The Unprinted Census: The Woman Who Saved the 1890 Records from the Fire

The 1890 United States Census is one of the great archival tragedies. In 1921, a fire ravaged the Commerce Department building in Washington D.C., and though the census schedules themselves weren’t entirely incinerated, the water damage and subsequent neglect sealed their fate. By 1935, what remained was officially destroyed by order of Congress. The event is often cited as a cautionary tale, a symbol of the fragility of physical records. But the story has a prologue, and a protagonist, that data preservationists should know about.

Her name was Lillian A. Hicks. She wasn’t an archivist by title, but a clerk—one of the hundreds employed by the Census Office under the stern direction of Superintendent Robert P. Porter. Her specific, meticulous task in the early 1890s was to transcribe the information from the handwritten census schedules onto blank cards, one for each individual, for the purpose of compiling statistical abstracts. This was the era before Hollerith’s tabulating machines were fully adopted; much of the work was still manual, human, and achingly slow.

What Hicks did, however, went beyond her duties. As she handled each card—each a single data point for a nation of 63 million—she began to see the larger shape. She recognized that the original, bound paper schedules were the sole, vulnerable source. The cards she was creating were a byproduct, destined for statistical analysis and then, presumably, the discard pile. So, she quietly, systematically, began to collate and preserve her own set of these individual cards. She didn’t save them for the statistics, but for the names, the addresses, the familial connections—the raw, human data that the originals held.

The Scrapbook of a Nation

By the time her work was done, Lillian Hicks had assembled a collection of over 63,000 cards. She didn’t have an institutional mandate to do this. She simply saw a potential for loss and created a redundant copy in a different, more durable format. She bound them into volumes, organized not by the abstraction of a statistical table, but by the tangible geography of enumeration districts. In essence, she created a parallel, searchable archive from what was meant to be ephemera.

When the 1921 fire struck, the original 1890 schedules were waterlogged and crumbling in a basement. Hicks’s personal card archive, however, was safe. It had been in her possession, then donated to a historical society. For decades, genealogists and historians have consulted “The Hicks Card Index” as the most complete surviving direct transcript of that lost census, a bridge across the archival void.

Hicks’s story reframes the 1890 Census disaster. It’s not just a tale of loss, but one of preemptive, personal preservation. She understood the core principle of modern digital preservation: redundancy in a separate, stable format. Her cards were the LOCKSS (“Lots of Copies Keep Stuff Safe”) protocol in pasteboard. Her work reminds us that the most crucial layer in any preservation stack isn’t the technology, but the human foresight to use it. Every web archivist today, crawling a site before it vanishes, is walking in the footsteps of a clerk who looked at her stack of cards and decided, against protocol, to keep them.

Notes & further reading

A few pages I came back to while writing this: