In almost any field of scholarly endeavor, there seems to be something—a source, an object, a place—that remains tantalizingly out of reach, unobtainable, even legendary. The Holy Grail and the “Lost Ark” of the Covenant fall into this category.  For librarians, the ancient library of Alexandria comes to mind.  For ardent Anglophiles, the location of Camelot has fueled years of speculation, scholarship, and archaeology.  In the field of genealogy, however, what do we have? Each family may have some elusive ancestors, the ones who seem to defy documentation, but what unites most, if not all, of us? Well, ask yourself this: how many times in your genealogy research have you caught yourself saying, “if only I had the 1890 Census!”?

Professor Jones, we “genies” have our Lost Ark.

The “lost census” of 1890 has stymied many a family history researcher, including myself.  Calling it “lost” is a misnomer: the population information was abstracted and published, ensuring that future generations would be able to understand the profile of America in 1890, but the loss (through both accident and bureaucratic option) of the schedules—the raw information, recorded at household level, has created a twenty-year void in the continuity of research using such information. Over the years, the 1890 Census has acquired a nearly mystical or mythic quality, as well as providing the excuse for a lot of genealogical frustration. In this article we will discuss the 1890 Census, its semi-demise and, on a practical note, ways by which determined family historians and others can compensate for the lost schedules through use of substitute sources.

The 1890 census was unlike previous censuses and touted many “firsts.” Most significantly, it was the first such canvass to employ automation for tabulation of population figures. The system was the creation of Hermann Hollerith, who had worked briefly for the 1880 Census: his experience led him to look for a better way to tabulate data than hand-counting.  Inspired by the Jacquard loom, a weaving machine which uses punch cards to control color patterns, Hollerith designed a machine which used the locations of holes on pasteboard cards to tabulate overall numbers, sort for individual characteristics and perform cross-tabulations. Hollertih tested his machine on a city census for Baltimore in 1887; the results were impressive enough to convince the Census Office to adopt Hollerith’s electric tabulating system for the 1890 survey, in order to speed the work of counting and analysis.

This was particularly important in light of changes to the Census for 1890. The 1890 census included a greater number of subjects than any previous census (and more than would be included in those immediately following it). New entries included questions about ownership and indebtedness of farms and homes; the names, as well as units served in, length of service and residences of surviving Union veterans and their widows. Another new question dealt with race, including "Japanese" as a category for the first time, along with "Chinese," "Negro," "mulatto," "quadroon," "octoroon," and "white." The population schedule was changed so that a separate sheet was used for each family, irrespective of the number of persons included. One other innovation was the preparation of maps which helped census enumerators stay within their assigned districts, discouraging duplicate counting! A final innovation was fateful: the 1890 Census abandoned the established practice of depositing a second copy of census schedules in county clerks’ offices throughout the nation.¹

Oops.

In March of 1896, the Special Schedules (mortality, crime, paupers, etc.) were damaged by a fire; subsequently, the Department of Interior (which at that time was the departmental “home” of the Census) ordered them destroyed. The general population schedules remained intact, however. In 1903, In 1903, the Census Office was moved to the newly created Department of Commerce and Labor and remained within Commerce when Commerce and Labor split into separate departments in 1913; the 1890 schedules were stored on wooden shelves in the basement of the Commerce Building.

On January 21, 1921, at five o’clock that afternoon, someone discovered smoke coming out of the Commerce building basement. District of Columbia firemen responded quickly, extinguishing the blaze by 9:45 p.m. However, they continued to pour water into the basement for an additional 45 minutes, to be on the safe side. The toll of fire and water damage was grave, but at the time, only 25 percent of the 1890 census materials were deemed totally lost. The remainder suffered water damage and, had the authorities acted quickly, might have survived. Workers moved the census materials to a large warehouse for temporary storage, an important first step and…there they stayed, until May, allowed to dry with no additional conservation or salvage efforts made.  That month, orders came from the higher-ups to have the records brought back to the Commerce Building, with no attempts to properly preserve the remaining documents. Debate on the fate of the 1890 Census schedules continued for the next thirteen years, until in 1934 the Department of Commerce cut its losses and ordered what remained of the 1890 Census destroyed. Workers carried out the “documentary execution” the following year. Some traces remained at large, though: in the 1940s and 1950s, fragmentary survivals of Census records from ten states (Alabama, Georgia, Illinois, Minnesota, New Jersey, New York, North Carolina, Ohio, South Dakota, Texas) and the District of Columbia, plus some Special Schedules enumerating Union Veterans and Widows of Union Veterans of the Civil War were located and added to the National Archives’ collections.²

* * *

So, is there anything that you can use for a substitute for “the 1890”? This depends on the state that is the target area for your research. Many states regularly paid for a census to be done in “off” years, and for ancestors in these states it makes the research much easier. Perhaps predictably, Arkansas did not commission such a special census, but Newport Jackson County) enumerated its citizens in 1887 by special census in 1887. This is the only known municipal census in Arkansas near the 1890 “target date,” so locating your relatives in the state can be more complicated. There are, however, other ways to fill in the 1880-1900 census gap, such as using county tax records and city directories.³

For the State of Arkansas, the county tax records constitute perhaps the best source for locating one’s ancestor during this census gap period. There are two forms of these records: personal and real-estate. As part of the Arkansas Sesquicentennial, the Arkansas Genealogical Society sponsored, in collaboration with county historical societies, a transcription project, using these resources to “reconstruct” the 1890 census. The resulting lists were deposited with the Arkansas State Archives and, when consulted along with the original documents (available on microfilm at the ASA), patrons can use these to locate their “lost” relatives. With these one can determine if the forbears lived in a particular county and if so, in which township they resided.  The real-estate taxes will only give researchers the heads of households while the personal property tax records will report anyone who confessed to owning anything of value (watch, horse, cows, etc.).

Another source by which to identify and locate one’s relative in an area is city directories. These exist for major cities in the state (Little Rock, Fort Smith, Pine Bluff, and Hot Springs and a surprising number of smaller communities as well.  If one’s ancestor was living in one of these places, they may appear in the directory. The directories help to reconstruct family units. Wives often are listed with husbands and since the address is indicated, researchers can see whether or not another person or persons are listed at the location. Patrons can access these resources in the Arkansas State Archives collection and on their free access to the Ancestry.com database in the ASA research room.⁴

For American genealogists and historians, the 1890 Census schedules constitute a sort of Holy Grail or Alexandria library— vital contents seemingly lost, stories doomed to never being told. In truth, though, some of the gap left by the destruction of the Census can be filled by patient sifting of other contemporary sources. Tax records, city directories and even special censuses may not render the exact information the census would have, but they can reveal and confirm locations as of a particular date; this provides a point from which to proceed, casting a wide net in search of the documentary traces of those elusive ancestors. available. The Arkansas State Archives staff stand ready, willing, and able to help patrons cast this net; we will be happy to assist in any of your research needs.

###