America in 1820

Each of the decades of the nation’s history prior to 1810-1820 saw at least one state abolish slavery. None would do so again until the 1860s. The importation of slaves ceased, legally, at the start of 1808. By then only South Carolina still did so, but smugglers took up some of the slack by dodging Navy patrols or running slaves across the border from Florida and Texas. The last year of the decade saw the Missouri Compromise, which continued earlier precedents for dividing the nation between slave and free territory. It accounted for all the land in the nation, until the nation took some more from Mexico, and so formed a kind of touchstone for those trying to settle sectional disputes in the generation to follow.

One can stretch the idea of turning points too far. Large subjects in history come down mostly to processes rather than single points in time, but certainly the second decade of the nineteenth century marked a kind of transition from a national order that imagined slavery’s inevitable and happy decline and to one based on a perpetual balance of power between sections divided by the legal status of slavery.

The North in the 1820 census. (Click for a larger version.)

The North in the 1820 census. (Click for a larger version.)

In 1820, the nation stood 15.15% enslaved, down from 16.45% in 1810. Of those 1,535,199 slaves, only 19,100 (1.24%) lived in the North., down from 2.13% in 1810. As before, New York and New Jersey account for vast majority (92.34%) of the nation’s slaves between themselves. Nowhere else had more than a thousand, though Illinois comes very close at 917 (4.80% of the Northern total and 1.66% of its population). By national standards, no northern locale is significantly enslaved. By the standards just of Northern demography, 0.91% enslaved counts as significant. New Jersey and Illinois hold that distinction in 1820, with late emancipation and a combination of geography and a free state constitution that allowed “apprentices” for life doing much to explain their divergence.

The South in the 1820 census. (Click for a larger version.)

The South in the 1820 census. (Click for a larger version.)

Down South lived 98.76% of America’s slaves. The upper band of the national norm for slavery in 1820 stands at 30.57%. The Southern regional demographics exceed that. As a whole, 33.84% of the section’s population lived in bondage. Alabama (32.91%), Georgia (43.89%), Louisiana (45.02%), Mississippi (43.49%), North Carolina (32.09%), South Carolina (51.35%), and Virginia (39.91) exceed the norm on their own. By Southern standards, anything above 45.89% counts as unusually enslaved. South Carolina maintains its place of distinction and by 1820 has recovered its colonial-era status as a majority-slave jurisdiction, further South demographically than the South itself.

The minimum level of slavery necessary to count as Southern in 1820 stands at 21.79% of the population. Just as in past censuses, the South has some under performers: Arkansas (11.33%), Delaware (6.20%), the District of Columbia (19.37%), Missouri (15.35%), and Tennessee (18.95%). The least of these, however, still exceeds New Jersey’s 2.72% by a healthy margin. As this pattern persists over all four censuses, and knowing how things work out in the 1860s, I feel confident in calling it a meaningful trend.

While the legal status of slavery might evenly divide the nation in two, the demographics argue for something more like a two and a half section model. The North stands wildly out of pace with the rest of the nation, far more free than the South as a whole and even more free at its most enslaved than the South’s least enslaved place. The South forms the second section, but its upper reaches are out of pace with the rest just as they don’t fall in step with the North. They come in too Southern to qualify for the normal in the North, but too Northern to qualify for normal in the South.

I call the border states/upper south bloc half a section because as a practical matter they tended to vote Southern in most cases, even if a few peeled away to vote Northern as part of the general pro-compromise Lower North/Upper South demographic. But then that demographic did not form its own coherent bloc that happily lined up as one for each compromise. Rather, architects of compromise like Henry Clay and Stephen Douglas found the few extra votes they needed within the group to join to the general unanimity of the sections otherwise. The cut and thrust of Southern loyalty (which in practical terms meant loyalty to slavery) politics created ample room for Upper South politicians to court the ultras to show up less dedicated opponents or shore up their own flagging fortunes.  Likewise when their local interests came into play, as they did with the Fugitive Slave Act, Upper South politicians could go more Southern than the Deep South, which for geographic reasons had far fewer runaways to reclaim to begin with.

