The Nebraska Phase In IowaMichael J. Perry
Office of the State Archaeologist
700 S. Clinton St.
University of Iowa
Iowa City, IA 52242-1030
Brief History of Major Investigations
Published descriptions of Iowa's earthlodge culture date to the 1880s (Dean 1883; Proudfit 1881a, 1881b, 1886a, 1886b). Amateur archaeologist Paul Rowe hunted the hills and ravines around Glenwood from at least the early 1920s through the early 1960s, leading to the location of over 200 sites (Green 1992:3). Rowe's efforts led to the accumulation of a major collection of Nebraska phase materials and several published descriptions of excavations he conducted (Rowe 1922, 1951, 1952a, 1952b; Davis and Rowe 1960; Billeck and Rowe 1992). The Rowe Collection is now housed in the Mills County Historical Museum and in Iowa City at the Office of the State Archaeologist. Through both personal visits and correspondence dating from the 1920s through the 1940s, Rowe brought to the attention of Charles Keyes the archaeological richness of the Glenwood area (e.g., Keyes 1924, Rowe 1933, 1942). As a result, Keyes arranged for his associate, Ellison Orr, to conduct an intensive survey of the Glenwood locality in 1938, which would include both site location, mapping, and excavation. The materials recovered by Orr are in the Keyes Collection, presently housed at the Iowa City office of the State Historical Society of Iowa, forming a second major collection of Nebraska phase materials. Orr completed a manuscript report of his survey (Orr 1942), which was later included in a microform publication (Orr 1963) of his activities for the Iowa Archaeological Survey. A number of sites identified by Rowe, Keyes, and Orr have recently been verified and correlated, providing an important link between modern site records and the archival information provided by these early researchers (Tiffany et al.1990; Billeck 1992a).
Anderson conducted excavations in the Glenwood locality in the mid 1950s as part of his graduate studies at the University of Iowa (Anderson 1954; Anderson and Anderson 1960), and again in 1969 with the onset of highway construction work on U.S. 34 west of Glenwood (Anderson 1973). Both Anderson's (1961) and Zimmerman's (1977a) graduate studies included analyses of ceramics recovered by Orr. Anderson and Zimmerman collaborated (1976) in an analysis of Glenwood locality settlement pattern that also employed Orr's site location data. Zimmerman (1977b) later expanded the Anderson-Zimmerman settlement model into a computer simulation model attempting to explain the factors that led to the known distribution of Nebraska phase components in Iowa at the time.
Recent archaeological investigations at Glenwood have been conducted under cultural resource management programs by government agencies. The proposed development of a number of erosion control structures in the Pony Creek watershed led to surveys and excavations by Brown (1967) and Fulmer (1974). Construction of U.S. 34 west of Glenwood led to further excavations at additional Nebraska phase earthlodges (Hotopp 1978a, 1978b, 1982). Billeck recently analyzed artifacts recovered by Hotopp's U. S. 34 excavations and conducted test excavations at 13ML360 and 13ML361 as part of his doctoral research (Billeck 1993). Most of the recent investigations in the Glenwood locality were prompted by proposed local road improvements, including grading along Mills County route L31 (Perry 1983, 1984, 1987; Morrow 1995) and bridge replacements along Pony Creek (Perry 1990).
Spatial and Temporal Range of the Nebraska Phase in Iowa
Nebraska phase utilitarian ceramics include globular jars, broad shallow bowls, and narrow-mouthed seed jars. The earthenware vessels were presumably produced by coiling or mass modeling a paste of clay tempered with grit, sand, or crushed mussel shell, followed by malleating with a cord-wrapped paddle. The cord-roughened pots were often partly or completely smoothed over. Decoration was applied to the rim exterior, consisting of incising, tool or fingertip impressing, or occasionally pinching or noding. Bodies were usually left undecorated. Billeck (1993) has suggested that pots with shell tempered paste appear both early in the occupation of the locality due to the influence of contemporary Mississippian/Steed-Kisker phase populations, and late in the occupation through contact with or influence by Oneota groups. The influence of Mississippian related, or later Oneota groups, is also reflected in body sherds with trailed/incised line motifs or chevrons that were placed on the shoulders of a few vessels at early or late sites (Billeck 1993:247-249). Sherds with cord impressed decoration similar to Late Woodland Missouri Bluffs Cord Impressed ware occur in low frequencies throughout the occupation of the region.
