...., PREHISTORIC SETTLEMENT PATTERNS IN SOUTHWEST OREGON by KATHRYN R. WINTHROP A DISSERTATION Presented to the Department of Anthropology and the Graduate School of the University of Oregon in partial ful'fillment of the requirements for the degree of Doctor of Philosophy December 1993 "Prehistoric Settlement Patterns in Southwest Oregon," a dissertation prepared by Kathryn R. Winthrop in partial fulfillment of the requirements for the Doctor of Philosophy degree in the Department of Anthropology. This dissertation has been approved and accepted by: //- :J-(,- 73 Date ii Committee in charge: D. Melvin Aikens, Chair Don E. Dumond Ann Simonds Patricia F. McDowell J.__ tCopyright 1993 Kathryn R. Winthrop iii An Abstract of the Dissertation of iv for the degree ofKathryn R. Winthrop in the Department of Anthropology to be taken Doctor of Philosophy December 1993 Title: PREHISTORIC SEn-LEMENT PATfERNS IN SOUTHWEST OREGON Approved: C. Melvin Aikens This study addresses the problem of prehistoric culture change in interior southwest Oregon as reflected in subsistence/settlement patterns. Eighty-three sites, excavated during cultural resource management projects, constitute the database. This study also demonstrates the applicability of cultural resource management data to questions of regional interest and of general importance to anthropology. Two contrasting subsistence/settlement regimes are modeled based on regional ethnographic and archaeological studies. One pattern is that of a mobile subsistence regime; the other is that of a more sedentary regime associated with permanent villages and the collection and processing of foods for over-winter storage. The first is reflected in the archaeological record by a settlement system consisting of seasonal camps and short-term task sites; the second is represented by a settlement system consisting of villages, seasonal camps, and task sites. To test these models against available data, sites were 1,000UJ ....J UJ 500 Alpine Zone Tsuga mertensiana Zone Abies magnifica shastensis Zone Abies coneD/or Zone CASCADE RANGE ROGUE RIVER VALLEY 23 FIGURE 2. Arrangement of Vegetation Zones in Southwestern Oregon (from Franklin and Dyrness 1988:131). p24 species (Franklin and Dyrness 1988:119). Oak woodlands vary from open savannahs of scattered oaks with grass understories to dense oak forests and mixed forests of oaks and coniferous species. The two main species of oak are Quercus kelloggii (California black oak) and Quercus garryana (Oregon white oak); associated conifers are most commonly Pseudotsuga menziessii (Douglas-fir) and Pinus ponderosa (ponderosa pine) (Franklin and Dyrness 1988:114-115). Chaparral, consisting of drought tolerant shrubs such as Ceanothus (buckbrush) and Arctostaphulos spp. (manzanita) occurs as a sub-climax species in the Interior Valley Zone. Today, these communities represent the northernmost extension of chaparral vegetation. The chaparral shrubs are frequently associated with oak, Douglas-fir, and ponderosa pine. Chaparral communities are indicative of warm, dry conditions, and may be dependent upon fire for their existence (Detling 1961 :356). The Interior Valley Zone also includes forests of hardwoods and conifers on the foothills surrounding the valleys. In addition to the oaks, Acer macrophyllum (bigleaf maple), and Arbutus mensiesii (Pacific madrone) are associated with Pseudotsuga mensiesii (Douglas-fir), Pinus ponderosa (ponderosa pine), and Abies grandis (grand fir) in these foothill communities (Franklin and Dyrness 1988:116). The Interior Valley Zone plant communities were especially important to native peoples, furnishing major dietary staples such as acorns and camas, a variety of other seeds and roots, and abundant forage for game. The grasslands and meadows, together with the open oak savannah and 25 pine/oak woodland, were in part a product of human interaction with the environment; cycles of regular burning carefully regulated by the native peoples kept the landscape open and promoted the growth of these plant communities (Boag 1992; Franklin and Dyrness 1988:122; Martinez 1993). Frequent fires also contributed to the growth of chaparral communities, which provided browse for game animals. Fire also increased the amount of area transitional between different vegetational zones, such as the savannah and forests; these transitional zones are places of increased biological diversity and especially productive of foods and materials used by native peoples (Boag 1992:21). Elevations above the Interior Valley Zone are characterized by forest communities which change with increasing altitude. The Mixed Conifer Zone occurs immediately above the Interior Valley Zone, with Pseutotsuga menziesii (Douglas-fir), Pinus lambertiana (sugar pine), Pinus ponderosa (ponderosa pine), (Libocedrus decurrens (incense cedar), and Abies concolor (white fir) as dominant species. Above these forests, a narrow belt of forest dominated by Abies concolor (white fir) occurs below the Abies magnifica shastensis (Shasta red fir) forests. At the highest elevations, Tsuga martensiana (mountain hemlock) and alpine vegetation succeed the fir forests. Wet and dry meadows and brushfields, important to native peoples for foods and game forage (e.g., Snyder 1987), provide openings in the forests at all elevations and occur in the valleys as well. These communities occur as a result of a number of factors, such as shallow soils, damp soils, and • T 26 normal successional patterns following fire or other disturbances. A number of varieties of berries and root crops, such as huckleberries and camas, are associated with these meadows and were of particular importance to native peoples. Major game species, such as deer and elk, are also attracted to these forest openings, which provide forage with nearby cover 'from predators. The fauna of the area includes large ungulates which browse on the forage found at the forest edge, as well as a variety of predators and small mammals. Black-tailed deer (Odocoileus columbianus columbianus) are the most common ungulate species, with limited numbers of white-tailed deer (Odocoileus virginianus leucrus) occurring along the Umpqua River (Mace and Smith 1970). Roosevelt elk (Cervus canadensis) also inhabit the area. Other large mammals, such as black bears, cougars, bob-cats and coyotes still roam the area; formerly the grizzly bear, bighorn sheep, antelope, and wolves also lived in the southwest mountains. Numerous smaller mammals, such as beaver, otters, rabbits, and squirrels, furnished prey species for food and materials. Game birds such as quail and grouse also occupy the valleys and foothills. Many of these birds and mammals were seasonally available to native peoples at different locations within the landscape. Deer and elk especially follow a migratory pattern, congregating in low elevation ranges dUring the winters, and dispersing into the uplands during the spring and summer as forage becomes available. 'I" 27 The rivers and streams provided a major staple to most of the people living in the research area (see Figure 3). Anadromous fish, consisting primarily of coho and chinook salmon, and steelhead, formed a significant part of the diet. These fish migrate annually up the rivers and their tributaries to spawn. These annual migrations, or "runs," provide excellent fishing opportunities, and occur every season of the year (Johnson 1993, personal communication; Loomis 1993, personal communication). On the Rogue River, spring chinook run from March to June; summer chinook from June to August, fall chinook from September to October or November; coho salmon from September to December; summer steelhead from March to October; and winter steelhead from October to March. On the North Umpqua River, Coho salmon run from September to January; spring chinook from March until August; fall chinook from August to December (today, this is a light run); winter steelhead from December to May, and summer steelhead from May to October. On the South Umpqua, coho run from September to January; spring chinook from March to August (this is a light run); fall chinook from August to December; and winter steelhead from December to May. There is no summer steelhead run today on the South Umpqua. Other fishes, such as trout, sculpin, lamprey, and dace, also occur in these rivers and their tributaries. It is difficult to assess the abundance of aboriginal fisheries. Current researchers agree that anadromous fish today have been severely affected by a number of historic factors, and that current fish runs are well below early historic levels. These factors include: over-fishing; habitat destruction \1O~'TH 28 RCNS I F A M I I A S o D ROGL'E RIVER Coho Sp. Ch. S. Ch F.Ch. S. Sthd. W.Sthd. NORTH lJMPQUA RIVER Coho Sp. Ch. F. Ch. S. Sthd W. Sthd. SOUTH UMPQUA RIVER Coho Sp. Ch. F.Ch. W.Sthd Coho co S.Ch. = Sp. Ch. = F.Ch. S. Sthd. = W. Sthd. co Coho salmon Summer chinook Spring chinook Fall chinook Summer steelh.ead Winter steelhead b FIGURE 3. Anadromous Fish Migration Periods (Johnson 1993: Loomis 1993). pi