The Environmental Origins of Shifting Cultivation: Climate ...shifting cultivation began with the burning of forest growth to release nutrients into the soil. After five or six years,

The Environmental Origins of Shifting Cultivation: Climate, Soils, and Disease in theNineteenth-Century US SouthAuthor(s): John Majewski and Viken TchakerianReviewed work(s):Source: Agricultural History, Vol. 81, No. 4 (Fall, 2007), pp. 522-549Published by: Agricultural History SocietyStable URL: http://www.jstor.org/stable/20454756 .Accessed: 26/12/2011 15:00

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

Agricultural History Society is collaborating with JSTOR to digitize, preserve and extend access toAgricultural History.

http://www.jstor.org

http://www.jstor.org/action/showPublisher?publisherCode=ahs

http://www.jstor.org/stable/20454756?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

The Environmental Origins of Shifting Cultivation: Climate, Soils, and Disease in the Nineteenth-Century US South

JOHN MAJEWSKI AND VIKEN TCHAKERIAN

Farmers and planters in the antebellum South held large tracts of unim proved land because they practiced shifting cultivation. Southern cultivators burned tracts of forest growth to quickly release nutrients into the soil. After five or six years, when the soil had been depleted, the old field was aban doned for as long as twenty years. Environmental factors such as poor soils, rugged topography, and livestock diseases accounted for the persistence of this practice, more so than slavery or the availability of western lands. Shifting cultivation slowly declined in the postbellum era, but southern farmers continued to improve a far smaller percentage of their land well into the twentieth century.

IN 1843 JULIAN RUFFIN TOOK CONTROL of Ruthven, a plantation in

Prince George County, Virginia. Hardly representative of the typical planter, Julian-the son of famous agricultural reformer Edmund Ruf

fin-sought to improve and rationalize his plantation operations. His

journal, though, hints at a traditional cultivation regime, one that would

remain an important part of southern agriculture throughout the nine

teenth century. From the start, Ruffin noted that only a small portion of

JOHN MAJEWSKI is an associate professor in the history department at UC-Santa

Barbara. His first book, entitled A House Dividing: Economic Development in Pennsyl vania and Virginia before the Civil War, was published by Cambridge University Press in

2000 and reissued in paper in 2006. He is currently working on a book-length project on

the economic vision of Confederate secessionists. VIKEN TCHAKERIAN taught for many years in the economics department at California State University-Northridge and is

now an independent scholar living in the Seattle area. He has published extensively on the

economic development of the nineteenth-century South.

? the Agricultural History Society, 2007

522

2007 The Environmental Origins of Shifting Cultivation

the farm consisted of improved acreage: "This farm consists of 351 acres ... only about 40 acres are cleared and a great part of this will require

heavy grubbing." For the next six weeks, Ruffin's workforce of six slaves worked hard to clear more land-burning trees and then laboriously removing their stumps and roots-while at the same time plowing fields, building fences, and hauling manure. Ruffin never indicated how much land he and his slaves managed to clear, but given the size of his labor

force it seems doubtful that he made significant inroads into his plan tation's large number of unimproved acres. This reform-minded planter continued to clear land for many years.1

The small percentage of improved land on Ruffin's plantation was typical of the antebellum South. In 1860 southerners cultivated only one out of every three acres of land on their farms, while northerners im

proved more than half of their acreage. Even the recently settled Mid west had a far higher percentage of improved land than Virginia, North Carolina, and other states along the South Atlantic coast. These differ ences resulted from two distinct agricultural regimes. Northern farmers generally practiced what might be called continuous cultivation. Ma nures and rotations kept a high proportion of land in constant use, either for crops or improved pasture, Southerners, on the other hand, practiced a highly commercialized form of shifting cultivation in which a substan tial portion of acreage rested in prolonged fallow. The basic routine of shifting cultivation began with the burning of forest growth to release nutrients into the soil. After five or six years, when the nutrients had been exhausted, the old field was abandoned to weeds, shrubs, and eventually trees. In the meantime, new fields would be burned and cropped. After fifteen to twenty years, the planter returned to the origi nal old field and began the process anew. Originating in the colonial Chesapeake, shifting cultivation remained an important part of southern agriculture for generations.2

Historians and economists have frequently noted the prevalence of shifting cultivation in the South, but there is no consensus as to why it predominated. Slavery, the availability of western land, ingrained cul tural traditions, and environmental constraints have all received blame. Environmental conditions-particularly poor soils, rugged topography, and livestock diseases-strongly correlated with low levels of improved land. While slavery and the availability of cheap western land may have

523

Agricultural History Fall

reinforced shifting cultivation in some areas, the quantitative evidence indicates that such factors played a secondary role. The southern envi ronment bore primary responsibility for the practice's widespread use. Environmental conditions, however, also contributed to southern dis tinctiveness.3

Understanding the environmental origins of shifting cultivation has important consequences for studying the overall development of south ern society. The presence of large tracts of unimproved land reduced

population densities. A dispersed rural population restricted the growth of markets for manufactured goods and urban services while hindering the organization of schools, voluntary societies, and other civic organi zations. In all of these areas, the presence of slavery exacerbated the

impact of shifting cultivation. With both slavery and shifting cultivation constraining social and economic development, it is hardly surprising that the South lagged behind other regions in most measures of long

range economic success. Our strategy was to analyze census data, first collected in 1850, on

improved and unimproved acreage in farms. The census statistics on these kinds of acreage give a rough measure of where shifting cultivation

was used. Counties in which farms and plantations contained high levels

of unimproved land almost certainly practiced some form of shifting

cultivation. We used a multivariate regression analysis of nearly seven

hundred southern counties in 1860 and 1890 to determine how soil types,

topography, urbanization, and railroads influenced southern land-use patterns.

The percentage of improved land on farms serves as our statistical

proxy for measuring the presence of shifting cultivation. The 1860 cen

sus defined improved land as acreage "reclaimed from a state of nature,

and which continues to be reclaimed and used for the purposes of pro

duction," including all land cleared for "grazing, grass, or tillage." In

every census year between 1850 and 1890, the South lagged far behind

the North, with the Cotton South having the lowest percentage of im

proved land (Table 1). In 1850 and 1860 the percentage of improved

land also declined as one moved from east to west. The Midwest, for

example, possessed a lower percentage of improved land than the

Northeast in 1850 and 1860. Yet as areas of the Midwest became more

settled, and farmers had more time to convert forests into farmland, the

524


Table 1. Percentage of Improved Land in Farms, 1850-1890

States 1850 1860 1870 1880 1890 1900 1910 1920

Northeast 61.6 63.8 65.5 68.2 67.5 59.5 58.1 56.8 Maine 44.8 47.2 50.0 53.2 49.3 37.9 37.5 36.4 New Hampshire 66.4 63.2 64.7 62.0 50.0 29.8 28.6 27.0 Vermont 63.1 66.0 67.9 67.3 60.4 45.0 35.0 39.9 Massachusetts 63.6 64.6 63.6 63.4 55.3 41.1 40.5 36.4 Rhode Island 64.4 64.3 57.5 58.0 58.5 41.2 40.2 40.1 Connecticut 74.2 73.1 69.6 66.9 61.2 46.0 45.2 36.9 New York 64.9 68.5 70.4 74.5 74.6 68.9 67.4 63.8 New Jersey 64.2 65.2 66.1 71.6 75.1 69.6 70.6 68.2

Pennsylvania 57.8 61.5 64.0 67.8 71.9 68.2 68.2 67.1 Midwest 44.8 53.3 60.9 70.4 74.3 73.1 75.4 74.3

Ohio 54.7 61.7 66.6 73.7 78.5 78.5 79.8 78.9 Indiana 39.5 50.3 55.8 68.2 74.2 77.15 79.5 79.2 Illinois 41.9 62.6 74.7 82.5 84.2 84.5 86.2 85.4 Kansas 34.8 50.1 73.8 60.1 69.0 67.4 Michigan 44.0 49.4 50.9 60.1 66.7 67.2 67.7 67.9 Wisconsin 35.1 47.5 50.4 59.7 58.3 56.6 56.5 56.2 Minnesota 17.4 20.5 35.8 54.1 59.6 70.3 71.0 71.1 Iowa 60.5 80.3 83.4 86.5 86.9 85.5