Despite its apparent short-comings for temporal seriation studies, the Gunnerson typology is still a useful means of describing Nebraska phase ceramics. McVey ware includes vessels with rims that are either vertical or flaring. McVey ware vessels occasionally display tabs or loop handles. Decorated McVey ware vessels display tool or fingertip impressions around the upper rim or lip (Figure 6), or incising or trailing on vessel shoulders executed in styles reminiscent of Mississippian or Oneota designs (Figure 7). Beckman ware vessels display a vertical rim profiles with a wedge-shaped exterior collar encircling the lower rim. If decorated, the decoration was usually applied to the collar panel area and consisted of tool or fingertip impressions (Figure 7), nodes, incised crosshatching, or incised parallel horizontal lines. Swoboda ware vessels display rims with S-shaped profiles due to the addition of an exterior collar and an interior channel along the upper rim. The distinction between Beckman and Swoboda ware rim sherds is often difficult, as large collections of collared rim sherds tend to exhibit a continuum from sherds that are strongly S-shaped to sherds that are vertical with only a slight collar or channel. Decoration on Swoboda ware pots is similar to that of Beckman ware. Debilka is something of a catch-all type consisting of the rather divergent vessel types of vertical or incurving rim bowls, and seed jars with their insloping profiles and narrow orifices. Decoration on Debilka bowls and seed jars usually consists of exterior trailing or incising.
Nonutilitarian ceramics found on Nebraska phase components include miniature pots, spoons or ladles, pipes, and figurines. Pipes were tubular or elbowed and occasionally decorated with incising or anthropomorphic or zoomorphic effigies. The pipes suggest the use of tobacco among Glenwood locality populations. Ceramic figurines have been found in human, waterfowl, and turtle forms. McNerney (1987) postulated complex relationships between Nebraska phase populations and Mississippian, Caddoan, and southwestern populations as the impetus for the production of the figurines.
Bone, Antler, and Shell Tools
Despite the generally good preservation conditions that have been noted for the loess hills area of western Mills County, Nickel's (n. d.) analysis of botanical remains from several of the U. S. 34 sites excavated by Hotopp (1978a) was the only Glenwood locality archaeobotanical study until the 1990s. Along with 13ML176 (this volume), the recently excavated sites 13ML175 (Asch and Green 1995) and 13ML361 (Billeck 1992b) are the most recently investigated sites that have yielded analyzed floral remains. Asch and Green (1995:68) suggest that the paucity of archaeobotanical studies was due in part to use of only dry screening excavated soils and feature fills over relatively course mesh, precluding the possibility of recovering most floral remains that could be used as indicators of subsistence resources.
Flotation processing of feature fills and other cultural deposits using fine mesh recovery screens is now common practice in archaeological research. These methods provide data useful for the reconstruction of both prehistoric diets and paleoenvironments. As a result, new discoveries of exploited plant remains are occurring with greater frequency. The analyzed U. S. 34 collection plus assemblages recently excavated sites 13ML175 and 13ML361 provide a glimpse into the kinds of plant resources used in the Glenwood locality (Table 1). Nebraska phase populations exploited a broad spectrum of floral resources that may be tentatively grouped into domesticates, cultivated wild species, and collected wild species.
Some of the species that were important to prehistoric Glenwood locality residents, such as little barley, are now extinct; others are considered weeds by modern people. Its likely that Glenwood sites are characterized by floral assemblages that include some species common to many of the other sites, while other species may occur uniquely or are found at only a few other sites (Asch and Green 1995:69). The commonly known prehistoric domesticates corn, beans, and squash, were undoubtedly important in the diets of Nebraska phase populations throughout the locality, but the seeds of cultivated and collected wild species also contributed significantly.
