back to back to Muller back to Anthropology

Please Note: This paper is posted for private viewing as a work in progress, and copyright is reserved by the author

Paper presented at the Southeastern Archaeological Conference Annual Meeting in Birmingham, Alabama, on 7 November 1996

The Lower Ohio Valley and its Frontiers

Jon Muller


The Lower Ohio Valley was a frontier zone between the Southeast and the Midwest in Mississippian times. While much of its population was usually concentrated in fertile river bottoms, research in upland environments has revealed that Mississippian settlements occur in these localities as well. While some shifts in settlement through time were related more-or-less directly to population size and density, there were also political and economic contradictions that parallel those in other localities. Comparisons will also be made to dispersal and aggregation among historic Southeastern peoples.


This paper is concerned with the interpretation of settlement patterns over space and through time. The paper consists of four parts. The first is a brief discussion of dispersion that redefines some of the terms employed following a broader development of these themes in another paper (Muller in preparation). The opposition of concentration and scattering of population that is often made is sometimes true, but masks deeper unity of these opposites. Either scattering or concentration&emdash;and the resulting patterns of site distributions in localities and regions&emdash;may reflect similar causes that are highly dependent upon situational variables.

The second part of this paper gives a few examples from historical data in the Southeast. These illustrate that "cycling"-like phenomena are not only characteristic of prehistoric Mississippian (as per Anderson 1994a, b), but also characterized the structures of historical societies. Historical population data&emdash;often interpreted in mostly biological terms&emdash;are also explainable in ecological and political terms. These historical cases suggest possibilities for understanding how Southeastern societies may have functioned; but I do not mean, thereby, that we can use them to "interpret" prehistoric cases.

The third section of the paper is a synopsis of what we know about Lower Ohio Valley variation in settlement form through time. Recent work has now made it possible to approach, even if somewhat tentatively, the issues of what happened within different phases as well as among the coarser periods usually characterized somewhat monolithically as "Late Woodland" or "Mississippian." For example, a recent paper by Berle Clay (in press) has discussed fairly fine-grain events as representing the rise and fall of autonomous polities in this region.

The final section of the paper presents some conclusions about the state of affairs in this region, but also goes beyond the existing data to suggest possible interpretations of how Mississippian utilization of the Lower Ohio Valley may have altered through time, reflecting a series of demographic, economic, and political factors. At first glance, this paper may seem to be a refutation of ecological definitions of Mississippian that have been presented (e.g., Muller 1978; Muller and Stephens 1991; Smith 1978). That conclusion, however, is mistaken since my purpose here is not to reject economic definitions of Mississippian, but rather to extend them in a political economic synthesis.

Dispersion: Aggregation and Dispersal

For convenience, I make a distinction between dispersion and dispersal . I shall use the term dispersion for a general measure of location in relation to resources, including both aggregation and dispersal as different kinds of dispersion. Dispersal, will be used here for the specific state of not being concentrated. Aggregation is the dispersion state of concentration. I develop the reasons for making these distinctions more fully in another paper (Muller in preparation), but I beg the indulgence of the reader.

Following Classical models, it has been stated that dispersal in the East was for easy access to farmlands and aggregation was a defensive response. An explicit expression of this view is Adair's description of the Choctaw:

Their country is pretty much in the form of an oblong square. The barrier towns, which are next to the Muskohge and Chikkasah countries, are compactly settled for social defense, according to the general method of other savage nations; but the rest, both in the center, and toward the Missisippi, are only scattered plantations, as best suits a separate easy way of living. A stranger might be in the middle of one of their populous towns, without seeing half a dozen of their houses, in the direct course of his path. (Adair 1775:281).

However, it is important to recognize that dispersal or aggregation are not so directly related either to defense or to access to farmland as assumed by Adair and others. In either case, there is an underlying strategic condition that makes dispersion a dialectical phenomenon&emdash;reflecting the unity of opposites. Settlement location, dispersion, is organized for easy access to critical resources at the time of settlement. If the critical resource at a particular time was access to scattered arable lands, then settlement might be dispersed for that reason. On the other hand, if some critical material resource were concentrated in limited locations, then settlement would have been aggregated for exactly the same underlying reason that it would have been dispersed in the first example. Aggregation for the purpose of defense is aggregation in relation to the resources for defense&emdash;the warriors that were important at that time and place. People and their social relations may be resources just as much as are more "material" things. We do not need to be idealists or "realists" to appreciate that human beings and their social relations are critical parts of the environment. Just as concentration on physical resources can lead to either dispersal or aggregation, the nature of the threat and the character of defensive resources can mean that the defense can lead to either aggregation or dispersal. For these and other reasons, it is best for archaeological interpretation to avoid a false dichotomy between defensive aggregation and peaceful dispersal. It is important that we understand and reassess interpretations of dispersion in this light. Major differences in responses to problems also exist according to the form of property that exists in a given social formation. For example, European colonists and American Indians responded differently to similar threats at least partly because of private real property in the first instance and community allocation of usufruct in the second. Each of these world views encompasses different values for personal control and use of land.

In general, settlement tends to fall either into regular (also called "repellent" or "uniform") or aggregated ("attractive") forms, but settlement may also be randomly distributed, reflecting independence of settlement episodes or "neutral" relationships. In practice, random distribution seems rarer than either hostile or attractive relationships, and one would expect it especially in the context of overabundance of dispersed resources. In some ways, Late Woodland times in the Lower Ohio may have seen a kind of superabundance and upland Late Woodland settlements seem random in distribution overall (but not so much so in bottomlands, see below). Mississippian on the other hand, was more often either uniform or concentrated. This suggests that competition or cooperation tended to be characteristic of the relations of adjacent Mississippian communities. The characterization of relationship, of course, applies only to the unit scale being measured, not necessarily to regional social and economic relations. That is to say, one should be cautious not to overinterpret "attraction" between homesteads as indicating regional cooperation. Such patterns are typical of complex circumstances in which the application of a relatively small number of principles may result in complex, bifurcated distributions of outcomes.

One must also define the scale on which either aggregation or dispersal is measured. However, the size of the sample quadrat and the nature of the sample can substantially affect the most common measures of dispersion such as comparisons of variance and mean, as demonstrated by Pielou (1977:136). Pinder, Shimada, and Gregory (1979) have noted the same problem for the archaeological application of nearest neighbor statistics. Nevertheless, there can be value in assessing degrees of dispersion so long as the context is properly specified. The need to define scale should not be taken as a cautionary tale against trying to interpret dispersion in the past. Mississippian societies had important elements of their organization organized at household levels (e.g., Mehrer 1995; Muller 1978, 1986, in press b; Rogers and Smith 1995). I think that many features of broader Mississippian settlement, at the very least on the local level, "emerge" from the strategies of dispersion of household productive units. Discussion of emergence of complex systems from simple principles may be found in the literature on complexity (e.g., Thompson and Stewart 1987; Waldrop 1992).

