Historians, geographers, economists, and other social scientists seem to have reached a consensus that urbanization in the antebellum cotton South never evolved into a central place system of regularly spaced and hierarchically arranged hamlets, villages, towns, and cities. Unfortunately, no one has yet rigorously tested this hypothesis and empirical studies remain all too meager. In particular, the lack of understanding of the role of smaller urban units demands a theoretical and methodological reexamination of notions about place development in the antebellum cotton South. This paper develops an urbanization framework, drawing upon the work of scholars in geography, rural sociology, anthropology, and history, and then tests that framework against the secondary literature and a community study of Hancock County, Georgia. The synthetic approach developed provides both interesting insights into southern urbanization and the potential for guiding future research and comparative analysis across space and time.
Social scientists have employed three basic approaches to the study of urbanization:[ref]See also Paul Wheatley, “The Concept of Urbanism,” in Ruth Tringham (ed.), Urban Settlements: The Process of Urbanization in Archaeological Settlements (Andover, Mass., 1973), 623‑4.[/ref]
(1) a “characterological” analysis defines urbanization as the transformation in the pattern of behavior and personality that individuals and groups undergo upon crossing some threshold from rural to urban places (or, alternately, changes in the pattern along an rural-urban continuum of places), reflecting strongly the “Chicago school” of Louis Wirth and Robert Redfield;
(2) a “demographic” analysis, seeking a more measurable, objective criteria, defines urbanization as “a process of population concentration” that “proceeds in two ways: the multiplication of points of concentration and the increase in size of individual concentrations,” following the work of Hope Tisdale Eldridge;[ref]Hope Tisdale Eldridge, “The Process of Urbanization,” in Joseph J. Spengler and Otis D. Duncan (eds.), Demographic Analysis: Selected Readings (Glencoe, Ill., 1956), 338.[/ref]
(3) a “functional” analysis, seeking a causal analysis of the conditions under which urbanization occurs, defines urbanization as the concentration of services, trades, activities, and other functions at a point, continuing the work of geographers Walter Christaller and August Losch.
Each of these approaches address different aspects of urbanization. Both the characterological and demographic frameworks treat more the effects than the causes of urbanization. Characterological studies also tend to be more relevant to the study of larger cities where the differences between urban and rural are more clear‑cut. Joseph A. Ernst and H. Roy Merrens point out that traditional reliance on demographic and characterological approaches (emphasizing urban size and form) distorts the functional importance of small villages in the colonial South.[ref]Joseph A. Ernst and H. Roy Merrens, “‘Camden’s Turrets Pierce the Skies!’: The Urban Process in the Southern Colonies during the Eighteenth Century,” William and Mary Quarterly, 3rd ser., XXX (1973), 552-7.[/ref] Of course, functional approaches can equally distort by equating an isolated general store with a town, ignoring the importance of townspeople to urban development.[ref]Hermann Wellenreuther, “Urbanization in the Colonial South: A Critique,” William and Mary Quarterly, 3rd ser., XXXI (1974), 660.[/ref] This paper eschews a characterological approach as inappropriate for the study of predominantly agricultural Southern counties where “non‑urban” hamlets, villages, and towns predominated and adopts for the most part a functional approach, while acknowledging the integral role played by demographic processes in the appearance and development of urban units.[ref]Nevertheless, most urban historians of the South follow a characterological approach concentrating on the largest cities, although David R. Goldfield, “The Urban South: A Regional Framework,” American Historical Review, 86 (1981), 1009‑1034, does emphasize the unique “rural” characteristics of southern cities.[/ref]
CENTRAL PLACE THEORY The most highly developed of the functional approaches to urbanization falls under the rubric of “central place theory,” originally developed in the 1930s by German geographer Walter Christaller.[ref]Walter Christaller (trans. Carlisle W. Baskin), Central Places in Southern Germany (Englewood Cliffs, N.J., 1966).[/ref] Central places are geographical sites which provide goods and services (economic, institutional, social, and cultural) for a surrounding population. Christaller posited three principles of central place development (pictured in Figure I):
(1) The “market principle” derives place development from two very simple assumptions about producer‑consumer relationships: consumers are only willing to travel so far to obtain a certain good or service and producers need to draw on a certain minimum size trade area in order to make a profit.
(2) The “traffic principle” “states that the distribution of central places is most favorable when as many important places as possible lie on one traffic route between two important towns, the route being established as straightly and as cheaply as possible.”[ref]Christaller 74.[/ref]
(3) The sociopolitical “separation/administration principle” assumes that communities are separated from each other either for defensive purposes, due to a strong “idea of community,” or for administration purposes in federalistic states. “The ideal of such a spatial community has the nucleus as the capital (a central place of a higher rank), around it, a wreath of satellite places of lesser importance, and toward the edge of the region a thinning population density‑‑and even uninhabited areas.”[ref]Christaller 77.[/ref]
Christaller found the market principle dominant in southern Germany but recognized that “in the competition of the three principles, one of the principles will not necessarily clearly triumph.”[Christaller 116.] Nevertheless, most followers of central place theory equate the theory solely with the market principle, ignoring the other two principles. Numerous studies‑-based on various measures of “centrality” (the degree of importance of any central place) estimated from statistics on population, occupations, types of goods and services, mail service, and frequency of consumer trips‑-have tested and verified the applicability of this market principle.[ref]Brian J. L. Berry and Allan Pred, Central Place Studies: A Bibliography of Theory and Applications (Philadelphia, 1961); John Urquhart Marshall, The Location of Service Towns: An Approach to the Analysis of Central Place Systems (Toronto, 1969); G. William Skinner, “Cities and the Hierarchy of Local Systems,” in idem (ed.), The City in Late Imperial China (Palo Alto, Calif., 1976); Bonnie Barton, “The Creation of Centrality,” Annals of the Association of American Geographers, LXVIII (1978), 34-44; Leslie J. King, Central Place Theory (Beverly Hills, Calif., 1984); Brian J. L. Berry et al., Market Centers and Retail Location: Theory and Applications (Englewood Cliffs, N.J., 1988).[/ref] As geographer Brian J. L. Berry concludes in his most recent summary of the evidence, whatever the initial impetus for place development, competition between incipient market centers, under conditions of differential “accessibility to the consuming population,” eventually results in a central‑place system. “At any point in time, the geographic distribution of retail and service business in central places approximates an equilibrium adjustment to the geographic distribution of consumers.”[ref]Berry et al., Market Centers, 4-5.[/ref]