Faunal remains, including mammal, fish, bird, and reptile bone and mussel shell, were usually recovered in past excavations, but analyses of the materials have been limited to identification of tools or ornaments (e.g., Anderson 1961). Few past investigators analyzed the faunal assemblages, but a growing number of researchers are focusing on Glenwood locality faunal data for both paleoenvironmental information and prehistoric subsistence practices (e.g., Bardwell 1981; Cordell et al. 1995; Hirst 1995). The 13ML175 faunal assemblage provides a preliminary impression Nebraska phase faunal exploitation (Table 2). As with floral remains, considerable intersite variability in species presence and frequency may be expected among faunal assemblages from Nebraska phase sites of the Glenwood locality. Clear evidence of bison exploitation has not been encountered in the Glenwood locality, but selected bison elements, such as scapulae, have been noted. Such elements were usually made into tools, and may have been acquired through trade.
Sites and Settlements
Nebraska phase sites in Iowa are usually associated with remains of houses termed earthlodges. Nebraska phase earthlodges were constructed using the typical Central Plains tradition pattern of four central support posts surrounded by shorter, closely spaced, outer wall posts. The central support posts were usually substantial oak, walnut, or elm timbers that were charred at the base and set into postholes. Charring the post bases may have helped prevent rot, and may also reflect the method of felling the trees used or sizing the posts (Hotopp 1978a). These vertical members supported large cross beams, and smaller rafters that were interwoven with twigs to form walls and a roof that were plastered with wattle and daub, a mixture of grass and mud. Access to the lodge interior was gained through a similarly constructed extended entryway, often south-facing. Lodges were built on both level and sloping topography. Excavation of the upper few inches of soil presumably provided the material for the wattle and daub wall chinking. Where the site of a lodge was sloping, such excavation involved the removal of considerable amounts of soil at the upslope end of the house area to create a more level living surface. Due to such site preparation, earthlodges are sometimes referred to as semi-subterranean structures.
The floor plan of a Nebraska phase earthlodge was square to rectangular in outline, with rounded corners. A central roof opening permitted the escape of smoke from a fire hearth located roughly in the center of the enclosed floor space. Several storage pits with vertical or bell-shaped profiles were usually contained within a lodge. Sleeping spaces were located along the perimeter walls. The nature of Nebraska phase people's use of space outside their earthlodges is poorly understood due in part to lodge-focused research (cf. Gradwohl 1969). Hotopp (1978b:115) tested lodge exterior areas during the U. S. 34 excavations, but only one exterior pit feature was encountered at 14 lodge sites. Abundant remains were recently encountered in lodge-exterior areas at 13ML175, including three features (Morrow 1995). Perry (1998) suggested that lodge exterior areas may contain important data on activities such as flintknapping, food processing, and refuse disposal. Considering the limited amount of useful space within a lodge for activities other than cooking or sleeping, it seems likely that many social and domestic functions took place outdoors.
Archaeologists have developed various formulas using measurements of lodge floor areas or perimeter lengths to estimate the number of occupants (Blakeslee 1989; Naroll 1962; Wedel 1979). Thirty-six excavated Glenwood locality earthlodges yielded floor areas ranging 306-1849 ft2 (28.4-171.8 m2) (Hotopp 1978b:113; Crismon and Green 1992). A majority of these lodges ranged 300-800 ft2 (27.9-74.3 m2), but a frequency histogram of lodge floor area ranges in 100 ft (9.29 m) increments yields a bimodal distribution pattern of both large (over 900 ft [83.6 m2]) and small (less than 700 ft2 [65.0 m2]) lodges (Figure 8). This bimodal distribution pattern has been noted throughout the Nebraska phase region (Blakeslee and Caldwell 1979; Hotopp 1982). At 5 m2 (Wedel 1979) per person, large Glenwood locality earthlodges may have housed at least 16 people, and small lodges could have held no more than 11 people. Such household populations suggest that the lodges were occupied by extended families.
The Nebraska phase settlement pattern may be generally described as earthlodges located on upland ridge summits and spurs or on footslopes or terraces above valley floodplains. The lodges occur both isolated and in clusters that are often arranged in linear patterns (Blakeslee and Caldwell 1979; Brown 1967; Sterns 1915; Wedel 1956, 1959, 1961; Wood 1969). Earthlodges located on uplands have historically received the most attention, since their locations were often marked by shallow depressions and were therefore easily located by researchers and local residents (e.g. Orr 1963). Sedimentation, both modern and prehistoric, has obscured lodge sites on lower-lying landforms. Hotopp (1978b:115) reported that none of the lodges located along new U. S. 34 contained obvious surface indicators.