T 29 through mining, logging, and road-building; pollution of water; blockage of migratory patterns with dams; and a number ofother effects attendant upon our modern way of life (FEMAT 1993:11-36-37; Kaczynski and Palmisano 1993; Netboy 1973). Anecdotal evidence from settlers and early fishery reports attest to the earlier richness of these streams, although these late nineteenth century reports may reflect a temporary surge in fish populations after the elimination or restriction of Indian harvests (Hewes 1947; 1973). Whatever the historic effects have been, however, fish runs in the Umpqua and Rogue Rivers and their tributaries at the time of historic contact were both abundant and predictable, and were a major source of food for the people who lived along them. The climate, topography, and vegetation created a seasonally varying resource base for the prehistoric inhabitants, with staple foods available throughout most of the year at different places and elevations. The rugged mountains hosted large game which migrated during the winter to lower elevations, returning to the uplands as the weather warmed. Berries and root crops--especially camas--provided spring, summer, and fall resources in both upland and lowland meadows. At lower elevations, oaks and grasses such as tarweed provided foods available in the warmer months. Major fish runs occurred in every season, at least in the North Umpqua and the Rogue Rivers, and, until recently, fish were available year-round in the rivers and streams (Spencer 1991 :vii). The topography of the area and the seasonality of the resources placed constraints upon the hunters and gatherers who lived V I h 30 here: adequate year-round provisioning required people to move among the various resources as they became available. .Palaeoenvironments During the time that people have inhabited southwest Oregon the global climate has experienced major shifts which have been expressed differently in different regions. At present, there is little direct information regarding either the climate or the effects of climatic change on vegetation, fauna, and hydrology for interior southwest Oregon. There is more information in adjacent regions, which suggests models for this area. Work in the Pacific Northwest has identified a climatic sequence with relevance to this project's stUdy area (Thompson, Whitlock, Bartlein, Harrison, and Spaulding 1993). Following deglaciation between about 14,000 BP and 10,000 Bp1, the northern hemisphere experienced an amplified seasonal cycle of solar radiation lasting until about 6,000 years ago. During this time solar radiation was greater and winter radiation was less than today, resulting in increased temperatures and decreased effective moisture (Whitlock 1992:16). This was a warm, dry interval, indicated by the expansion of open forests or savannahs into the Puget Trough in Washington. This vegetation, with Pseudotsuga (Douglas-fir), Alnus (alder), and Pteridium, as well as Quercus (oak), Chrysolepis (chinkapin), and herbs, was similar to vegetation now characteristic of the Willamette Valley in Oregon. After about 5,000 or 6,000 years ago, this xeric period began to 1BP = Before Present. m31 moderate. Throughout the region, modern vegetation patterns began to appear. This climatic scheme is summarized as follows: 14,000 - 10,000 BP: Appearance of temperate taxa during Deglaciation. 10,000 - 5,000 BP: Introduction of xerothermic communities in the early Holocene. 5,000 BP - Present: Establishment of modem vegetation patterns in the late Holocene. A small amount of research in the central part of western Oregon has been conducted to date, consisting of a pollen core from Gold lake Bog (east of Cottage Grove) and a core from Indian Prairie Fen (east of Eugene), both in the Western Cascades and north of the study area (Sea n.d.). Another core comes from little lake, in the central Coast Range (Worona 1993). These three studies support the general pattern outlined above, indicating a warmer and possibly drier interval during the early Holocene, followed by a wetter and cooler climate in the later Holocene. At Indian Prairie Fen, development of the present-day forest occurs after 6900 BP; at little lake on the coast this appears after 5,000 BP. The critical elements of this Holocene climatic scenario are the contrasts between a xeric early Holocene period with more limited moisture and higher summer temperatures than today with the more mesic--wetter and cooler--conditions which follow. These mesic conditions characterize the modern environment. The earlier xeric conditions produced distributions of plant and animal species different from those typical of this region today. Both the nature of the xeric and mesic Holocene periods as well as the 32 timing of the transition between them provide important background to the present study. The timing of the transition from the xeric to the mesic period fluctuated throughout the American West. There is as yet little direct evidence for southwestern Oregon; it is assumed that this region would fit the broad patterns characterizing more studied areas in western Washington or California. These patterns, however, do not define a consistent theme readily applicable to southwest Oregon. For example, one interpretation of fossil pollen data for a location on the northern Washington coast (Hoh- Kalaloch) notes that the warmest was prior interval to 8,000 BP, followed by colder and wetter conditions which peaked between 5,000 to 2,000 BP (Heusser 1985:160). On Mt. Rainier in Washington, however, warmer conditions prevailed into the fourth millennium BP, and modern vegetation was established only after 3,500 BP (Whitlock 1992:18). Similar fluctuations are apparent to the south. Analysis of fossil pollen from Clear Lake, for example, in coastal, central California, together with fossil pollen from an off- shore ocean core, suggest a period of disequilibrium between 15,000 and 5,000 years ago, with essentially modern characteristics established by about 7,000 BP (Baker 1983:118; Gardner, Heusser, Quinterno, Stone, Barron, and Poore 1988:181). Pollen from Osgood Swamp in the western Sierras, however, indicates a warmer mid-Holocene climate lasting until about 2800 BP (Baker 1983:118). A recent discussion of climate change in the American West during the last 18,000 years models climate at three-thousand year intervals. This '7 effort synthesizes information from numerous diverse sources, including general circulation model simulations based on various boundary conditions (e.g., the presence and size of the ice sheet during the earlier periods), and direct evidence of past climates derived primarily 'from pollen studies, plant macro-fossils from pack-rat middens, and lake-level records. At 9,000 BP the climate in the Pacific Northwest was at its driest, based on data from Washington and the Columbia Basin. In California, similarly arid conditions prevailed in the Sierras as well as along the coast. By 6,000 BP, drought conditions were less severe than earlier but still more arid than at present, both in the Pacific Northwest (Washington and the Columbia Plateau) and California. Additional data from Clear Lake (i.e., growth-increment widths from fossil tule perch scales) in California suggest that warmer conditions also prevailed here. Although this climate study does not present data for 3,000 BP in any detail, a map showing periods of maximum effective moisture shows the region west of the Cascades and Sierras as reaching this peak 3,000 years ago (Thompson et al. 1993:Fig.18.14). Although there is as yet little direct evidence for conditions in southwest Oregon, comparison with gross climatic conditions in northwest Washington and California imply the following scenario. The warmest and driest period was at about 9,000 years ago. This period lasted at least until 6,000 years ago, at which time the climate was still drier than at present. Some time after 6,000 years ago, the climate began to turn cooler and moister. By about 3,000 years ago, the climate had reached maximum effective moisture. 