Nevertheless, some social scientists have questioned the unrealistic assumptions of the theory. They note that neither consumers nor producers behave as purely rational “economic men”; perfect competition does not exist; few consumer‑producer relationships stand in sociocultural isolation; periodic markets occur in low demand areas; consumers make multi-purpose trips with distance not the only factor. Also, central place theory only emphasizes retail operations, ignoring important non‑central place functions like manufacturing, and becomes less useful as urban areas grow to the point where they can generate their own demand.[ref]For further discussion of central place theory and its limitations, see Wheatley, “Concept of Urbanism,” 613‑9; Carol A. Smith, “Regional Economic Systems: Linking Geographical Models and Socioeconomic Problems,” in idem (ed.) Regional Analysis (New York, 1976), I, 3-63; D. J. Walmsley and G. J. Lewis, Human Geography: Behavioural Approaches (London, 1984), 81-88.[/ref]
The greatest challenge to central place theory comes from historians, historical geographers, and other social scientists who find that such an overly mechanical and deterministic theory simply fails to describe the historical dynamics of urbanization in the Western Hemisphere and other areas of European colonization. Geographer James E. Vance, Jr. outlines a “mercantile” model, stressing the importance of historic exogenic forces on urbanization (especially wholesaling), the role of general stores, and institutionalized initial advantage.[ref]James E. Vance, Jr., The Merchant’s World: The Geography of Wholesaling (Englewood Cliffs, N.J., 1970).[/ref] Economist E. A. J. Johnson and anthropologist Carol Smith believe urbanization in developing nations better fits a “dendritic” rather than central place pattern, as “peasant-produced goods flow directly from rural areas to urban centers or major ports and in the process leave the domestic or peasant economy poorly serviced and undersupplied.”[ref]E. A. J. Johnson, The Organization of Space in Developing Countries (Cambridge, Mass., 1970); C. Smith, “Regional Economic Systems,” I, 34.[/ref] Unfortunately, this helpful corrective too often leads to the dichotomization of exogenic and endogenic forces and the too ready dismissal of central place theory.
Overall, recognition of the “real world” deficiencies of central place theory does not diminish the theory’s heuristic value and indeed has contributed to greatly improved theory.[ref]C. Smith, “Regional Economic Systems”; Michael P. Conzen, “The American Urban System in the Nineteenth Century,” in D. T. Herbert and R. J. Johnson (eds.), Geography and the Urban Environment: Progress in Research and Applications (Chichester, England, 1981), IV, 295-347.[/ref] Students of urbanization have moved far toward developing a synthetic model that describes how initial exogenic dendritic development gradually gives way to endogenic central place development.[ref]Vance, Merchant’s World, 164-5; Michael P. Conzen, “A Transport Interpretation of the Growth of Urban Regions: An American Example,” Journal of Historical Geography, I (1975), 361-82; Edward K. Muller, “Selective Growth in the Middle Ohio Valley, 1800-1860,” Geographical Review, LXVI (1976), 178-99; Edward K. Muller, “Regional Urbanization and the Selective Growth of Towns in North American Regions,” Journal of Historical Geography, III (1977), 21-39; Ronald Eugene Grim, “The Absence of Towns in Seventeenth‑Century Virginia: The Emergence of Service Centers in York County,” unpub. Ph.D. diss. (University of Maryland, 1977); Berry et al., Market Centers, 149.[/ref] One need not be a central place determinist in order to use the theory for it can provide simply a useful deductive model to compare and test against the historical record. While offering no definitive proof of the theory, such an exercise helps focus attention on possible processes or general behavioral tendencies at work and can help guide research. This paper begins with central place theory as an overly simplistic model that provides a good base for examining the complexities of historical context.
COMMUNITY STUDIES APPROACHES Contrary to many applications which focus on “top-down” urbanization, central place theory actually has its greatest heuristic value at the rural end of the rural‑urban spectrum. When so employed, the study of “bottom-up” urbanization blends well with the “community studies” approaches of human ecology, network analysis, and rural sociology that focus on the complex vertical and horizontal interrelationships between people, places, and institutions. All these approaches share some notion of centrality. Sociologist Amos Hawley’s ecological approach defines community as “the area, the population of which, however widely distributed, regularly turn to a common center for the satisfaction of all or a major part of its needs.”[ref]Amos H. Hawley, Human Ecology: A Theory of Community Structure (New York, 1950), 246.[/ref] Historian Darrett B. Rutman’s network analysis centers on the regular networks of person‑person and person‑place relationships occuring through “nodal points,” affected by landform, distance, and technology.[ref]Darrett B. Rutman, “Community Study,” Historical Methods, XIII (1980), 29‑41.[/ref] Central place theory simply makes a few basic assumptions about the nature of a particular set of social relations‑-those between providers and consumers of goods and services‑-in order to develop a hypothetico-deductive model.
Indeed, American rural sociologists, building on their analysis of the symbiotic relationship between farmstead and service center, published findings remarkably similar to and predating Christaller’s first German writings by twenty years. Charles J. Galpin introduced the concept of circular trade ranges for services and goods based on supply-demand economics with distance as a critical factor. John H. Kolb extrapolated these principles into a hierarchical nested arrangement of different order services towns along with an analysis of the interrelationship between numbers of functions and town population as sophisticated as what geographers would end up doing forty years later.[ref]Charles J. Galpin, The Social Anatomy of An Agricultural Community (Madison, Wis., 1915); John H. Kolb, Service Relations of Town and Country (Madison, Wis., 1923).[/ref]
Community studies employing the concepts of human ecology, network analysis, or rural sociology thus show many similarities to a central place approach to urbanization. Each can contribute greater theoretical sophistication to an urbanization framework as well as methodological techniques and empirical data for verification of theory.
URBANIZATION IN THE ANTEBELLUM COTTON SOUTH Almost all historians and social scientists studying antebellum United States urbanization seem to agree that the antebellum cotton South was a classic example of an exogenic, dendritic pattern.[ref]Robert E. Baldwin, “Patterns of Development in Newly Settled Regions,” Manchester School of Economics and Social Studies, XXIV (1956), 161-79; Douglass C. North, The Economic Growth of the United States 1790‑1860 (Englewood Cliffs, N.J., 1961); Vance, Merchant’s World; David Charles Weaver, “The Transport Expansion Sequences in Georgia and the Carolinas, 1670-1900: A Search for Spatial Regularities,” unpub. Ph.D. diss. (University of Florida, 1972); Carville Earle and Ronald Hoffman, “Staple Crops and Urban Development in the Eighteenth‑Century South,” Perspectives in American History, X (1976), 7‑78; Ralph V. Anderson and Robert E. Gallman, “Slaves as Fixed Capital: Slave Labor and Southern Economic Development,” Journal of American History, XLIV (1977), 24‑26; Goldfield, “Urban South.”[/ref] The current consensus divides the lower South into four general regions dominated by seaport entrepots (Charleston, Savannah, Mobile, New Orleans). Each regions contained major “break‑of‑bulk” centers located at strategic points such as the fall line (Columbia, Augusta, Macon, Columbus, Montgomery). Merchant factors in these major cities serviced the plantations, collecting cotton and delivering supplies. Planters and local merchants handled the cotton production of smaller producers through these same factors. General stores located in rural neighborhoods and county seats supplied self-sufficient yeoman farmers with their limited consumer needs.