Historic Examples of Dispersion

The examples of historical dispersion given here are not presented as a direct- historical interpretation of Mississippian in the Lower Ohio, but rather as illustrations of the settlement practices that one might expect if Mississippian and Historic groups were organized in similar ways. While I think that a good argument can be made for similarities and connections (see Muller in press b), that is not my intention here. Rather, I am seeking merely to warrant discussion of a dialectical interpretation of dispersion within the context of societies at the chiefdom level. It should be noted that models suggesting "hegemonistic" structures in Mississippian societies (e.g., Pauketat 1994) or employing "core-periphery" models (e.g., Dincauze and Hassenstab 1989) give quite different views of Mississippian society than that taken here. Of course, one implication of interpreting Mississippian political power in such terms is that it implies major organizational discontinuities between historical Eastern societies and their ancestors. While this is by no means impossible (see Galloway 1995 for a reasoned presentation of the case), I think the case for such a gap has been more often asserted than argued.

Figure 1. Cherokee Total Population Fluctuation through Time.


Historical groups illustrate bifurcation of dispersion quite clearly (see also Snow 1995). Figure 1 shows the fluctuation in reported Cherokee population through time (also see Thornton 1990). Both Eastern and Western Cherokee are combined in this chart; but, if they are separated (as in Figure 2), the Eastern Cherokee numbers decline very rapidly at the time of removal and the total population numbers and the Western Cherokee numbers converge. This illustrates that decline in one location is often followed by increased numbers elsewhere. It should be clear that historical population reports can be used to estimate actual mortality from disease or warfare only with great caution. Many fluctuations reflect the character of dispersion at a given time rather than actual deaths and births. The historical population estimates from which these data are derived have many biases, but it is a mistake to assume from such oscillations that the data are necessarily inaccurate. On the contrary, we need to recognize the social nature of these records and abandon the idea that these are documenting purely biological changes. When we do so, many of the reasons to reject the reports as inaccurate disappear. In many cases, the data simply indicate states of dispersion of one kind or another. In some cases, losses in population are emigration from one center of concentration to another, beyond the ken and power of a colonial enumerator. In other cases, it seems more likely that actual dispersal into small, mostly unenumerated groups occurred, to be followed by reaggregation as favorable conditions for aggregation were reestablished. Epidemic disease did often cause direct and substantial losses in population, but examination of even these cases often reveals "recoveries" in populations as a result of joining together with survivors from other groups (see Galloway 1995; Muller in press b; M. Smith 1987). Most of the historic "tribes" and "confederacies" of the East display diverse multiethnic origins in their recent pasts, and I think that the internal dynamics of prehistoric populations may have presented similar, if less drastic, circumstances (Muller in press b).

Figure 2. Eastern and Western Cherokee Population.


Some historical episodes of alteration in settlement form reflect actual expulsions and displacements by aggressive external forces, as in the Eastern and Western Cherokee. These are too well documented in Euroamerican relations with Indian populations to need more than a reminder. I can see little evidence for similar asymmetries of power in prehistoric populations, but there are traditional (and some archaeological) indications of replacement of one group by another, as in the upper Tennessee drainage in protohistoric times where the Creek were replaced by the Cherokee.

Figure 3. Total Reported Creek Confederacy Population.


The Creek are another illustration of how the conglomerate and the individual segments of the Eastern societies may display similar patterns. Figure 3 illustrates the aggregate Creek Confederacy data. The smoothed line averages reported figures if a data point was within 5 years of the previous report. As in the Cherokee population data, this figure illustrates the nearly universal tendency for Eastern population estimates to show considerable fluctuations over relatively short time periods, with rapid declines and rapid "recoveries."


Figure 4. Variation in Some Creek Confederacy Subgroups.


Figure 4 illustrates a series of finer grained data for different Creek Confederacy subgroups and related groups. The purpose of this figure is to illustrate that the populations of individual "Creek" towns and those of the Creek chiefdom as a whole did not necessarily co-vary, although they often shared many common historical occasions for decline, increase, aggregation, dispersal, immigration, and emigration. As in all of the historical data, some variations are probably from reporting error, but it would be a serious mistake to dismiss all of these data merely because of these rapid fluctuations without seeking to understand whether the data may reflect underlying conditions of dispersion. I would suggest that these fluctuations are one of our best "symptoms" for identifying the social responses of these peoples, as opposed to what the largely European sources claim for "their" Indians. While fluctuations in archaeological cases cannot usually be defined to the same precision, I suggest that political economic factors played similar roles in prehistoric population and political bifurcations of dispersion.

There are still other kinds of occasions and causes for dispersion. Historical records attest gaps or "no-man's lands" between settlements, right back to our earliest records for the interior (e.g., de Biedma 1544). This remains a valuable tool for identifying some kinds of political boundaries, if it is used cautiously and with full attention to other reasons for apparent "gaps" in the archaeological data.

Some settlements were described as being dispersed, but also dense (de Biedma 1544 [1993:241]). Other historical records describe settlement as consisting of multiple towns and individual settlements scattered along river valleys (e.g., Hawkins 1799). The interplay between autonomy and interdependence is commented on explicitly in the historical accounts. Sometimes these conditions are related to political circumstances. Beverley, for example noted:

The method of the Indian settlements is altogether by Cohabitation, in Townships, from fifty to five hundred Families in a Town, and each of these Towns is commonly a Kingdom. Sometimes one King has the command of several of these Towns, when they happen to be united in his Hands by Descent or Conquest; but in such cases there is always a Viceregent appointed in the dependent Town, who is at once Governor, Judge, Chancellor, and has the same Power and Authority which the King himself has in the Town where he resides. This Viceroy is obliged to pay to his Principal some small Tribute, as an acknowledgment of his submission, as likewise to follow him to his Wars whenever he is requir'd. (Beverley 1705 [1968:174])

The duplication of ranks such as these and the general parallels of functions among local and regional authorities is certainly one of the factors involved in the "cycling" phenomenon (e.g., Anderson 1990, 1994a, b). "Twinkling on and off" of different centers can reflect both the lack of strong authority on the regional level and the potential competition for authority by every "town king" (see also Hally 1993).