Border South 38.1 39.0 45.4 56.8 61.4 65.8 68.4 68.6

Maryland 60.4 62.1 64.6 65.3 68.9 68.0 66.3 65.9 Delaware 60.8 63.4 66.3 68.5 72.2 70.2 68.7 69.1

KentuCky 35.2 39.9 43.4 49.9 55.2 62.5 64.7 64.7 Missouri 30.2 31.3 42.1 60.1 64.3 67.4 71.1 71.4

Upper South 31.74 32.76 33.5 36.3 41.7 45.6 48.3 51.2 Virginia 39.6 36.8 45.0 42.9 47.8 50.7 50.6 51.0 West Virginia 30.1 37.2 44.1 51.6 55.1 57.7 North Carolina 26.0 27.4 26.5 29.0 34.6 36.6 39.3 41.0 Tennessee 27.3 32.9 34.9 41.1 46.4 50.4 54.3 57.3 Arkansas 24.5 29.8 36.8 41.8 46.4 52.8

Cotton South 29.5 30.8 29.9 32.3 38.9 41.3 47.0 51.5 South Carolina 25.1 28.2 24.9 30.7 39.9 41.3 45.1 49.8 Georgia 28.0 30.3 28.9 31.5 38.0 40.2 45.6 51.3

Mississippi 32.8 32.0 32.1 32.9 39.0 41.6 48.5 51.3 Alabama 36.5 33.4 33.8 33.8 38.8 41.8 46.8 50.5 Louisiana 28.8 29.1 29.1 33.1 39.5 42.2 50.5 56.2

OVERALL NORTH 53.4 57.7 62.6 70.4 72.4 71.7 71.6 70.8 OVERALL SOUTH 32.1 33.3 31.1 39.9 45.6 49.1 53.0 56.0

SOURCE: Census statistics at Geostat Center, University of Virginia Library, http:// fisher.lib.virginia.edu/collections/stats/histcensus/ (hereafter Geostat Center).

525


percentage of improved land correspondingly increased. The differences between the North and the South, on the other hand, stubbornly per

sisted into the twentieth century. Southerners faired somewhat better when the measure was changed to improved land as a percentage of

total area, even so, the South still trailed far behind the North. The

North-South ratio of improved acreage to total area was 1.47 in 1850,

1.62 in 1860, and 1.72 in 1870. Like Table 1, these results exclude rela

tively unsettled southern states such as Florida, Texas, and Arkansas.

Adding these states would significantly widen the already large regional differences in land-use patterns.4

Land-use patterns might also help explain the high rates of unim

proved land; it is certainly possible that Southerners held large amounts of unimproved land for reasons entirely unrelated to shifting cultivation. Lumber production appears to be the most likely alternative. According to the 1870 census, which broke down unimproved acreage into the

categories of "woodland" and "other," some 80 percent of uncultivated

land on southern farms was forested. While southern farmers undoubt

edly utilized some of their unused woodland for firewood and fencing

northern farmers did the same-unimproved land on southern farms

generally was not used for commercial lumbering. In 1859 the South

produced only 17 percent of the nation's lumber, which suggests that

southern farmers burned, rather than cut, their forests. When the South

became a prominent producer later in the nineteenth century, large

scale lumber camps-not individual farms-became the center of pro

duction, suggesting that shifting cultivation accounted for the large num

ber of unimproved forest lands on southern farms. Planters and farmers

might also have used their unimproved woodlands to feed livestock, but

the value of southern livestock in the antebellum period remained quite low relative to the North. The fact that southern planters let their live

stock forage on low-quality forest growth also indicates shifting cultiva

tion. Cattle grazing on nutritious grasses in improved pastures produced

manure to revitalize fields, but livestock subsisting on poor-quality for

est vegetation provided relatively little in the way of usable fertilizer.5

Plantation journals and travelers' accounts suggest that shifting cul

tivation was widespread in the colonial period; many farmers in both the

North and South used shifting cultivation to rapidly clear land. Whereas most northern farmers eventually made the transition to continuous

526


cultivation, Southerners generally did not. Seventeenth-century tobacco planter Robert Cole planned his farming operations around the basic fact that after six years "the land had to lie fallow for twenty years

before its fertility returned." As a result, Cole owned a large reserve of land that would be cleared once cultivated fields were exhausted. In the antebellum decades many Chesapeake planters and farmers still utilized what historian Jack Temple Kirby calls "fire culture." Among them was Daniel Cobb, a Southampton County, Virginia, planter who recorded in January 1853 that "I fired a parsel of logs in oald land with 2 hands

where I am going to put Cotton as I wish to rais $250 extry." Shifting

cultivation was quite common in the Lower South as well. Using plan tation journals and other evidence, historian Steven Stoll has docu

mented how planters scarred much of South Carolina through its indis criminate use. Not surprisingly, South Carolina ranked near the bottom in the percentage of improved land throughout the nineteenth century.

When South Carolinians emigrated to the West, they frequently brought shifting cultivation to their new locales. In antebellum Mississippi, his torian John Hebron Moore notes, "wise cotton growers anticipated the destruction of land by acquiring tracts larger than they planned to cul

tivate at the time they were setting up plantations, so that they would

have a reserve of virgin soil to exploit in the future."6

Farmers in southern Appalachia also practiced shifting cultivation, setting fires to clear wooded hillsides and slopes. As was the case in plantation districts, the ash from burnt trees and shrubs effectively fer tilized the land. In many cases, Appalachian farmers let the tree stumps remain, planting their corn in untidy mounds scattered throughout the charred landscape. Although such practices were particularly unsightly, they made ecological sense: the root systems from the deadened trees delayed the onset of soil erosion. Once the fertility of the land was

exhausted either through cropping or erosion-the farmer abandoned the tract to long-term fallow and burned another section of his farm.

This form of shifting cultivation was still practiced in the Ozark Moun tains as late as the 1980s. In contrast to slaveholding planters, who

produced staple crops for national markets, many farmers in western Virginia and other parts of southern Appalachia used shifting cultiva tion to grow corn and wheat for household use and local trade. Higher

transportation costs, which dampened the incentive to produce sur

527


pluses for eastern markets, undoubtedly discouraged these farmers from adopting continuous cultivation even when they owned relatively more fertile soils.7

Historians have frequently argued that a combination of slavery and

cheap western lands led to shifting cultivation. According to this inter pretation, planters simply "mined the soil" before migrating with their mobile labor force to inexpensive western lands. Other historians have similarly argued that the South's "relative abundance of land and com

paratively high price of labor" created low land prices, which encour

aged shifting cultivation. Still another interpretation posits that South erners, influenced by a Celtic cultural heritage, avoided the hard work of

building barns and improving pastures, preferring instead to let their cattle and swine roam in the woods. Although intuitively appealing, these explanations do not fit all of the available evidence. If cheap land

led to shifting cultivation, why did relatively unsettled states such as Michigan and Wisconsin have a far higher percentage of improved acre

age than Mississippi, Alabama, and Georgia, which contained much of

the nation's most valuable cotton lands?8

Even more importantly, stressing slavery, western lands, or cultural traditions ignores major variations within the South. Significant areas of the South, such as the Kentucky Bluegrass region and the Nashville

Basin, had large plantations, large slave populations, and many at

tributes of the South's distinctive cultural heritage. Planters and farmers

in these regions nevertheless cultivated a higher percentage of land than

even many northern farmers. Planters and farmers in some bluegrass

counties, in fact, often improved more than 80 percent of their land.9

Nature blessed the Kentucky Bluegrass region and the Nashville Ba

sin with rich soils and a temperate climate ideally suited for continuous

cultivation. Most other regions of the South lacked such favorable en

vironmental conditions. Southern soils, as recent studies have suggested, were particularly ill-suited for continuous cultivation. Scientists have classified most of these soils as ultisols, an order characterized by leach

ing, high acidity, and poor fertility. The poor fertility of ultisols helps explain why Southerners burned forest and undergrowth to fertilize

their land. The ash quickly provided important nutrients and its calcium

content helped neutralize acidity. Agricultural reformers highly re

garded ash because its application increased the productivity of soils

528


lacking "calcareous matters." Given that ash was a cheap, effective fer tilizer-at least before it is leached out of the soil-it is quite expected

that scientists have found a strong correlation between shifting cultiva tion and ultisols in today's developing world.10