33 34 The possible effects of these climatic shifts on those aspects of the environment which were important to the native peoples have not been analyzed for southwest Oregon, due to the absence of direct evidence for past climate and vegetation patterns. It is possible, however, to make some inferences regarding these effects, and there are several studies which model past conditions based on knowledge of current environments. These studies and inferences allow discussion of possible changes in the environment in southwest Oregon during the Holocene. For areas such as interior southwest Oregon, where today many species are at their physiologic limits (such as the California black oak), the past environment is particularly difficult to predict. The vegetation patterns in the Pacific Northwest consist of "loose associations composed of species independently adjusting their ranges to environmental changes on various time scales" (Whitlock 1992:22). Simple zonal shifts of intact communities of plants and animals, either altitudinally or latitudinally, were unlikely. That is, the specific constellations of plants and animals present today probably did not migrate to higher elevations or more northerly regions as the climate warmed. Rather, certain species migrated, and others disappeared, producing con'figurations of plants and animals which are somewhat different than today. Furthermore, during the early warm and dry interval, wildfire was probably more common; fires would have positively affected those species which are fire tolerant or dependent, such as oak and chaparral communities, and placed further stress on those which are not, such as conifers (Detling 1961 ). •35 In southwestern Oregon, already warm and dry in comparison with the rest of the Northwest, the warm interval of the early Holocene would probably have seen the coniferous forests d,iminish and much of the cold zone hemlock forests eliminated (Franklin, Swanson, Harmon, Perry, Spies, Dale, McKee, Ferrell, Means, Gregory, Lattin, Schowalter, and Larsen 1991 :243). Elimination of the cold zone also implies changes in hydrology, with diminished snow-packs. Non-forest type ecosystems, today represented by grasslands, chaparral, and oak savannah, probably covered a greater part of the study area than today. The specific effects of these palaeoclimatic changes on the human environment are difficult to gauge. Several models exist, however, which predict conditions for the early to mid-Holocene xeric period. These models examine the effects of environmental constraints on resource productivity, and then predict the potential effects of environmental changes on those resources. The potential effects of a warmer, drier climatic regime on anadromous fish, other game resources, and staple plants such as oak and camas, as expressed in these studies, are summarized below. Recent attempts to predict changes to anadromous fish runs in the event of global warming draw upon the warm period of the Holocene as a model (Chatters, Neitzel, Scott, and Shankle 1991; Neitzel, Scott, Shankle, and Chatters 1991). These studies are particular to the Columbia Basin fisheries; another study concerning salmon in the Rogue River drainage also presents a model for changes in fish populations during this warmer interval ·r, .. ~J 36 (Spencer 1991). These works suggest that significant effects on the fish populations were possible. One study for the Columbia River Basin provides an assessment of the effects of a warmer and drier climate on fish populations (Neitzel et aI. 1991). Based on palaeoenvironmental studies, the authors estimate a 1-2 degree centigrade increase in temperature, with a decrease in effective moisture of between 33-38 percent during the xeric period. They assess the effects of increased temperature and decreased precipitation on four hydrologic variables which affect salmon production: duration of peak flow, amount of sedimentation, stream temperature, and annual flow (decrease or increase in annual surface runoff). For the Columbia basin as a whole, they conclude that the climate changes estimated would not adversely affect the rivers and streams west of the Cascades, in terms of their ability to support anadromous fish, but would have a generally detrimental effect on streams east of the Cascades. They note that proximity to the ocean as well as differences in vegetation patterns and hydrologic regimes account in part for the differences postulated between the east and west Cascade streams. They also conclude, however, that changes in climate would affect various species of fish differently. Spring and summer chinook would be affected negatively in most streams, for example, due to changes in timing and volume of the spring freshet (peak flow). Steelhead, however, have a tolerance for warmer water and intermittent streams, and would have been unaffected or possibly helped in certain areas. ¥f' ,... ! 37 In another study these researchers examine the effects of conditions resembling the warm, dry Holocene interval on spring Chinook in the Yakima River. They conclude that a climate change to such xeric conditions could significantly reduce current fish runs. They base their conclusions on a computer model which calculates the effects of climate change (using the same figures for temperature and effective moisture noted above) on hydrologic variables, and the effects of changes in these variables on fish survival. Speci'fically, they model the effects of changes in stream temperature, sedimentation, 'flow volume, and timing of the peak annual flow on three critical life stages of anadromous 'fish and a stream's capacity to produce juveniles (smolt capacity). The important life stages are: egg-to- smolt survival rate; smolt-to-smolt survival rate; pre-spawning (adult fish) survival rate. (A smolt is a young fish ready to migrate to the ocean from the home stream.) The computer model also calculates cumulative survival over several generations. Changes in the hydrologic variables would have the following effects on fish. Increased changes in water temperatures decrease pre-spawning survival by increasing the incidence of disease in adult fish. Higher sedimentation causes low egg-smolt survival; however, climate induced changes in sedimentation would vary with stream gradient and watershed type. Changes in the timing of the annual peak flow would adversely affect smolt-smolt survival, and would be most likely to affect upstream areas. Finally, smolt capacity depends upon stream volume; a 33 percent reduction in volume correspondingly reduces smolt capacity, except in those streams p. ..,. I 38 which become intermittent and loose their smolt capacity entirely. The authors conclude that a climate change such as that modeled for the early Holocene (Le., a 2 degree centigrade increase in temperature and a 33 percent decrease in precipitation) could reduce spring chinook salmon production by 60 percent (Chatters et al. 1991). Both of the Columbia Basin studies stress the complexity of 'the factors which interact to provide good habitat for anadromous fish. It would be inappropriate to apply the findings of these studies for the Columbia Basin directly to southwest Oregon, where the present-day climate, vegetation, and stream environments are different. These studies do, however, highlight the possibility of significant differences in anadromous fish populations during the xeric interval of the Holocene. In a less elegant but equally intriguing paper, Spencer (1991) analyzes the possible effects of the warm, dry interval on the Rogue River and its tributaries, and hypothesizes effects on the salmon inhabiting these streams. According to Spencer, a lesser snowpack during this period would have produced a peak-flow period in the winter, rather than the spring. Lower stream flows in the spring and summer, in turn, made steeper gradients and low falls effective barriers to migrating fish. Furthermore, some streams which are perennial today would have been intermittent, further limiting salmon populations. Warmer s'tream temperatures would also have inhibited salmon populations. The upper reaches of the Rogue and its tributaries already are at the further end of the anadromous fish migration routes, where runs are less abundant and fish more e'xhausted than in those areas closer 7 R 39 to the coast. The postulated climatic effects--warmer waters, reduced seasonal flows, effective migration barriers, intermittent flows--could have had a greater effect on fish in the upper Rogue River drainage than for areas closer to the coast. Overall, under this scenario, a possible effect on fish of a more xeric climate would 'have been to limit major runs to the winter months, and to limit the geographic extent of those runs, particularly at upper reaches of the major rivers and tributary streams. Another recent study, by Nan Hannon (1992), constructs a model for prehistoric availability of critical resources based on the study of plants today and inferences regarding past conditions. Hannon argues that during the early Holocene xeric period oaks expanded, but their productivity was low. Camas may have disappeared from valley floors, and the major plant species available to people were seeds from various grasses. Chaparral expanded and, together with oak, provided increased forage for deer, elk, and other mammals such as rabbits and squirrels. She argues that this xeric period would have fostered a highly mobile subsistence regime focused more on hunting and less on the acquisition of valuable plant foods such as acorns and camas, which were not as available as during later times. Grass seeds may have supplemented the diet, but may not have been used as staple foods. Based on her ten year study of acorn production, Hannon argues that acorns were not likely to have ever been an abundant and predictable crop in interior southwest Oregon. In southwestern Oregon, two main species of oak have nutritionally valuable acorns: 'the Oregon white oak and California ---""fpp---- 40 black oak. Black oak was the