I emphasize political causes in the preceding paragraphs because of a tendency to try to explain much of the reported population variations in the historical East in biological, especially disease, terms (e.g., Ramenofsky 1987, but also see 1990; Thornton 1987). Such political competition, of course, is also grounded in particular economic and biological conditions at any given time. These economic and biological factors are as important as political ones. I think that a purely political explanation is just as incomplete as a purely biological one. Not the least important non-political factor is the possibility that soil exhaustion was often more important in the prehistoric and historic East than we have sometimes felt, given settlement on rich alluvial soils (see Baden 1995; but also a response in Muller in press b).

Settlement in the Lower Ohio

Settlement in the Lower Ohio Valley has been studied from many perspectives. The work of the Lower Ohio Valley Project began from a settlement pattern and ecological emphasis (Muller 1986). Other studies attempted to integrate ecological and locational principles (especially Clay 1976, but see Clay in press also).


Figure 5. Numbers of Mississippian Radiocarbon Dates in each 50-year Period for the Lower Ohio Valley.


The number of radiocarbon dates per 50-year period from the Lower Ohio Valley show oscillations (Figure 5; also see Muller 1993a). There are, however, far from enough data to treat the frequency of radiocarbon dates as necessarily indicating temporal frequency of settlement; but this at least serves as a reminder of the need to consider the problem. It is certainly the case the there was a peak in Mississippian settlement in the second half of the 13th century, but there I would guess that earlier settlement is underrepresented. It is also possible that what we are seeing here is actually the distribution of dates around the roughly A.D. 1250 mean. Clay's recent reassessment of the archaeological sequence on the Lower Ohio is a worthy effort to place such "cycles" into a regional perspective (in press) and while I would argue with individual points made in that paper, it presents an working framework for discussions of dispersion in this region. In all, there are indications of variation in settlement through time, although recent dates such as those on Dillow's Ridge, Dogtooth Bend and Millstone Bluff (Butler personal communication; Cobb 1995, personal communication; Stephens 1996) suggest that it is going to be very difficult to see clear chronological markers in the ordinary ceramics of this region.

Late Woodland and Early Mississippian

Lewis phase settlement in the Black Bottom (Figure 6) is characterized by relatively uniform settlement, but with a high concentration of sites in the immediate vicinity that later became the Kincaid site. That area and the adjacent ridges are relatively more fertile and free of flood than are other areas in the Black Bottom. This shows land utilization that foreshadows the later Mississippian pattern, but was balanced by a high degree of other sites away from the bottomlands proper, shown by settlement on the terraces (MacNeish 1944; Rudolph 1981). In addition to the sites shown for this locality, of course, there are a huge number of upland sites, including rock shelters, open sites, and even a series of hilltop stone "forts" (Muller 1986). At least some of the sites shown in the figure may date to fairly late in the period, however, so it is difficult to sort out the chronology of different parts of the Late Woodland period ranging from roughly A.D. 500 to 800. Unfortunately, we have no coverage for upland areas that is quite comparable, although transect survey in upland stream valleys (Canouts et al. 1984) and older survey data all indicate that upland Late Woodland is among the most frequent occupations detected in systematic survey, second only to Middle Archaic in this region. The Black Bottom pattern of common location of Mississippian and Late Woodland sites on the same locations is typical of the entire Lower Ohio (see Butler et al. 1979). Within the Black Bottom, some Late Woodland settlement is fairly uniformly distributed, but there was already substantial aggregation in the locality that later became the Kincaid mound center in Mississippian times. Outside the Black Bottom, there seems to be near random distribution of Lewis components, although detailed study of these distributions has not yet been undertaken. Surveys such as that of the Smithland Pool (Butler et al. 1978) indicate little reason to see local divisions in Lewis components throughout the region.

Figure 6. Late Woodland Settlement in the Black Bottom.


The situation is a more complicated for the "terminal Late Woodland" or "Emergent Mississippian" phases in the region. In the Kincaid locality, this period is represented by the Douglas phase (MacNeish n.d., 1944; Muller 1986, treated as "Lewis" in Cole et al. 1951), also marking the beginning of platform mound construction at the Kincaid site (see Muller 1986). The distribution of Douglas phase components in the rest of the Black Bottom corresponds very closely to that of later Mississippian phases, excepting only that the horizontal spread of the Douglas components seems to be slightly wider than in later times&emdash;probably related to the increasing use of maize after A.D. 900 and selection of residential sites on the very most fertile soils. The introduction of shell-tempering in the ceramics occurred at roughly the same time. It is difficult to sort out Douglas components from Mississippian phases because early Douglas sherds are essentially Mississippian without the shell temper, and in a locality where most of the shell has leached away, it can be hard to tell fine shell-tempered sherds from those that never had shell temper at all. The Douglas components at ca. A.D. 800, like the later Kincaid (a.k.a. Angelly) components, were scattered across the best soils in the bottomland, but do not occur on still-fertile, but lower rank soils in the floodplain. If the boundaries of the sample units are defined in terms of the top ranking soils for maize production, then settlement within the zone is generally uniform. If, however, the bottomland as a whole is considered, then there is a high degree of aggregation, probably resulting from selection of favorable soils on relatively higher ground.

Upstream, Douglas phase is distinguished from the similarly "Emergent Mississippian" complex known as Duffy (Winters 1967; Muller 1986) primarily by the presence in the latter phase of a distinctive "bar stamp" ceramic decoration. It is notable, however, that the Duffy complex is identified in the area near the major salt sources close to the confluence of the Wabash with the Ohio (Butler et al. 1979). The major difference between the Duffy phase and Yankeetown still higher on the Ohio is that the latter phase has some vessels with an indented fillet decoration (Blasingham 1953, 1965; Muller 1986). It would be foolhardy to make much of the minor, and rare, decorative differences among these local variations of formerly Late Woodland, but not-quite-yet Mississippian peoples.

The Early Mississippian Jonathan Creek phase (Butler 1991; Clay 1963, 1979, in press; Muller 1986; Riordan 1975) has been given dates of A.D. 1000-1100, although I suggest that the rather arbitrary boundary between Douglas and Jonathan Creek phases could be placed earlier at A.D. 900 to 950. The collections on which this phase are defined do not provide us with clear temporal markers, however, to distinguish Jonathan Creek settlement from the later Kincaid phase (which ends around A.D. 1350). On the basis of what happened in Douglas and Kincaid times, however, one can suppose that the trend toward concentration on the very best soils for maize cultivation continued. At Tinsley Hill and other places we have evidence for occupations in this period, but it would be going far beyond our present evidence to try to interpret local and regional boundaries or dispersion.


