Journal of Southern History

Every United States census results from a fraught political process and a mix of intentions. Because the census is tied to political power and access to federal tax dollars, Americans watch it carefully and argue over the results. Yet many forget or refuse to fill out the forms. Some undercounting is expected and even tolerated. Every ten years, as mandated by the Constitution, Congress votes to authorize and fund the census, and while the counting is underway, many of those same members of Congress run for reelection. Every twenty years, the census coincides with a presidential election. The president appoints the person who administers the census. Overseeing a census count gives sitting presidents the opportunity to leave a legacy, to shape policy after they are gone.

Of the many factors that affect the accuracy of the count, two stand out: what questions are asked, and who is doing the counting. It will be some time before we know the outcome of U.S. president Donald J. Trump’s attempt to add a citizenship question to the 2020 census schedule for the benefit of his party. There is a chance that it will work—that is, it could still lead to undercounting among Latinx communities and communities of color, in general—even though the question, feebly defended as an effort to gather data on voting rights, did not appear on the schedule.¹ If it fails, then [End Page 35] the result may be a more accurate count, an outcome that future historians studying the early twenty-first century will no doubt applaud. But between a failed count and a successful one, there are many more possible outcomes, some of them predictable and many of them not. Once the counting begins, there is no guarantee that the party doing the counting will benefit. In 1870, last-minute questions were added to the census with the goal of shoring up the power of the party in office. It did not work. An instrument of the state, the census reflects the agenda of those who authorize it, but it also relies on the cooperation of those who will be counted. This participation cannot be taken for granted, as was perhaps particularly so in the aftermath of a civil war when the contours of citizenship in the postwar nation were just beginning to take shape. The Republicans did not reap the benefits of overseeing the 1870 count, and neither did those whom they most wanted to count, the southern freedpeople. The story of how the Ninth Census managed to fail both in its noble intentions and in its counting can give historians new means with which to question the census in general and reveals what is at stake when the census asks new questions.

In the months leading up to the count, Ohio congressman James A. Garfield thought a lot about the 1870 Census and the importance of getting the numbers right. As a U.S. Army veteran who had fought against the rebellion that had tried to break apart the nation, Garfield saw the Census as a tool for putting the United States back together; he called it “a muster-roll of the American people,” a phrase that made the process sound like the people of the nation would line up in rows and report for counting.² As a Republican, Garfield was deeply committed to Reconstruction and to enforcing the constitutional amendments that were essential to Reconstruction, including the Thirteenth Amendment that in 1865 abolished slavery and the Fourteenth that in 1868 federalized citizenship and made crucial changes to the way the people of the nation would be counted. The Fourteenth Amendment’s second section revised the Constitution’s original census mandate, which, as written in Article I, Section 2, stated that each state’s representation in the House of Representatives would be “determined by adding to the whole Number of free Persons, . . . three fifths of all other persons.” With the abolition of slavery, this three-fifths clause was defunct, replaced by the Fourteenth Amendment’s new directive that representation...