincaid 1947, 274-281).
MAPPING LOST HISTORICAL ROUTES
Pioneer surveyors and hunters established frontier routes that they deemed most practicable while also recognizing elementary requirements for potable water, sensible river crossings, and available fodder for pack and draft animals. Contingencies such as security and weather also influenced migrants' travel. Detailed knowledge of historical trails suggests questions about how pioneer hunters, surveyors, and migrants perceived and experienced the frontier. This knowledge may also encourage map readers to posit ideas about pioneer way-finding practices in difficult, unfamiliar country. By employing GIS software to visualize terrain we were able to plot historical road routes that linked well-documented road-related sites. We were also able to map conjectural, best-choice routes linking scattered sites by interpreting gradients as recorded by topographic map contour lines and visually plotting best-choice routes. This approach will permit researchers to assess the broad range of considerations, such as distance, direction, and land cover, that frontier folk had to make in choosing routes while also trying to follow a path that minimized changes in elevation. Routes mapped at large scales can also be assessed and compared through sinuosity and relief indexes.
Reestablishing the location of historical trails at large scales can proceed by utilizing objective data points and lines obtained through archival research and field survey in combination with interpretation of best-choice routes through GIS visualization and computer graphics. Lost historical routes can be reconstructed in a manner that will invite further refinement of route and historical site locations. The methods outlined here yield a high level of correspondence between conjectural route segments and identified landmarks, including documented pioneer road sections. Following these methods and recognizing practical cartographic requirements, we created GIS layers of Kentucky's primary frontier trails: the Warriors' Path, Boone's Trace, Skagg's Trace, and the Wilderness Road. These layers contain attribute information detailing the reference sources and a description of the confidence level we have in the accuracy of each road segment's alignment. For portability and reference, we also drafted a hard-copy frontier-trails map using the GIS data files at an intermediate scale of 1:341,000. We accomplished this by overlaying the trails and points on top of statewide shaded relief in a graphics program. Our map is available at state and national parks and regional libraries.
Selected historical routes are transductive, in that routine route use led to roadside development, which in turn reinforced the alignment of subsequent roads along the same general corridor. Across Kentucky, from Cumberland Gap to the Falls of the Ohio, for example, eleven county seat towns grew up along or near sections of the Wilderness Road (Whebell 1969, 1-7). (2) Knowledge of these routes is also important to public concerns about family history and heritage. In October 2008, Cumberland Gap National Park administrators inaugurated a new program entitled "The Pioneer Roots of Our National Destiny ..." to commemorate the eighteenth-century migration of thousands of settlers into Kentucky via Cumberland Gap. Park administrators invited the descendants of migrants to reenact the experience by walking through the recently restored gap at sundown. Today, some 48 million Americans have ancestors who moved through the gap and into Kentucky along pioneer trails (NPS 2008). Such popular heritage commemoration reminds us that historical trails are an important part of a regional and national legacy. Detailed large-scale maps can provide basic information about trail routes and conditions that will reinforce scholarly research and underpin interest in popular geographical history.