The most successful northern cultivators, on the other hand, farmed soils classified as alfisols, which contained an abundance of phosphorus, potassium, calcium, and other essential plant nutrients. Northern farm ers used a highly developed system of mixed husbandry to make the most of their excellent soils. Field rotations usually included clover and other legumes, which added nitrogen to fields depleted by wheat and corn during previous seasons. Cattle grazing in improved pastures, meanwhile, provided tons of high quality manure that recycled key nu trients to maintain the farm's ongoing productivity. Nutrients, however, cannot be recycled unless they are initially present in the soil. That is why the most productive and populous agricultural hinterlands of major northeastern cities-southeast Pennsylvania (Philadelphia), central and western Maryland (Baltimore), and upstate New York (New York City)-all contained significant stretches of alfisol land ideally suited for continuous cultivation."1

In the South, however, alfisol soils were not always associated with continuous cultivation. Substantial portions of Mississippi and Tennes see-two states with fairly low ratios of improved land-contained large

areas of alfisol land. The problem for these areas was topography. Even

the best of soils cannot support continuous cultivation if they are subject to intense erosion, located on inaccessible mountain slopes, or lack proper drainage. Many of the counties along the Mississippi River con tained fairly low percentages of improved land, despite rich alluvial soils, because frequent flooding inhibited cultivation. Erosion prevented in tensive farming practices in the fertile Black Belt of Mississippi and Alabama, where row cropping created channels that carried away valu able top soil. Erosion was particularly severe in these areas because of

the South's unique rainfall patterns. Soil erosion experts have developed a "Rainfall and Runoff Factor" (R-factor) to measure the intensity and

duration of storms. The higher the R-factor, the more likely that local

rainfall will intensify soil erosion. The Southeast has the highest R-factor of the nation, with the intense, pelting storms of the Gulf Coast states

creating a particularly high rate of soil erosion.12

529


Southern farmers suffered other disadvantages as well. Key fodder crops, such as hay and clover, that supported continuous cultivation could not thrive in the warm and humid southern climate. Southern

planters and farmers fed their cattle the best they could, most often with

cow peas and low-quality grasses, but the relatively low nutritional value

of such feed meant southern cattle produced far less dung than northern

cattle. The warm southern climate also created a hospitable environ

ment for the ticks that spread bovine babesiosis, also known as southern

cattle fever or Texas fever. Southern cattle exposed to such infections at

an early age developed immunity to the worst effects of the disease.

Southern cattle fever, however, prevented Southerners from improving the quality of their animals because high-quality breeds from Europe or

the North-without immunity-quickly succumbed to the disease. Southern cattle thus lagged behind northern livestock in terms of weight and milk production. Under these adverse conditions, most southern agriculturalists devoted little time and energy to tending to their live stock. Continuous cultivation posed far more difficulties without thriv ing livestock that could efficiently recycle key nutrients.13

To statistically test the importance of soils, climate, and disease, we

estimated a series of county-level regressions in which the dependent variable is the percentage of improved land for 1860 (Table 2). Our

independent variables (on the left side of the table) classify every county in eight southern states according to soil type, topography, and level of

economic development. For soil type, we used data from the State Soil

Geographic Database-STATSGO-to determine if alfisols were the

primary soils of a county or at least important secondary soils. Out of

698 southern counties, 110 counties contained mostly alfisol soils, while another 232 contained smaller stretches of alfisols. To capture the im

pact of topography, we grouped southern counties into seven major geographical areas (please see the appendix). We also included a vari

able to account for counties that most likely suffered from Texas fever.14

Most of our independent variables are dummy variables that are

basically "yes" or "no" responses to a particular question. For the rail

.,road variable, for example, a county with a railroad receives a value of

1 (a statistical "yes"), while a county without a railroad receives a value

of 0 (a statistical "no"). The coefficient is the impact of a "yes" response.

To use the railroad variable as an example, the coefficient reported in

530


Table 2. Determining the Percentage of Improved Land: County-Level Regression Results for Nine Southern States, 1860

Variable or Geography Addition of RRs Addition of Diagnostic Statistic and Climate and Cities Crop Choice

Intercept (Piedmont Counties) 39.7** (31.57) 37.7** (29.2) 34.02** (23.48)

Coastal Flatwoods -19.3** (-8.96) -18.7** (-8.78) -15.0** (-7.16) Coastal Plain -6.9** (-4.76) -6.39** (-4.48) -6.1** (-4.42) Mountains and Ridges -16.4** (-11.51) -15.3** (-10.70) -13.5** (-9.41) Hills and Valleys -11.8** (-6.50) -10.6** (-5.91) -7-53** (-4.21) Miss. Alluvial and

Black Prairies -9.0* (-4.23) -8.71** (-4.20) -9.27** (-4.49) Limestone Basins 15.8** (7.49) 16.3** (7.84) 18.23** (9.01)

Primary Alfisol 12.3** (7.72) 10.9** (6.84) 10.02** (6.55)

Secondary Alfisol 4.2** (2.75) 3.74** (3.49) 3.2** (3.11) Southern Cattle Fever -4.2** (-3.88) -4.0** (-3.08) -5.4** (-4.21) Railroad County 5.31** (5.05) 3.8** (3.80) Urban County - 0.01 (0.30) 1.8 (0.73) Wheat - 8.10** (6.74) Cotton 8.99** (6.30) Tobacco -1.3 (-0.07) Adjusted R-Squared 0.476 0.500 0.556 F-Statistic 71.29 63.13 61.19 Degrees of Freedom 688 686 683

*Significant at 5 percent level **Significant at 1 percent level

Dependent Variable: Percentage of Improved Land in Farms at the County Level, 1860

(t-statistics in parentheses). SOURCE: Census statistics at the Geostat Center. Soil orders classified using Soil Survey

Staff, Natural Resources Conservation Service, USDA, "US General Soil Map (STATSGO)," http://soildatamart.nrcs.usda.gov (hereafter STATSGO). Please see the appendix for topographical classifications. We classified most counties in Mississippi, Ala bama, Georgia, and South Carolina as southern cattle fever counties.

Table 2 is 5.3, which means that a railroad added 5.3 points to a county's

percentage of improved land. A county without a railroad that culti

vated 40 percent of its improved land, in other words, would have cul

tivated 45.3 percent of its land if a railroad had been built. The result

suggests that railroads made commercial agricultural more feasible, thus encouraging farmers and planters to cultivate more land. Regression analysis is so valuable because it holds the impact of all the specified

531


variables constant. In this case, the coefficient of the railroad variable

measures the impact of railroad access after taking into account soil

type, topography, urbanization, and crop choice. The regression results highlight the powerful impact of environmen

tal factors on levels of improved land. Counties primarily composed of alfisols, in fact, added 10 to 12 percentage points to the percentage of

improved land. The effect for counties that contained only partial strips of alfisols, not surprisingly, was somewhat less than for counties com

posed primarily of alfisols, but both variables were statistically signifi cant. As for topography, the limestone basins, which include the Ken

tucky Bluegrass region and the Nashville Basin, contained a higher per

centage of improved land than the other areas. A limestone county, on

average, had 15 to 16 more percentage points of improved land than the

typical Piedmont county. The large negative coefficient for the Coastal Flatwoods-an area well known for its pine barrens and poorly drained

marshlands-is particularly striking because many of the South's largest port cities were located in that region. The negative coefficient for the

Texas fever variable is also statistically significant, suggesting that live stock disease made continuous cultivation far more difficult in the Cot

ton South. Since the ticks that carried babesiosis thrived in heat and

humidity, this variable could also be picking up the negative impact of

the Deep South's climate on fodder crops and improved pastures. The

ticks within one hundred miles of the Gulf Coast-known as Rhipi

cephalus microplus (previously known as Boophilus microplus)-were more pernicious than Rhipicephalus annulatus (previously known as Boophilus annulatus) ticks that infected the rest of the South, which may

have worsened the large, negative coefficients for the coastal flatwoods

and coastal plain variables. Taken all together, the environmental vari

ables seriously suggest that the southern climate and geography directly

resulted in shifting cultivation.15 In contrast, two indicators of general economic development-the

presence of railroads and cities over five thousand residents-had a

relatively small impact on levels of improved land. The railroad variable

is positive and statistically significant, suggesting that better transporta tion opened markets, which encouraged farmers to improve more land.

