This site is outdated: I've updated the map and posted it here, along with newer projects I'm working on.    

Introduction and Context:
    Culture manifests in countless ways. Toponyms, the names given to places, features, or locations by the people that use them, carry hidden cultural information in their etymology, definition and usage. Generic toponyms, those used to describe a general type of feature such as lake or pond, tend to be more culturally stable; a change in a generic toponym indicates a larger process at work. Streams provide a good outlet for the study of toponym generics due to their ubiquity and importance to the cultures that use them. This project maps the distribution of stream generics in the United States, and attempts to draw some conclusions about their distributions. 
    There have been several academic studies conducted on generic stream toponyms in North America, from the scale of the state (Raup 1957, Miller 1969, Kuethe 1935) to that of the region (West 1954, Zelinsky 1955), to more recently that of the nation (Campbell 1991). This project attempts to build on the previous research by using an untapped source for toponymic data (the USGS National Hydrography Dataset, or NHD) and a unique mapping technique (the use of line data) to gain a fresh persepective on the distribution of stream toponyms in the contiguous US. Stream toponyms vary widely across America, in their etymological roots, the types of features they refer to, and the degree to which they are used in different ways from region to region. Toponyms are inherently cultural artifacts, but stream terms in particular vary due to both physical and cultural differences. This project will hopefully serve to clarify the patterns of these variations, as well as the processes driving them. These connections have been explored before, but advances in the availability and quality of data, combined with an increased ease of mapmaking, have allowed the distributions to “crystallize”, standing in sharper focus, and allowing more precise interpretation than has been possible in previous studies.
Data Sources 
    The data for this project come from the publicly available National Hydrography Dataset, or NHD. The NHD was developed by the USGS as part of the larger National Map, and is designed to provide a comprehensive database of surface-water features in the US for researchers using geographic information systems. (Simley and Carswell 2009) The data for this project were downloaded from the NHD servers at medium resolution (a scale of 1:100,000); SQL queries were then used to extract records only for streams which had toponyms attached to them. The toponymic portion of the NHD is derived from a preexisting project undertaken by the federal government: the Geographic Names Information System, or GNIS (USGS 2000). GNIS data were used by Campbell in his study of national variation in US hydronyms (Campbell 1991) – the advantage of the NHD data comes from the fact that each toponym is attached to the geographic line data of a specific stream, as opposed to the isolated Lat/Long points that are defined by the GNIS. Thus, using NHD data allows the geographical examination of toponyms as integral elements of the streams they describe, rather than as single points associated with the s tream at some point along its route (usually at the geographic center or at the mouth of the stream) as in the GNIS (USGS 1981). NHD data allows far greater flexibility in examining the physical path of a given stream, ultimately allowing more precise conclusions to be drawn. The data for base shapefiles, including hydrography, bathymetry, national and state borders were acquired from the National Atlas of the United States.
Map and Methods
    After downloading the national stream data from the NHD, ESRI ArcMap software was used to join each watershed segment into a single shapefile representing the entire conterminous United States. The data queries used to form categories for each stream generic followed the general format used in a previous study by Campbell (1991), searching full names for an ending in a chosen generic term. A few terms, such as rio and bayou, were searched for at the beginning and end of the full stream names to account for the lingering effects of French or Spanish word order. 
    The generic terms extracted from the NHD data and ultimately mapped were chosen according to the terms mapped in previous studies (such as Campbell 1991 and Zelinsky 1955), commonality across the United States, and to highlight significant local concentrations. The line data that resulted from queries for each stream term were mapped together, displaying more geographically precise distributions for each term than has been possible in previous studies using only point data; more specific maps were created using the same methods to emphasize the distributions of certain terms. The finished, merged maps were colored subjectively. The black base color of the map was chosen to allow the colors of even small streams to “pop” out of the background. The terms branch, prong, and fork were colored similarly to emphasize their shared definitions describing a side flow of a larger stream. Rio, arroyo, and cañada were colored similarly as well, emphasizing their Spanish etymology. Most other generic terms were colored to emphasize contrast, with the exception of creek and river: these terms represent universals, present across the United States, and so were desaturated to allow the viewer to more easily discern distributions of the less common generic terms.

