In January 1844, Dr. Edward Jarvis of Dorchester, Massachusetts, checked the freshly printed Sixth Census tables town by town, comparing every entry against the population it claimed to describe. The 1840 census had, for the first time, counted the nation's insane and idiotic by race, and the published totals were already read as proof that freedom caused madness: one insane or idiotic Black resident for every 144 people in the free states, against one for every 1,558 in the slave states, about eleven times the rate wherever slavery had ended. Jarvis found the first crack in Worcester, Massachusetts, credited with 133 "coloured lunatics and idiots" at public charge, the same number as the white patients already at the state hospital there.
He then matched every free state's insane count against its own inhabitants. In Scarborough, Maine, the census reported six Black residents insane in a town it separately recorded as having none at all. In Perry County, Pennsylvania, twenty-four of 154 Black residents, about one in six, were listed insane or idiotic. The errors ran only one direction: toward more madness among free Black Americans than the returns could support.
By the time Jarvis published, the false numbers had already traveled: the Southern Literary Messenger and Hunt's Merchant's Magazine ran them in 1843, reprinted across Europe as settled fact. On April 18, 1844, Secretary of State John C. Calhoun cited the same census to the British minister in Washington, defending slavery against pressure from London: wherever a state had changed the old relation between the races, he wrote, free Black people had invariably sunk into vice and pauperism, with the bodily and mental inflictions incident thereto. The census was not a footnote to that argument. It was the argument's evidence.
Congress asked its own executive to check. On February 26, 1844, John Quincy Adams moved a House resolution asking whether the printed census held gross errors, and if so, how, and what had been done to fix them. Calhoun's answer came back on May 6: that no such errors had been discovered. Adams objected on the floor, insisting the errors were real. Congress pressed further, and the State Department hired William A. Weaver, who had overseen the census, to investigate. Weaver's report, which Calhoun forwarded to the House in February 1845, found the census's count of Black insanity fully sustained. A year after Jarvis's first count, the department's own final word was the same one it started with.
A census makes one promise: that this many people, in this condition, were actually found and counted, one town at a time. Once a government prints that promise, it travels on the government's own authority, cited by cabinet secretaries and repeated by editors as settled fact. Jarvis had no power to strike a false number from a federal document; disproving it took a private citizen's arithmetic against an official seal, and proof did not move them. What changed, over the next six years, was the machine itself: the 1850 census began naming every free inhabitant individually instead of tallying a household by its head, and Jarvis was later appointed, by the Secretary of the Interior, to tabulate the government's own mortality statistics for the 1860 census.
A census is only as true as the last person willing to count it again.