In fact, neither historians nor geographers have seriously attempted to analyze central place development in the antebellum cotton South with any of the sophistication of the synthetic endogenic-exogenic framework applied to other times and places. The case against central place theory in an antebellum southern context seems to rest on cursory examinations of maps and attempts to apply a nebulous rank‑size rule.[ref]John B. Sharpless, City Growth in the United States, England and Wales, 1820-1861: The Effects of Location, Size and Economic Structure on Inter-Urban Variations in Demographic Growth (New York, 1977); Allan R. Pred, Urban Growth and City Systems in the United States, 1840-1860 (Cambridge, Mass., 1980).[/ref] Students of antebellum southern urbanization regularly ignore hamlets, villages, towns, and county seats for the “bright lights” of the largest cities.[ref]For a similar critique of typical attitudes of historians toward smaller urban units, see Darrett B. Rutman and Anita H. Rutman, “The Village South,” Small Worlds, Large Questions: Explorations in Early American Social History, 1600-1850 (Charlottesville, Va., 1994), 231-3.[/ref] This ready dismissal seems all the more remarkable when one considers the excellent work of historical geographers in tracing the evolution of central places from older dendritic systems in the colonial South and the number of Southern community studies undertaken by rural sociologists in the 1920s, 1930s, and 1940s all of which emphasized towns, villages, and hamlets and most of which incorporated historical analysis.[ref]For the colonial South, see Harry Roy Merrens, Colonial North Carolina in the Eighteenth Century: A Study in Historical Geography (Chapel Hill, 1964); Robert D. Mitchell, “The Shenandoah Valley Frontier,” Annals of the Association of American Geographers, LXII (1972), 461-86; Ernst and Merrens, “‘Camden’s Turrets”; Robert D. Mitchell and Edward K. Muller, “Interpreting Maryland’s Past: Praxis and Desiderata,” in idem (eds.) Geographical Perspectives on Maryland’s Past (College Park, Md., 1979); Grim, “Absence of Towns”; James O’Mara, An Historical Geography of Urban System Development: Tidewater Virginia in the 18th Century (Downsview, Ont., 1983). For rural sociologists, see Augustus W. Hayes, Some Factors in Town and Country Relationships (New Orleans, 1922); Carle C. Zimmerman and Carl C. Taylor, Rural Organization: A Study of Primary Groups in Wake County, N.C. (Raleigh, N.C., 1922); Luther C. Fry, A Census Analysis of American Villages (New York, 1925); Edmund deS. Brunner, Gwendolyn S. Hughes, and Marjorie Patten, American Agricultural Villages (New York, 1927); Maynard C. Conner, An Economic and Social Survey of Patrick County, Virginia (Charlottesville, Va., 1937); Irwin T. Sanders and Douglas Ensminger, Alabama Rural Communities: A Study of Chilton County (Montevallo, Ala., 1940); Harold Hoffsommer and Herbert Pryor, Neighborhood and Community Areas in Covington County, Mississippi (Washington, D.C., 1941).[/ref]
Nevertheless, at least two scholars have seriously considered a bottom-up theoretical framework for Southern urbanization. Sociologist Rudolf Heberle‑-whom historian David R. Goldfield accuses of applying “ahistorical geographic models to the Southern urban system”‑-developed a framework highlighting the importance of local trading and service centers founded by enterprising businessmen as the population grew and moved westward. For Heberle,
the size and density of these trading centers is essentially a function of the spatial extent of their trading area, of the density of population in this area, and of economic factors such as the value of farm products produced in the area and the aggregate purchasing power of the rural population and other factors such as the location of highways, railroads, and waterways. The geographical distribution of these places at any given time corresponds roughly to the distribution of the population.[ref]Rudolf Heberle, “The Mainsprings of Southern Urbanization,” in Rupert B. Vance and Nicholas J. Demerath (eds.), The Urban South (Chapel Hill, N.C., 1954) 19; Goldfield, “The Urban South,” 1032n.[/ref]
Geographer Wilbur Zelinsky, in his study of the historical settlement patterns of Georgia, offered a highly similar interpretation that he directly linked to Christaller’s central place theory. Zelinsky concluded that it was
impossible to venture an opinion on whether this theory adequately describes the siting of Georgia’s towns without an exacting and time-consuming analysis. Nevertheless, it is obvious from the overall evenness in the distribution of trading centers that Georgia offers an excellent laboratory for the Christaller hypothesis.[ref]Wilbur Zelinsky, “The Settlement Patterns of Georgia,” unpub. Ph.D. diss. (University of California, 1953), 42.[/ref]
Later research in antebellum southern urbanization by geographer Howard G. Adkins and historians Darrett B. Rutman and Anita H. Rutman provides empirical verification for Heberle’s and Zelinsky’s “central place” pattern of smaller urban places, emphasizing more specifically the strong correlation between the number of towns and the density of white farm population.[ref]Howard G. Adkins, “The Historical Geography of Extinct Towns in Mississippi,” unpub. Ph.D. diss. (University of Tennessee, 1972), 76-8; Rutman and Rutman, “Village South,” 231-72.[/ref]
COUNTY SEATS AS CENTRAL PLACES Both Heberle and Zelinsky acknowledged that county seat towns represented an important sub‑group of trading centers.[ref]Zelinsky, “Settlement Patterns” 307-8; Heberle, “Mainsprings,” 19.[/ref] Social scientists have, however, yet to develop the implications of county formation and county seat location on southern urbanization. Many social scientists, as well as novelists and popular writers, have noted the importance of the county in southern society, finding the southern county coextensive with “community.”[ref]Conrad M. Arensberg and Solon T. Kimball, Culture and Community (New York, 1965), 106-8; Glenn V. Fuguitt, “County Seat Status as a Factor in Small Town Growth and Decline,” Social Forces, XLIV (1965), 250; Richard V. Francaviglia, “County Seat Centrality as a Regional Trait,” Geographical Survey, II (1973), 6.[/ref] A 1920s study of St. Matthew’s County, South Carolina, and other southern counties, found county lines “effective barriers.”[ref]Brunner et al., American Agricultural Villages, 86-7.[/ref] Nevertheless, some rural sociologists have noted three or more well‑defined communities within southern county boundaries.[ref]Sanders and Ensminger, Alabama Rural Communities; Hoffsommer and Pryor, Neighborhood and Community.[/ref] More correctly, Darrett B. Rutman points out that community boundaries do not necessarily coincide with political boundaries, but the local political structure does provide a nodal point for those within the political boundaries that tends to draw people into the county seat.[ref]Rutman, “Community Study,” 41n22.[/ref]