Peak Mississippian population and settlement in the Lower Ohio Valley came in the 13th century. Nearly every site with any Mississippian components, has some occupation from this time period, and it appears also to have been the period of maximum utilization of the Great Salt Spring as well (Muller et al. 1992; Muller and Renken 1989). Figure 7 shows the location of Mississippian sites in the Black Bottom locality. Much the same zones were occupied as in Lewis times, but the pattern is more linear and is uniformly distributed on the particular soil types (largely Armiesburg Silty Clay Loam) chosen by Mississippian settlers. Of course, the selection by a Mississippian horticulturist may have been made on any number of visible characteristics such as the presence of cane brakes, but the end result was the same. The sites shown in the figure are mostly farmstead sites, although there are a fair number of hamlets (groups of circa 15 structures). Kincaid itself is shown in Figure 7 as clusters of hamlets and homesteads, not as the circa 50 to 70 ha site that is constituted of many smaller site units either within the palisade or adjacent to the palisaded area.

Figure 7. Mississippian Site Distribution in the Black Bottom.


In his recent paper, Clay (in press) has reassessed his earlier discussions of pioneering and other settlement in the Lower Ohio (1976). In fact, Clay was one of the first in the entire Southeast to see that many of the settlements categorized as being hierarchical subordinate sites might have been essentially autonomous (see also Muller 1993a, in press b). Assumptions of hierarchy and hegemony are still reflected in such terminology as "primary centers," secondary centers," and so on (e.g., Kreisa 1991, 1995). These terms and the assumptions that underlay them now have come to be seen as questionable as we come to understand Mississippian political organization as distinctive, chiefly structures, not merely as pale shadows of Mesoamerica. Clay has argued that, even though our dating of these complexes is imperfect, there is evidence that the growth of "secondary" centers in the Lower Ohio, such as the Twin Mounds site, occurred at times when older centers were in "decline" (in press).

When only the larger Mississippian "centers" are considered (Figure 8) no clear gaps in settlement can be seen going up the river to the north until the Saline River is reached, although the narrow floodplains south of the Saline have only a few Mississippian sites recorded. With so little bottomland available for settlement, it is not surprising that Mississippian settlement is rare in some localities. The area around the mouth of the Saline to the Wabash, in the areas adjacent to the Great Salt Spring and the Half Moon Lick salterns is, however, also without much permanent Mississippian occupation, and it is notable that the Southwind site (Munson n.d.) is a nucleated and fortified village, perhaps suggesting that there was some kind of border between, say, Kincaid and Angel polities at the time of its occupation.

Figure 8. Mississippian Settlement in the Lower Ohio Valley from the Wabash to the Mississippi Confluence.

Figure 9. Mississippian Settlement from the Saline to the Arkansas Confluence.

South of the Ohio-Mississippi confluence (Figure 9), there are no obvious gaps in distribution; and "centers" were relatively regularly spaced well into the Lower Mississippi Valley (see, for example, Lewis 1990; Stephens 1995a; Stout and Lewis 1993). Most of these mound centers are relatively small by comparison to Kincaid or Angel, and I find it difficult to see any one of them as being politically or economically dominant over the others (also Stephens 1985a). I suspect that the pattern is one of more, but smaller, polities in this part of the broader region (see Stephens 1995a).

Distance from one "mound center" to another in the Lower Ohio is about 28 km. The range is from 6 km apart (in sites separated by the Ohio River) up to 48 km. In the confluence area to the south, the distance between one mound center and another is smaller, usually a dozen or so kms (Stephens 1995a). Of course, polity organization in the region could have fluctuated through time between multiple small polities and larger polities containing several major town-centers within a paramount chiefdom. Each small polity may have surrounded a single small town-center. Paramount chiefdoms may have covered considerable geographic areas, but were more fragile and perhaps existing for relatively brief periods. The figures for Hally's Georgia contemporaneous "centers" are similar to these (a mean of 36 km, median about 42 km, Hally 1993). In Hally's computations, the mean distance between contemporaneously occupied mound sites without synchronous mound construction is about the same. This may be another indication of the "cycling" kinds of emergence and disappearance of single centers over time in Southeastern societies under these conditions (see Hally 1993 for a discussion of the conditions of distance and effective administration; also see Wright 1977 and compare Moundville in Steponaitis 1991).

The Lower Ohio Mississippian centers were not all contemporary, but most sites seem to have at least some occupation in the mid to late 13th century. The evidence seems to support relative autonomy for each center rather better than it does the older idea of different levels of subordination for "primary" and "secondary" centers. Of course, the "antagonistic" nature of the site-location pattern does not necessarily mean political opposition, but may reflect dispersal for easier access to rich agricultural lands. This interpretation seems consistent with the aggregation of centers on relatively flood-free ridges in the wide reaches of the Mississippi floodplain in some localities. It is true that many, perhaps most, of the mound centers shown in Figure 8 had palisades at one point or another in their histories. However, data at large and small sites alike suggest that constructions of palisades was a "sometime" thing, and that even sites like Kincaid and Angel were more often "open" than they were enclosed.

Baden's suggestion of (periodic) abandonments because of soil exhaustion makes much sense in upland areas (e.g., Hargrave et al. 1983), but I have argued (in press b) that the Baden's time scale for soil exhaustion in bottomland localities seems at odds with both archaeological data and known historical European practice. Bottomland Mississippian settlement really does seem to have been more stable than Baden's estimates, based on upland soil exhaustion, would suggest. Nonetheless, it is well to keep the possibility of production variables having been key factors in apparent "political" cycling. My own guess is that much Mississippian occupation was fairly continuous, even though a given ridge may not have had farmsteads all the time. Some evidence of fallow field vegetation has been found, but these species also grow on the margins of field still in use (Blakeman 1974; Muller 1986: 218-19).

Another line of evidence for the nature of relationships among Mississippian centers may be found in exchange in Mississippian societies. Once upon a time, it was largely supposed that Mississippian "trade" was controlled by either "specialists" or&emdash; more often&emdash;by elites who monopolized control of needed resources, both internally and externally (e.g., Dincauze and Hassenstab 1989). It becomes increasingly difficult to accept such levels of control, not least because so much Mississippian exchange can be easily "explained" by simple mechanisms such as "down-the-line" kinds of relations (see also Welch 1996). For example, Figure 10 shows that the numbers of stone hoes found at different intervals away from the source of the raw material in Union County, Illinois, have very high correlations to the natural logarithm of the distance from the source, indicating no evidence to support the idea of intermediate control of these exchanges.

Figure 10. Mill Creek Stone Hoes Distributions (data from Brown, Kerber, and Winters 1990).