America in 1810

To keep the images manageable and marginally legible without clicking through, I’ve split them by section.

Ten years rolled by and 1810 brought the third census, a snapshot of America during the Madison years. Between last census and this one the nation expanded through the Louisiana Purchase. For the first time since the Northwest Ordinance passed, the United States had territory not yet spoken for with respect to slavery. As of 1804, New Jersey adopted gradual emancipation. As of January 1, 1808, the United States forbade the importation of slaves.

The nation stood 16.45% enslaved, down from 1800’s 16.84% and 1790’s 17.75%. Of those slaves, only 2.31% lived in the North, down from 4.05% in 1800 and 5.75% in 1790.

The North in the third census. (Click for a larger version.)

The North in the third census. (Click for a larger version.)

In 1790, the North enslaved 59.72% of its black population. In 1800, it enslaved 43.44%, nearly reversing the proportion in just ten years.  Come 1810, it enslaved only 26.03%. The bottom fell out of Yankee slavery in both metaphor and geography: in advance of emancipation dates, many slaveholders sold their human property South.

Though officially free, only five locales lacked slaves entirely: Maine, Massachusetts, New Hampshire, Ohio, and Vermont. Connecticut,  Illinois, my native Michigan, Pennsylvania, Rhode Island join them below the typical minimum national rate of enslavement, 1.54% in 1810. As before, nowhere in the North exceeds the typical national rate.

Setting aside national standards and just looking at the North on its own terms tells a different story just as it has before. Anything beyond 1.58% enslaved puts a state or territory out of pace with Northern norms. New York, which adopted emancipation in 1799, has fallen into line with the rest of the section. In 1800, the Empire State enslaved 3.52%. Now it held only 1.57%. Fellow 1800 misfit Illinois also falls into line, dropping from 4.35% to 1.37%. New Jersey, late to emancipate, now stands alone with 4.45% enslaved.

While not all sunset and light, the arc of history seems to point toward freedom…in the North.

The South in the third census. (Click for a larger version.)

The South in the third census. (Click for a larger version.)

Down south, much the opposite story played out. Between adding enslaved territory, twenty-five years of slaves coming in across the Atlantic, and human husbandry to grow the slave population (with a significant amount of a female slave’s value coming from the expectation of her producing more slaves, called her increase), much of the section comes in significantly above the national norm: Georgia (41.68%), Louisiana (45.27%), Mississippi (46.69%),  South Carolina (47.30%), and Virginia (40.27%). Among those national over-performers, Louisiana, Mississippi, and perennial favorite South Carolina exceed even the higher Southern norms. You can see the Deep South getting its start.

The lower end of normal for the South comes at 22.08% enslaved. Arkansas (12.81%…and all of two free black people), Delaware (5.75%), Maryland (3.02%), Missouri (14.43%), and Tennessee (17.02%) all fall out of the South. The District of Columbia (22.97%) comes close.

As in the two previous censuses, each region has its misfits but even the most enslaved Northern area comes in below the least enslaved Southern area. New Jersey (4.42%) and Delaware (5.75%) come very close, but even as they have converged the rest of the North has moved farther toward freedom. But while everywhere in the North sees slavery on the decline, only a few places in the South see the same: Delaware, the District of Columbia, and Maryland.

America in 1800

In 1800, John Adams moved into the unfinished White House. He moved out in 1801, giving way to a Virginian as a Virginian had given way to him. Thomas Jefferson owed his victory in the close election of 1800 to the special accommodations the Constitution gave to slavery: absent the extra electoral votes given primarily to the South for 3/5 of its slave population, otherwise counted by the law as no more than livestock with two legs, Adams would have won. No non-Virginian would again take up residence in the Executive Mansion until Adams’ son came after Jefferson, Madison, and Monroe all left.

The Constitution required an election in 1800. It also required a census:

The second census. (Click for a larger version.)

The second census. (Click for a larger version.)

Unfortunately, as American territory expands and states and territories proliferate, the spreadsheets will only grow and so become less and less legible when reduced to a size that fits within the margins of the blog. I put it in the caption as well, but in case I ever forget you can always click the image for a version kinder to the eyes.

