In an effort to collect my earlier writings in one place, I am planning to republish many pieces here. This article originally appeared in my online column “African American Genealogy Examiner” (no longer available online except through the Wayback Machine), posted 16 November 2011. Some links, as well as some text, have been updated to reflect changes since the original publication.
It usually does not take long before beginning genealogists discover the usefulness of the U. S. federal census. Not long after this discovery, they learn about the destruction of the 1890 census.
Because of this destruction, quite a few books have been published as “1890 census substitutes.” These generally consist of tax lists, state censuses, voter registration lists, city directories, and other lists of people from the years around 1890, as “replacement” censuses.
Certainly the loss of the 1890 federal census was devastating, as would the loss of any record group as universal and informative as the federal census. But why has so much effort been taken to “reconstruct” this census? (It should be noted that no effort to actually reconstruct the census as it existed has actually been made. So-called reconstructions are generally mere transcriptions or extracts of the information contained within the alternative source material.)
The federal census is useful because, after 1850, it (in theory) provides the names and other personal information concerning every resident of the United States. Created every ten years, one is able to follow most of their ancestors through decennial “snapshots” of their lives.
However, the census being taken every ten years is a fact of the record group, but not a benefit. In fact, it is a distinct disadvantage of the federal census. Most of the record groups used as “1890 census substitutes” were not created decennially. Some were created annually. Yet, because of the absence of this decennial record, only those records surrounding 1890 have been published and, in many cases, genealogists only consult these published 1890-era records.
Let us consider one common “substitute” record as an example. City directories were created annually in Baltimore, Maryland, since 1864–65, and biennially or sporadically for many years prior. Various editions are available online in several locations, including Ancestry, Fold3, Internet Archive, and the Archives of Maryland Online. The same is true for many other cities across the United States. Genealogist Miriam J. Robbins has created the “Online Historical Directories Website,” providing links to many of these city directories.
City directories, however, do not provide the same information as the 1890 federal census. In general, they provide names, occupations, and residential addresses, and occasionally the names of spouses or racial identity (through the use of separate sections for white and “colored” residents). They do not report ages, places of birth, immigration status, family relationships, or any of the other details common to the federal census.
Furthermore, city directories, tax lists, and other annually created record groups are often best used by looking at a large number of years across time, to accentuate changes in the local population. The correlation of this multi-year evidence can be an important source of indirect evidence concerning an individual’s identity or relationship. Looking only at 1890 is virtually useless.
These annual records are not limited by the decennial nature of the federal census, and we are only doing ourselves a disfavor by limiting our access and use of these records as a substitute for another record. Instead, we should take advantage of the benefits these annual records hold.