For most classes of both fancy and plain goods, there is really no reason to postulate "elite control" in most Mississippian exchange relations (although Welch argues that there may have been some at Moundville, 1991, 1996). I have elsewhere shown how both stone hoes and shell gorgets are distributed across Mississippian space in relation to distance from their sources (see Muller 1995, in press b). I predict that as we are better able to detect the origins of various other Mississippian objects of exchange (as per Steponaitis, Blackman, and Neff 1996), we shall see similar patterns of fall-off of numbers of goods as the square of their distance from their sources. Even the data we already have make it unlikely that we shall be able to explain Mississippian chiefly power in terms of their control of exchange.

As we come to understand the contradictions between Mississippian interdependence and Mississippian autonomy in the period of peak occupation, we also see how evidence for exchange of goods, multiple centers of different sizes, and other features need not be interpreted as indications of political hierarchies. The evidence can be understood as reflecting the dynamic interplay of control and independence through time. While some periods undoubtedly were characterized by more widespread regional integration, other periods may have been relatively unorganized, without much changing what survives to be seen in the archaeological record. What we do not see, even in the period of peak population, is much evidence for strong control or hegemony.


Late Mississippian in the Lower Ohio has been characterized as the Tinsley Hill phase (A.D. 1350-1450; Butler 1991; Clay 1979). Like earlier Mississippian phases in this region, it has proved difficult to identify diagnostics for the Tinsley Hill phase. However, the still later Caborn-Welborn&emdash;extending up to the protohistoric&emdash;is marked by suite of highly distinctive traits including shoulder-trailed vessels, effigy head pots, Nodena points, end scrapers, and catlinite objects. Some of these traits indicate exchange relations to the north and with the Memphis locality.

By the late 13th century, Mississippian settlers in bottomland towns in the Lower Ohio faced a series of problems and opportunities. Regardless of whether the length of time required to deplete bottomland soil fertility was short (Baden 1995) or longer (Muller in press b), cultivation for 4 or 5 hundred years in the bottomlands at Kincaid and Angel would surely have taken its toll. In addition, basically monocultural exploitation of the same soils would have also meant probable buildup of pests and parasites of many kinds. And not only for the plants&emdash;humans too had been living on the same ridges for centuries. I have also argued that there may have been a slow and gradual drop in population in many Mississippian populations leading up to this period (Muller in press a, in press b). I have also suggested the possibility that mechanisms of reaggregation like those of the Historic period may have been long-established Mississippian practices in response to the more gentle depopulations of the Late Prehistoric period (see also Milner 1986).

To some extent, lower populations could have compensated for lower soil fertility, by allowing more fallow fields, and by reducing the competition for non-cultivated resources like nuts and game. However, the labor force needed to exploit some of these resources would also have been smaller, and perhaps more dispersed. By the 14th century, we see that there were fewer people living in the formerly densely populated lowlands, and we now know that there was an increase use of upland areas that had not seen much occupation since the middle of Late Woodland times. Cobb and Butlers' paper in this symposium has already described how we have come to change our view of the "disappearance" of Mississippian in this region. It is still likely that there were considerably fewer people in Lower Ohio Valley, but at the same time, these persons were probably more dispersed into upland areas, thus reducing their archaeological visibility still further (Muller 1986:254-5). Similar kinds of reductions and dispersals seem to have occurred slightly earlier in the Mississippi Valley north of the Ohio confluence (e.g., Kuttruff 1972, 1974), although we also have evidence of 14th- and 15th- century Mississippian occupation at floodplain centers like Dogtooth Bend at the Mississippi- Ohio confluence (Stephens 1995b, 1996). A scattering of dates in this time period from sites like Kincaid and Angel may sometimes merely indicate the statistical nature of radiocarbon dating, but these have become more acceptable to regional specialists as it has become obvious that other sites were occupied into the 15th century. Clay has proposed that small groups continued to use the mound centers, but without any new mound construction. While this evidence reduces the "gap" between Kincaid/Angel phase and the protohistoric Caborn-Welborn occupations at the mouth of the Wabash (Green and Munson 1978), it remains true that we have no evidence of Caborn-Welborn presence between the Saline River and the Mississippi (Muller 1986). On the whole, the pattern of Late Mississippian settlement is far more dispersed and less aggregated than settlement prior to A.D. 1300. I also rather suspect that the lack of Caborn-Welborn diagnostics in these sites suggests that Caborn- Welborn might not have begun until later than the 15th century date we formerly accepted, perhaps even in the 16th century. We do know that Caborn-Welborn societies were exchanging goods with the Lower Mississippi Valley. It is hard to see how such goods could have moved through the lower portions of the Ohio Valley without leaving traces in local settlements. Another possible consequence of depopulation would have been an overall increase in the distances between polities, a trend I think will be seen in the later Late Prehistoric into the Historic period as we gain increasing control of the chronology.

It seems likely that Late Mississippian dispersal was a result of several different factors. New productive technology, including the common bean (see Smith 1992 on the development of cultivation and cultivars), developed throughout the Mississippian period. In addition, lowered population levels seem likely (Muller in press a and b). These and other factors may have reduced the interdependence of Mississippian households (Muller 1986) to an extent that many were willing to "pioneer" upland stream valleys. Lowered fertility of the floodplains after so many years of cultivation may have further reduced the attractiveness of floodplain environments for settlement. As the autonomy of households was translated into dispersed settlements, declines in central authority may have also led to an increased level of competition in regard to more limited resources in upland areas. Although we have little evidence of permanent settlement at the Great Salt Spring, Dillow's Ridge and Dover occupations suggest that late Mississippian settlers may have sometimes reaggregated at prime resources (e.g., Cobb 1995; Gramly 1992). The 14th century dates at the Dover quarries may result from similar late period "staking out" of key resources (Gramly 1992). As we have heard, similar dates have been obtained for Mississippian settlement at the easily defended Millstone Bluff.

We may propose that increased individual family economic autonomy was translated into dispersal from cooperative settlement in floodplains, but entailed increased risks of raiding and conflict as the number of more or less independent political entities increased. Given the small size of the remaining Mississippian populations in the area, however, the options of reaggregating in the main valley seems not to have been considered. In fact, the clearest indication of a probably "political" boundary in the entire record for the Lower Ohio Valley is from Caborn-Welborn times when most of the area between the Saline River and the Mississippi seems to have been what has been called a "vacant quarter" (Williams 1980, 1990). However, the political and sociological implications of such a gap are not clear, since there were clearly smaller and more scattered polities during this period. It is also unclear what implications the connections between Caborn-Welborn and Nodena have for Dhegiha Siouan legends of origins up the Ohio River Valley (e.g., Swanton 1946:176).