By 1800, every state of the North save New Jersey had abolished slavery by means of gradual emancipation. They still had slaves, as you can see. New York’s 20.613 slaves 56.92% of the total Northern slave population. New Jersey’s 12,422 make up 34.30% of the section’s bondspeople. But the remainder hold a few surprises. Probably every American schoolchild learns that the Northwest Ordinance prohibited slavery in the future states of Ohio, Indiana, Illinois, Michigan, and Wisconsin. That land appears in the census for the first time and 135 slaves. More curious still, 107 of those lived in Illinois.

What happened to make the Land of Lincoln into a slave state? Quite a few things and the story deserves its own post. Without looking at the microdata, which I do not have, I can’t tell how many of Illinois’ slaves we can attribute to any particular cause but at least two different stories come into play here. First, the Old Northwest had plenty of French traders living in it. They originally came when France claimed the territory and with British and then American authority distant and local British and Americans thin on the ground, saw little reason to leave. For that matter, it took until 1795 for the British to see cause to leave despite ceding the land to the United States at the end of the Revolution. Those French traders sometimes had slaves with them and the Northwest Ordinance barred slavery from the land but took no action against any slaveholders already there.

The second story flies in the face of the usual received narrative about an active, always forward-pressing North and a laid back, some might say ossified, genteel South. When it came to populating the West in the early years of the nation, Yankees tended to stay at home. Southerners filled up Kentucky and got it statehood in 1792. Not satisfied, they filled up the Southwest Territory too and Tennessee came into the Union in 1796. Only Vermont, the fourteenth state, came in before them. Those same energetic Southerners did not see the Ohio River as the Great Wall of China. They came over into the areas immediately adjacent to Kentucky. Probably some of them brought slaves along on the sly. Trying to lure more, Congress suspended the Northwest Ordinance’s ban on slavery for ten years.

But let’s look at the statistics. In 1800, the nation was 16.84% slave, down from 1790’s 17.75%. The sections’ shares of the white population have changed less than a percentage point. But with gradual emancipation on the books in every Northern state that has slaves to emancipate, save tardy New Jersey, its share of the nation’s slave population has dropped from 5.75% to 4.05%. The corresponding rise comes from the South, which still had Constitutional and national license to import slaves from Africa and the Caribbean as well.

Nationally, the typical enslavement rate runs from 1.75% to 31.93%. Of the Northern locales that still have slaves, Illinois (its small population helping skew things) and 1790’s two Northern misfits, New Jersey and New York, fit into the low end of that range. Of the remainder, only Indiana comes close at 1.06%. Both returning misfits have seen declines in their percents enslaved as well. Two data points do not make a trend worth taking very seriously, but so far as they go we can say New Jersey and New York lean more Northern now than they did a decade prior.

By the standards of just the North, the typical percent enslaved falls between -1.55% and 1.55%. The negative number is an artifact of the math; obviously you can’t have negative people. So many places with no slaves at all skew the range. But as slavery declines, the norms pull more and more toward no slavery at all. By the 1800 benchmarks, Connecticut, Indiana, Pennsylvania, and Rhode Island practice slavery on a small enough scale to fit within the North. Illinois, New Jersey, and New York belong no more by the sectional norms than by the national norms.

Down South, Alabama, Georgia, Mississippi, South Carolina, and Virginia all exceed the national norms for slavery. Restricting ourselves to Southern demographic standards, the high end of normal slavery comes at 41.15% enslaved. Once again South Carolina stands alone at 42.29%, a very slight decline from its 1790 performance at 43.00%. (They come roaring back later, of course.)

But on the low end, locales with less than 24.24% enslaved populations don’t belong either. (Neither, of course, do the places too enslaved to fit with the Northern norms. High degrees of slavery by Northern standards doesn’t get a place close enough to see the low end of slavery by Southern standards.) Those include the perennial southern misfit Delaware, Kentucky, and Tennessee just as in 1790. The newly incorporated District of Columbia almost slips out of the demographic South, weighing in at 25.44% enslaved. That constancy masks the beginning of another trend: In 1790, the lower bound for slavery in the South stood at 23.15%, a 1.09% difference. One percent doesn’t make the ground shake and cause lots of wailing and gnashing of teeth, but as early as we have two sets of statistics to compare, the South leans to growing slavery and the North to shrinking slavery.