Archaeologists have debated the nature of Nebraska phase settlement pattern throughout the history of archaeological inquiry in the area (Perry 1998). At issue is whether the clusters of lodge represent villages, hamlets, or simply the result of replacement of single lodges during a lengthy occupation of a localized area (Anderson and Zimmerman 1976:153; Blakeslee 1990). Researchers such as Proudfit (1881b) simply assumed that clusters of lodge depressions represented villages without any objective means of verifying contemporaneity among the lodges. With reference to the large lodge cluster at Kullbom hollow, later investigators (e.g., Anderson 1961:67) were just as lax. Orr (1963) had difficulty locating lodges anywhere but on uplands, where surface artifact scatters or depressions were easily discerned. Those that Orr investigated were never closer together than about ¼ mile (400 m), and he therefore argued that the settlement pattern could not have included villages. Anderson and Zimmerman (1976), and Zimmerman (1977b), used Orr's site distribution map to argue that the settlement pattern was dispersed, and that clusters of earthlodges were much more likely to represent serial occupation of favored locations.
Perry (1998) showed that the past settlement pattern analyses were hampered by several factors. Perhaps the biggest problem, one that still plagues settlement pattern research in the region, is that of establishing contemporaneity of the lodges in a cluster. Additionally, the sample of known lodge sites used by the past researchers was very incomplete and biased in favor of dispersed upland locations. The data generated by Paul Rowe, including maps of the numerous site locations, were overlooked or not readily available. The density of Nebraska phase sites in areas subject to sedimentation was not adequately evaluated (c.f. Bettis 1990).
A thoroughly satisfactory means of establishing contemporaneity may never emerge, but cross-matching artifacts between lodges may be the best evidence one could hope for, as the excavations along U. S. 34 accomplished at two sites near the mouth of Pony Creek (Hotopp 1978a). The cross-matched sherds provided the first solid indications of a contemporaneously occupied lodge cluster in the locality. Perry (1990) used archival data from the Keyes Collection, including maps by Paul Rowe, and extensive test excavations, to locate additional buried sites near the mouth of Pony Creek. The lower Pony Creek valley was thus shown to be one of the most extensive occupational areas in the Glenwood locality. The lower Pony Creek valley data were later used in the context of Billeck's three subphase temporal model to suggest the presence of an early subphase village, and to suggest a developmental model of the settlement of the Glenwood locality (Perry 1998).
Temporal and Spatial Distributions
Anderson (1961) produced the first seriation of Glenwood locality ceramics by comparing vessel type (ware group) frequencies at sites located throughout the locality that were excavated by Orr, and later by himself. The result was the postulated three sequential phases called Keg Creek, Pony Creek, and Kullbom. Working without any direct radiocarbon evidence, a long temporal span for the Nebraska culture in Iowa was suggested, ranging roughly A. D. 800-1500. From earliest to latest, Keg Creek phase sites contained high percentages of Beckman and Swoboda wares; McVey ware predominated among the Kullbom phase sites and also contained shell tempered Oneota or Oneota-like ceramics. Sites in Pony Creek valley were intermediate between Keg Creek and Kullbom in terms of collared versus direct rim vessel frequency ratios. The analysis suggested spatial as well as temporal arrangement of Nebraska phase sites in the Glenwood locality, with the sites of the Keg Creek phase located near the mouth of Keg Creek, the Pony Creek phase sites lying in the middle reaches of Pony Creek valley, and the Kullbom phase site clustered in the Kullbom hollow and extending southward along the Missouri bluffs to the mouth of Pony Creek.
Using somewhat different seriation approaches, Brown (1967) and Zimmerman (1977a) questioned Anderson's temporal span and phase ordering. Brown's analysis of sites in the Pony Creek valley led him to suggest two subphases for the recently defined Nebraska phase, early and late. Brown disagreed with Anderson's interpretations, placing sites like Kullbom into the early subphase and those from Keg Creek into the late subphase. As dates for the Glenwood locality began to accumulate (Brown 1967; Hotopp 1978a, 1978b), the temporal span of the occupation of the region compressed, and Anderson?s phases, which were based on ceramic seriation, could not be considered valid.