Implications and Conclusions

This survey has examined issues of dispersion, especially in the Lower Ohio. The historical record of population organization shows patterns of aggregation and dispersal and I think that it is warrantable that similar mechanisms of dispersion operated both in historical and prehistoric times. The task of placing settlement use and history in the Lower Ohio into a chronological framework is just beginning. What was presented here was only a rough measure at fairly coarse intervals&emdash;the phases defined for the region. In the long run, much finer control of chronology will be necessary to more accurately assess the differences in settlement tactics and strategies during any given period. For example, early Late Woodland had different dispersion than did even late Lewis phase, much less the transitional Douglas, Duffy, and Yankeetown phases.

In the earlier sections, I tried to present the basic information on dispersion. In closing, I want to explore some of the possible implications of these data. Despite problems from the coarse resolution of the known facts, a general picture emerges. It can be suggested that early Late Woodland distributions reflected lack of intergroup competition. The standard view of Late Woodland as presenting a fairly uniform cultural and social field with relatively few "boundaries" is probably true. At the same time, those upland "stone forts" exist, I would guess mostly from later in Late Woodland times, and they suggest some regional integration and competition at least after the early 8th century.

By the 9th century, the transitional phases often called Emergent Mississippian were present throughout the Lower Ohio River Valley. There are at least three regional variants, based on ceramic attributes, but none of these traits seem likely to mark any major political differences. One important feature of the period is the beginning of mound construction, at least at Kincaid. Below the Wabash-Ohio confluence, at least, the distribution of Emergent Mississippian sites appears to be largely congruent with later Mississippian patterns, but somewhat broader. In the vicinity of the Wabash and Saline Rivers, Duffy phase sites occur in areas where middle period Mississippian sites are absent.

The peak use of the Lower Ohio occurred in the 13th century, followed by a rapid decline in local populations. During this period, there are two interesting aspects to site distributions. The first is the rarity of upland Mississippian sites of any kind until after A.D. 1300. I suspect this was a result of the integration of bottomland economics and politics. This period appears to have been the peak for hierarchical political expressions in the region, and cooperation and control in the floodplains were an integral part of those developments.

The second aspect is the rarity of Kincaid or Angel phase sites in the Wabash and Saline confluences and downstream. These might be indications of a political division between Angel and Kincaid, but there are reasons not to accept this argument too readily. The first is the fact that the area to the south of the Saline River that lacks intensive Mississippian use is a section of the valley with little floodplain for cultivation. The lack of Mississippian settlement may be only a reflection of the unsuitable character of this part of the valley for normal Mississippian activities. The broad lowland areas between the Saline and Wabash Rivers, on the other hand, are at least superficially like the broad floodplains to the south. However, the soil types in this locality are not really the same as those downstream soils that proved attractive to Kincaid phase settlement. Secondly, and more importantly, this is a locality that is adjacent to the important salterns at the Great Salt Spring and Half Moon Lick. Evidence at the Great Salt Spring indicates largely transient use of the site in the 13th century by substantial numbers of people from all over the region. In that sense, the communal use of the salterns may have been a factor in the suppression in this locality of more ordinary horticultural settlement. On the other hand, the presence of fortifications and an aggregated settlement at the Southwind site does suggest the possibility of a political boundary in the Wabash locality at some times.

Late Mississippian as defined here is only now becoming recognized in this region. At one time we saw the entire region as a vacant quarter, but have come to understand that the "emptiness" of the main valley sites, reflects both a decline in population through time and a dispersal of that population into upland areas that had seen little use since Late Woodland times. It is too early to be sure of the sequence of events here, but I suggest that the hilltop, really mesa top, settlements will prove generally a little later than the scattered farmsteads of the upland stream valleys. The focus of late upland settlement may have shifted from soil fertility to protection against raids. I suspect that the development of the historical "Illinois Country"&emdash;a zone used by many different peoples, but controlled by none may have started after the mid-15th century. Such intrusions into the region might have encouraged aggregation for defense on the part of the relatively small remaining populations.

The Caborn-Welborn phase may be directly descendent from the dispersed populations of the 15th century, but it is also possible that they were an immigrants from other regions (see Muller 1986 for a discussion of these issues). It is clear, however, that the areas between the Saline River and the Mississippi on the Ohio did not have Caborn- Welborn occupations, although goods found in Caborn-Welborn sites in the Wabash Confluence show that there were connections between that locality and the Memphis locality. Such goods would surely have passed through the lower sections of the Lower Ohio, and the absence of any evidence suggests that the area was very probably a "vacant quarter" by that time. Such a pattern may reflect mechanisms of reaggregation and reformation that resulted in societies that were similar in size to those of earlier Mississippian times, but more widely scattered across space (Muller in press b).


Send them to: (This page last modified on 29/vi/05)



Adair, James

1775 The History of the American Indians, particularly those nations adjoining to the Mississippi,... Edward and Charles Dilly, London. [reprinted 1930 Society of Colonial Dames, facsimile of 1930 edition by Promontory Press, New York]

Anderson, David G.

1990 Stability and Change in Chiefdom Level Societies: An Examination of Mississippian Political Evolution on the South Atlantic Slope. In Lamar Archaeology: Mississippian Chiefdom in the Deep South, edited by M. Williams and G. Shapiro, pp. 187-213. University of Alabama Press, Tuscaloosa.

1994a Factional Competition and the Political Evolution of Mississippian Chiefdoms in the Southeastern United States. In Factional Competition and Political Development in the New World, edited by Elizabeth Brumfiel and J. Fox, pp. 61-76. Cambridge University Press, Cambridge, U.K.

1994b The Savannah River Chiefdoms: Political Change in the Late Prehistoric Southeast. University of Alabama Press, Tuscaloosa.

Baden, William W.

1995 The Impact of Fluctuating Agricultural Potential on Coosa's Sociopolitical and Settlement Systems. Paper presented at the Southeastern Archaeological Conference, Knoxville. [Formerly posted at or directly from a version of Baden's dissertation which covers the material in greater detail is available at ]

Beverley, Robert

1705 The History and Present State of Virginia in Four Parts. R. Parker, London. [1968 reprint of 1705 edition, The History and Present State of Virginia. Dominion Books, The University Press of Virginia, Charlottesville. ]

Biedma, Luys Hernandez de

1544 Relation of the Conquest of Florida Presented by Luys Hernandez De Biedma in the Year-1544 to the King of Spain in Council. [translated from the original document by Buckingham Smith in 1904. Narratives of the Career of Hernando de Soto, edited by Edward G. Bourne. pp. 3-40 (Volume II) A. S. Barnes, New York. translation by John E. Worth 1993 in Clayton, et al.]

