How Many Slaves Were Refugeed to Confederate Texas?

In his new book, Freedom National, historian James Oakes notes that by the middle of 1862, slavery was already crumbling in the lower Mississippi Valley—the “throne” of the South’s “cotton kingdom.” In Louisiana, northern Mississippi, and western Tennessee, “everyone knew that runaway slaves and Union armies formed a nearly irresistible magnetic attraction.” One result, writes Oakes, was that many planters began trying to “refugee” their slaves by moving them west out of the path of Union armies—sometimes by traveling as far as Texas.¹

I’ve spoken before about this wave of refugeed slaves to Texas, and contemporaries noticed and spoke about it, too. Arthur J. Fremantle, for example, a famous British traveler who journed from Mexico to Pennsylvania in time to witness the Battle of Gettysburg, commented on the movement of slaves from western Louisiana into Texas when he passed through the region in the spring of 1863. “We met several planters on the road, who with their families and negroes were taking refuge in Texas, after having abandoned their plantations in Louisiana,” Fremantle wrote, adding two days later near Monroe, Louisiana, that “the road to-day was alive with negroes, who are being ‘run’ into Texas. … We must have met hundreds of them.”²

That numerous slaves were refugeed to Texas is by now an established fact. But just how many slaves were “run” into Texas to remove them from the magnetic pull of Union armies?

One frequently cited estimate is that 150,000 slaves entered the Lone Star State. In a recent essay on Juneteenth, Henry Louis Gates claimed that 150,000 slaves had “made the trek west,” citing Leon Litwack’s Been in the Storm So Long as his source. Litwack himself said the number applied only to those slaves “sent out of Louisiana and Mississippi.” Historian Thavolia Glymph, in another Juneteenth essay several years ago, cited 150,000 as the number of slaves taken to Texas just in the aftermath of the battle of Vicksburg in July 1863. And other historians, too, have used this number.³

In all of these works, the 150,000 estimate usually serves one of two purposes: either to emphasize the significance of Juneteenth by drawing attention to the large number of enslaved people in Texas at the end of the war, or to emphasize how desperate Confederate planters were to move away from the “magnetic attraction” that Union armies exerted on their slaves. Viewed from another angle, however, a number this large raises questions about the relative power of Union armies and slaveholders in the lower Mississippi Valley. If planters were able to forcibly move 150,000 people to Texas—a number slightly larger than the number of runaway slaves who were enlisted in the Union army after January 1, 1863—we may well have to reconsider the general claim that the approach of Union armies after 1862 always worked to disrupt the power of slaveholders.⁴

Indeed, a smaller estimate of “refugeed” slaves to Texas would do more to strengthen the common argument that slavery in the lower Mississippi Valley was collapsing by the summer of 1862, rendering slaveholders unable to control the movements of bondspeople in the region. And, in fact, two Texas historians have offered much smaller estimates of refugeed slaves. In 1989, Randolph B. Campbell used Texas county tax records from 1862 and 1864 to estaimte that there were at least 32,000 more slaves taxed in 1864 than there would have been given natural rates of increase. While Campbell admitted that 32,000 was likely an underestimate since some refugee planters would have avoided paying their taxes, his conservative estimate was still a far cry from the 150,000 cited by other scholars.⁵

In a more recent 2008 essay, historian Dale Baum has offered some persuasive reasons why Campbell’s estimate should be revised upward. By once again mining county tax records from 1862 and 1864, while changing some of Campbell’s premises about uncounted refugeed slaves and the antebellum natural rate of increase in the slave population, Baum suggests a number closer to 47,800, and speculates that the total number could “have easily reached 51,000.” Factoring in the mortality rate of Texas slaves could make that estimate even higher, especially considering ample contemporary evidence that refugees suffered high rates of disease on the road, but even Baum’s higher estimate falls well short of six digits.⁶

Yet if Baum’s statistical estimate of refugeed slaves brought to Texas is approximately correct, then where did the oft-cited figure of 150,000 come from in the first place?

The earliest source for that number appears to be Confederate general John Bankhead Magruder. After serving for about a year as commander of the District of Texas, New Mexico, and Arizona, Magruder was transfered to Arkansas in August 1864, and in November he wrote a letter to one of Arkansas’s senators that touched on the subject of refugeed slaves. “I am told,” Magruder said, “that over 150,000 negroes have gone from Missouri and Arkansas into Texas.”⁷

Two things are worth noting about Magruder’s estimate. First, if accurate, it would probably require historians to raise their estimates of refugeed slaves to Texas even higher than 150,000. After all, Magruder referred only to those taken into the state before November 1864, and he claimed that this was the number taken only from Missouri and Arkansas, leaving the more heavily populated slave districts of Louisiana totally out of the equation. No scholar that I have found has come close to suggesting that such a large number of refugeed slaves came from Missouri and Arkansas alone, even though Magruder appears to be the source for the 150,000 estimates used by Leon Litwack and others.⁸