Galpin stressed the importance of the initial advantage of the county seat in decisions for locating other services, consequently reinforcing the seat’s centrality.[ref]Cited in John H. Kolb and Edmund deS. Brunner, A Study of Rural Society: Its Organization and Changes (Boston, 1940; 2nd ed.), 112-3.[/ref] Adkins believes that “as a result of the cotton economy very few towns would have developed in any of the inland areas had it not been for the need for area administration.” But, once established, the “centripetal force of the courthouse” soon attracted economic and social as well as political functions.[ref]Howard G. Adkins, “The Geographic Base of Urban Retardation in Mississippi, 1800‑1840,” West Georgia College Studies in the Social Sciences, XII (1973), 40-2.[/ref] With time, through a positive multiplier effect, the county seats in the South gained even greater centrality as consumers increasingly made multi-purpose trips into the county seat and county-seat merchants took advantage of external economies to absorb the trade of smaller places.[ref]Kolb and Brunner, Study of Rural Society, 9-10; Zelinsky, “Settlement Patterns,” 308; Weymouth Jordan, Ante‑Bellum Alabama Town and Country (Tallahasee, Fla., 1957) 22-40; William H. Baker, “The Economics of a Small Southern Town,” unpub. Ph.D. diss. (University of Alabama, 1963), 164; William Warren Rogers, Ante‑Bellum Thomas County 1825‑1861 (Tallahassee, Fla., 1963); Arensberg and Kimball, Culture and Community, 106; James H. Soltow, The Economic Role of Williamsburg (Williamsburg, Va., 1965), 51; Adkins, “Historical Geography,” iv; Wellenreuther, “Urbanization in the Colonial South,” 665; William Warren Rogers, “The Way They Were: Thomas Countians in 1860,” Georgia Historical Quarterly, LX (1976), 131‑44; John T. Schlotterbeck, “Plantation and Farm: Social and Economic Change in Orange and Greene Counties, Virginia, 1716 to 1860,” unpub. Ph.D. diss. (Johns Hopkins University, 1980), 15; O’Mara, Historical Geography, 193; Rutman and Rutman, “Village South,” 263-4.[/ref]
Many scholars of Northern urbanization have also noted the importance of county seats. Geographer Michael F. Dacey has even gone so far to suggest that a “county seat model”‑-based on the axiomatic assumption of the political division of American states into counties, with one city or town in each county designated county seat‑-can explain much of the distribution and arrangement of American cities and towns.[ref]Michael F. Dacey, “A County‑Seat Model for the Areal Pattern of an Urban System,” Geographical Review, LVI (1966), 527‑42. Edward Stephan and Douglas R. McMullin, “The Historical Distribution of County Seats in the United States: A Review, Critique, and Test of Time‑Minimization Theory,” American Sociological Review, XLVI (1981), 907‑17, present evidence that the distribution of county seats from 1790 to the present has been a function of population density raised to the two‑thirds power.[/ref] Much of Dacey’s proof rests on the recognition that county seats tend to be the most populous towns within each county, but his model never really explains why a county seat should not just remain a small political unit. The framework developed here suggests that county seats simply provide a focus, a convenient, fairly stable (owing to its political function) central place, attracting both providers and consumers of goods and services. Although initially only a political unit, the county seat quickly takes on permanent socioeconomic functions. While other central places arise in the county rise and fall, the southern county seat usually retains its initial advantage, with its economic and political centrality constantly reinforced in the face of changing modes of transportation (the rise of railroads, for example).[ref] Edmund deS. Brunner and J. H. Kolb, Rural Social Trends (New York, 1933); Selz C. Mayo, “Two Population Characteristics of County‑Seat Towns in North Carolina,” Rural Sociology, XII (1947), 423‑6; Zelinsky, “Settlement Patterns,” 40-1; A. Alexander Fanelli and Harold A. Pederson, Growth Trends of Mississippi Population Centers, 1900‑1950 (State College, Miss., 1956); Fuguitt, “County Seat Status”; Adkins, “Historical Geography.” Rutman and Rutman, “Village South,” 269, note an exception in Houston County, Georgia, where the town of Fort Valley won the route of the Southwestern Railroad away from and subsequently surpassed the county seat Perry.[/ref]
The county seat was clearly the major central place in almost every southern county from the antebellum era to the present. Moreover, county seats were generally in the geographic center of most southern counties. Initially sited by a state‑appointed commission shortly after the establishment of a new county, the county seat immediately became a matter of competition between various intracounty interests and a plurality of voters could force a relocation.[ref]Richard Pillsbury, “The Morphology of the Piedmont Georgia County Seat Before 1860,” Southeastern Geographer, XVIII (1978), 115.[/ref] Inevitably, for most southern counties, the geographic center became the only politically (and in certain states like Mississippi the only legally) feasible choice.[ref]Guion Griffin Johnson, “The Ante‑Bellum Town in North Carolina,” North Carolina Historical Review, V (1928), 373; Adkins, “Geographic Base,” 42; Francaviglia, “County Seat Centrality”; O’Mara, Historical Geography, 194. Those counties with extreme peripheral seats, usually located on major rivers or lakes (reflecting the dominance of early modes of transportation), were usually established trade centers with predominant political influence over the surrounding hinterland well before the county formed.[/ref] But politics, not geography, created and maintained the centrality of the county seat. Decisions to relocate county seats normally led to the demise of the old site.[ref]Adkins, “Historical Geography,” 94; Susan Yelton, “Newnansville: A Lost Florida Settlement,” Florida Historical Quarterly, LIII (1974), 319‑31; Pillsbury, “Morphology,” 116; Rutman and Rutman, “Village South,” 264, 266.[/ref]