Yet the small size of the coefficient, 3.8-5.3 percentage points, depend

ing on the specification, indicates that railroads had far less impact than

532


environmental variables. For cities, there was no statistically significant relationship between urbanization and the percentage of improved land after controlling for other factors. This comprised a major difference between the North and the South. The North's vibrant urban and manu facturing sector, as well as its rapidly improving transportation network, created larger markets for farmers, which gave them a greater incentive to improve more land. A self-reinforcing dynamic emerged in which intensive cultivation encouraged urban growth and transportation im provements, thus allowing farmers to improve an even higher percent age of their land. Some antebellum observers argued that more manu facturing and urban growth in the South would eventually foster more intensive cultivation practices, but such prescriptions failed to recognize the severity of geographic and climatic restraints in the South.16

The low level of improved land within urban counties demonstrates the tenacity of these environmental constraints. In 1860 twenty-seven southern counties had cities with a population greater than five thou sand people. The percentage of improved land within these urban coun ties barely budged in the 1850s-moving from 34.5 percent to 37.3 per cent-and remained far below national averages. In the nine counties containing the South's major port cities, including Norfolk, Charleston, and New Orleans, planters and farmers cultivated less than 20 percent of their land. These statistics confirm the regression results showing that urbanization made no impact in determining levels of improved land. Historian Ulrich B. Phillips made a similar point in regards to Charles

ton more than a century ago: "To maintain Charleston's commercial

eminence, nothing would suffice, in fact, but to make the wilderness behind her blossom; but railways and all things else have thus far failed

to convert the pine-barrens into any semblance of a garden."17

Some historians have maintained that southern staples, such as cot ton, tobacco, and corn, depleted the land more rapidly than wheat and

other crops that were more likely to be grown in the North. To test this

proposition, we included dummy variables representing counties in the top quartile of wheat, cotton, and tobacco production. The wheat vari

able has a statistically significant and relatively large coefficient, which suggests that grain farmers in Virginia and other areas of the Upper

South were more likely to integrate livestock and crops into a mixed

farming regime. The argument that the special characteristics of cotton

533


and tobacco accounted for shifting cultivation, however, receives little

support. Cotton, in fact, was associated with more, not less, improved land. The large coefficient for the cotton variable indicates that planters

brought more marginal land into cultivation in response to the high

prices of the 1850s. Notice that even with the inclusion of the crop choice variables that almost all of the coefficients of the geographic and

climatic variables remain statistically significant.18

However, the analysis thus far has ignored a vitally important ques tion: what about the relationship between slavery and shifting cultiva

tion? As a first step to understanding the impact of slavery, Figure 1

1.2 . .......... ~

0.8 ..*... ...,... .

4 ~~~~~~~~~~~~~~4 ,

4 .,s~~~~~. 4 ~

WO:W 0.4~~~~~~~~"4

L. 4~~~~~~~~~~~.... .. 0 44~~~~~~~ 024 '?~~*~* 4'vc.'.. W44'', ~~~~~~~ -~ ~~~~~~~~~.... .

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Percent Enslaved

Figure 1. The Relationship Between Slavery and the Percentage of Improved Land in

Farms, 1860. Source: Calculated from census statistics at the Geostat Center.

charts the percentage of improved land and the percentage of the en

slaved population for each county. On the bottom left-hand corner there

is a large clustering of counties with both a low percentage of improved land and a low percentage of slaves. Almost all of these counties, not

surprisingly, were located in the mountainous upcountry. In the middle

of the graph are counties with intermediate to high percentages of slaves

and intermediate to high levels of improved land, including counties in

the Alabama and Mississippi Black Belt, the southern Piedmont, and the

limestone areas. Levels of "intermediate" and "high," of course, are

based only in relation to the South; planters in many of these counties

534


improved less land than most northern farmers. Located on the bottom right-hand corner is a smaller group of counties with high concentrations of slaves but low levels of improved land. Most of these counties were located along the coast of the Carolinas and Georgia, where rice plant ers intensively cultivated a small number of acres and left most of their poorly drained "provisioning" lands unimproved. Overall, Figure 1 in dicates that antebellum slavery was a flexible institution suitable for both shifting cultivation and continuous cultivation.

Ideally, we would econometrically test the relationship between sla very and levels of improved land in a regression that holds soil quality, topography, and climate constant. When slavery is inserted into the equations in Table 2, the coefficient shows a strongly positive-and statistically significant-relationship with higher levels of improved land. Simply put, counties with more slaves contained more improved land when all other environmental and economic variables are taken into account. Adding slavery to the regressions, though, is bedeviled with what econometricians call an endogeniety bias. To put it simply, it is impossible to tell with county-level data whether slavery caused more land to be improved or whether slaveholders simply preferred to locate in areas with the best soils and best access to transportation. We have

run statistical tests to correct the endogeniety bias, but they do not work well given the limitations of census data.19

The non-linear relationship between slavery and levels of improved land increases our skepticism about the econometric evidence. It makes little sense to suppose that the lack of slave labor was a major reason for

low levels of improved land in the southern uplands; likewise, it seems unlikely that Carolina rice planters improved a small percentage of their land because they somehow possessed too many slaves. Our interpre tation of the evidence, then, is that the slavery coefficient reflected underlying environmental conditions, rather than acting as an important variable in its own right.20

Although little direct evidence exists on this point, it is certainly possible that slavery and cheap western lands discouraged Southerners from finding sustainable alternatives to shifting cultivation. With a bound labor force available to clear land, slaveholders had less incentive

to develop fertilizers, find suitable fodder crops, or combat southern cattle fever. The same logic holds true of inexpensive western land; it did

535


not cause shifting cultivation per se, but it certainly might have discour aged efforts to find alternatives. Although agricultural reformers in older, eastern seaboard states worried about the developmental conse quences of shifting cultivation, planters only intermittently supported long-term agricultural research. Southern state legislatures hesitated to spend even small sums on basic geological surveys, and they often re

fused to subsidize experimental farms, agricultural professorships, and research stations. Many planters belonged to state and local agricultural societies, which made headway in developing marl (a mix of clay and

fossilized seashells) and other fertilizers that provided the soil with cal cium, thereby reducing its acidity. Marl and associated fertilizers, though, were risky and expensive. As frustrated agricultural reformers like Edmund Ruffin noted, planters making substantial profits using shifting cultivation had little economic incentive to use marl or other

means of reducing soil acidity. Why adopt continuous cultivation when highly mobile planters and their slaves could use shifting cultivation to quickly convert new land into workable plantations?21

In the postbellum era, the continued development of the railroad

network-combined with greater investment in agricultural research gradually raised the percentage of improved land on southern farms.