    Most generic terms mapped for this project follow a fairly clustered pattern. River and creek are the most common endings for stream names in the US: these two terms provide the toponymic “substratum” for the rest of the study, and along with some less common terms such as slough, creek and river are found throughout almost the entire country. As one moves westward from the Eastern Seaboard, the variety of generic terms declines, with creeks and rivers dominating the toponymic landscape. One reason for this change may have been the historic standardization of the settlement process undertaken by the newly-minted United States: the decline of more unique generic terms coincides interestingly witha map (Figure 2) showing the extent of the Public Land Survey System (PLSS), adopted in 1785 to control settlement to outlying territories (Thrower 1966). The homogeneity of stream terms in the continental US may have been a direct result of the shift from naming as a subconscious, social act to naming as a more deliberate act by frontier explorers in the late 18th century. A map showing the extent of human settlement in 1830, shortly after the beginning of the PLSS, provides further evidence for this hypothesis (Figure 3).
    Political factors are just one set of variables that affect the generic terms given to streams. The name that a stream ends up with depends on an interaction of both cultural and physical influences, often with blurred lines between the two. While some generic terms may refer to specific types of streams, this is a flexible, purely cultural distinction. The distribution of the generic term bayou provides an example of a term that spread to encompass a range of stream types due to cultural forces. While bayou originated as a French bastardization of a Choctaw word for a slow-moving stream, Figure 4 shows how French settlers spread the word up the river system from its origins on the Gulf, slowly changing its meaning along the way (West 1954). The naming process is inherently a cultural act, but one guided by the character of one’s environment.
Cultural Influences:      
    The colonial settlement history of the United States defines much of the toponymic variation that can be seen in generic stream toponyms today. Colonists from European countries brought their names with them to the Americas; many of these names have remained despite successive changes in the ethnic and cultural composition of the people using them. Geographer Wilbur Zelinsky explains the persistence of toponyms and other cultural traits using a concept called the Doctrine of First Effective Settlement: “Whenever an empty territory undergoes settlement, or an earlier population is dislodged by invaders, the specific characteristics of the first group able to effect a viable, self-perpetuating society are of crucial significance for the later social and cultural geography of the area, no matter how tiny the initial band of settlers may have been” (Zelinsky 1973, 13). Using this doctrine as a guideline, many of the variations in stream generics can be explained by examining the history of settlement and colonization efforts in the US.
    Any cursory examination of stream names in the US reveals the overwhelming number that end in creek. Creek as a generic term is an import from England, carried with settlers into Tidewater Virginia during the early 17th century. As settlers explored the low-lying area, following the long arms of the Chesapeake Bay inland, they named streams as they had in England. According to George Stewart, “A smaller stream where the tide washed back and forth they called a creek (…) [f]ollowing up such a creek, they came also to a place where the water flowed always one way, and since it seemed the same, they kept the same name for it. Thus a creek came to mean a flowing stream, although in England it meant, and still means, a tidal channel” (Stewart 1945, 60). Combined with the Doctrine of First Effective Settlement, this may explain the failure of creek to gain a toponymic foothold in New England, where streams were named (generally as brooks) by English settlers without any opportunity to mistake a stream for a tidal channel. Kill is another New England generic, used almost exclusively in the Hudson River Valley area; the term is derived from the Dutch word kil, and the modern extent of its usage corresponds almost perfectly with the area of Dutch settlement in the mid-17th century (Zelinsky 1955). As the Dutch spread throughout the Valley, land grants given to individual farmers prompted a dispersed pattern of settlement; when the English took over the area, most of the Dutch farmers stuck around (Trewartha 1946). The Dutch were sufficiently entrenched in the region that the use of kill as a stream name stuck, ultimately being passed to a prominent physical feature in the area, the Catskill Mountains.
    The old Spanish colonial hearths of New Mexico and coastal southern California provide another example of cultural influences on generic stream toponyms. These areas represent the oldest branches of Spanish colonization efforts in North America (Figure 5), where missionaries and settlers from New Spain/Mexico dominated the indigenous peoples for almost 200 years. Although according to historian David Weber “[t]he number of Spainards in New Mexico in the 1600s probably never exceeded 3,000” (1992, 90), this initial core was apparently enough to imprint the landscape with persistent Spanish placenames: arroyo, rio, and cañada dominate this area. (Figure 6)

  Physiographic Influences:       

    Wash is another stream generic that is also common throughout the Spanish-influenced southwestern US, however, this term isn’t derived from a Spanish word. The predominance of wash, although certainly culturally influenced, owes its distribution to the physiology of the area. Wash labels an intermittent streambed that is dry most of the year – the term refers to the “washing” away of topsoil that occurs during rainy periods. When they are dry, these streambeds are commonly used as trails in the region (USACE). The largely homogenous clustering of wash in Arizona, Nevada, and New Mexico provide an example of a term distribution based almost entirely on physical factors, although the apparent dropoff in usage at the Utah border perhaps raises questions about cultural influences, as well.