Secondly, however, the circumstances surrounding Magruder’s 1864 letter make it highly plausible that the general exaggerated for rhetorical effect. Magruder was writing Johnson in 1864 partly to complain about his need for additional enslaved laborers to build fortifications and support his troops’ movements in Arkansas. In February 1864, the Confederate Congress had authorized military commanders to impress up to 20,000 slaves to perform labor for the army, and in keeping with that legislation, Magruder’s superior in the Trans-Mississippi Department, Kirby Smith, had recently ordered Magruder to impress slaves in Arkansas to support an upcoming campaign to Fort Smith. But Magruder balked at the order, partly because of his past experience with Confederate slaveholders who were reluctant to comply with slave impressment orders and legislation. In Texas during the previous year, Magruder had battled constantly with state officials and private citizens in his attempts to impress slaves to build coastal fortifications in Galveston and elsewhere, and some of the fiercest resistance had come from refugee planters who believed they should have been exempted from impressment until they were settled in a new home.⁹

In view of these experiences, Magruder was convinced that any effort at mass impressment in Arkansas in November 1864 would lead to backlash among local planters, especially since (he claimed) the slave population of the state had already been so depleted by flight to Texas. His motives become clearer in the complete sentence containing his estimate of refugeed slaves:

I am told that over 150,000 negroes have gone from Missouri and Arkansas into Texas, and leaving out all considerations of fairness, I fear that if the few that remain here are impressed, those who would otherwise sow and plant would emigrate to Texas, and through the depopulation of the country we should not be able to support an army through another season, though otherwise successful.

Clearly, then, Magruder’s vague report (“I am told”) of over 150,000 refugeed slaves in Texas stemmed partly from his view that Smith’s orders were unfair and partly from his view that Arkansas planters would not abide a general impressment. Magruder hoped instead that Smith would attach some Texas counties to the Arkansas district in order to give Magruder access to a larger pool for impressment, especially since he believed that many of the refugeed slaves in Texas had been settled between Harrison County (just across the border from Shreveport, Louisiana) and Indian Territory to the north. He hoped that Johnson would help persuade Smith “as to the impolicy of impressing negroes in Arkansas.”¹⁰

In sum, I believe that Magruder’s estimate owed less to a careful empirical survey than to the politics of slave impressment in the waning days of the Confederacy. Further support for that view may be found in the fact that the next month, in a similar letter on his impressment orders, Magruder asserted that “from 100,000 to 150,000 slaves had gone to Texas from Arkansas and Missouri,” offering a much wider margin of error for his estimate. According to historian Ira Berlin, there is another letter in the National Archives, bearing the same date as Magruder’s second letter, which also uses the 150,000 figure. I have not been able to see this letter yet, but given that it was written by Theodore Herman, a Confederate engineer in the District of Arkansas, it seems likely that Herman was also writing in the context of a discussion on impressment and fortifications, and he may well have been repeating what he had heard earlier from Magruder.¹¹

This brief inquiry into the origins of the estimate that 150,000 slaves were refugeed nonetheless has two potentially very different implications for historians of emancipation and slavery in the lower Mississippi Valley. On the one hand, since Magruder’s immediate political purposes were clearly best served by a large estimate of refugeed slaves, we perhaps should be more suspicious of the number that he used. It may well have been hyperbole intended for rhetorical effect. If so, the lower estimates offered by Campbell and Baum may strengthen the argument that by 1863, the power of slaveholders’ to control the movements of enslaved people was dissolving rapidly across the region.

On the other hand, Magruder’s estimate does give valuable evidence of slave movements in 1864—a period that is beyond the reach of the statistical methods used by Campbell and Baum to arrive at their smaller estimates. As Baum notes, the Union army’s Red River campaign in the spring of 1864 likely brought a new surge of refugees into Texas, and none of these slaves would have been assessed for tax purposes until the first day of 1865. In other words, large numbers of slaves arriving from Arkansas, Missouri, Louisiana, or farther away in 1864 and 1865 would not be accounted for in many wartime county tax rolls. And because of Juneteenth and the end of the Confederacy in the first half of 1865, this influx of refugeed slaves never would have left a significant trace on public records.¹²

Magruder’s estimate, despite being a likely exaggeration, therefore remains useful as evidence that a significant number of slaves were refugeed into Texas in the very last months of the war. While their numbers may not have have reached the 150,000 high water mark often cited by historians, they may well have been higher than the low water mark estimated by Baum. Either way, the ability of refugee slaveholders to increase the slave population of Texas by 25 to 50 percent suggests that despite the “magnetic” pull of Union armies in the lower Mississippi Valley, a large number of slaveholders retained enough power or pull of their own to move thousands of enslaved people farther away from freedom.¹³

Notes