The region's railroad network expanded dramatically after the Civil

War; 90 percent of the southern population lived in a county that had a

least one railroad station by 1890. The quality of rail service also im

proved as large trunk lines integrated individual lines into coherent

networks with standardized gauges. The expansion of the region's rail

road network contributed to a dramatic increase in urbanization, par

ticularly in upland areas that had been relatively isolated before the

Civil War. Within agriculture itself, phosphate emerged as a relatively

inexpensive fertilizer for southern planters and farmers. Federal and state experiment stations provided advice and technical support about fertilizers and other cultivation practices for those wishing to adopt

more intensive agriculture. In light of these changes, shifting cultivation

gradually declined and the percentage of improved land within southern farms gradually increased between 1880 and 1920 (Table 1). By the turn

of the century, Southerners improved a higher percentage of their land

than farmers in New England, where a sharp drop in wool prices led

farmers to abandon many of their improved pastures.22

536


These important technological developments and institutional changes, however, did not suddenly sweep aside the South's severe en vironmental constraints. The regional differences in land-use patterns remained, as farmers in the cotton states still lagged far behind their

mid-Atlantic and midwestern counterparts. In 1920 farmers in North Carolina, South Carolina, Georgia, Alabama, Mississippi, and Louisi ana-despite the spread of railroads, the growth of cities, and the in

troduction of new fertilizers-still cultivated a lower percentage of their land than Ohio farmers had in 1850. Advocates of continuous cultivation within the South, according to historian Gilbert Fite, often expressed keen disappointment at the slow rate of progress. Fite attributes the southern failure to fully adopt continuous cultivation to ingrained tra ditionalism and the South's reliance on cotton, but he also notes that

environmental factors continued to hinder progress:

The natural grasses and forage in much of the South were less nutri

tious than those found in the Midwest. To get satisfactory pastures and

hay in the Deep South the soil had to be plowed, planted, and fertil

ized at considerable expense.... In addition, southern livestock was

subject to diseases that caused heavy losses. Swine and fowl cholera

were common, and Texas fever became so widespread that in the early

1880s it was being referred to by a veterinarian in the USDA as

"Southern Cattle fever."23

Regression analysis for 1890, using many of the same variables as the

1860 specifications, confirmed Fite's observations (Table 3). Initial soil quality, despite the introduction of cheaper fertilizers, still had a signifi

cant impact. Counties composed mostly of alfisols continued to improve

10-12 percent more land than counties possessing ultisols or other soils.

The coefficients for the 1890 topographical variables generally remain statistically significant and have the same signs as those in the 1860

regressions. Farmers on limestone plains continued to cultivate a far

higher percentage of their land than farmers living in other topographi

cal regions, while the marshy Coastal Flatwoods lagged especially far

behind. Texas fever, which had spread to several areas in the Upper

South during the Civil War, grew even worse in 1890. Farmers in coun

ties with babesiosis cultivated 8 to 12 percent less land after taking into

537


Table 3. The Continuing Importance of Environmental Factors in the Postbellum Era: County-Level Regression Results for Nine

Southern States, 1890

Variable or Geography Addition of RRs Addition of Diagnostic Statistic and Climate and Cities Crop Choice

Intercept (Piedmont Counties) 50.6** (32.0) 45.0** (25.2) 43.2** (22.8)

Coastal Flatwoods -19.6** (-9.4) -19.5** (-9.7) -15.2** (-7.3) Coastal Plain -7.4** (-5.1) -6.6** (-4.6) -4.7** (-3.2) Mountains and Ridges -12.5** (-7.9) -11.5** (-7.44) -9.5** (-6.0) Hills and Valleys -5.5** (-2.8) -4.6** (-2.6) -3.9* (-2.2) Miss. Alluvial and

Black Prairies -2.0 (-0.9) -1.9 (-0.9) -4.2* (-2.15) Limestone Plains 16.6** (7.9) 15.5** (6.8) 16.5** (7.5)

Primary Alfisol 13.3** (8.4) 11.6** (7.4) 11.4** (7.5)

Secondary Alfisol 3.2** (3.0) 2.6** (2.5) 2.5* (2.4) Southern Cattle Fever -7.9** (-6.5) -9.2** (-7.6) -12.0** (-9.5) Railroad County 6.9** (5.6) 6.68** (5.6) Urban County 6.6** (4.1) 7-7** (4.8) Wheat 1.4 (1.2) Cotton 9.8** (7.5) Tobacco 0.8 (0.71) Adjusted R-Squared 0.48 0.52 0.55 F-Statistic 81.3 75.8 67.9 Degrees of Freedom 766 764 761

*Significant at 5 percent level

**Significant at 1 percent level

Dependent Variable: Percentage of Improved Land in Farms at the County Level, 1890

(t-statistics in parentheses). SOURCE: Census statistics at the Geostat Center. Soil orders classified using STATSGO. Please see the appendix for topographical classifications. Counties with southern cattle

fever classified according to Bureau of Animal Industry, Map Showing Boundary Line of

District Infected with Splenetic or Southern Cattle Fever, as Defined in Order of Hon. J. M.

Rusk, Secretary of Agriculture (Washington, DC: GPO, 1891).

account other variables. In Virginia, for example, the percentage of

improved land in twenty-three newly infected counties declined from 44 percent to less than 40 percent between 1860 and 1890, whereas the

percentage of improved land within the state as a whole increased from

37 percent to 48 percent.24

In contrast to the 1860 results, the regressions show that cities and

538


railroads in the 1890s significantly influenced land-use patterns. In 1860 railroads and cities had only a small impact on the percentage of im

proved land, but by 1890 farmers in a county with a railroad connection and a city improved nearly 14 percentage points more land than their non-railroad, non-urban counterparts. What accounts for this change? Southern railroads and urbanization before the Civil War had focused on linking plantation districts to port cities. The postbellum extension of the rail network to the southern uplands, working in conjunction with increased urbanization, seems to have "unlocked" high quality alfisol soils in Appalachia. Indeed, the percentage of improved land between 1860 and 1890 increased faster in the southern uplands than any other region. Notice, too, that the dummy variable representing a county in the top quartile in cotton production is statistically significant. The con tinued expansion of the South's cotton culture might well have incor porated previously marginal acreage that had gone unimproved. The completion of levees on the Mississippi, in particular, created rich new land for cotton in the Black Belt. Within the state of Mississippi, for example, farmers and planters in the eleven counties bordering the river increased the percentage of improved land from 36 percent in 1860 to 52.5 percent in 1890, a rate of increase far greater than the South as a

whole.25 The use of cheap phosphate fertilizers represented another important

change in the postbellum period. To what extent did these fertilizers increase the percentage of improved land? Data from the 1890 census shows that farmers in the Coastal Flatwoods, Piedmont, and coastal plains spent significantly more on fertilizer than the South's other to pographical regions: eighty-nine cents, forty-nine cents, and thirty cents per improved acre, respectively, compared to a combined average of eighteen cents. All of the areas that relied on phosphate fertilizers also cultivated relatively low levels of improved land. The correlation be tween low levels of improved land and fertilizer usage is hardly surpris ing. Farmers with the worst soils, and hence the lowest levels of im proved land, might have been more likely to use fertilizers than farmers working the best lands. Phosphates, in fact, may well have prevented complete disaster in the piedmont and coastal flatwood regions. Because of increased erosion and competition from more fertile regions, farmers in these areas might have cultivated less land, in absolute terms, in 1890

539


than in 1860 without the use of artificial fertilizers. As it was, the struggle to achieve even modest increases in improved acreage with the most advanced fertilizers shows how the struggle to conquer nature was still incomplete in 1890.26

Our evidence suggests that southern farmers and planters utilized shifting cultivation because of difficult environmental conditions. From the standpoint of an individual planter or farmer, shifting cultivation

worked reasonably well as long as there was enough reserve land. Hav

ing large tracts of unimproved land in reserve, though, posed a major impediment to southern development. Shifting cultivation, in essence, created huge swaths of economic "dead space" that generated little or

no economic activity. Adding unimproved acres meant larger farms and larger plantations without increasing the size of the workforce, thus reducing population densities. One of the overlooked facts of the ante bellum period is that the South trailed far behind the North not only in

overall population density, but rural population density as well. In 1860,

for example, the South Atlantic region-Virginia, North Carolina,

South Carolina, and Georgia-contained 20.3 total rural residents per

square mile and 11.7 free rural residents per square mile, far behind New

England (33 rural residents per square mile), the Middle Atlantic states

(49 rural residents per square mile), and the Old Northwest (35 rural

resident per square mile). The impact of shifting cultivation in dispersing rural populations was well understood, even for the colonial period. "By

the middle of the eighteenth century," writes historian T. H. Breen,

"most [Virginia] planters accepted that dispersed settlement was an

inevitable product of a particular type of agriculture." In the North, low

population density in rural areas was a temporary phenomenon associ

ated with recent settlement. In the South, shifting cultivation created the

demographic equivalent of a permanent frontier in which vast amounts

of land remained uncultivated every generation.27 A number of scholars have noted how the southern economy gener

ated far smaller markets for manufactured goods and urban services,

thus hindering the growth of industry and cities. In the North, continu

ous cultivation created a prosperous and deeply rooted rural population

that helped stimulate early industrialization. As population increased

and transportation improved, manufacturing firms competing in larger

540


markets had greater incentives to expand output and improve produc tivity. Dense networks of farms in southern New England, upstate New York, and southeastern Pennsylvania provided particularly rich markets for manufactured goods. These local markets spurred manufacturing in cities such as New York and Philadelphia. The chapter titles of David R. Meyer's recent synthesis of early northern industrialization-"Prosper ous Farmers Energize the Economy" and "Agriculture Augments Re gional Industrial Systems"-testify to the importance of deep and rich rural markets to northern manufacturers.28