Neither of those trends had to continue and neither had to make civil war inevitable, but both started at least seventy years before P.G.T. Beuregard looked down the barrel of a cannon at Fort Sumter. I wish we had comprehensive colonial censuses to tell us how far back the divergence goes. Already the regional norms have precious little overlap, but when did they last? Did they ever?

A Better Look at 1790

As previously stated, standard deviation tells us less than we’d like about data which do not fit into a bell curve. Average deviation tells us more. The specific observations about 1790 from before still hold, but looked at through a better lens, the regions don’t have quite the same ambiguity.

The first census with average deviations. (Click for a larger version.)

The first census with average deviations. (Click for a larger version.)

By national standards, in 1790 the normal enslaved population of a place should come between 4.65% and 30.86%, still a large range but not so huge as the standard deviation gave. Most of the North falls below that by a healthy margin, but New Jersey and New York exceed it. They fall into the lower extremity of the normal range.

But that’s by national standards. A normal Northern place has an enslaved rate between 0.06% and 4.02%. By that standard, New Jersey and New York don’t belong. Nor, of course, do the places with no slaves at all: the future Maine, Massachusetts, and Vermont. Those three locals have too much freedom, if by a small margin, to fit with the more typical Northern slavery of Connecticut, New Hampshire, Pennsylvania, and Rhode Island. One could call them the North within the North and New Jersey and New York the South within the North.

Down South, nowhere has a notably low rate of slavery. On the far side of the national range, Georgia, Maryland, South Carolina, and Virginia stand out. All four have significantly greater numbers of their people enslaved. But the South doesn’t amount to a home away from home for New Jersey and New York. The region’s normal range runs from 23.15% to 43.86%. South Carolina beats that top band by a narrow margin, keeping its place of distinction as the most enslaved of the most enslaved. But Delaware, Kentucky, and Tennessee all under perform by the region’s norms. They don’t quite belong in the South any more than New Jersey and New York belong in the North, demographically speaking.

Screwing Up by the Numbers

I screwed up. Sorry.

Yesterday I ran the standard deviations of 1790 census numbers. I did not compute them wrongly or interpret them wrongly, but I did use the wrong measure entirely given the nature of the data. The standard deviation works best for data that fit into a bell curve arrangement. The first standard deviation is thirty-odd percent in either direction from the peak of that curve. The census refused to give me data that cooperated with that. I still think that North and South had significant overlap, if not as much as I did when I wrote yesterday’s post. None of the state-level data changes. But because the census data has, in effect, two peaks (one for the North and one for the South) the standard deviation is unhelpfully large. It expresses a mathematical fact with limited real world utility.

I noticed the large normal range and a few other oddities when writing the post, but chalked them up to my general ignorance of statistics.  When working over the 1800 statistics, I saw the same pattern developing. I reached out to a friend who did chemistry and physics in college and now teaches junior high science. He put in a fair amount of time finding out where I went wrong and getting the right tool for the data. Unlike me, he knows his math.

For heterogeneous data like the censuses, average deviation fits better than standard deviation. Luckily, my spreadsheet can do that too. I’ll put the new data up in their own post shortly.

Misfit Places and Standard Deviations

Continuing yesterday’s theme, I want to use the 1790 census to look at misfit places. We can take some averages and say that a typical Northern state enslaved 2.04% of its population. Free black people constituted 1.37% of the population. White people made up the remaining 96.74%. Likewise a typical Southern state enslaved 27.81%, had a free remainder of 1.8%, and white people accounted for the other 70.93%. But individual place really fits the average. Nor does any fit the national averages. We can look at each region and see some clear trends: the North had a much whiter populace and far fewer slaves, both in absolute terms and as regional and state-level percentages, than the South did. But there’s a lot of heterogeneity within both regions even allowing for those trends.