Blakeman, Crawford

1974 The Late Prehistoric Ethnobotany of the Black Bottom, Pope and Massac Counties, Illinois. Unpublished Ph.D. dissertation, Department of Anthropology, Southern Illinois University, Carbondale.

Blasingham, Emily J.

1953 Temporal and Spatial Distribution of the Yankeetown Cultural Manifestation. unpublished M.A. thesis, Department of Anthropology, Indiana University.

1965 Excavation of Yankeetown. Report to the National Park Service, NER-815. Loyola University.

Brown, James, R. Kerber and H. Winters

1990 Trade and the Evolution of Exchange Relations at the Beginning of the Mississippian Period. In The Mississippian Emergence, edited by B. Smith, pp. 251-280. Smithsonian Press, Washington.

Butler, Brian M., Glen P. Harrell, and Mary C. Hamilton

1979 An Archaeological Reconnaissance of the Illinois Portions of the Smithland Pool of the Ohio River. Southern Illinois University, Center for Archaeological Investigations, Research Paper No. 5.

Canouts, Veletta, E.E. May, N.H. Lopinot, and J. Muller

1984 Cultural Frontiers in the Upper Cache Valley, Illinois. Southern Illinois University, Center for Archaeological Investigations, Research Paper No. 16.

Clay, R. Berle

1963 Ceramic Complexes of the Tennessee-Cumberland Region in Western Kentucky. Unpublished M.A. thesis, Department of Anthropology, University of Kentucky.

1976 Tactics, Strategy, and Operations: The Mississippian System Responds to Its Environment. Midcontinent Journal of Archaeology 1(2):138-62).

1979 A Mississippian Ceramic Sequence for Western Kentucky. Tennessee Anthropologist 4(2):111-28.

1981 Kentucky: An Introduction to State-wide Research Design. Office of State Archaeology. Lexington, Kentucky.

in press The Mississippian Succession on the Lower Ohio. Southeastern Archaeology.

Cobb, Charles R.

1995 1995 SIUC Field School Investigations at the Dillow's Ridge Site (11-U-635). Ms. On file, Southern Illinois University at Carbondale, Center for Archaeological Investigations.

Cole, Fay-Cooper et al.

1951 Kincaid: A Prehistoric Illinois Metropolis. University of Chicago Press, Chicago.

Dincauze, Dena F., and Robert Hassenstab

1989 Explaining the Iroquois: Tribalization on a Prehistoric Periphery. In Centre and Periphery: Comparative Studies in Archaeology, edited by Timothy C. Champion, pp. 67-87. Unwin Hyman, London.

Galloway, Patricia K.

1995 Choctaw Genesis: 1500-1700. University of Nebraska Press, Lincoln.

Green, T. J. and C. A. Munson

1978 Mississippian Settlement Patterns in Southwestern Indiana. In Mississippian Settlement Patterns, edited by B. Smith, pp. 293-330, Academic Press, New York.

Hally, David J.

1993 The Territorial Size of Mississippian Chiefdoms. In Archaeology of Eastern North America: Papers in Honor of Stephen Williams, edited by J. B. Stoltman, pp. 143- 168. Mississippi Department of Archives and History, Archaeological Report No. 25.

Hawkins, Benjamin

1799 A Sketch of the Creek Country in the Years 1798 and 1799. ms. in possession of Georgia Historical Society. published in 1848 in Georgia Historical Collections III, Part I. pp. 13- 83. Another copy of ms. in Peter Force Papers, Manuscript Division, Library of Congress. [1938 reprint of 1848 with additions in The Creek Country, as by "Hawkins, et al.", pp. 19-80. Americus Book Company, Americus, Georgia; also reprint by Kraus Reprints Co., 1971 facsimile reprint of the 1848 publication)].

Hargrave, Michael L., G. Oetelaar, N. Lopinot, B. Butler and D. Billings

1983 The Bridges Site (11-Mr-11): A Late Prehistoric Settlement in the Central Kaskaskia Valley. Research Paper No. 38 Center for Archaeological Investigations, Southern Illinois University at Carbondale.

Kreisa, Paul P.

1991 Mississippian Sites of the Lower Ohio River Valley in Kentucky. Western Kentucky Project, Report 9. Department of Anthropology, University of Illinois, Urbana. (Submitted to Kentucky Heritage Council).

1995 Mississippian Secondary Centers along the Lower Ohio River Valley: An Overview of Some Sociopolitical Considerations. In Current Archaeological Research in Kentucky: Volume Three. edited by John Doershuk et al., pp. 161-177.

Kuttruff, Carl

1972 Late Woodland and Mississippian Settlement Systems in the Lower Kaskaskia River Valley. NSF Grant Report. Mimeographed.

1974 Late Woodland Settlement and Subsistence in the Lower Kaskaskia River Valley. Unpublished Ph.D. dissertation, Department of Anthropology, Southern Illinois University at Carbondale.

Lewis, R. Barry

1990 The Late Prehistory of the Ohio-Mississippi Rivers Confluence Region, Kentucky and Missouri. In Towns and Temples along the Mississippi, edited by D. Dye and C. Cox, pp. 38-58. University of Alabama Press, Tuscaloosa.

MacNeish, R. S.

n.d. An Early Middle Mississippi Horizon at the Kincaid Site. Ms. on file at the Center for Archaeological Investigations, Southern Illinois University - Carbondale.

1944 The Establishment of the Lewis Focus. Unpublished Master's thesis, Department of Anthropology, University of Chicago.

Milner, George R.

1986 Mississippian Period Population Density in a Segment of the Central Mississippi River Valley. American Antiquity 51(2):227-238.

Muller, Jon

1978a The Southeast. In Ancient Native Americans, edited by Jesse D. Jennings, pp. 280325. W. H. Freeman, San Francisco.

1978b The Kincaid System: Mississippian Settlement in the Environs of a Large Site. In Mississippian Settlement Patterns, edited by B. Smith, pp. 269-292, Academic Press, New York.

1983 The Southeast. In Ancient North Americans, edited by Jesse D. Jennings, pp. 372419. W. H. Freeman, San Francisco.

1986 Archaeology of the Lower Ohio River Valley. Academic Press, New York.

1993a Lower Ohio Valley Mississippian Revisited: An Autocritique of "The Kincaid System." In Archaeology of Eastern North America: Papers in Honor of Stephen Williams, edited by James B. Stoltman, pp. 128-142. Archaeological Report No. 25, Mississippi Department of Archives and History, Jackson, Mississippi.