In sharp contrast, scholars have used the term "disarticulation" to describe a southern economy that lacked market towns and local manu facturing. It seems reasonable to suppose that shifting cultivation, along with slavery, contributed to the small size of southern markets. Recall that farmers and planters even near relatively large cities such as Nor folk and Charleston cultivated less than 20 percent of their land. Lacking nearby markets, southern ports specialized in the collection and expor tation of staple crops and supported a modest manufacturing base rela tive to northern cities.29

Low population densities also made it more difficult for Southerners to create institutions to cultivate and disseminate productive knowledge.

A region composed of isolated farms and plantations generated fewer subscribers for periodicals and newspapers, possessed fewer potential members for mechanics institutes or literary associations, and provided fewer students for schools and colleges. Economic historians have found a strong correlation between low population density and illiteracy in the antebellum period, whether one considers the North or the South. In the North, low population density in rural areas was temporary, whereas in

the South it became the norm.30 The South, as many historians have noted, remained behind the

North in literacy rates, schools, libraries, and other indicators of educa tional achievement. Southerners themselves often felt overly dependent on northern newspapers and literature; calls for an authentically "south ern" literature became commonplace in the antebellum period. Cultural and political factors-including racism and anti-democratic attitudes undoubtedly contributed to southern backwardness in education and the

541


arts, but the combination of slavery and shifting cultivation decreased the number of possible readers for southern newspapers, periodicals, and books, thus making the publication of such works even more diffi

cult.31

Northerners noticed the impact of shifting cultivation on southern society. To Northerners, southern agriculture lacked the order and re finement that characterized their neat and carefully maintained farms. The unsightly nature of southern farms and plantations-the recently burnt fields, the seemingly endless forests of pine, and the shockingly neglected livestock-all accentuated the region's relative lack of devel opment. The South's uncultivated landscape and dispersed population created a ramshackle air about the region; its public buildings seemed less impressive, its farms and plantations less permanent, and its com

munity institutions (including libraries, schools, and churches) less de veloped. For Northerners who believed in the economic and moral su periority of free labor-what historians have called free-labor ideol

ogy-the desultory state of southern agriculture and the region's general

underdevelopment became a powerful indictment of slavery. The lazi

ness that seemed to pervade southern agriculture provided compelling evidence that slavery devalued hard work and suppressed the spirit of enterprise. During the Missouri Crisis of 1819-1820, New York Con gressman John Taylor, for example, emphasized the differences across

the border between Maryland and Pennsylvania:

the dividing line between farms highly cultivated and plantations lay

ing open to the common and overrun with weeds; between stone barns

and stone bridges on one side, and stalk cribs and no bridges on the

other; between a neat, blooming, animated, rosy-cheeked peasantry on the one side, and a squalid, slow-motioned, black population on the

other.32

No writer solidified the northern view of southern agriculture and its impact on slavery more than Frederick Law Olmsted. As a landscape

architect-he would go on to design New York City's Central Park

Olmsted had an eye for telling detail as he traveled through the South

in the 1850s. When in Tidewater Virginia, he famously noted that "'old

542


fields'-a coarse, yellow, sandy soil, bearing scarce anything but pine trees and broom-sledge" dominated the countryside. Olmsted became lost along the poorly marked roads and isolated homesteads. Getting directions was almost impossible, and Olmsted reported that one was more likely to see gangs of wild hogs, "long, lank, bony, snake-headed, hairy, wild beasts," than people when traveling through the pine forests of the state. The only dwelling that he passed was "a house, across a

large, new old-field . . . there was no distinct path leading towards it out

of the wagon-track we were following." When he finally reached "the Court House," the local designation for the county seat, he noted that it consisted of thirty or so buildings, including several stores, a law office, a saddler's shop, and two public houses. For the hub of local commerce, "the Court House" was slim pickings indeed.33

In line with current historical theory, Olmsted blamed slavery for shifting cultivation. Like other antebellum writers, he conflated the abil ity to grow staple crops such as cotton, tobacco, rice, and sugar with fertile soils and a hospitable climate. The reality was far different. In the

long run, the South's soils, climate, and topography prevented intensive agriculture and slowed long-term development. Shifting cultivation was not the sole cause of southern underdevelopment, though. While slavery did not create shifting cultivation, it may well have accentuated the developmental consequences of the land-hungry agricultural regime. A number of scholars have argued that slaves consumed little in the way of

manufactured goods or urban services, thus contributing to the limited

nature of southern markets. Slavery, in essence, accentuated the "lack of demand" that hurt southern industry and commerce. Our brief and suggestive analysis of shifting cultivation's economic impact indicates that historians and other scholars might well see slavery and shifting

cultivation working together to create an economy with low population densities, little manufacturing, and low levels of urbanization. Slavery, of course, also made it even more difficult to support educational institu

tions. Slaves could not subscribe to newspapers, join voluntary organi

zations, or attend schools. Rather than approach the question of south

ern underdevelopment in dichotomous terms-either the "environ ment" or "institutions" (slavery)-future research should fruitfully focus on the interaction of the two.34

543


Appendix: Explanation of Topographical Variables

Southern Piedmont: Acting as the intercept, this region supported a mix of soils and landscapes that was generally favorable to agriculture but nevertheless susceptible to erosion.

Coastal Flatwoods: Generally infertile region dotted with pine barrens and marshes.

Coastal Plain: Highly variable region between the coastal flatwoods and piedmont regions.

Mountains, Valleys, and Ridges: Includes counties within the Cumber land Plateau, Appalachian Plateau, Appalachia Hill and Ridge re gion, and Blue Ridge Mountains. Soil quality is variable, but steep slopes and narrow valleys posed severe erosion problems.

Hills and Valleys: Includes the Pennyroyal region, Sandstone and Shale Hills, Sand Hills, and Fall Line Hills. Variable in terms of soil drain age and erosion.

Limestone Plains and Valleys: Includes the Bluegrass region of Ken tucky, the Nashville Basin of Tennessee, and the northern Shenan doah Valley of Virginia. All are well-drained landscapes ideal for continuous cultivation.

Mississippi Alluvial and Black Prairies: Includes Mississippi Alluvial

Valley, Mississippi Valley Silty Uplands, Mississippi Black Belt, and Alabama Black Belt. These were the premier cotton-growing regions of the Lower South, but often poorly drained and subject to severe erosion.

NOTES

1. We would like to thank Jarad Beckman for excellent research assistance. For com

ments on earlier drafts, we thank Jay Carlander, Dennis Halcoussis, Jeff Hummel, Lisa

Jacobson, Naomi Lamoreaux, Jean-Laurent Rosthenthal, Roger Ransom, Ken Sokoloff, Gavin Wright, and the participants of the Von Gremp Workshop in Economic History at

UCLA and the All-UC Economic History Group Conference on agricultural history at

Davis, Calif. We also benefited from the comments of three referees.