I intended to just point at that in the table and comment on a few places that leaped out to me, a prospect which left me wondering if the post deserved writing since you can read the columns as well as I can, but in the course of writing this post ancient memories stirred. Back in high school I did briefly learn how to figure standard deviations and what they meant. Determining the standard deviation of a set of data lets you know how different from the mean it needs to be before that difference counts for more than just statistical noise. Crucially, my spreadsheet software also does the math for me.

The first census with standard deviations. (Click for a larger version.)

The first census with standard deviations. (Click for a larger version.)

What does all of that mean? Enslaving any percent of the population between 2.29% and 33.21% of the population counts as normal for 1790. Coming in beneath 2.29% or above 33.21% makes a territory notably free or enslaved, respectively. Of the Northern states and territories, only New York and New Jersey don’t fall into the notably free category.

I don’t know about New Jersey’s history to say, but colonial observers noted how Southern New York often seemed. Like the Chesapeake and the Carolina lowcountry, the colony had a bad case of the ubiquitous New World problem: too much land and not enough people to commercially exploit it. The Dutch, like the English further south, could find plenty of their countrymen who wanted to come over, strike it rich, and hurry home but few who wanted to stay for good. Faced with the same problem, both colonies turned to the same solution: bringing people who did not want to come and working them against their will. When the English took over New Amsterdam and renamed it New York, they saw little reason to change course.

On the other side of the probability curve, Georgia, South Carolina, and Virginia share the 33.21%+ enslaved club. We could call them a surrogate Deep South. Virginia tobacco plantations hardly require an introduction, even if the industry had seen much better days by 1790. South Carolina does deserve a bit more. The lowcountry, which extends down the Carolina coast and into Georgia, had the New World problem with an extra complication: few Europeans even wanted to try their hand in its coastal swamps. Precious few crops they knew grew in such terrain and tropical diseases gave them little inclination to experiment. But many of the early Carolinians did not come from Europe, but from overpopulated and deeply enslaved Barbados. They brought Barbadian slavery, already generations old, with them. Land that most white men did not want just meant land ideal for black men and women to work, whether they liked it or not. But what to grow there?

Here the slaves probably led the way. Europe knew few crops that thrived in swampy lowlands, but Africa had one: rice.  Cultivating rice required a lot of know-how, infrastructure, and constant maintenance to keep the fields flooded when they needed it and dry when the did not. All of that required a lot of labor, but they had a ready supply of cheap, unwilling labor. The low country had the perfect traits to become the most heavily enslaved section of the country in the years before the cotton gin.

Of course, that does not mean that 1860 simply recapitulates 1790. In 1790, no state or territory had a black or slave majority. South Carolina came closest, with Virginia a near second. Before the revolution, South Carolina did have a slave majority. In fact, South Carolina probably had that majority by the early 1700s. Between an upcountry frontier and a lowcountry most whites, save planters, wanted nothing to do with, one could expect little else. The colony lost an estimated 20,000 slaves during the Revolution and refused to ratify the Constitution unless it had twenty-five years to rebuild its economy with the brick and mortar of lives stolen from Africa.

The extremes have novelty to draw the eye, but the statistics don’t lie: while Northern areas tend to cluster toward freedom and Southern toward slavery, most places fit within one standard deviation of the mean.

The Ambiguity of North and South in 1790

Update: This post is not as accurate as it could be. None of the numbers are wrong, but the standard deviation does not give us measures as useful as one would hope. This doesn’t change any facts, but does have interpretive consequences. Short version: ambiguity remains but the regions do have larger differences.

Given the length, I’ve opted to cease using Slavery by the Numbers as part of the post title. It remains in the categories over on the sidebar.

With the exception of the immediately preceding longitudinal post, I have drawn all Slavery by the Numbers data from the 1860 census.  But the United States conducted censuses every decade, on the decade, from 1790. Those census figures highlight continuities that one might expect, but also interesting differences. To tease those out, I’ve decided to do some comparisons. In addition, I plan to use the data compiled in this post as a starting point for future posts with more detailed analysis.

