Methods

First, a note on reading the maps:

The maps show estimated slave migration for each county, both by net number and by rate. First, these are net numbers of forced migrants per county, that is, the number of slaves who left that county minus the number entering. Net numbers don’t reflect the total numbers of people moving back and forth. For example, if 500 people moved out of a county and 400 others moved in, the net number of out-migrants is 100.

 How I came up with these numbers and mapped them:

Calculations are based on U. S. Census Bureau population schedules (ICPSR files, proofread against published volumes). For areas affected by boundary changes, I created “supercounty” blocs, summing the totals for all counties involved in each bloc and treating them as a single unit for that decade. In these cases, I had to assume rates of migration were equal across the supercounty area, with net numbers following each county’s proportion of the supercounty’s slave population.

For the war years, I used figures tallied for tax and conscription purposes by the Virginia Auditor of Public Accounts, 1861-1863. I inflated these to approximate the population as counted by the 1860 U. S. Census and to include slaves under 12 years old, who were not counted for taxation.

I used “growth-rate analysis,” a fairly simple demographic method described most clearly by Frederic Bancroft in Slave Trading in the Old South (1931), pp. 384-395, and by Michael Tadman in Speculators and Slaves: Masters, Traders, and Slaves in the Old South (1989), pp. 225-227, and “Slave Trading in the Antebellum South: An Estimate of the Extent of the Inter-Regional Slave Trade,” Journal of American Studies 13 (1974): 195-220.

First I calculated the growth rate of the entire southern slave population as the “natural” growth rate for each decade (adjusting downward in the first two decades to negate African imports). Then, for each decade, I applied that rate to the initial population of each Virginia county to find the expected “natural” population for the end of that decade, if no migration had occurred. The difference in the expected and the actual end-of-decade population (after scaling back for migrants’ post-move offspring) represents the net migration. If the actual population fell short of the expected number, the county experienced net outmigration; if the actual number was higher than the expected, there was net in-migration.

Natural growth rates for the African American population south-wide ranged roughly from 25-30% per decade. That means that slaveholders in counties showing “export” rates of 25-30% were basically selling off the natural increase for that decade (though they were actually selling off and moving young adults, mostly). Counties with rates of export of more than 25-30% experienced overall decreases in the slave population that decade.

In either case, these were remarkable demographic shifts that would be quite visible to people on the ground. Herbert Gutman put it this way: At least one-third of all African American marriages in slavery were broken up by this forced migration. An enslaved child living in the upper South in 1820 stood a 30 percent chance of being sold south by I860.

More on net numbers

Net migration figures also include “migration” out of the slave population by several means: flight, manumission, sale, and migration with moving planters. Flight and manumission, important as they were to the individuals involved, altered overall patterns only slightly, except perhaps in western Virginia, where small changes in numbers could change rates substantially and where even net numbers were sometimes significant. Net numbers may also hide illegal slave “imports,” although by the time of the Revolution, few were likely to have been smuggled into the Chesapeake, according to Alan Kulikoff (“Uprooted Peoples,” in Berlin and Hoffman, eds., Slavery and Freedom), due to the general labor glut there.

The proportion of enslaved African Americans who were sold off and taken south by slave traders vs. those relocated by planters was once subject of debate, with the assumption that planter migrations carried entire communities intact. Michael Tadman (Speculators and Slaves, 1989) successfully discredited the low (17%) estimate of trader-migration forwarded by Robert Fogel and Stanley Engerman in Time on the Cross (1974); Tadman’s estimate ranges from 60% to 80% of the total migration. We now know that this trader/planter dichotomy itself is somewhat false, as planters broke up many families as well, often taking south only the most fit to clear fields, establish crops, and start bearing children.

And of course, planters were the ones selling to slave traders in the first place. All these migrations, then, were forced migrations driven by the slave market. The slave market and slavery were one and the same.

Copyright 2017 © Phillip Troutman. Unmodified not-for-profit usage is freely granted with acknowledgement. Cite as: Phillip Troutman,Virginia’s Slave Trade blog, virginiaslavetrade.wordpress.com (your access date).