Farm Journal of Julian Ruffin, Ruf fin Papers, Mssl:R838a 818, Virginia Historical

Society, Richmond, Va. On Ruffin's continued use of shifting cultivation, see, Jack Temple

Kirby, Poquosin: A Study in Rural Landscape and Society (Chapel Hill: University of

North Carolina Press, 1995), 111.

2. Shifting cultivation in the South often had a pronounced commercial orientation in

which planters produced cotton and tobacco for international markets. The southern

544


variant is thus somewhat different than the shifting cultivation regimes used in the tropics

today, where it is used "for producing basic foodstuffs and meeting subsistence and local

market needs." National Research Council, Sustainable Agriculture and the Environment

in the Humid Tropics (Washington, DC: National Academy Press, 1993), 78.

3. For discussions of shifting cultivation, see, Ulrich B. Phillips, Life and Labor in the

Old South (Boston: Little, Brown, and Company, 1929), 9,136; Lewis Cecil Gray, History

of Agriculture in the Southern United States to 1860, Vol. 1 (Gloucester, Mass.: Peter Smith,

1958), 448; Paul W. Gates, The Farmer's Age: Agriculture, 1815-1860 (New York: Harper & Row, 1960), 112-13; Julius Rubin, "The Limits of Agricultural Progress in the Nine

teenth-Century South," Agricultural History 49 (Apr. 1975): 362-73; Douglas Helms, "Soil

and Southern History," Agricultural History 74 (Fall 2000): 723-58; Steven Stoll, Larding the Lean Earth: Soil and Society in Nineteenth-Century America (New York: Hill and

Wang, 2002); and Kirby, Poquosin, 95-125. Phillips, Rubin, Helms, and Kirby generally

provide geographic and climatic explanations for shifting cultivation, while Gates, Gray, and Stoll generally focus on slavery, western lands, and southern traditionalism.

4. Improved and unimproved acreage taken from census data available at Geostat

Center, University of Virginia Library, http://fisher.lib.virginia.edu/collections/stats/ histcensus/ (hereafter Geostat Center). According to special instructions to the 1860 cen

sus, marshals were to exclude "irreclaimable marshes" and large bodies of water from total

acreage. Jeremy Atack and Fred Bateman, To Their Own Soil: Agriculture in the Ante

bellum North (Ames: Iowa State University Press, 1987), 118-19.

5. The South's share of lumber production was calculated from census statistics found

in J. M. Edmunds, Manufactures of the United States in 1860: Compiled from the Original Returns of the Eighth Census, Under the Direction of the Secretary of the Interior (Wash

ington, DC: GPO, 1865), 682, 695, 708, 716. On lumber production in the postbellum

period, see, Edward L. Ayers, The Promise of the New South: Life After Reconstruction

(New York: Oxford University Press, 1992), 124-25. On the relatively poor quality of

southern forage and livestock, see, John Hebron Moore, The Emergence of the Cotton

Kingdom in the Old Southwest: Mississippi, 1770-1860 (Baton Rouge: Louisiana State

University Press, 1988), 26-27; Stoll, Larding the Lean Earth, 82-83.

6. Lois Green Carr, Russell R. Menard, and Lorena S. Walsh, Robert Cole's World:

Agriculture and Society in Early Maryland (Chapel Hill: University of North Carolina

Press, 1991), 39; Allan Kulikoff, Tobacco and Slaves: The Development of Southern Cul

tures in the Chesapeake, 1680-1800 (Chapel Hill: University of North Carolina Press,

1986), 48; Kirby, Poquosin, 95-125; Daniel W. Crofts, ed., Cobb's Ordeal: The Diaries of a Virginia Farmer, 1842-1872 (Athens: University of Georgia Press, 1997), 71. According to Crofts, "Cobb replicated patterns of slash-and-burn agriculture that had been practiced in Virginia long before the arrival of Europeans" (64). Stoll, Larding the Lean Earth,

120-66; Moore, The Emergence of the Cotton Kingdom in the Old Southwest, 86.

7. J. S. Otto and N. E. Anderson, "Slash-and-Burn Cultivation in the Highlands South:

A Problem in Comparative Agricultural History," Comparative Studies in Society and

History: An International Quarterly 24 (Jan. 1982): 141-42. On the farmers of the southern

upcountry, see, Harry L. Watson, "Slavery and Development in a Dual Economy: The

South and the Market Revolution," in The Market Revolution in America: Social, Political,

and Religious Expressions, 1800-1880, ed. Melvyn Stokes and Stephen Conway (Char lottesville: University Press of Virginia, 1996), 43-73.

545


8. For a recent statement of this view, see, Roger G. Kennedy, Mr. Jefferson's Lost

Cause: Land, Farmers, Slavery, and the Louisiana Purchase (New York: Oxford University

Press, 2003), 11-25. See, also, Gray, History of Agriculture in the Southern United States,

448; John Solomon Otto, Southern Agriculture During the Civil War Era, 1860-1880

(Westport, Conn.: Greenwood Press, 1994), 4-6; Grady McWhiney, Cracker Culture:

Celtic Ways in the Old South (Tuscaloosa: University of Alabama Press, 1988), 51-79.

9. Census marshals, in fact, reported no unimproved land in Bourbon and Clark coun

ties, part of the heart of the Kentucky Bluegrass. Census data is available at the Geostat

Center. For more on the Kentucky Bluegrass and Nashville regions, see, Stephen Aron,

How the West was Lost: The Transformation of Kentucky from Daniel Boone to Henry

Clay (Baltimore: Johns Hopkins University Press, 1996), 124-49; Helms, "Soil and South

ern History," 733-34; Phillips, Life and Labor, 80-83.

10. For excellent explications of the many problems of ultisols, see, S. W. Buol et al.,

Soil Genesis and Classification (Ames: Iowa State University Press, 1997); Helms, "Soil

and Southern History"; National Research Council, Sustainable Agriculture, 53-57. On the

fertilizing properties of ash, see, Kirby, Poquosin, 111-14; "Leached Ashes as a Manure,"

Southern Agriculturalist (Sept. 1841): 479-80; Henry D. Foth and John W. Schaffer, Soil

Geography and Land Use (New York: John Wiley & Sons, 1980), 179.

11. On the importance of livestock to intensive land use among northern farmers, see,

Donald H. Parkerson, The Agricultural Transition in New York State: Markets and Mi

gration in Mid-Nineteenth-Century America (Ames: Iowa State University Press, 1995),

94-98; Stoll, Larding the Lean Earth, 49-54.

12. Moore, Emergence of the Cotton Kingdom, 31-34; Phillips, Life and Labor, IS, 11,

102-103; Walter H. Wischmeier and Dwight D. Smith, Predicting Rainfall Erosion Losses:

A Guide to Conservation Planning, USDA Handbook No. 537 (Washington, DC: USDA,

1978), 5-6; Stanley Wayne Trimble, Man-Induced Soil Erosion on the Southern Piedmont, 1700-1970 (Ankeny, Iowa: Soil Conservation Society of America, 1974), 12-13; Stoll,

Larding the Lean Earth, 134-43.

13. Tamara Miner Haygood, "Cows, Ticks, and Disease: A Medical Interpretation of

the Southern Cattle Industry," Journal of Southern History 52 (Nov. 1986): 551-64; Claire

Strom, from a forthcoming book manuscript tentatively entitled "Making Cat Fish Bait out

of Government Boys: Politics, Class, and Environment in the New South," 1-15, in pos session of author.

14. Our eight southern states include Virginia, Kentucky, Tennessee, North Carolina,

South Carolina, Georgia, Alabama, and Mississippi. We excluded all states west of the

Mississippi, as well as Florida. The 1890 table includes West Virginia as well. Soil Survey

Staff, Natural Resources Conservation Service, USDA, "US General Soil Map

(STATSGO)," http://soildatamart.nrcs.usda.gov. We classified all counties in Alabama,

Mississippi, South Carolina, and Georgia as likely to have been infected. Portions of North

Carolina and Virginia were infected in the late nineteenth century, but that was probably the result of a large influx of cattle from the Deep South during the Civil War. Before the

Civil War, North Carolina banned cattle from the Lower South from entering the state

between April 1 and November 1, which "largely kept cattle fever out of Virginia and

states to the north before the Civil War." G. Terry Sharrer, A Kind of Fate: Agricultural

Change in Virginia, 1-861-1920 (Ames: Iowa State University Press, 2000), 18.