My categorization of North and South rests on the familiar 1860 definitions, not on 1790 understandings of the regions. I have marked non-state areas with a (T) for territory, regardless of their then-current legal status.

America according to the first census in 1790

America according to the first census in 1790 (Click image for a bigger version.)

Immediately continuity leaps out: 94.25% of the nation’s slaves lived South of the Mason-Dixon line. One could take away from that the notion that slavery had only a transient, irrelevant existence in the North. Massachusetts, Maine (then a part of Massachusetts) and Vermont had no slaves at all. New Hampshire had less than two hundred and Rhode Island less than a thousand. All in all, only 5.75% of the nation’s slaves lived in Northern states or territories, where they amounted to 2.04% of the population.

But in 1860, Delaware enslaved only 1.60% of its people. In 1790, New Jersey and New York both enslaved almost six and a quarter percent of their populations. Those two states alone account for 32,616 (81.37%) of the North’s 40,086 slaves in 1790. Rhode Island (1.39%) and Connecticut (1.11%) come quite close to 1860 Delaware. While it feels strange, a century and a half later, to call barely-enslaved Delaware a slave state, we ought to find it just as strange not to think of 1790 New Jersey and New York as slave states. While neither matches 1860 Missouri, the second least enslaved state of the then-South, they get within two-thirds of its 9.72% enslaved population.

All of the previous, however, looks at slavery from a distinctly white perspective. White people enslaved this percent or that percent of other people. We at least implicitly measure their virtue in the size of that percentage and then look askance at the South for exceeding it. The data do support it; one cannot get around the fact that the South had slavery on a far larger scale, enslaving vastly more people even when Northern law still permitted slavery. Of course the South also had slavery longer and much of the South fought a bitter war to keep it.

But what did that noble North look like to black people? Across the North, 59.72% of the black population lived in bondage. Place by place that varies from 100% in the states with no slaves at all to a mere 18.09% in New York. The regional average comes to 40.28% free, but that figure includes areas with rather few people of any color or condition. The three places with no slaves, Maine, Massachusetts, and Vermont account for only 6,174 free black people, a mere 9.20% of the North’s free black population.

To contextualize this a bit more, 1790 New York and 1790 New Jersey’s enslaved black populations (81.91% and 80.53%, respectively) exceed 1790 Delaware’s  69.51%. A random black person from those states is more likely enslaved than one from Delaware. For that matter, the same is true if we compare 1970 New York and New Jersey to 1860 Maryland. The 1860 Border States had 21.56% (maximum of 91.69% in Delaware, minimum of 3.01% in Missouri) of their black populations living free, at least in the eyes of the law.

All of this touches the edges of a related story. White people accounted for 96.56% of the 1790 North’s population and the white North comprised 59.89% of the nation’s white population. The rest of the white people, 40.11%, lived with 91.14% of the nations black people. Of that third of the South’s population, they owned 95.29%. Whatever virtues and vices each section had, and each had its fair share of both, the South had far more black people on hand to hold as slaves. Taking into account pervasive racism, that also translated into far more reason to adopt and maintain far harsher means of racial control.

To turn things around a little and touch on a point from this blog’s infancy, since slavery forms the common ground on which the states we call Southern stand, in 1790 the South included at least New York and New Jersey. Both still practiced de jure slavery. Residual slavery persisted in the form of eighteen people in New Jersey up into 1860. Despite its early emancipation law, Pennsylvania still had some residual slavery until 1847. Slaves also appear for Connecticut in the 1840 census.

Colors on a map give the impression that people live in worlds of bright lines and clear distinctions, but those exist for our convenience in making generalizations. Those generalizations have many uses and we could not make any sort of progress at anything of consequence without making some kinds of generalization. But North and South do not reduce to colors on the map or islands in the sea, all one or the other with never the two meeting or mixing. Instead both coexist to varying degrees and in varying amounts from place to place and time to time.