Geographer Richard V. Francaviglia finds “county seat centrality” a distinctive southern cultural trait, reflecting a strong agrarian, egalitarian, legalistic tradition, traceable to the colonial “Virginia plan” of county government.[ref]Francaviglia, “County Seat Centrality.” This southern emphasis on centrality also manifests itself in the unusual number of “round towns” in Georgia and the Carolinas and the central courthouse square. See Zelinsky, “Settlement Patterns,” 340-5; Howard A. Schretter, “Round Towns,” Southeastern Geographer, III (1963), 46‑52; Edward T. Price, “The Central Courthouse Square in the American County Seat,” Geographical Review, LVIII (1968), 29-60.[/ref] Perhaps this Virginia plan evolved out of numerous battles in eighteenth-century Virginia over efforts by commercially‑oriented leaders to relocate county seats to more economically strategic but less geographically central locations. Such moves often provoked an uproar from the citizens of the more remote parts of the county, as in Fairfax County, Virginia.[ref]Wellenreuther, “Urbanization in the Colonial South,” 665n.[/ref] When certain “progressive-minded” citizens attempted to move the seat of Spotsylvania County, Virginia to Fredericksburg, inhabitants of the western part of the county raised such a fuss that eventually a new county (Orange) was created and Fredericksburg independently incorporated.[ref]E. Lee Shepard, “‘The Ease and Convenience of the People’: Courthouse Locations in Spotsylvania County, 1720‑1840,” Virginia Magazine of History and Biography, LXXXVII (1979), 279‑99.[/ref] Later, the western third of Orange County agitated against the “‘incalculable inconvenience and consequent oppression from their remote situation from their Court House'” and separated off to form Greene County.[ref]Schlotterbeck, “Plantation and Farm,” 226-8. See also William H. Gaines, Jr., “Courthouses of Amelia and Dinwiddie Counties,” Virginia Cavalcade, XVIII (1969), 17‑28, on courthouse sites and county boundary battles in Prince George, Dinwiddie, and Amelia Counties, Virginia.[/ref]
Combined with the tendency for southern county borders to follow river and creek valleys (also traceable to eighteenth-century Virginia practice), the predisposition to areal centrality necessarily forced county seats away from waterways onto high ridges.[ref]Johnson, “Ante-Bellum Town”; Weaver “Transport Sequence,” 135. The map of Yoknapatawpha County in William Faulkner, Absalom, Absalom! (New York, 1936), strikingly shows the county seat centrally located in the county on the ridge between the Tallahatchie and Yoknapatawpha Rivers and the courthouse in the dead center of town.[/ref] General concerns about the salutary benefits of situating towns away from unhealthy rivers and swamps, as well as spring floods, mutually reinforced these geographical and political factors. The town of Gaston in Sumter County, Alabama, was located twelve miles from the Tombigbee River in hilly terrain in order to provide a healthier climate but was still considered “a river town.” For these Alabamians, the benefits from a healthier climate outweighed any inconvenience due to transportation difficulties.[ref]Louis Roycraft Smith, Jr., “A History of Sumter County, Alabama, through 1886,” unpub. Ph.D. diss. (University of Alabama, 1988), 86, 91.[/ref]
This forcing of the county seats away from the most economically advantageous location must have had an overall detrimental effect on southern urbanization. The case of East Texas demonstrates the great impact that alternative county formation patterns could have on subsequent urbanization. Counties influenced by early Spanish empresario land grants, which lay on both sides of rivers with the centrally located county seat consequently on the river, urbanized much more rapidly than those more traditionally formed counties.[ref]Dan Dillman, “The Relationship of Early Settlement on the Gulf Coastal Plain to the Use of Rivers as County Boundaries in Texas,” Papers of the Michigan Academy of Science, Arts, and Letters, XLVII (1962), 429‑36.[/ref]
NON-COUNTY SEAT CENTRAL PLACES While county seats were key links in Southern central place development, other smaller central places also developed. Whereas county seats came into existance for basically political reasons, smaller urban units arose mainly due to the efforts of entrepreneurs.[ref]Zelinsky, “Settlement Patterns,” 307-8.[/ref] Often general stores formed the nuclei of new towns.[ref]Adkins, “Historical Geography,” 57; James Oakes, The Ruling Race: A History of American Slaveholders (New York, 1982), 93-4. John C. Hudson, Plains Country Towns (Minneapolis, 1985), 28, also emphasizes the entrepreneurial role of general merchants who purposefully tried to create “centrality.”[/ref] A few backcountry hill towns with a good supply of spring water, like Holly Springs, Mississippi, and Madison Springs, Georgia, arose as small resort towns.[ref]E. Merton Coulter, “Madison Springs, Georgia Watering Place,” Georgia Historical Quarterly, XLVII (1963), 375‑407; James H. Stone, “The Economic Development of Holly Springs During the 1840’s,” Journal of Mississippi History, XXXII (1970), 341‑61.[/ref] Factory towns arose at dam sites on creeks and rivers. Antebellum southern railroads created some new central places. Several antebellum Georgia towns centered around an academy of higher education.[ref]Adiel Sherwood, A Gazeteer of the State of Georgia (Athens, Ga., 1939; orig. pub. 1827), 62, 79, 88, 100; Elbert W. G. Boogher, Secondary Education in Georgia, 1732‑1858 (Philadelphia, 1933).[/ref]
When purposefully siting a service center‑‑whether by virtue of the government establishing a post office, an entrepreneur a general store, or a local committee an academy‑‑centrality to service area and distance from competing institutions always figured highly. Some services, whether through traditional custom or economic circumstances, went directly to the customer on an intermittent basis rather than remain fixed in central places. For example, daguerrotype artists annually travelled a county seat circuit staying a week or two at a time. Dentists and other professionals, while typically settled most of the year in one place, regularly visited outlying villages and hamlets. In the South, insufficient demand may have forced many craftsmen to practice their trade only part‑time.[ref]Dale L. Couch, “John Riley Hopkins: A Nineteenth‑Century Georgia Cabinetmaker,” Atlanta Historical Journal, XXVIII (1984), 43‑55.[/ref] Such intermittent services, however, when performed in an urban unit contributed to the centrality of that unit.
