artment there is a location pattern of individual operations and pieces of equipment, such as ial disposition of the units in question must be considered: industries, d a ations of actual or desirable locations at differ e t l te m s methods used. Thus, it is necessary to eir enticements ne outlets in a the store, or ome location re outlets)are arried a one. Abou te r4n hr d so on. For iaeym a pe at car be considere arage, or fraternity inquire into its location noted later, we can specity pre e of desirability alternative consider in ever ery rtia-even on costs are irm. may be a a change aken, as a nt location) aisal of site are relativ se scrutiny of new site atl gencrata who g 0 le locations for many kinds of n very much the same e garages, the problem is garages, storing and servicing the buses, and getting them to and fi The correspondence between public and private decisions is less close where the product is not marketed with an eye4 within one building; within one department there is a location pattern of individual operations and pieces of equipment, such as punch presses, desks, or wastebaskets. At each of the levels indicated, the spatial disposition of the units in question must be considered: industries, plants, buildings, departments, wastebaskets, or whatever. Although determinations of actual or desirable locations at different levels share some elements,1 there are substantial differences in the principles involved and the methods used. Thus, it is necessary to specify the level to which one is referring. We shall start with a microscopic but not ultramicroscopic view, ignoring for the most part (despite their enticements in the way of immediacy, practicality, and amenability to some highly sophisticated lines of spatial analysis) such issues as the disposition of departments or equipment within a business establishment or ski lifts on a mountainside or electric outlets in a house. Our smallest location units will be defined at the level of the individual dwelling unit, the farm, the factory, the store, or other business establishment, and so on. These units are of three broad types: residential, business, and public. Some location units can make independent choices and are their own "decision units"; others (such as branch offices or chain store outlets) are located by external decision. Many individual persons represent separate residential units by virtue of their status as self-supporting unmarried adults; but a considerably larger number do not. In the United States in 1980, only about one person in twelve lived alone. About 44 percent of the population were living in couples (mostly married); nearly 30 percent were dependent children under eighteen; and a substantial fraction of the remainder were aged, invalid, or otherwise dependent members of family households, or were locationally constrained as members of the armed forces, inmates of institutions, members of monastic orders, and so on. For these types of people, the residential location unit is a group of persons. In the business world, the firm is the unit that makes locational decisions (the location decision unit), but the "establishment" (plant, store, bank branch, motel, theater, warehouse, and the like) is the unit that is located. Further, the great majority of such establishments are the only ones that their firms operate. In general, a business location unit defined in this way has a specific site; but in some cases, the unit's actual operations can cover a considerable and even a fluctuating area. Thus, construction and service businesses have fixed headquarters, but their workers range sometimes far afield in the course of their duties; and the "location" of a transportation company is a network of routes rather than a point. Nonprofit, institutional, social, and public-service units likewise have to be located. Though the decision may be made by a person or office in charge of units in many locations, the relevant locational unit for our purposes is the smallest one that can be considered by itself: for example, a church, a branch post office, a college campus, a police station, a municipal garage, or a fraternity house. 2.2 OBJECTIVES AND PROCEDURES FOR LOCATION CHOICE Let us now take a locational unit—a single-establishment business firm, as a starting point—and inquire into its location preferences. First, what constitutes a "good" location? Subject to some important qualifications to be noted later, we can specify profits, in the sense of rate of return on the owners' investment of their capital and effort, as a measure of desirability of alternative sites. We must recognize, however, that this signifies not just next week's profits but the expected return over a considerable future period, since a location choice represents a commitment to a site with costs and risks involved in every change of location. Thus, the prospective growth and dependability of returns are always relevant aspects of the evaluation. Because it costs something to move or even to consider moving, business locations display a good deal of inertia—even if some other location promises a higher return, the apparent advantage may disappear as soon as the relocation costs are considered. Actual decisions to adopt a new location, then, are likely to occur mainly at certain junctures in the life of a firm. One such juncture is, of course, birth—when the initial location must be determined. But at some later time, the growth of a business may call for a major expansion of capacity, or a new process or line of output may be introduced, or there may be a major shift in the location of customers or suppliers, or a major change in transport rates. The important point is that a change in location is rarely just that; it is normally associated with a change in scale of operations, production processes, composition of output, markets, sources of supply, transport requirements, or perhaps a combination of many such changes.2 It is quite clear that making even a reasonably adequate evaluation of the relative advantages of all possible alternative locations is a task beyond the resources of most small and medium-sized business firms. Such an evaluation is undertaken, as a rule, only under severe pressure of circumstances (a strong presumption that something is wrong with the present location), and various shortcuts and external aids are used. Perhaps the closest approach to continuous scientific appraisal of site advantages is to be found in some of the large retail chains. Profit margins are thin and competition intense; the financial and research resources of the firm are very large relative to the size of the individual store; and the stores themselves are relatively standardized, built on leased land, and easy to move. All these conditions favor a continuous close scrutiny of new site opportunities and the application of sophisticated techniques to evaluate locations. Still more elaborate analysis is used as a basis for new location or relocation decisions by large corporations operating giant establishments, such as steel mills. These decisions, however, are few and far between, and involve in general a whole series of reallocations and adjustments of activities at other facilities of the same firm. Within the limitations mentioned above we might characterize business firms as searching for the "best" locations for their establishments. This calls for comparison of the prospective revenues and costs at different locations. What has been said about the choice of location for the business establishment will also apply in essence to many kinds of public facilities. Thus a municipal bus system will (or, one might argue, should) locate its bus garages on very much the same basis as would private bus systems. Since the system's revenues do not depend on the location of the garages, the problem is essentially that of minimizing the costs of building and maintaining the garages, storing and servicing the buses, and getting them to and from their routes. The correspondence between public and private decisions is less close where the product is not marketed with an eye