preferred species, producing an acorn that has a higher fat content than that of the white oak. However, the black oak is at the northern limit of its range in southwest Oregon, and is neither an abundant nor reliable producer. Yields may fluctuate widely from year to year or place to place. Though both oaks are moderately drought and 'fire resistant, the black oak appears to need more water than the white oak to be a good producer; in the drought years of the 1980s and early 1990s, the black oaks in Hannon's study were poor producers. The white oaks observed in the study were more consistent producers, but production was nonetheless highly variable. During the recent drought years, monitored acorn production from white oaks in natural settings was abundant only one year out of five (Hannon 1993, personal communication). Both white oaks and black oaks were better producers in swales or near irrigated areas, where they received additional moisture, indicating that drought may affect acorn production in both species. Furthermore, though the white oak is drought and fire resistant, it nonetheless needs moisture to establish seedlings and may therefore have been restricted to riparian zones and north-facing slopes during the xeric interval. Hannon concludes that black oaks were unlikely to do well during the xeric phase. White oak may have expanded its range, especially as conifers retreated, but these trees may have been restricted to specific locations and were not necessarily reliable or abundant producers of acorns. Camas, formerly abundant in the meadows of the interior valleys (Le., prior to modern agricultural practices), provided a significant carbohydrate to rJr _ 41 the diets of prehistoric peoples. Camas today grows in moist meadows, and is most productively harvested in those areas where the ground is damp most of the year. Hannon's experiments harvesting this crop showed that in moist areas the camas bulb grows closer to the surface, and is easier to dig. She argues that the hot summer regime of the xeric period would have effectively eliminated camas as a food crop from the dry, low-lying valleys, which even today experience high summer temperatures. Camas may have been available at higher elevations, however, where damp meadows existed. Although these significant plant foods may have been restricted, the xeric climate of the ear1y Holocene may have enhanced the availability of certain game species. Open environments, such as grasslands, wet and dry meadows, oak savannah, and chaparral communities were probably more characteristic of this period. These environments provide browse, seeds, nuts, and cover for numerous game species, including deer and elk: rabbits, squirrels, and other small mammals: and birds such as grouse. Although high value, easily processed vegetable foods may have been limited during this period compared to later times, high value game species--which feed on vegetation less appealing to humans--may in fact have flourished. If the above inferences are valid, the xeric climate of the ear1y Holocene would have produced a different distribution of staple resources than the mesic climate of the later Holocene. The Interior Valley Zone's biota would have expanded, with larger areas of grassland and chaparral. Oaks may have replaced conifers at the valley edges. Fisheries may have been more restricted than they were later, with abundant runs only during the crtz---- 42 winter season. Acorn production may have been important. but under conditions of moisture stress minor fluctuations in local moisture regimes may have contributed to unpredictable and annually fluctuating harvests. Camas may have been restricted to moist meadows at higher elevations. Game animals may h,ave flourished. Winters were not as long or harsh. and use of the uplands was possible from earlier in the spring until later in the fall than during later more mesic times. both for game animals and their human predators. The transition to a more mesic interval brought about changes significant to the prehistoric inhabitants. Additional rainfall led to an expansion of oak trees and acorn production; anthropogenic burning maintained the oak woodlands. and was necessary to keep conifers from encroaching upon the oaks. Cooler and damper conditions fostered growth of camas at lower elevations. More rainfall and a winter snowpack contributed to better conditions for anadromous 'fish. Harsher winters and heavier snows also kept people at lower elevations for longer periods during the year. Coniferous forests expanded, possibly limiting the lower elevation habitat beneficial to those game species important to people. Summary The environment of interior southwest Oregon at the time of historic contact promoted a seasonal round of subsistence activities, in which native people provisioned themselves from the resources available at different times and places throughout the year. The abundance and predictability of ___..tz _ 43 anadromous fish runs throughout the year provided a stable resource. Availability of fish was complemented by acorn harvests from low-elevation oak groves probably maintained against colonization and replacement by conifers through anthropogenic burning. Cold and wet winters restricted human movement into the uplands, but also drove important game animals to low elevations. As elsewhere in the Pacific Northwest, these factors operated to promote a semi-sedentary way of life, with stable winter settlements along fish-bearing streams and summer camps at locations of seasonally abundant foods in both the lowlands and the uplands. Gross differences in the climate of the early Holocene engendered a different constellation of resources available to the prehistoric inhabitants, and may have influenced different ways of life. A xeric interval in the early to mid-Holocene may have limited the availability of staple foods such as anadromous fish, acorns, and camas, but may have permitted movement throughout the countryside for longer periods during the year, due to milder or shorter winters. It may also have enhanced the availability of game throughout the year. The timing of the transition from an earlier xeric period to a later more mesic one is as yet unclear, but it probably occurred sometime between 6,000 and 3,000 years ago. In addressing the question of culture change as seen in the subsistence and settlement patterns of prehistoric inhabitants, the environmental context assumes great importance. Hunter-gatherer economies are inextricably tied to the resources available in their local territories. The potential resources of those localities, as well as --",tp_-..~ environmental changes within them, provide the most basic explanations of subsistence/settlement patterns and changes within those patterns. Other factors which condition these cultural configurations may have great importance, but environmental possibilities and constraints are fundamental. 44 .Jz __ 45 CHAPTER III DEFINITION OF SUBSISTENCE AND SETrLEMENT PATrERN MODELS Introduction In addressing the question of culture change in prehistoric southwest Oregon, it is useful to examine contrasting modes of subsistence and settlement established for hunter-gatherer societies in general. Local ethnographic and archaeological work can then assist in determining how such contrasts apply to this area. The task of this chapter is to review these contrasting subsistence/settlement modes, and to use the ethnographic and archaeological evidence to formulate models expressing these contrasts which are appropriate to this region and discernible in the archaeological record. Hence, the intent of this chapter is to present two alternative models for prehistoric subsistence/settlement systems in southwest Oregon, based on distinctions generally recognized in hunter-gatherer societies. The first pattern, termed here the "Collector Model," represents a more sedentary way of life in which people established themselves at permanent villages for at least the winter months, and at which they stored foods collected and processed throughout the year for use during that time. The people living in this area at the time of historic contact followed this way of life, and descriptions of their way of life help define and identify the archaeological 46 elements of this regime. The contrasting model, here termed the "Mobile Model" represents a hypothetical pattern in which people led a mobile existence, moving among resources as they became available, and relatively independent of collection and processing of foods for provisions over winter. Previous archaeological work in this region permits the hypothesis that this pattern existed early in this area's prehistory. Differences in hunter-gatherer subsistence/settlement modes are reviewed below, as are the ethnographic and archaeological data, in order to define the expression of these two models in this area and the types of sites of which they are constituted. The chapters following this discussion focus on the methods used to discern these subsistence/settlement patterns in the archaeological record. Subsistence/Settlement Contrasts in Hunter-Gatherer Societies The archaeological analyses presented in subsequent chapters reveal a difference between the subsistence/settlement systems of the earlier and later prehistoric periods. In order to interpret these distinctions, it is useful to review contrasts noted for hunter-gatherer subsistence/settlement systems more generally. The contrasts reviewed here provide a specific framework for interpreting the archaeological record. and for developing the two models used in this study. Distinctions in hunter-gatherer subsistence/settlement systems are sometimes expressed as differences in mobility/sedentism and in the degree of reliance on processed and stored foods. Three examples of such r.