As already described for county seats, the initial impetus for urbanization in these lower-order central places also has a positive multiplier effect.[ref]Zelinsky, “Settlement Patterns,” 308-10.[/ref] Post offices quickly attracted general stores (or vice versa) since people regularly combined a shopping trip with a chance to pick‑up their mail.[ref]Sam T. Bratton and W. O. Smith, “Historical Geography of Salt River Community, Audrain County, Mo.,” Missouri Historical Review, XXIII (1928), 91‑98.[/ref] The presence of an academy in a town strengthened all the trade relationships.[ref]Kolb and Brunner, Study of Rural Society, 112-3.[/ref] Although most sawmills and gristmills, churches, and schools remained highly decentralized in the antebellum South, frequently they concentrated at some particular place in various combinations.[ref]Hoffsommer and Pryor, Neighborhood and Community Areas; Glenn T. Trewartha, “The Unincorporated Hamlet: One Element of the American Settlement Fabric,” Annals of the Association of American Geographers, XXXIII (1943), 39-40, 51-2; Schlotterbeck, “Plantation and Farm,” 214.[/ref] Any good model of historical central place development must explain not only the initial impetus but why certain places grew and others did not.
HANCOCK COUNTY, GEORGIA, 1860: A TEST CASE Although a regional study provides the best framework for testing central place models, a study of a single community can provide helpful insights, particularly toward an understanding of bottom-up urbanization. The county chosen for study‑-Hancock County, Georgia‑-could hardly be mistaken for an “urban” county in the traditional sense. Located in what contemporaries called Middle Georgia and what historians call the eastern part of the Lower Piedmont or Black Belt–characterized in the antebellum era by large cotton plantations and slaves–the largest urban unit had a white population of only 676 individuals in 1860. This paper makes no claim that Hancock County was representative of either the Black Belt, Georgia, or the South. The county merely serves as a first test for the incipient model thus far developed and provides a good example of how the theory can be tested in other parts of the antebellum South.[ref]Hancock County was chosen for study because of the great number of extant antebellum merchant records, a necessary resource for understanding changes in trade relationships over time.[/ref]
As shown in Table I, however one calculates centrality‑-whether based on population, number of merchants, the value of merchandise in stock, postmaster compensation (a function of the volume of post office business), or an index based on the prevalence of non-agricultural occupations‑-Sparta (the county seat of Hancock County) displayed a high degree of primacy over other urban units in the county.[ref]Following Marshall’s framework, Hancock County’s urban units fall into natural categories of town (Sparta), village (Mount Zion, Powelton, Linton, Culverton), and hamlet (Shoals of Ogeechee, Island Creek, Mayfield), although Culverton’s population seems rather small in comparison with other measures of centrality. Rutman and Rutman, “Village South,” 269, divides Houston County’s urban units along similar lines.[/ref] Figure II shows Hancock County and the surrounding counties, with identified urban units scaled by measures of postmaster compensation, and primary interurban transportation routes.[ref]Richard W. Helbock, “Postal Records as an Aid to Urbanization Studies,” Historical Methods Newsletter, III (1970), 9‑14, and Rutman and Rutman, “Village South,” 244-6, find a strong correlation between postmaster compensation and town population. For other studies of urbanization employing postal service information, see Allan R. Pred, Urban Growth and the Circulation of Information: The United States System of Cities, 1790‑1840 (Cambridge, Mass., 1973) 78-103; John A. Alwin, “Post Office Locations and the Historical Geographer: A Montana Example,” Professional Geographer, XXVI (1974), 183-86; James R. Shortridge, “The Post Office Frontier in Kansas,” Journal of the West, XIII (1974), 83-97; Skinner, “Cities and the Hierarchy,” 348.[/ref] A brief comparison with Christaller’s central place patterns in Figure I would seem to indicate that central place principles are at work in Middle Georgia, but exactly which principles and in what proportions is not quite clear. Determining the forces behind the pattern in Figure II requires a combined deductive theoretical and inductive historical approach.
Clearly this pattern reveals the predominance of county seats over each county and the region as a whole, confirming geographer Richard Pillsbury’s conclusion that “the county seat was the most important type of urban center on the Georgia Piedmont before 1860.”[ref]Pillsbury, “Morphology,” 155. Penfield in Greene County, site of Mercer University, and Midway in Baldwin County, site of Oglethorpe University, have exceptionally high populations and post office activity for non‑county seat towns due to the presence of the universities.[/ref] Certain county seats in newer and smaller counties such as Crawfordville (Taliaferro County established 1825) and Gibson (Glascock County established 1858) remained relatively small, while Milledgeville (Baldwin County) as the state capital grew exceptionally large. However, no clear central place hierarchy developed. Overall, the county seats appear to form a regularly spaced (20-30 miles equidistant) network of urban units of approximately the same order of centrality.
Only Christaller’s separation/administration principle emphasizes the centrality of administrative centers like county seats. Indeed, the separation/administration principle appears ideal for explaining the Middle Georgia urban pattern since it hypothesizes “the creation of virtually complete districts, districts of practically equal area and population, in the center of which lies the most important place. The borders lie in the thinly settled regions and should follow closely the natural borders and barriers.”[ref]Christaller, Central Places, 77-8.[/ref] And indeed, as that principle predicts, rings of smaller urban units do appear at a range of about 10‑15 miles from the county seat.[ref]Rutman and Rutman, “Village South,” 234-6, find similar rings in Houston County, Georgia. See also Roger Raymond Van Dyke, “Antebellum Henry County,” West Tennessee Historical Society Papers, XXXIII (1979), 48‑80.[/ref] (Compare Figure I(c) with Figure II.) But the separation/administration principle has many practical problems when applied to Hancock County. Firstly, the rings do not always seem to recognize county boundaries. Secondly, contrary to theoretical implications, nowhere do urban units appear directly across from each other on opposite sides of county boundaries. Indeed along the Ogeechee River, several hamlets such as Shoals of Ogeechee, Rock Mill, and Mayfield grow up on the county border on both sides of the river. Thirdly, rural population densities may be greatest near county seats, but there is no evidence that county borders are “thinly settled regions.” Fourthly, constant complaints of political alienation from outlying regions of counties were constantly leading to the formation of new counties like Glascock County in 1858. Fifthly, in Georgia individuals frequently had their county line property switched from one county’s tax rolls to the other’s, thus creating highly crenellated (and hardly sacrosanct) county borders.[ref]Burke G. Vanderhill and Frank A. Unger, “Georgia’s Crenelated County Boundaries,” West Georgia Collge Studies in the Social Sciences, XVI (1977), 59-72.[/ref] Overall, as Rutman and Rutman note, although the county seat provides a nodal point for all who live within the county boundaries, these invisible borders determine very little.
