The Freedmen's Bureau files for Texas are the fullest record of the years after the war, and the general histories barely open them

The report is dated at the end of a month in 1867, from a town on the Brazos, and it is addressed to a superior in Galveston. It runs to four pages in a clerk’s hand. The first page counts labor contracts approved that month. The second lists grievances received, with the name of the person who brought each one, the name of the person it was brought against, its nature and its disposition. The third reports on schools, with enrollment. The fourth is the agent’s own remarks, and in this case they run to a paragraph about a planter who had refused to pay wages at the end of the season and a sheriff who had declined to intervene. The document is a form, filled in. That is what makes it valuable.

The files of the Bureau of Refugees, Freedmen and Abandoned Lands for Texas cover roughly five years, from the end of the war to 1870, and they are the most detailed account that exists of daily life in the state during the years that Richard Coke and his contemporaries would spend the rest of their careers arguing about. For most of the twentieth century they were consulted by a handful of specialists and skipped by nearly everyone else. I want to describe what they are, how I came to use them, and what they change, because a research problem that looks particular to Texas is a general problem about which records get read.

What the files are

The Bureau was a federal agency, run by the army, with agents posted in towns across the state. Its agents approved and recorded labor contracts between freedpeople and employers, heard grievances, kept registers of violence, oversaw schools, distributed rations in the first years and reported up the chain monthly. Everything went on paper because the army ran on paper, and the paper went to Washington, where it survived in the National Archives long after the state’s own records of the same years had burned, been discarded or never been made.

So the Texas files contain things no other source contains. Labor contracts by the thousand, naming the worker, the employer, the wage, the term and the conditions, signed or marked. Registers of grievances that name both parties and say what was done. Reports of violence with the names of victims and, often, the names of those responsible, written by an officer who expected to be believed by another officer. Lists of schools and teachers. Correspondence with local officials, some of it courteous and some of it a record of a sheriff or a judge refusing to act.

How they are read

For decades the practical barrier was that the files were in Washington on microfilm, organized by office and by date rather than by county or by name, and that reading them for one Texas county meant scrolling through every county. Digitization removed that barrier within the past fifteen years or so. The Texas files are now imaged and, in large part, indexed by name, which means a researcher can search for a person, a plantation or a town and go straight to the page. The change is hard to overstate. A record that once cost a trip and a month now costs an evening, and the evening usually turns up something that was never in a Texas history before.

The way I read them is the way I read any government form: for the boxes first and the remarks last. The counts and the registers are the reliable part, because the agent was reporting numbers his superior could check against last month. The remarks are the interesting part and the part to be careful with, because the agent had opinions, sometimes about the people he was supposed to protect and always about the people who were making his life difficult. An agent’s paragraph about a planter is a claim. The contract the planter signed and the grievance the worker brought are facts, and the two together let you weigh the claim.

What they change

The claim I will make, and it is one that some who write the history of these years will resist, is that a Texas history of Reconstruction that does not rest on the Bureau files is a history of the white politics of Reconstruction and should say so in its title. The state’s own records for those years were kept by men who did not regard the freedpeople as a subject, and the newspapers of the period were parties to the argument. The Bureau files are the one place where the majority of the events of those years, the contracts, the wages, the schools, the assaults and the refusals to prosecute, were written down at the time by someone whose job was to write them down. Coke’s political career was built on ending the regime those agents worked for. To understand what he was ending, you have to read what the agents recorded, and for a long time almost nobody did.

That conclusion carries a warning, and I want to state it plainly rather than bury it. The files are a federal record of a federal program, and they record what came to the agents. A county with a diligent agent has a thick file and looks violent and troubled. A county with a lazy agent, or no agent, has a thin file and looks quiet. The thin file proves nothing about the county. It proves something about the agent. I have caught myself twice reading a quiet file as a quiet place, and both times the county newspaper, read for the same months, corrected me. The absence of a record is a fact about the record.

There is a second warning. The people in these files appear, for the most part, only here. A worker who signed a contract with a mark in 1866 does not appear in the deed index, the brand book or the newspaper, and after 1870 he does not appear in the Bureau files either, because the Bureau was gone. The files give a researcher five years of a life and then silence, and the temptation to fill the silence with inference is strong. The honest sentence says the record ends. It is a poor sentence and it is the true one.

If you are beginning work on any Texas county in the years after the war, open the Bureau file for that county before you open the county history. The county history was written by the people who won. The file was written by the people who were counting.