··z " I 47 distinctions provide a basis for analyzing the local materials: Bettinger and Baumhbffs (1982) traveler/processor distinctions; Woodburn's (1988) immediate versus delayed-return conceptions; and the forager/collector contrasts used by Binford (1980). Each of these three examples arise from different purposes and have different research orientations, but they share a common perspective in recognizing mobility, intensive use of resources, and storage, as key elements in contrasting hunter-gatherer systems. Bettinger and Baumhoff (1982) describe two hunter-gatherer systems which are distinguished from one another on the basis of the intensity of resource use, resulting in differences in subsistence mobility and in processing and storage of foods. Travelers, they argue, focus on "high quality" foods--such as game--which do not take much time and effort to process, but which do take time and effort to procure. Such groups move their camps frequently, and send hunting groups out from camps for long distances to procure these high-return items. Processors, however, focus on foods--such as seeds--which take considerable time and energy to obtain and make palatable, but which do not require as much time and energy to locate. Processors also use high return items, resulting in a broader subsistence base and more intensive subsistence strategy than travelers. Since they focus on labor intensive items, processor groups are not as mobile in the subsistence quest. In attempting to explain culture change in the Great Basin, Bettinger and Baumhoff (1982) argue that the processor strategy will out-compete the traveler one, when the two regimes come in contact. Processors have larger III"-------------------r 48 populations and eat not only what travelers do, but other foods as well, giving them a competitive advantage over travelers. Bettinger and Baumhoff also argue that cultural differences would make it difficult for a traveler society to shift rapidly into a processor mode, even when faced with competition from such groups. Woodburn approaches the differences between hunter-gatherer societies from a different orientation. Working primarily with modern hunter- gatherers, Woodburn expresses major differences among hunter-gatherers as differences in understanding and intention. Immediate-return societies are those in which activities are focused on the present; delayed-return societies are those in which activities are oriented to the past and future as well as to the present. More fully expressed, immediate-return systems are those in which "people deploy their labor to obtain food and other resources which will be used on the day they are obtained or casually over the days that follow"; have "simple, portable, utilitarian, easily acquired, replaceable tools and weapons," and are not dependent upon assets which have delayed yields based on labor invested (Woodburn 1988:32). In delayed-return systems, however, people do hold assets which provide a return on their labor (Woodburn includes some hunter-gatherers and all other societies in this system). There are four main types of such assets for hunter-gatherers, often found together in mutually reinforcing arrangements (Woodburn 1988:32): (1) Valuable technical facilities used in production: boats, nets, artificial weirs, stockades, pit-traps, beehives and other such artefacts which are a product of considerable labour and from which a food yield is obtained gradually over a period of months or years. ~.. c , k T.. ·· ~J ---rs4Irnn--- 49 (2) Processed and stored food or materials usually in fixed dwellings. (3) Wild products which have themselves been improved or increased by human labour: wild herds which are culled selectively, wild food- producing pla.nts which have been tended, and so on. (4) Assets in the form of rights held by men over their female kin who are then bestowed in marriage on other men. Immediate versus delayed return systems are reinforced and further differentiated by a number of correlated aspects of the social organization of each. Immediate return systems, for example, have flexible social groupings which change constantly in composition; social relationships stress sharing and mutuality, resulting in leveling mechanisms in terms of accumulation of wealth; social relations do not include long-term, binding commitments; and distinctions in wealth, power, and status are consequently eliminated. Delayed return systems depend upon "binding commitments and dependencies between people" in order for people to "build up, secure, protect, manage and transmit the delayed yields on labour" or other assets which are part of a delayed-yield system (Woodburn 1988:33). Binford (1980) offers a third example of hunter-gatherer subsistence contrasts. He uses the concepts of foragers and collectors to explain variation in the ethnographic and archaeological record in hunting-gathering societies. Forager societies are those which "map on" to resources, moving people among different resources to obtain their subsistence needs. These groups do not engage heavily in storage of seasonally available foods, but rather circulate, often through large territories, on an annual foraging round. Collectors, in contrast, do have a stable home base where foods are collected and stored, provisioning a more sedentary way of life. Such groups are logistically organized; they send specially organized task groups to --_...._--- 50 resource patches where foods are processed and brought back to the home base. Binford presents these concepts not as stark contrasts, but as concepts which can help explain the variation evident in hunter-gatherer societies, many of which employ both strategies at various times, depending on local circumstances. Binford presents these differing subsistence strategies as strongly correlated to environmental constraints. He argues that conditions which constrain mobility foster a collector strategy. A seasonal climate, for example, constrains mobility due to weather factors, and gives rise to temporal (seasonal) and spatial differences in the availability of resources. Storage becomes necessary to meet subsistence needs for at least part of the year; stored foods in turn decrease a group's options for mobility by tethering them to the place of storage. Furthermore, a seasonal climate produces a variety of desired resources available more or less simultaneously at different places, but only during part of the year. In order to harvest all desired resources, it thus becomes necessary for a group to carefUlly plan and organize its subsistence strategies, with members of the larger group frequently engaging in different tasks. Foragers, however, are typically found in the tropics, where seasonal limits are not as pronounced and where resources are more spatially and temporally homogeneous. Under these conditions foragers simply move from place to place, meeting subsistence needs until the surrounding territory is depleted and a new camp is made. I " 51 The three sets of contrasts just presented are not entirely equivalent, yet for the purposes of this study it is possible to derive two altemative patterns which assist in the definition of subsistence/settlement models for prehistoric southwest Oregon. Processors, delayed-return systems, and collectors all share an emphasis on processed and stored foods and a more sedentary settlement regime. These types share a number of other characteristics which stem from these factors, such as higher densities of population, non-portable facilities which represent an investment in time and labor, and carefully organized strategies for food procurement accompanied by labor intensive food processing and "binding" social ties. Woodburn's enumeration of assets (listed above) is especially interesting from an archaeological view, since all but the last are potentially visible in the archaeological record. Where such things are found, therefore, a delayed- return system is indicated. Translated into an archaeological idiom, such groups would have stable villages, with substantial architecture, storage facilities, tools for processing, and possibly distinctions in wealth and status evident among the people. In contrast to the above, travelers, people engaged in immediate- return systems, and foragers follow a more