Nothing on Glowing Wounds at Shiloh

Past posts on this subject: The Strange Tale of Glowing Soldiers, A Tiny Update on Glowing Soldiers

My inquiries on period sources for the story that some soldiers at Shiloh had glowing wounds, attributed to a possible infection by bio-luminescent bacteria, have hit a dead end. I found nothing in the Official Records. The author of the Mental Floss piece on the story had the science, but had not checked contemporary accounts. (I don’t mean that as a criticism, by the way. There’s nothing wrong with focusing on one aspect of the story instead of another and the science is interesting in itself.) I used the web contact form and emailed the Shiloh National Military Park.

They got back to me today and have not been able to find any contemporary source. If anybody ought to know, they ought to know. A phenomenon so conspicuous and unusual would generate at least some paper trail which should to have survived, but one proves elusive. A key part of the story, at least to me, is that the soldiers with the glowing wounds had a better survival rate than those without. That fits the science, but also implies at least an informal study of outcomes.

Since I started looking into this, I’ve finished Drew Gilpin Faust’s This Republic of Suffering: Death in the Civil War. She spends a lot of time on how disorganized and overwhelmed field hospitals were by the sheer numbers of wounded and dead. A formal ambulance corps did not exist until late in the war. The armies issued no forms of personal identification like dog tags. Even something as simple as formally notifying the families of the deceased fell on officers and doctors with more pressing duties. That often meant men reported wounded when dead, dead when wounded, or even either when they emerged unscathed. (At least one soldier reported dead visited his own grave yearly after the war.) Shiloh came early on, when the disorganization and chaos would be greatest. It hardly seems like the place for even a casual study of who lived and died.

Some text could appear tomorrow and show otherwise, but this story looks more like folklore that grew up afterward than memories of actual events.

Slavery by the Numbers: Discrepancies

This post is a little inside baseball. It doesn’t meaningfully impact any previous posts in Slavery by the Numbers but will matter for future longitudinal work. I write mostly in the interests of full disclosure and leaving myself a reminder.

The University of Virginia’s Census Browser has returned to my internet and so I dug down into how numbers came up differently. So far I’ve only posted national totals, but I got those totals from adding up UVa’s state-level totals, which also include some, but not all territories. For every census, the UVa-derived totals and Gibson & Jung’s totals did not match. I knew that UVa excluded the District of Columbia but previously considered the pre-territorial populations of most places negligible. In doing that I forgot that the usual progression from unorganized territory to organized territory to part of the organized territory getting statehood did not always work out in practice. Ohio became a state direct from the Northwest Territory, without any Ohio Territory in between. So a population I would neglect because it did not appear on UVa’s summaries actually amounted to a number of people I would otherwise think fit to count.

I went back to my spreadsheets with Gibson & Jung in hand and confirmed that the state-level totals at UVa match. Likewise, with one exception where I entered a formula incorrectly, so do the percentages. That left only the territories and the District of Columbia for me to miss people in. Initially, I thought most of the missing people had to be in the District, because I did not remember the Ohio situation and related matters. (I did find some in the District, but far from all of them.) As my analysis to date has focused on 1860, when all the land in question constituted states, that those details just flew past me. In future longitudinal work I will keep the greater complexities in mind.

Slavery by the Numbers: Errata and More Longitude

The table I posted yesterday had a serious error and a lot of small ones which I should have spotted but did not. Right after I post this I’ll be editing to fix it. The serious error comes in the 1800 census figures, where the black population leaps from 19.25% to 24.01% of the nation’s total. I noticed the unusual spike when I made the table, but did not notice also that the white proportion of the population did not also decline. Somehow, I ended up with 105% of all Americans. That can’t be right, even accounting for some rounding. I also did not notice that the share of the population made up by black people increased but the percentage of slaves and free blacks did not. Legally and logically, all black people the census counted had to fall into one of those categories.  The second discrepancy did not become apparent to me until I worked up the charts for today’s post.

At first, I thought the peak might represent a rush of slave imports in advance of the prohibition that hit at the start of 1808, but then those people would appear as a spike in the percent slave column. They do not. Likewise such would require a reduction in the white proportion of the population. No such luck. I went to check my figures against the University of Virginia’s census browser, thinking I might have mistakenly carried a figure over from one column to another, made a regular typographical error, or something like that. Unfortunately, I haven’t been able to raise the site all day.