An analysis of the account books of T. I. [Thomas Irby] Little & Co. (a general store located in Hancock’s county seat of Sparta) for the years 1859 and 1860, shows that the centrality of the county seat, as measured by the percentage of households with credit accounts with Little, distinctly declined with distance from the seat. (See Figure III.) County seat centrality came mostly from business generated by customers within an approximate ten-mile radius.[ref]Robert C. Kenzer, Kinship and Neighborhood in a Southern Community: Orange County, North Carolina, 1849‑1881 (Knoxville, Tenn., 1987), and Gregory S. Rose, “Reconstructing a Retail Trade Area: Tucker’s General Store, 1850‑1860,” Professional Geographer, XXXIX (1987), 33‑40, confirm the importance of distance to the retail trade of a general store. Arensberg and Kimball, Culture and Community, 106, define the limits of a southern community in terms of the roundtrip distance a man could travel in one day.[/ref]
Almost every head of household and every person with an occupation in Sparta, apart from the mill workers, had a credit account with T. I. Little. From the planter buying cigars and brandy, to the carpenter buying paints and varnishes, to the physician buying drugs and medicines, to the little girl buying nuts and candy, everyone in Sparta eventually went into Little’s. Indeed, such a relationship with the general store was probably an essential component of being a Spartan. Even Sparta’s other merchants all had credit acounts at Little’s. In the nearer districts, as many as 50% of the heads of household had credit accounts; in the outermost districts the percentage dropped to as low as 7%. The great majority of these “rural” customers were planters and farmers, although a few overseers and farm laborers had accounts. A quarter of the heads of household from the village of Mt. Zion (seven miles north of Sparta) and the hamlet of Culverton (five miles east) and half of the heads from the new village of Linton (twelve miles south) had credit accounts despite the existence of local general stores.
In the village of Powelton (fourteen miles northeast of Sparta), on the other hand, only 3.6% (1 out of 28) of village households and 7.1% (3 out of 42) of surrounding rural households (Militia District 106) had a credit account with Little. Thus, not only were county borders invisible, but the influence of the one common nodal point weakened considerably over distance. Indeed, as a political expedient, county officials visited Powelton and other outlying districts several times a year rather than expect the citizens to have to travel to Sparta.
The essential point is not that counties defined political or community boundaries, but that certain places were given an initial advantage by a political process at critical moments in the urbanization process of a region. Since the administrative function of county seats did not define their centrality and county boundaries were not trade boundaries, the separation/administration principle cannot explain the distribution of smaller non‑county seat central places.
In Hancock County, the larger central places began mainly as “academy towns” (Powelton, Mt. Zion, Island Creek, Linton, Culverton), while smaller “mill towns” arose on the Ogeechee River (Mayfield, Rock Mill, Shoals of Ogeechee).[ref]Sherwood, Gazetteer, 79, 88; Boogher, Secondary Education, 368-9; Elizabeth Wiley Smith and Sara S. Carnes, The History of Hancock County, Georgia (Washington, Ga., 1974), I, 23-6, 101-13.[/ref] A long tradition as a county devoted to higher education, combined with the absence of a railroad, channeled entrepreneurial and reformist talent‑-usually in the form of committees of prominent local men‑-into founding academies. Just as the county seats attracted other goods and services, so did the academies and mills, by offering foci for previously dispersed activities. Usually the entrepreneurs carefully considered access to a potential clientele when choosing a site. Academies furthermore provided many social benefits to the community that motivated many rural citizens to move to the towns, villages, and hamlets that frequently grew up around the schools. The better academies took on non-central place functions, attracting students from far outside the local area, students who brought in significant revenues to Hancock County in the form of tuition and room and board.
Continuing to follow Christaller, one should look to either the market principle or traffic principle (or both) for explaining the distribution and arrangement of such lower-order central places. A brief glance at Figure II suggests the traffic principle, as most non-county seat places of any importance (Linton, Mount Zion, Culverton, White Plains) lay on the main transportation routes between county seats. The market principle hypothesizes that the next order of places below county seats would be located near the midpoint of three county seats. Powelton, midway between Sparta, Warrenton, and Crawfordville would appear to fit this description. But the evidence shows that in fact Powelton‑-which had thrived in the 1810s and 1820s as an academy town with a strong backcountry trade‑-steadily declined after the creation of Taliaferro County in 1825 and the rise of Crawfordville as county seat and railroad town in the 1830s.[ref]Smith and Carnes, History of Hancock County, I, 111.[/ref] Indeed, following the traffic principle, Powelton’s early trade advantage may have been due more to its strategic position at the crossroads of the Greensboro-Warrenton-Augusta and Washington-Sparta-Milledgeville roads, a position fundamentally undermined by the coming of the railroad in the 1830s.[ref]See Marion R. Hemperly, Georgia, Early Roads and Trails, circa 1730-1850 (Atlanta: Georgia Dept. of Transportation, n.d.). Washington was the county seat of Wilkes County, off of Figure III to the northeast. Island Creek, midway between Milledgeville, Eatonton, and Sparta, was also an academy town but never really grew and had no merchant or general store in 1860.[/ref] Overall, the traffic principle best explains the location of the order of central places below the order of county seats (together encompassing almost all the general stores as well as post offices).