mobile existence, generally unencumbered by the accoutrements of a group that is dependent upon processed and stored foods. Such mobile groups have smaller populations, move frequently about the countryside, and are unlikely to display great distinctions in wealth or to invest in substantial facilities for housing or storage. Archaeologically, such groups ·would lack the stable settlements 52 postulated for the alternative regime, as well as the tools and facilities associated with a heavy dependence on processed and stored foods. Settlements would be small, relatively temporary, and frequently moved. Tools would be useful and expedient, and wealth items would not be particularly important. In the review that follows, the ethnographic record demonstrates the existence in southwest Oregon of the processor/delayed-return/collector mode (here termed "collector mode" for simplicity). Ethnographic peoples followed a way of life with the distinguishing hallmarks of the collector mode: stable communities with a significant investment in labor; processed and stored foods; limited mobility with movements tied to a central base; and a labor intensive subsistence regime which required centralized planning and logistical organization. In southwest Oregon, the particular variant of this pattern was expressed in a semi-sedentary subsistence/settlement pattern in which the stable home base was occupied for part of the year, with the remainder of the year devoted to forays aimed at obtaining, processing, and storing foods for the winter. If there was a difference between the way of life expressed in the ethnographic record and that of an earlier time, as argued here, then the contrast to the collector regime poses the likely altemative for the earlier period. Here termed the "mobile" pattern, this alternative hypothetically consists of a subsistence/settlement regime with the following characteristics: small, mobile groups which do not depend heavily upon processed and ___.A.nn _ r__A.nrz _ 53 stored foods, with no major investments in architecture or facilities, with fluid social organizations and little distinction in wealth and status. The above contrasts are illustrated in the ethnographic record and expressed in previous archaeological work for the study area. The following sections review the ethnography and previous archaeology for this area, as sources for the collector/mobile models used to interpret the archaeological materials in the subsequent analyses. The Collector Model: Ethnographic Example Ethnographic research complements archaeological studies in many ways. It provides a deeper understanding of the cultural reality of which the archaeological materials were a part, and in this study gives specificity to the concept of a collector subsistence/settlement pattern. The ethnographic evidence available for the people living in the study area supports the inference that these groups participated in a collector regime. This information also assists in the definition of archaeological site types which characterized that regime, and helps describe the archaeological assemblages and features which identify these types. The following review introduces the diverse groups who lived in the study area, and presents information from the ethnographic record which is directly relevant to the description of subsistence/settlement patterns. The ethnographic record for interior southwest Oregon is limited, and scattered among ethnographic summaries and notes, historic accounts, oral histories, and recent analyses and summaries (e.g., Beckham 1971, 1983a, ~.. --_..._--~ 54 1983b, 1986; Beckham and Minor 1992; Gray 1987; Kendall 1990; Lalande 1989; Miller and Seaborg 1990; O'Neill 1989b; R. Winthrop 1993). While it is possible to use this material to illustrate different parts of the collector regime, such as descriptions of winter villages and resources used, it is more difficult to obtain an integrated picture of the seasonal round. In order to provide a fuller picture of a collector way of life, therefore, I provide a brief description of the seasonal round employed by the Yakima, at the end of this section. The Yakima are a Plateau group living along the central reaches of the Columbia River, in an environment which shares certain essential characteristics with southwestern Oregon, including a seasonal climate, mountain and valley topography, and fish-bearing rivers. These people remained in their homelands and maintained their subsistence traditions beyond the period of historic contact. This brief description adds unity to the disparate pieces of information available from the local material, and emphasizes the utilization of strategy and planning which accompanied the annual round. Though some of the staple foods were different than for the southern Oregon groups, the annual rhythm was similar, and this example illustrates the timing of various subsistence tasks undertaken, and underscores the hard work which was part of a collector way of life. The tribes who inhabited interior southwest Oregon were distinguished from one another mainly on linguistic grounds but were connected through all the usual ties of social concourse, including intermarriage, trade, and warfare. They shared furthermore a common approach to the land in terms rTT'T~rl' I 'to .~... ' '·C· 55 of economy and settlement, and the material remains from their sites bear no easily recognized ethnic signatures. The people who inhabited the study region consisted of the Takelma, the Cow Creek Band of Umpqua, the Applegate Athapascans, the Molala, and the Shasta (see Figure 4). The Takelma spoke a Penutian language and inhabited the Rogue Valley. Their linguistic kin, the Cow Creek Band of Umpqua Indians, lived immediately north of them along the South Umpqua River and its tributaries. The southwestern part of the Rogue Valley was inhabited by Athapascan-speaking peoples, who lived along the Applegate River. Other Athapascan groups lived along the lower reaches of the Rogue, including Galice Creek, and along the Umpqua and lower portions of the North Umpqua Rivers. The Molala, speaking a language in the Penutian family (Rigsby 1969:79), inhabited the uppermost reaches of the North Umpqua and Rogue Rivers, in the Cascades. The Shasta, speaking a Hokan language, maintained a hold in the southern Rogue Valley from their main homeland in the Shasta Valley of northern California. The Takelma The Take/ma Indians inhabited the Rogue Valley, with a territory extending from about the confluence of Grave Creek and the Rogue River on the west to the crest of the Cascades on the east, and along the Rogue Umpqua divide on the north to about the present town of Ash/and on the I r -, ...- Klamath ./ ,.- - / / / M:rloc / 'l. • Kl 1., --- ' f ,r/ amath ""'- Lake ~ A} ~. /' I . ( o~e~ ,"~/r'.,. r,r) , . l . ""'- 01. ""\... l ~ 5 Athapa~an FIGURE 4, Ethnographic map of study area, 01 m r... ······ I: ';YI 57 south.1The Takelma, along with the other Indians in this area, fought a series of battles with the historic (mainly Euro-American) invaders, who flocked to this region after the discovery of gold in the ea.rly 1850s. This period, known as the time of the "Rogue Indian Wars," was disastrous for the Indians, many of whom were killed or removed from the region to distant reservations by 1856. As a result of this traumatic history, the ethnographic information which exists is very limited. It is based largely on interviews with a few informants near the turn of the century (Dorsey 1890; Sapir 1907, 1909) or a few decades later (Drucker 1937; Gray 1987:10). Gray (1987) has extensively reviewed the extant data on the Takelma, including the unpublished fieldnotes of J. P. Harrington which pertain to these peoples. Unless otherwise noted, the following brief discussion draws largely upon Gray's synthesis. Gray distinguishes two and possibly three divisions of the Takelma, recognized by the informants of the early decades of this century (cf. Sapir 1907:252). These divisions consist of the Lowland Takelma, occupying the western part of Takelma territory; Upland Takelma in the eastern part of the territory; and another band whom Gray refers to as the "Northern Takelma" in the northeastern area. These groups shared a common way of life, though local differences in the availability of certain resources may have engendered variations in the seasonal subsistence regime. The Lowland Takelma were situated along better fisheries on the Rogue, for example, and 11n this and in other territorial distinctions I am following Gray (1987), who has extensively reviewed the literature,·much of it conflicting, regarding local ethnographic territories in this region. In'r"r' , ~,.! 