1993b Eastern North American Population Dynamics. In Highways to the Past, edited by T. Emerson, A. Fortier, and D. McElrath, pp. 84-99. Illinois Archaeology 5(1&2).

1995 Regional Interaction in the Southeast. In Native American Interactions: Multiscalar Analyses and Interpretations in the Eastern Woodlands, edited by M. Nassaney and K. Sassaman. pp. 317-353. University of Tennessee Press, Knoxville.

1996 Dispersion in the Southeast. ms.

in press a Native Eastern American Population Continuity and Stability. In Integrating Archaeological Demography, edited by Richard R. Paine. Center for Archaeological Investigations, Southern Illinois University at Carbondale, Occasional Paper.

in press b Mississippian Political Economy. Plenum, New York.

in preparation Dispersion in the Southeast. ms.

Muller, Jon, editor, assisted by Lisa Renken, George Avery, and others

1992 The Great Salt Spring: Mississippian Production and Specialization. Draft Report of March, 1992, on File with the Shawnee National Forest, U. S. Forest Service, Harrisburg, Illinois. and the Center for Archaeological Investigations, Southern Illinois University, Carbondale.

Muller, Jon, and Lisa Renken

1989 Radiocarbon Dates for the Great Salt Spring Site: Dating Saltpan Variation. Illinois Archaeology 1(2):150-160.

Munson, Cheryl, ed.

n.d. The Southwind Site: A Mississippian Village in Southwestern Indiana. Prepared for the Indiana Port Commission by the Glenn A. Black Laboratory, Indiana University. ms.

Pauketat, Timothy R.

1994 The Ascent of Chiefs: Cahokia and Mississippian Politics in Native North America. University of Alabama Press, Tuscaloosa.

Pielou, E.C.

1977 Mathematical Ecology. John Wiley, New York.

Ramenofsky, Ann F.

1987 Vectors of Death: The Archaeology of European Contact. University of New Mexico Press, Albuquerque.

1990 Loss of Innocence: Explanations of Differential Persistence in the Sixteenth Century Southeast. In Archaeological and Historical Perspectives on the Spanish Borderlands East. Columbian Consequences. Vol. 2, edited by David H. Thomas. pp. 31-48. Smithsonian Institution Press, Washington.

Riordan, Robert

1975 Ceramics and Chronology: Mississippian Settlement in the Black Bottom, Southern Illinois. Unpublished Ph.D. dissertation, Department of Anthropology, Southern Illinois University, Carbondale.

Rogers, J.D. and B. Smith, editors

1995 Mississippian Communities and Households. The University of Alabama Press, Tuscaloosa.

Rudolph, Teresa Perry

1981 The Distribution of Late Woodland Sites in the Black Bottom Area, Pope and Massac Counties, Illinois. Unpublished M.A. Thesis, Department of Anthropology, Southern Illinois University at Carbondale.

Smith, Bruce D.

1978 Variation in Mississippian Settlement Patterns. In Mississippian settlement patterns, edited by B.D. Smith, pp. 479-503. Academic Press, New York.

1992 Rivers of Change: Essays on Early Agriculture in Eastern North America.. Smithsonian Press, Washington, D.C.

Smith, Marvin T.

1987 Archaeology of Aboriginal Culture Change in the Interior Southeast: Depopulation during the Early Historic Period. University of Florida/ Florida State Museum, Gainesville.

Snow, Dean R.

1995 Microchronology and Demographic Evidence Relating to the Size of Pre- Columbian North American Indian Populations. Science 268:1601-1604.

Stephens, Jeanette E.

1994 Near the Confluence: The Dogtooth Bend Mound Center and Its Settlement System. Paper presented at the joint Southeastern Archaeological Conference and Midwest Archaeological Conference, Lexington, Kentucky.

1995a The Political Geography of Mississippian Mound Centers at the Mississippi-Ohio River Confluence. Paper presented at the Society for American Archaeology Annual Meeting, Minneapolis.

1995b An Archaeological Survey of Dogtooth Bend on the Mississippi River in Alexander County, Illinois. U.S. Army Corps of Engineers, St. Louis District Historic Properties Management Report No. 45.

1996 Archaeological Mitigation at the Dogtooth Bend Site (11-Ax-31), Alexander County, Illinois. Southern Illinois University at Carbondale, Center for Archaeological Investigations Technical Report 1996-1.

Steponaitis, Vincas P.

1991 Contrasting Patterns of Mississippian Development. In Chiefdoms: Power Economy, and Ideology, edited by T. Earle, pp. 193-228. School of American Research.

Steponaitis, Vincas P., M.J. Blackman, and H. Neff

1996 Large-scale Patterns in the Chemical Composition of Mississippian Pottery. American Antiquity 61(3):555-572.

Stout, Charles and R. Barry Lewis

1993 Toward a Functional Grammar of Mississippian Town Composition in Kentucky. Paper presented at the Society for American Archaeology. St. Louis, Missouri.

Swanton, John R.

1946 The Indians of the Southeastern United States. Bureau of American Ethnology, Bulletin 137. [reprint 1979, with some minor changes in illustrations, Smithsonian Institution Press, Washington.]

Thornton, Russell

1987 American Indian Holocaust and Survival. University of Oklahoma Press, Norman.

Thornton, Russell with the assistance of C.W. Snipp and N. Breen

1990 The Cherokees: A Population History. University of Nebraska Press, Lincoln.

Thompson, J.M.T. and H.B. Stewart

1987 Nonlinear Dynamics and Chaos: Geometrical Methods for Engineers and Scientists. John Wiley and Sons, New York.

Waldrop, M. Mitchell

1992 Complexity: The Emerging Science at the Edge of Order and Chaos. Simon & Schuster, New York.

Welch, Paul D.

1991 Moundville's Economy. The University of Alabama Press, Tuscaloosa.

1996 Political Economy in the Late Prehistoric Period in the Southern Appalachians. Paper presented at the Conference on Integrating Appalachian Highlands Archaeology, Albany, New York.

Williams, Stephen

1980 Armorel: A Very Late Phase in the Lower Mississippi Valley. Southeastern Archaeological Conference Bulletin 22:105-110.

1990 The Vacant Quarter and Other Events in the Lower Valley. In Towns and Temples along the Mississippi, edited by D. Dye and C. Cox, pp. 170-180. University of Alabama Press, Tuscaloosa.

Winters, Howard D.

1967 An Archaeological Survey of the Wabash Valley in Illinois (revised edition). Illinois State Museum Reports of Investigations 10.

Wright, Henry T.

1977 Recent Research on the Origin of the State. Annual Review of Anthropology 6:379-397.