15. Strom, "Making Cat Fish Bait," 1-15.

546


16. For a good summary of this view, see, Sarah T. Phillips, "Antebellum Agricultural

Reform, Republican Ideology, and Sectional Tension," Agricultural History 74 (Fall 2000): 799-822.

17. Percentage of improved land in urban counties calculated from census data avail

able at the Geostat Center. Ulrich B. Phillips, History of Transportation in the Eastern

Cotton Belt to 1860 (1908; repr., New York: Octagon Books, 1968), 355.

18. Carr, Menard, and Walsh, Robert Cole's World, 39.

19. We tried to run various Hausman tests to correct for endogeniety, but all conceiv

able instrumental variables?farm size, per capita personal wealth, per capita total wealth, or some measure of per capita agricultural output?are themselves associated, perhaps

endogenously so, with high levels of improved land. This makes any result from such a test

highly suspect.

20. A related problem is a possible missing variable bias, as any positive association

between slavery and levels of improved land might be the result of variations in soil quality and commercial development not captured in our original specifications. Given the varia

tions of soils and topography within our broad topographical categories, the possibility of

a missing variable bias is quite strong.

21. In South Carolina, for example, the state government spent $40,367 on geological

surveys and agricultural research in the 1840s and 1850s, which amounted to a paltry 0.4

percent of total state expenditures in the period. Even Edmund Ruffin's well-known

geological survey of South Carolina received only modest state support. Ruffin, for ex

ample, had no staff to assist his efforts to survey the entire state. See, William M. Mathew,

Edmund Ruffin and the Crisis of Slavery in the Old South: The Failure of Agricultural

Reform (Athens: University of Georgia Press, 1988), 37-40. Expenditures on geological

surveys and the South Carolina State Agricultural Society were taken from the annual

session laws published annually in Columbia, usually titled Acts of the General Assembly

of the State of South Carolina. State budget data was taken from Lacy K. Ford, Origins of Southern Radicalism: The South Carolina Upcountry, 1800-1860 (New York: Oxford Uni

versity Press, 1988), 311.

22. David F. Weiman, "The Economic Emancipation of the Non-Slaveholding Class:

Upcountry Farmers in the Georgia Cotton Economy," Journal of Economic History 45

(Mar. 1985): 83-88; Ayers, Promise of the New South, 3-13, 55-62; Gavin Wright, Old

South, New South: Revolutions in the Southern Economy since the Civil War (New York:

Basic Books, 1986), 39-42; Michael M. Bell, "Did New England Go Downhill?" Geo

graphic Review 79 (Oct. 1989): 460-61.

23. Gilbert C. Fite, Cotton Fields No More: Southern Agriculture, 1865-1980 (Lexing ton: University Press of Kentucky, 1984), 85-86.

24. To determine infected counties, we consulted the Bureau of Animal Industry, Map

Showing Boundary Line of District Infected with Splenetic or Southern Cattle Fever, as

Defined in Order of Hon. J. M. Rusk, Secretary of Agriculture (Washington, DC: GPO,

1891). 25. For improvements along the Mississippi, see, Albert E. Cowdrey, This Land, This

South: An Environmental History, rev. ed. (Lexington: University of Kentucky Press,

1996), 122-23; Ayers, Promise of the New South, 194-95. Changes in the percentage of

improved land calculated from census data available at the Geostat Center.

547


26. This fact fits well with Carville Earle's negative assessment of phosphates in "The

Price of Precocity: Technical Choice and Ecological Constraint in the Cotton South,

1840-1890," Agricultural History 66 (Summer 1992): 55-58. Census data on money spent

on fertilizers is available at Geostat Center.

27. Rural population densities calculated from "Urban and Rural Population: 1790 to

1970," US Census of Population: 1970 (Washington, DC: GPO, 1970), Vol. I, Pt. 1, Sec. 1,

Table 18; T. H. Breen, Tobacco Culture: The Mentality of the Great Tidewater Planters on

the Eve of Revolution (Princeton: Princeton University Press, 1985), 43; Lee Soltow and

Edward Stevens, The Rise of Literacy and the Common Schools in the United States: A

Socioeconomic Analysis to 1870 (Chicago: University of Chicago Press, 1981), 166-76.

28. Scholars have long documented the importance of prosperous rural markets in

stimulating early manufacturing. David R. Meyer, The Roots of American Industrialization

(Baltimore: Johns Hopkins University Press, 2003), 15-54, 162-88. See, also, John Ma

jewski, A House Dividing: Economic Development in Pennsylvania and Virginia before the

Civil War (New York: Cambridge University Press, 2000), 141-72; Diane Lindstrom,

Economic Development in the Philadelphia Region, 1810-1850 (New York: Columbia

University Press, 1978); John Majewski and Viken Tchakerian, "Markets and Manufac

turing: Industry and Agriculture in the Antebellum South and Midwest," in Global Per

spectives on Industrial Transformation in the American South, ed. Susanna Delfino and

Mich?le Gillespie (St. Louis: University of Missouri Press, 2005), 131-50.

29. Peter Coclanis, The Shadow of a Dream: Economic Life and Death in the South

Carolina Low Country, 1670-1920 (New York: Oxford University Press, 1989), 143-50;

David L. Carlton, "Antebellum Southern Urbanization," in The South, the Nation, and the

World: Perspectives on Southern Economic Development, ed. David L. Carlton and Peter

A. Coclanis (Charlottesville: University of Virginia Press, 2003), 39-AV, Mathew, Edmund

Ruffin and the Crisis of Slavery in the Old South, 162.

30. Soltow and Stevens, The Rise of Literacy and the Common Schools in the United

States, 166-76.

31. Ibid., 155-66; James M. McPherson, Ordeal by Fire: The Civil War and Reconstruc

tion, 3rd ed. (New York: McGraw-Hill, 2001), 27-29. On the lackluster state of the south

ern publishing industry, see, Alice Fahs, The Imagined Civil War: Popular Literature of the

North and South, 1861-1865 (Chapel Hill: University of North Carolina Press, 2001), 21-41

and Drew Gilpin Faust, The Creation of Confederate Nationalism: Ideology and Identity in

the Civil War South (Baton Rouge: Louisiana State University Press, 1988), 16-18.

32. Richard L. Bushman, The Refinement of America: Persons, Houses, Cities (New

York: Vintage Books, 1992), 390-98; Eric Foner, Free Soil, Free Labor, Free Men: The

Ideology of the Republican Party before the Civil War (New York: Oxford University

Press, 1970), 40-72; Foner, "Free Labor and Nineteenth-Century Political Ideology," in

The Market Revolution in America, ed. Stokes and Conway, 99-127; John Ashworth, "Free

Labor, Wage Labor, and the Slave Power: Republicanism and the Republican Party in the

1850s," The Market Revolution in America, ed. Stokes and Conway, 128-46.

33. Frederick Law Olmsted, A Journey in the Seaboard Slave States, with Remarks on

Their Economy (New York: Mason Brothers, 1859), 65, 66, 74-76.

548


34. For these basic divisions, see, Sam Bowers Hilliard, Atlas of Antebellum Southern

Agriculture (Baton Rouge: Louisiana State University Press, 1984), 10-12. We also con

sulted the various state maps available at USDA, Natural Resources Conservation Ser

vice, "Southeast Coastal Plain and Caribbean: Soil Survey Region #15," http://

www.mol5.nrcs.usda.gov/states/index.html; Richard E. Lonsdale, Atlas of North Carolina

(Chapel Hill: University of North Carolina Press, 1967), 2-5; William Thorndale and

William Dollarhide, Map Guide to the US Federal Censuses, 1790-1920 (Baltimore: Ge

nealogical Publishing Co., Inc., 1987).

549

Documents

The Environmental Origins of Shifting Cultivation: Climate ...shifting cultivation began with the burning of forest growth to release nutrients into the soil. After five or six years,