The Pacific Coast Nearest Larger Neighbor Gradient

The Pacific Coast Nearest Larger Neighbor Gradient Harold M. Elliott* In traditional central place studies it is often necessary to group urban places into discrete classes before attempting to identify hierarchical relationships (Christaller 1966, p. 158; Berry 1967, p. 27; Brush 1953, p. 393; Kenyon 1967, p. 741). These classes are usually given names, such as hamlet, village, town, and city, based on their populations or some other criteria. Frequently the end result is the designation of urban areas as central places of the first, second, and third order. After classifying cities, many investigators have grouped them into hexagonal networks extending at regular intervals in all directions. Much effort has gone into finding empirical evidence to support the existence of these hexagonal hinterlands. When such evidence is found, the networks are used to test theoretical systems like the Christaller and Lösch models. In most cases the correlation between these theoretical models and actual landscapes has been low. A major reason for this lack of fit has been the necessity of first classifying cities into higher and lower orders. Although in some areas like southern Germany and southern Ohio the fit is fairly good, most such exercises in spatial taxonomy yield inconclusive results. In his study of central places in Georgia, for example, Kenyon (1967) found little evidence of discrete size groupings among places. *Dr. Elliott is an associate professor of geography, Weber State College. This is a revised version of a paper presented at the annual conference in Moscow, Idaho, Tune 1981. 30ASSOCIATION OF PACIFIC COAST GEOGRAPHERS Another problem involves the process of change. As has often been observed, central place systems are not static or fixed. Instead, they are subject to change and development (Ullman 1941, p. 861; Huff and Lutz 1979, p. 211; Murphy 1974, p. 44). In the United States the urban hierarchy is a thing in motion. It is evolving from something toward something. Hierarchial relationships have changed radically in many parts of the country even though ranksize relationships have remained fairly constant. Correspondences between theoretical and actual patterns of central places may not mean very much if the actual patterns were quite different several years earlier. To use an example from the Christaller model, if what was originally a K = 4 hierarchy is changing toward a K = 3 hierarchy or something entirely different, it should not be surprising that an experiment designed to test the validity of Christaller's scheme during any given year produces negative results. The problem is compounded when one recognizes that the transition might be operating at different rates in different parts of the hierarchy. Furthermore, if there is no way of measuring an actual hierarchy through time, then it is not possible to test any dynamic conception of central place evolution. The Nearest Larger Neighbor Method One solution to such problems of measurement is the Nearest Larger Neighbor method, which is based on two important generalizations from range and threshold studies and from studies of economic hinterlands. These are: (1) higher-order functions are concentrated in successively higher-order centers; and (2) a greater range of goods and services can be found in larger central places than in smaller ones (Berry and Garrison 1958, p. 145; Berry 1967, p. 13; Murphy 1974, p. 58; Eliot Hurst 1972, p. 204; Stafford 1963, p. 171). By treating these two findings as axioms, the question is then asked "Where does one go if something is not available locally?" The answer, in most cases, must be the Nearest Larger Neighbor (NLN). An individual may either travel to the NLN in person and search for the locally unavailable higher-order product or he may conduct YEARBOOK · VOLUME 44 · 1982 31 the search by telephone or mail. He also may have the higher-order product (or service) transported to him from the larger place via delivery truck, mail, telephone, or television. Although there is some disagreement (Eliot Hurst 1972, p. 207; Golledge et al. 1966, p. 271), many investigators hold that consumers tend to patronize the closest centers offering the goods and services they need (Berry 1967, p. 10; Bunge 1966, pp. 151, 159). Keep in mind, however, that this tendency is best stated in...