    The size and speed of a stream are the two physical variables that most obviously affect the name given to it. Put simply, a primary purpose of naming is to “distinguish one place from other places” (Stewart 1945, 6); generic terms are especially bound to the type of feature being named, often stating “what” a feature is while specific terms describe “which”. Rivers are larger than creeks; larger streams have different names than smaller ones. Describing the stream names of Maryland, one scholar observes the spatial progression in generics: “it would seem that run is used in the mountains for the action, the speed, which it denotes; branch is used in the plateau because of the greater number of divisions or ‘branches’ in the streams of that section; and the term creek predominates in the coastal area because of the influence of the original meaning of the word, an inlet or tidal estuary” (Kuethe 1935, 269). A similar physiological hierarchy of stream names is especially recognizable in the distribution of the term swamp in tidewater Virginia, North Carolina, and South Carolina (Figure 8). Swamp is used in this sense to refer to slow-moving streams at low elevations; comparison with a map of the Fall Line (Figure 7), the border between the rolling Applachian Piedmont to the gently sloping Coastal Plain, shows that the distribution of swamp almost perfectly matches changes in the stream morphologies.


    Toponyms have been intensively studied by geographers, historians, and linguists for centuries, resulting in a large collection of literature related to the subject. Most of the toponymic research uncovered over the course of this project seemed to be decades old: intensive toponymic study is less common now than in the first half of the 20th century. This project shows the value of applying newer mapping and data management technologies to the field of toponym study. There are still downsides, of course: the NHD data are inherently limited in scope. Folk names, for example, aren't encompassed by this study; only official names given by the US Board on Geographic Names appear on the map. The NHD is also simply unfinished, part of an ongoing USGS project pushing towards increasing accuracy with each new revision.  However, there is value in using the NHD data for other toponym research, such as on the occurance of certain specific toponyms, the etymology of stream names, or the occurance of compound generics such as Bushkill Creek. Finally, the methodology used for this research can be expanded as similar data become available for other countries.  One example that is already online is Canada's National Hydro Network, a dataset almost identical to the NHD, ripe for toponymic research.


Campbell, Jon C. 1991. “Stream Generic Terms as Indicators of Historical Settlement Patterns.” Names, 39(4): 333-365.
Kuethe, J. Louis. 1935. “Runs, Creeks, and Branches in Maryland.” American Speech, 10(4): 256-259.
Miller, E. Joan Wilson. 1969. “The naming of the Land in the Arkansas Ozarks: A Study in Culture Processes.” Annals of the Association of American Geographers, 59(2): 240-251.
Raup, H. F. 1957. “The Names of Ohio’s Streams.” Names, 5: 162-168.
Simley, J.D. and W.J. Carswell Jr. 2009. The National Map - Hydrography: U.S. Geological Survey Fact Sheet 2009-3054. (Accessed November 18, 2009).
Stewart, George R. 1945. Names on the Land. New York: Random House. (Accessed November 18, 2009).
U.S. Geological Survey. 1981. U.S. Geographic Names Information System (GNIS). U.S. Geological Survey, Reston, VA.
U.S. Army Corps of Engineers. Desert Pattern/Feature Summaries. Army Geospatial Center. (Accessed November 18, 2009).
Weber, David J. 1992. The Spanish Frontier in North America. New Haven: Yale University Press.
West, Robert C. 1954. “The Term ‘Bayou’ in the United States.” Annals of the Association of American Geographers, 44(1): 63-74.
Zelinsky, Wilbur. 1955. “Some Problems in the Distribution of Generic Terms in the Place-Names of the Northeastern United States.” Annals of the Association of American Geographers, 45(4): 319-349.
–––––. 1973. The Cultural Geography of the United States. Englewood Cliffs: Prentice-Hall.

Image Sources
Figure 1: Generic Toponyms for US Streams. Self-made.
Figure 2: US Bureau of Land Management. 1988. "Principal Meridians and Base Lines." Acquired from US BLM website. (
Figure 3: USGS. 1970. "Exploration and Settlement: 1820-1835." The National Atlas, 137. Acquired from the online Perry-Castadeña Library Map Collection, UT Austin. (
Figure 4: From West (1954), page 67, fig. 3: "Types of Bayous."
Figure 5: From Weber (1992), page 66, Map 2: "The Spanish frontier, circa 1550-1600."
Figure 6: Swamps. Self-made.
Figure 7: USGS. 1999. Illustration from Developing Landscape-Indicator Models for Pesticides and Nutrients in Streams of the Mid-Atlantic Coastal Plain, online publication FS-157-00. Acquired from .
Figure 8: Spanish generics. Self-made.
All original maps and text by Derek Watkins (derek.watkins at
Project originally completed for a Cultural Geography class taught by Joby Bass.