Advertisements

Mapping Virginia’s Slave Trade, 1790-1860

These maps represent enslaved migration across county boundaries within and out of Virginia, decade by decade, 1790-1860. Red shades represent out-migration; blues represent in-migration. The maps on the left represent rates; on right are net numbers. The key is at the bottom. Note that in western Virginia, the rates were sometimes very high when the actual numbers were relatively low. (See Methods.)

1790-1800
While planters and traders moved thousands of slaves out of the tidewater and into  piedmont and western counties, the state as a whole was already a net “exporter” to other states, as Virginia planters turned to corn and wheat, which required far fewer laborers than tobacco:

RATE790 NET1790_1800

1800-1810
Piedmont and northern Virginia counties, too, became net “exporters” to other states:

RATE800 NET1800_1810

1810-1820
The “export” market expands further west, and some tidewater counties see an actual drop in the slave population, indicating very high rates of forced removal.

RATE810 NET1810_1820

1820-1830
The pattern largely continues. Note that the exodus from northwestern counties may also represent escape across the Ohio River, while salt mining operations in the Kanawha River valley continued to demand new enslaved labor:

RATE820 NET1820_1830

1830-1840
The vast expansion of cotton and sugar in the deep south created a speculative bubble that drove slave prices up and slave traders into virtually every county, with planters everywhere looking to cash in:

RATE830 NET1830_1840

1840-1850
The bubble bursts in the Panic of 1837, but Virginia’s forced out-migration continues, even amidst a tobacco revival in the southern piedmont.

RATE840 NET1840_1850

1850-1860
Numbers remained high (over 80,000), with western counties now emptying out, perhaps as much due to escape as to the slave trade.

RATE850 NET1850-1860

Key

Key_Rate                                    Key_Net

See Methods.

All maps copyright 2017 © Phillip Troutman. Unmodified not-for-profit usage is freely granted with acknowledgement. Cite as: Phillip Troutman,”Mapping Virginia’s Slave Trade, 1790-1860,” Virginia’s Slave Trade blog, virginiaslavetrade.wordpress.com (your access date).

Graphing Virginia’s Slave Trade, 1790-1860

These graphs break down the numbers of enslaved migrants from Virginia in different ways.

“Virginia is the mother of slavery.”
Louis Hughes made this observation in his 1897 autobiography–where he described being sold away from his family in Virginia to Memphis, Tennessee. Virginia in fact represented some 45% of the total number of enslaved African Americans forced to move across state lines between 1790 and 1860:  550,000 from Virginia out of 1.1 million total forced migrants. In early decades Virginia’s share was the clear majority:

   

Destinations
As the United States seized western and southern lands from Indian nations and opened them up to American settlement, planters, traders, and bankers expanded and shifted the slave market to supply the vast numbers of laborers they demanded to transform forests and swamps into cotton and sugar plantations. This graph indicates the shifting destinations of enslaved African Americans from Virginia. For each decade, the numbers represent Virginia’s “export” divided according to each “import” state’s proportion of the overall slave market:

VaSlaeTradeDestinations

Tidewater, Piedmont, Valley, West
These graphs represent the shifting participation of each of Virginia’s sub-regions in the slave market. First net numbers, then rates (numbers below 0 represent in-migration to that region):

graph4net

graph4rate

Subregions
Breaking each sub-region down a bit further, we can see that the northern Tidewater dominated in the early decades, while the southern Piedmont overtook it in the 1830s.

graph7net

Here’s another way to see it, with each region and subregion shown as a proportion of the whole for each decade:

graph7prop

See Methods.

All graphs copyright 2017 © Phillip Troutman. Unaltered non-commercial use is granted, with acknowledgement. Cite as Phillip Troutman, “Graphing Virginia’s Slave Trade, 1790-1860,” Virginia’s Slave Trade blog, virginiaslavetrade.wordpress.com (your access date).