The Poor on the Hilltops? The Vertical Fringe of a Late Nineteenth-Century American City

Abstract

Features of the physical urban site merit more attention than they have traditionally received in models of city form, but in bestowing it the interrelation of social and natural features must be recognized and a neoenvironmental determinism avoided that would see the roles played by site features as always and everywhere the same. In American cities today, the affluence of residents, as a rule, increases with elevation. Yet in the “walking city” of the nineteenth century and earlier, high land's difficulty of access might have outweighed its attractions and made it the home of the poor and not the rich. The possibility is investigated through a study of upland residential patterns in Worcester, Massachusetts, in 1891, just before the city's first electric trolley line was installed. Though a simple inversion of today's pattern did not appear, working-class residents indeed predominated on the highest land. They shared it with pockets of upper-class estates and with other land uses—such as parks, large residential institutions, and extractive and nuisance industries—typically associated with the premodern horizontal urban fringe and apparently drawn to the vertical fringe as well by the cheapness of land.

Key Words:

• site
• situation
• terrain
• urban geography
• Worcester

Acknowledgments

I am grateful to the section editor and the anonymous reviewers for their helpful criticisms and comments, and to Anne Gibson for producing the maps.

Notes

1. James (1967, 6) deplored the “common misuse” of the word “topography”—which denotes place or areal description of all sorts—to mean merely patterns of elevation. Such available terms as “relief, landforms, terrain,” he pointed out, amply capture the narrower meaning, if it is intended. The point is still worth insisting upon.

2. The term “preindustrial city' was given wide circulation by Sjöberg (1960), but his focus was on cities where an elite exercising an authority at once cultural, religious, and political resided in the center and those of lower status on the margins. Abbott (1974, 37–38) suggested that such cities are better termed “premodern” than “preindustrial.” Commercial but preindustrial cities not dominated by such an elite, he observed, tended to display the same overall residential pattern. The shift in residential patterns that occurred in many European and American cities during the early and mid-nineteenth century will be termed here a change from a premodern pattern in Abbott's sense to a modern one.

3. City directories have been criticized as sources of data on residential patterns for their presumed incomplete coverage and bias in favor of the well-to-do and the long-settled. Researchers have found them useful nonetheless, and on detailed comparison they have not always turned out to be inferior even to the chief alternative, the manuscript schedules of the decennial federal census (Thernstrom and Knights 1970, 13; Kellogg 1982, 26). In any case, the destruction by fire of the 1890 census schedules leaves only the directories as records for Worcester population patterns at the end of the pre-trolley era, ca. 1891. Moreover, as in a similar study (Kellogg 1982), if poor residents of the neighborhoods in question (here, the Worcester uplands) were indeed undercounted in the directories, the conclusions reached as to their substantial presence there would only be strengthened.

4. Jordan and Rowntree (1990, 392) offer a different explanation, based solely on construction costs (higher on hilly than on level land) rather than transportation costs and/or environmental amenities, for the tendency of hills to house wealthy or poor city dwellers more commonly than middle-income ones. Either, they write, “the increased cost of building may be passed on to the consumer—meaning that those who buy the houses pay more, and the area will be occupied by higher-income groups,” or cheap houses of poor quality on small lots may be built to cut the excessive costs of construction, which “means that lower-income groups will probably occupy the area.” Yet in each case, the houses built would be competing in the market with similar ones of equal quality built at lower cost on level land. The higher cost of building could not be passed on to the buyer, and the only result is that the owners of building lots on easier terrain would reap an economic rent. Increasing or cutting costs for houses built on hills would simply raise or lower the class of houses with which they would compete. As exactly the same would happen with houses suitable for middle-income (or any other) groups, it does not follow that upper- and lower-income groups would be more than ordinarily likely to live on hills.

5. In many other nineteenth-century American cities, inadequate pressure in the city water system added to the difficulties of living on high ground (or even on the upper floors of buildings) (Blake 1956, 269, 277), but not in Worcester in 1891. The city had years before established a “high-service” component to its water system, using the uppermost of its two supply reservoirs, fully capable of meeting the needs of the most elevated neighborhoods (Sanborn-Perris Map Company 1892, 1–2).