'f 1 ~---- 58 the Upland Takelma were probably more dependent upon deer and elk for animal foods (Drucker 1937:294; Gray 1987:32). The staple foods of the Takelma consisted of acorns, camas, anadromous fish (particularly salmon), deer and elk. Manzanita berries, tarweed seeds, pine nuts and cambium, wild plums, small mammals such as rabbit and squirrel, other fish, eels, and mussels, and certain insects are also listed in the ethnographic record as foods (Gray 1987:30-34; Sapir 1907:257- 260). These foods became available at different times of the year: Acorns were gathered in the late summer or fall, as were camas and the pine-bark cambium; other seeds and fruits became available in the summer and fall. Fishing occurred during seasonal spawning runs, which are noted for summer, winter, and spring, although not all fish-bearing streams had runs of fish every season (Gray 1987:32). Favorite fishing locations were at falls and rapids along the Rogue and its tributaries; fishing along the Applegate was remembered as particularly productive (Gray 1987:32-33). Hunting deer and elk was primarily an upland pursuit, generally associated with the warmer months of the year (Gray 1987:33; Sapir 1907:260). Many of the foods listed above required preparation for eating and storage. Acorns needed to be pounded into meal which was leached of its natural tannic acid to make it palatable; camas was roasted in earth ovens, mashed, and formed into cakes for winter use (Sapir 1907:258). Manzanita berries were pounded into flour, mixed with sugar-pine nuts, and stored for future use (Sapir 1907:259). Salmon were split and dried, and the meat sometimes pulverized for storage (Drucker 1937:294); baskets of roasted ~,. ! 59 -. salmon were kept for winter use (Sapir 1907:260). Deer products were also processed for winter storage; Sapir notes that cakes of deer fat were put away for winter use (1907:260). The Takelma occupied permanent villages, situated along the major waterways and at lower elevations2• These habitation sites were occupied over the winter, and as needed during the warmer months (Gray 1987:38- 39). These villages were home-places, and provided the locus of an individual's social identity (Sapir 1907:267). Villages consisted of substantial houses, built of poles and planks with the floor excavated up to two feet into the ground; structures were approximately 12 feet wide and 15 to 20 feet long (Gray 1987:37). A village would contain one sweat-house, a semi- subterranean structure which was covered with earth and sUfficiently large enough for six men (Sapir 1907:263). In terms of the annual cycle, villages were probably the locations at which inhabitants spent the most time. Goods were stored there, and the dead were brought back to the village for burial if the death took place elsewhere (Gray 1987:42). Villages may have varied considerably in size, depending upon location. Sapir notes that they were "generally insignificant" (1907:267), though Peter Skene Ogden (a trapper in the area in 1827) noted a village of "six large houses" sufficient to contain upwards of "100 Indians" (lalande 1989:22). During the warmer months, the need to gather foods available at other locations frequently took villagers to the uplands (Spier 1927:359), to be near such resources as oak groves and game. Seasonal camps could occur at 2 These villages are listed and mapped in Gray, 1987. r 60 .. any location, however, such as along the rivers at fishing stations and near meadowland resources such as camas. A more mobile existence in the summer is noted by one informant, who stated that "... in summer Indians travelled all around" (Gray 1987:38). Summer shelters were temporary and minimal; Sapir notes that "in summer the Indians dwelt in a brush shelter built about a central fire" (1907:262). The ethnographic record for the Takelma does not directly indicate several features of the subsistence/settlement system which are important to this study, though such features may be inferred. Warm-season camps are assumed to represent generally more specialized activities than those occurring at the main village, since these locations are specific to the acquisition of certain foods or other resources. It is also inferred that small groups or individuals took part in specialized tasks outside the winter or summer habitation areas, such as hunting. Support for these inferences comes from ethnographies of the Shasta Indians, who inhabited the land directly south of the Takelma and who were, according to Sapir, closely allied in terms of cultural patterns (Sapir 1910:673). The many similarities in the environment, staple subsistence foods (acorns, salmon, deer), as well as the close proximity of these two peoples supports this assertion. The Shasta survived the period of contact better than did the Takelma, and the ethnographic data are correspondingly richer. A brief review of the Shasta data helps augment the scanty Takelma record. __.Ann _ r.···,·,·,· ,. j scin-.__ 61 Like the Takelma, the Shasta inhabited villages along waterways3 during the winter and dispersed into the mountains during the warmer months to gather supplies. Winter villages could be small, consisting of only two to three families. Winter houses were substantial semi-subterranean pole and plank affairs, inhabited for about five months of the year and sometimes large enough to accommodate several related families. Houses contained baskets for storing acorns and dried meat and fish, and for cooking. A communal house was built only in villages containing several families, and was used for gatherings, games, and other social purposes; a men's sweat-house might also be built in the larger villages (Holt 1946:344; Silver 1978:215). Menstrual huts and small family sweat-houses used mainly by women were also part of the village pattern. As the weather warmed, the Shasta would move into brush shelters, not far from the village (Dixon 1907:413-422). According to Holt (1946:308), they lived in these shelters through the summer salmon season. When acorns were ready, they moved higher up in the hills to seasonal camps in the oaks, where they lived in bark shelters. Later in the fall, when further into the mountains for the fall hunt, they camped in the open. Smaller parties would depart from these sites to accomplish discrete tasks. Dixon, for example, notes fall or winter hunting parties composed of men and women, or only men (1907:431). In early summer a group of men and women might go to the mountains to prepare arhese villages are listed and mapped in Heizer and Hester (1970). s62 lumber and other materials for house construction back at the village (Holt 1946:306). The Shasta seasonal round is summarized by Holt (1946:312) as follows: The type of activity changed with the season. In the summer the people lived in brush houses by the river and almost their entire attention was turned to fishing and its attendant activities. In earty fall when acorns were ripe, they moved up on the hills among the oaks, leaving a few old people in the village, put LiP their bark houses, and set about gathering the year's supply of acorns. While the women gathered acorns the men hunted deer, singly at this time, with bow and arrow. Then they came down and late in the fall went high up in the Siskiyous for the last big fall deer hunt. It was at this time they had the big drive, encircling the deer with fire. This was a busy time, occupied entirely with hunting and cutting up and drying the meat. ... After this hunt, the acorns, left stored where they had been gathered among the oaks, were brought in by the people, who hurried to get them in [to the village] before the storm. . .. The people gathered wood, shelled acorns, and generally prepared for winter. At the onset of the first snowstorm all prepared their snowshoes ... after the storm settled, there came the hunt in the snow ... and in earty spring came the hunting at the deer lick. The Applegate Athapascans The people living along the Applegate River, in the southern part of the Rogue Valley, spoke an Athapascan language and were differentiated from the Takelma mainly on that basis. The literature on this group is very scant. In 1904 Pliny Goddard collected information on these Athapascans, as did Melville Jacobs in the 1930s. These studies are reviewed in Gray (1987). The Applegate Athapascans followed a cultural pattern similar to the Takelma (Drucker 1937:284). There were perhaps no more than three villages in the Applegate drainage (Gray 1987:56). Villages were reported to 63 consist of two to ten houses; archaeological evidence suggests that these winter "villages" could be as small as one house inhabited by an extended family (Gray 1987:56). The houses themselves were substantial wooden constructions built over excavated, semi-subterranean floors. Houses contained beds and mats, hearths and stools. Winter drying structures were associated with family dwellings, and sweat-houses were part of the village architecture. Sweat-houses varied in size, depending on the size of the community, and were used by the men. Like the Takelma, the Applegate Athapascans would bring the cremated remains of people who died elsewhere back to the Village for burial, where the dead were interred in graveyards. Like the other peoples in this region, the Applegate Athapascans followed a seasonal round, gathering and processing foods for over-winter storage at the home village. Summer