How maps created for fire insurers show the evolution of cities.
Sanborn maps are crowded with detail and color. So is their history.
Daniel A. Sanborn created these maps for one, very specific (and kind of dry) reason: to provide insurers a catalogue of city structures that could be fire risks. But over the years, these maps came to serve another purpose. Flipping through a series of maps of the same location, you can see mushrooming buildings, shops, and churches and deduce who lived, worked, and prayed in these structures. So, apart from insurance companies, historians, genealogists, and scholars started looking them them up for the moving pictures of urban growth that they offered.
"It was accidental in some respect," says Chris Genovese, general manager at Sanborn, which still offers mapping services with updated technology. When they started the maps, they had never imagined they would be of use to anyone else but their insurance-company patrons.
Before he started his company in 1866, Daniel Sanborn was commissioned by Aetna to map several cities in Tennessee. (That company had been impressed with his work on this atlas of Boston). When his own company took off, Sanborn expanded his mapping coverage across the entire U.S. The company's surveyors—more than 300 in the field at the peak of operations—infiltrated all the states. Across offices in New York, Chicago, and San Francisco, another 400 employees carefully worked to translate the data on linen-backed sheets. Using wax stencils, artists then colored the maps by hand.
In 1905, the company standardized the maps' presentation, releasing a 100-page manual with comprehensive instructions and symbol keys.
This standardization ensured that the artists rendered the information uniformly. They never signed their work, but they did add a bit of personal flair when they designed the title pages, Genovese says.
During the 1930s, Genovese says the company would revise whole maps when necessary to reflect changes in a given city, but it was a "major undertaking." Soon, the speed of cities' expansion outpaced the company's ability to make revisions. To be more efficient and cut costs, they started redoing just the portions of the maps that needed updating. They would then paste the updated patches onto the outdated sections of the old map.
Benjamin Grant, an urban-design and planning expert at SPUR, remembers using the maps in graduate school in the 1990s. He says he loves the way that the maps—cut, patched and color-coded by hand—physically bear the evidence of a city's evolution.
From one version to the next, you can deduce a lot. If the housing units increase in a certain region, get subdivided, or if the number of stories in a building increased—you can make see that the population in that region is increasing, says David Kaplan, author and professor of geography at Kent State University. Subtle changes like the types of churches or community centers in certain corners signal what sort of people are settled in a neighborhood. The materials that make up a building announce what kind of establishment it is: Is it affordable housing? Is it a fancy, upscale apartment complex?
This kind of evolution is evident in the two maps of Washington, D.C., below, one from 1888 and the other from 1913. Kaplan points out that there are more occupied blocks North of the Capitol in 1913. "Knowing the development of D.C. during this period, I would expect this to be the case." says Kaplan.
More zoomed-in maps (like the map of Dupont Circle below) reveal much more, particularly the number of stories and types of units. Compared to the bustling Dupont Circle of today, "the sparseness of structures here really stands out," Kaplan says of the 1903 map below.
Both historic leaps (like replacing stables with railroads) and slow crawls (like the sprouting of steel corporations and mining districts) are visible in different states around the U.S.
Demographic hypotheses drawn from the maps can then be confirmed or ruled out using other historical sources, Kaplan says. In his classes, he couples the maps with census data.
His colleagues Andrew Curtis and Jacqueline Curtis use Sanborn maps to look at how diseases move over time and space. For example, they looked at the death records for yellow fever cases in New Orleans and took down addresses from the maps. Using Geographic Information Systems, they then plotted clusters of the disease.
The sheer volume of information Sanborn maps contain make them useful over so many disciplines. But what about what the maps don't include?
This blurb from the University of Virginia Library about the 1920 Charlottesville Sanborn maps points out that there might have been a selection bias:
"... these maps were produced for insurers who did not particularly care about areas of town their offices would not cover. Therefore, while the maps depict most of Charlottesville's major business district and many of its residential areas, they do not show all of Charlottesville in 1920. Areas that were not heavily developed are not depicted. Local historians also will note that predominantly African-American communities such as Fifeville (on the northside of town between the university and downtown) are not shown on the maps, although that area of town was well built-up at the time."
Genovese says that the company was pretty comprehensive in detailing everything within the city's contiguous boundaries. But if there was, in fact, a selection bias, that in itself reveals a lot.