The excavation of 15 sites along the new Highway 34 corridor provided a suite of radiocarbon dates for the locality which prompted Hotopp (1978a, 1978b) to reject both Anderson's and Brown's temporal phases. Billeck's (1993) analysis of the radiocarbon evidence and review of previous ceramic seriation studies led him to conclude that ceramic seriations based on vessel form typology would never yield an ordering of the sites that coincides with an ordering based on radiocarbon ages. Billeck proposed three new temporal subphases for the Glenwood locality, designated early, middle and late (Table 3).
Relying instead on detailed attribute analyses of ceramics, backed with a seriation of projectile point notching styles, Billeck found a variety of vessel form and decorative attributes that changed during the radiocarbon dated span of occupation in the locality. Those attributes included collared jar (Beckman and Swoboda wares) panel height, direct jar (McVey ware) rim height and protrusion ratio, position of tool or fingertip impressions, frequency and type of decoration on collared vessels, percentage of shell tempered vessels, and the percentage of bowl forms and sherds representing trade wares (Table 4). The attributes changed continuously over the course of the occupation of the locality, so the division between subphases based on ceramic attribute frequencies is rather arbitrary.
Ceramics of the early subphase were dominated by direct jar vessels with short, flaring rims and by collared jars with low panel and rim heights. Tool or fingertip impressed decoration was usually placed at the lip-rim juncture on direct rim jars, and collared rim jars were more frequently decorated with incising or trailing on the collar panel. During the middle and late subphases, rim heights increased and protrusion ratios on direct rim jars diminished; the rim and panel heights of collared rim vessels also increased. The use of tool and fingertip impressed decoration increased, and was placed at the lip, at the lip-rim juncture, or on the side of the rim of direct rim jars. Collared rim jars became more frequently decorated with tool impressions rather than incising. These trends are paralleled by changes in projectile point notching styles, with multi-notched points occurring more frequently during the early subphase. Middle and late subphase projectile points display fewer notches.
Billeck also noted that the early subphase sites were located in the southern part of the region near the mouth of Pony Creek. Late subphase sites were located in the northern portion of the region near the Kullbom hollow, and middle subphase sites were located along Keg Creek. This pattern suggests that occupation of the Glenwood locality moved progressively northward through time. Reasons for the south to north migration of the occupation of the locality are yet uncertain, but the temporal pattern, with its supporting artifact attribute trends, is the first workable model of the history of the numerous Nebraska phase sites in the Glenwood locality.
Perry (1998) extended the temporal model to the Glenwood locality settlement pattern, postulating the presence of a village at the mouth of Pony Creek. Pioneering Nebraska phase populations moving into the Glenwood locality from the west, where the culture was already well established (Blakeslee and Caldwell 1979:108), would have encountered the lower reaches of valleys like Pony Creek and Keg Creek first, and may well have found conditions suitable to settlement. The problem of establishing the contemporaneity of lodge occupations was minimized by noting that all of the dated bottomland sites at the mouth of Pony Creek fell into the early subphase (Billeck 1993:180), along with several others with ceramic assemblages that are within the range of the early subphase. The high concentration of sites at the mouth of Pony Creek may have been the result of replacement of several contemporaneously occupied lodges over the course of the early subphase. If so, the random arrangement of houses, typical of the Nebraska phase, may have comprised a village similar to Chang's (1958) unplanned village type.
Perry (1998) further postulated that settlement expanded into both upland and upvalley locations, with a northward progression, through the middle and late subphases. The topography of the uplands and the narrow valley floor along the upper reaches of Pony Creek precluded the establishment of closely spaced contemporaneous lodges, and the settlement pattern changed to one consisting of more dispersed homesteads.
Knowledge about the culture of a prehistoric population inhabiting earthlodges in the hills of western Mills County has grown steadily over the past century. What was once thought to be an ancient society that remained in the territory for centuries is now recognized to be a relatively recent culture that lasted a rather brief 150 years. They took advantage of the richness and diversity of some of the most scenic landscapes in the Midwest. We don't know yet why they left the Glenwood locality. But during their stay they left behind a record of their activities indicating the considerable complexity of their lives. That record has captured the attention of both archaeologists and non-archaeologists, resulting in scholarly debate, museum construction, earthlodge replication, grade school study programs, and dozens of reports, articles, and books in both the popular and scientific realms.
Additional Information on Prehistoric Iowa
Mary De La Garza, Web Master