But just what exactly does the traffic principle imply about the entrepreneurs who founded such places? The traffic principle does not negate the market principle since the spacing of central places along the main traffic arteries still depends on sufficient consumer demand. Christaller simply stressed that the traffic principle comes into play when (1) the cost of establishing and maintaining routes stymies general road development and (2) the difficulty and expense of transportation inhibits traffic on inferior roads forcing most major traffic onto a few superior roads, generally the shortest routes possible between two larger order central places.[ref]Christaller, Central Places, 74.[/ref] Certainly this scenario fits the antebellum South, where even the superior roads were, by unanimous consent, notoriously bad. With limited funds, counties put the most money into those roads connecting the county seat with major outside centers, namely other county seats. Thus the major roads normally follow the routes connecting county seat to county seat. This pattern was reinforced by postal routes that followed the major roads, which indeed were often called “post roads.”[ref]James Silk Buckingham, travelling from Macon to Columbus in 1842, found post offices spaced along the road every ten to twelve miles, usually combined with a tavern, general store, blacksmith shop, and sometimes a tannery, church, or mill. Cited in Rutman and Rutman, “Village South,” 250.[/ref] Major collection points for staples would also tend to lie on major transportation routes. The introduction of the railroad strengthened the importance of transportation routes as small central places off the railroad declined at the expense of new railway stations creating a highly linear urban pattern.[ref]Charles E. Allred, Benj. H. Luebke, and Jas. H. Marshall, Trade Centers in Tennessee 1900-1930 (Knoxville, Tenn., 1936); Zelinsky, “Settlement Patterns,” 40.[/ref] All things considered, any southern entrepreneur would choose a place on a major road if a sufficient market could be tapped, and thus rural central places in the antebellum South would tend to fall evenly spaced along the major roads connecting county seats, which is the pattern noted in Figure II.[ref]Obviously much work remains before such a conclusion can be reached for the rest of the antebellum South. The map of towns in 1840 Mississippi in Adkins, “Geographic Base,” 41, shows far more randomness than that described here.[/ref]
Furthermore, the market principle (as well as less economically “rational” principles) undoubtedly contributed to the distribution of more decentralized services like mills, churches, and schools, frequently located along minor county roads in neighborhoods close to the consuming population. Whether in the antebellum era such neighborhoods formed “communities” around non‑county seat central places‑-such as those identified by rural sociologists in the inter-war period‑-is difficult to determine from the evidence available. However, the greater self‑sufficiency of farms, the difficulty of transportation, and the reduced commerce with the county seat in the outlying parts of the county would all have increased the tendency to promote separate communities within a county.[ref]Robert Kenzer’s Cane Creek “neighborhood” in Orange County, North Carolina, totally isolated from the county seat Hillsborough, would fully qualify as a “community” in the rural sociological sense of the term.[/ref]
THE MERCANTILE MODEL IN ANTEBELLUM GEORGIA The analysis of urbanization presented so far overlaps only slightly with previous top‑down approaches of southern urbanization. Although this study can only offer local evidence to substantiate the point, Vance’s mercantile model appears basically correct for describing the growth of places above the limited order of county seats. U. B. Phillips and J. William Harris have both placed antebellum Hancock County within Augusta’s hinterland and T. I. Little’s account books confirm that Augusta wholesalers predominantly supplied Sparta merchants.[ref]Ulrich Bonnell Phillips, A History of Transportation in the Eastern Cotton Belt (New York, 1968; orig. pub. 1908), 126-7; J. William Harris, Plain Folk and Gentry in a Slave Society: White Liberty and Black Slavery in Augusta’s Hinterlands (Middletown, Conn., 1985), 5. Little purchased cigars and tobacco from J. Volger and medicinal drugs from Haviland & Chichester, both identified as Augusta firms from advertisements in John P. Campbell, The Southern Business Directory and General Commercial Advertiser (Charleston, 1854).[/ref] Since none of the outlying general stores in Hancock had any account with Little, one can presume that the same or similar Augusta wholesalers also supplied the “country” stores. As the secondary literature reports, wholesaling centers reached out to the smallest general stores and individual plantations and farms across the South and enhanced their own development at the expense of county seats or other intermediate central places above the level of the general store.[ref]Fred Mitchell Jones, Middlemen in the Domestic Trade of the United States 1800‑1860 (Urbana, Ill., 1937); Lewis E. Atherton, The Southern Country Store, 1800‑1860 (Baton Rouge, 1949); Harold D. Woodman, King Cotton and His Retainers (Lexington, Ky., 1968), 76-83; Rita Moore Krouse, “The Germantown Store,” North Louisiana Historical Association Journal, VIII (1977), 53-64; Kenneth R. Wesson, “The Southern Country Store Revisited: A Test Case,” Alabama Historical Quarterly, XLII (1980), 157-66; Doris D. Fanelli, “William Polk’s General Store in Saint George’s, Delaware,” Delaware History, XIX (1981), 212‑228. Nevertheless, there are problems with Vance’s analysis. He simply assumes that all general stores are identical and provide the same goods and services, thus allowing him to dismiss lower-order urbanization above the level of the country store. In reality general stores did differ in the quantity and quality of goods offered and did specialize. Vance also ignores the role of tradesmen, professionals, and others who provide endogenic services, along with other central place services like churches and schools and other non‑central place functions (like manufacturing) which create much diversity among the lower and middle orders of urbanization.[/ref]
The larger break‑of‑bulk centers and seaports probably did not owe much of their growth to retail trade from outside their county, although the railroad likely increased such trade. In 1858, newspapers in Macon, Georgia, reported that “delegations of ladies came for shopping from Forsyth, Griffin, Barnesville, Thomaston, Milledgeville, Perry, Fort Valley, Americus, Albany, Eatonton, and Oglethorpe,” but most farmers must have let their local retailer make such purchasing trips.[ref]William Thomas Jenkins, “Ante Bellum Macon and Bibb County, Georgia,” unpub. Ph.D. diss. (University of Georgia, 1966), 256. One little studied aspect of southern retailing, however, concerns the mail order operations of large wholesalers who regularly advertised in local newspapers.[/ref] Those county seats in antebellum Georgia that grew to higher-order towns‑-like Atlanta, Marietta, Griffin, and Albany‑-did so mainly through wholesaling and manufacturing rather than retail trade.[ref]David F. Weiman, “Urban Growth on the Periphery of the Antebellum Cotton Belt: Atlanta, 1847-1860,” Journal of Economic History, XLVIII (1988), 259-72; Rutman and Rutman, “Village South,” 255-7. Kenneth Weiher, “The Cotton Industry and Southern Urbanization, 1880‑1930,” Explorations in Economic History, XIV (1977), 120‑40, identifies major central place development in the postbellum South with the rise of intermediate central place towns like Augusta, Macon, and Athens due to the advent of the new cotton press and oil mill industries. But Weiher fails to take proper account of the pre‑1880 urban system which he declares “virtually nonexistent.” Too literally accepting the insights of August Losch, Weiher fails to incorporate the insights of Vance and thus ignores the fact that these “new” towns had always existed as an urban order above the level of the “cotton gin” county seats.[/ref] Much work still needs to be done to understand the nature of relationship between retailing, wholesaling, and manufacturing in the antebellum cotton South, but the fact that so few higher-order central places existed per county seat shows that in this region urbanization beyond the level of county seat did not rest primarily on retail trade.
CONCLUSIONS As Christaller concludes, his central place principles are not mutually exclusive and his own work, along with that of August Losch and Carol Smith, finds much interaction between the principles in actual practice. Many historical geographers of American urbanization outside the antebellum cotton South acknowledge overlap between exogenic dendritic development and endogenic central place development. The model of urbanization presented here‑-combining a county seat model, a traffic principle, and a mercantile model‑-continues in this tradition. Moreover, this paper reminds social scientists that urbanization is a complex historical process that demands much of those wishing to apply modern theory. While social science theory helps us to frame questions and to seek commonalities, the answers must always remain historical, built solidly on empirical observations and an awareness of historical peculiarities.
[cite]