But census figures also exist at the Census Bureau, oddly enough. Therein I found Historical Census Statistics on Population Totals By Race, 1790 to 1990, and By Hispanic Origin, 1970 to 1990, For The United States, Regions, Divisions, and States by Gibson and Jung. They both draw on census data, so I used Gibson and Jung to check my national totals. Sure enough, I found errors. Most figures did not vary by a great deal, but essentially every one had small variances from what I transcribed earlier. Relying on Gibson and Jung, I replaced the data and the sudden surge in anomalous black Americans vanished. I don’t know how many of the errors come down to my poor transcription and how many belong to the poor UVa students who transcribed from the original tabulations. I have most of the 1860 tabulations as scans of the original published tables and they do not make for easy reading with tiny type closely packed. The older census scans I have seen have similar issues. In either case, Gibson and Jung’s national figures make sense and mine and/or UVa’s do not.

The Longitude of Slavery. (Click image for a larger version.)

The Longitude of Slavery. (Click image for a larger version.)

I’ve said before that I do not have any special gift for numbers. I can see the trends over time, but the degree of change and its relation to changes in the other variables leave my innumerate brain with a bit too much to keep track of. The following charts present the same data much more effectively. I planned to include them with yesterday’s post, but could not get the spreadsheet to generate what I wanted. A friend helped me out with that. (Thanks, bud.)

The national population broken down by race and slave status.

The national population broken down by race and slave status.

Right away, some of the human cost of slavery shows through. White population growth follows the total trend closely, as one would expect from a nation never less than 80.73% white. But white and black lived in the same places, usually at very close proximity. They suffered the same food shortages, natural disasters, diseases, and so forth. Given that, we would expect both lines the follow quite similar curves. They do not. However, this measure does not quite compare apples to apples as the white population also grew due to immigration.

Apologists for American slavery will sometimes go on about how in the United States, and here alone, a forcibly transplanted slave population successfully replaced its own numbers and even grew. So far as that goes, one can’t argue. Caribbean and South American slavery rarely ever managed that and Caribbean sugar plantations worked through people at a prodigious rate. But not matching the most brutal and murderous slave system in the hemisphere hardly seems like something to brag about. Separately, population growth requires a fairly even distribution of the sexes. If one puts enough people of the right age and sexes together for any length of time, babies will ensue no matter what the situation.

Knowing the legal importation of slaves would likely end shortly after Congress gained the power to end it, American slaveholders had plenty of time to buy enough women to increase the numbers of their human herds. Living in an agricultural world, they probably required no advice on the matter at all. If you want more horses, you make sure you have sufficient breeding stock. The same works for people. Given the starkly commercial interest animating the purchase of women who would go on to birth more slaves, I don’t see anything to admire here. The slaveholding class might have practiced good husbandry, but they practiced it on people. The demographics tell us just how well owners as a class treated their human property: well enough not to run out.

The confounding effect of immigration appears more strongly when one looks at proportions.

The total black, free black, and slave populations as percents of the whole.

The total black, free black, and slave populations as percent of the whole.

Whatever their natural growth, black people formed a smaller portion of the population in 1860 than they ever had before. Immigration swelled the numbers of white Americans (at least by modern understandings, to the nineteenth century plenty of Europeans did not fully count as white) but black people arrived only when taken by force, whether they came openly or got smuggled across after the ban on importing slaves became law.

I did not previously know that the free black population peaked, proportionately, in 1810. It makes sense in hindsight, though. A wave of emancipation followed independence, even in states that maintained slavery to the bitter end, but it petered out as state legislatures made manumission harder and probate courts more eagerly set aside wills providing for it. Given the increasing difficulty in granting freedom and essentially zero immigration by free blacks, combined with the numerous obstacles they faced despite their freedom, a decline seems quite reasonable. Of course many who freed themselves did not stop until they reached the Canadian border. Many who did stop short went the rest of the way, at least for a while, after the Fugitive Slave Act, after Kansas-Nebraska, and especially after Dred Scott.