dwellings were temporary affairs consisting of "brush or grass walled shelters with a flat roof of fir boughs surrounding a centrally located campfire" (Gray 1987:56). Subsistence tasks followed a seasonal pattern. In the autumn deer and elk were hunted specifically to dry meat for the winter, and quantities of salmon were dried and pulverized for winter use (Gray 1987:49). People had a camp in the early fall "at the foot of the mountain to snare deer" and earlier in the year lived at a summer camp at the mouth of the Applegate River where they fished. In fall they hunted on "the big round mountain," and packed their kill back down to temporary camps in the upper Applegate Valley (Gray 1987:50). ~--- ~ .' , 64 ~J.';.,; J, ---- The Umpqua The North and South Umpqua Rivers were home to several different tribes. The lower reaches of the North and South Umpqua Rivers were inhabited by a group known as the Umpqua, speaking an Athapascan language. The upper reaches of the South Umpqua were inhabited by a group known today as the Cow Creek Band of Umpqua, who were a people speaking a Takelman language. The Molala lived along the upper reaches of the North Umpqua River as well as the upper reaches of the Rogue. This section considers the two Umpqua peoples; the following section discusses the Molala. There are no published ethnographies directly pertinent to either the Athapascan Umpqua or the Cow Creek Band. The information available for their histories comes from the records of early settlers and explorers, Bureau of Indian Affairs agents, linguistic studies and archival sources pertinent to the larger Athapascan-speaking group in Oregon, and later oral histories conducted as part of a federal recognition treaty (Beckham 1983a and 1983b) or cultural resource projects (e.g., R. Winthrop 1993). The early information is largely anecdotal, and has been analyzed and summarized in several recent works (Beckham 1983a, 1983b; Beckham and Minor 1992; O'Neill 1989c; R. Winthrop 1992). More inclusive studies of Athapascan- speaking peoples in southwest Oregon provide firmer ethnographic documentation on a broader scale for the Athapascan-speaking Umpqua (e.g., Drucker 1937; Miller and Seaburg 1990). -- ,... I I crrbz_....__ 65 The Athapascan Umpqua depended on a constellation of staple foods similar to those of their neighbors to the south: anadromous fish, acorns, deer, and camas, though camas may have been more emphasized and acorns somewhat less so, due to possible differences in availability of these resources in the Umpqua and Rogue River drainages. The annual subsistence/settlement cycle was apparently similar to that of the other peoples in the area and other Athapascan-speaking groups in southwest Oregon. Semi-permanent villages provided a place for winter habitation and storage of winter foods. Habitation at these sites alternated with movements to seasonal camps in the countryside as annual resources became available. Shelters at these camps were made of grass or thatch and temporary in nature (Drucker 1937:279) An early explorer noted a village of two houses containing about 25 people: another observer noted that the lodges were about 15 or 20 feet long and made of cedar planks. These houses contained baskets, mortars, and pestles (Beckham and Minor 1992:107). Sweat-houses resembled those along the lower Rogue River (Drucker 1937:279). The seasonal round, described generally for the Athapascans of southwest Oregon, was probably applicable to the Umpqua. In June, women gathered roots such as camas, and berries, which were processed for storage. In July men fished; in August the old people stayed in the village while the younger people departed for summer camps to hunt and to dry the meat. Early fall was spent at fishing camps; men fished, women processed the fish, and gathered ~.;:,:H'"~r;:1 . r I sthr_.......__ 66 acorns, nuts and seeds. In winter, people settled into the villages (Miller and Seaburg 1990). The Cow Creek Band of Umpqua consisted of several bands living along the South Umpqua and its main tributaries. Though enemies of the Takelma, they spoke a closely related language (Beckham and Minor 1992:111) and followed a similar way of life. Beckham, who conducted the historical studies in support of the Cow Creek Tribe's federal recognition application, has examined the historical literature and oral histories associated with these peoples. In a recent summary of their ethnology, he presents a view of the subsistence/settlement regime which is broadly in keeping with that of the other peoples reviewed here. Beckham defines three zones related to subsistence/settlement activities. The Lowland zone exists between 400 - 800 feet in elevation, and has river terraces and extensive meadows of camas and tarweed, and oak groves. Cow Creeks had permanent winter villages in this area. The Uplands, from 800 - 1,800 feet, was also used for winter villages, especially along the South Umpqua and its tributaries Elk Creek and Cow Creek. This area encompassed forested hillsides with hunting and gathering areas, and the river and creeks provided fish and other aquatic foods. The High Mountains, from 1,800 to 5,500 feet provided extensive huckleberry patches and excellent hunting. Beckham notes that the Cow Creeks, like other Indians in the region, utilized all three environmental settings throughout the course of the seasonal round. A good description of a summer camp, for example, comes ---------- ------ ,.------------------ ! 67 from the reports of summers at the huckleberry fields. Families would rendezvous at these camps in August and September; the men would hunt, the women and children would gather berries, and everyone would enjoy themselves with socializing and games (Beckham 1983a:44-55). As elsewhere, the annual round also included sorties from the village and seasonal camps by specialized groups and individuals for specific and focused purposes. Beckham (1983a) lists a number of such specialized activities, inclUding hunting, herb and medicine gathering, and spirit quests. The Molala Data on the Molala of the southern Oregon Cascades are even more meager than for the Umpqua. Sources of information on this group have been most recently reviewed in R. Winthrop (1992) and Beckham and Minor (1992). These authors draw upon ethnographic studies by Leo Frachtenberg in 1910-11 (unpublished), as well as on studies by Harold Mackey (1972) and Bruce Rigsby (n.d.; 1966; 1969). The Molala, divided into several groups, inhabited the Western Cascades from the Rogue Valley in the south to Mt. Hood in the north. Based on Frachtenberg, Rigsby (n.d.:2) states that The Molalas wintered in sites located along streams in the lower elevations, usually west of the Cascades, and they exploited the higher country for roots, berries, and larger game (deer, elk, and bear) at other times of the year. Also based on Frachtenberg, R. Winthrop (1992:3-35) notes that winter houses were built of cedar and were six to eight feet wide and 20 to 30 feet long; summer shelters consisted of a roof of fir boughs with no walls. A b ....._-- 68 winter house might contain several families, and villages were small, consisting of a few families (R. Winthrop 1992:3-36). The complex of important resources consisted of the familiar deer, fish, camas, acorns and berries. However, the interior location of this group suggests a greater reliance on hunting, since anadromous fish runs would not be as abundant as for those living further west along the main fish- bearing streams. The Molala are noted for trading smoke-dried meat for other goods in the Willamette Valley, and one informant stated that "all the Molala people did was huntl" (quoted in R. Winthrop 1992:3-35). In reviewing a Molala myth, Winthrop argues that the theme of the myth is to provide a cultural charter for the Molala's identity as a hunting people. Based on this evidence, he suggests that these people may have been somewhat more mobile in the food quest than were their neighbors, with hunting a greater focus than other more stationary pursuits (R. Winthrop 1992:3-35). The Yakima of the Plateau Eugene Hunn's contemporary work with the Yakima Indians of the mid-Columbia River presents a useful portrait of a collector regime in an area with environmental parameters similar to those of interior southwest Oregon (Hunn 1982; 1990). The brief description of that regime rendered here provides a more coherent picture of the collector way of life than can be attained from the fragmentary data available for the groups just discussed. Hunn's work has the further advantage of an ecological focus, which is ~... , ---I', , 69 lacking in ethnographic work compiled for the southern Oregon groups, and provides a fuller picture of the planning and hard work that are part of the logistically organized collector regime. The Yakima inhabit the middle section of the great Columbia River, occupying a territory which incorporates the fish-bearing river and its tributaries, the valleys of these streams, and the forested mountains which rise above them. This landscape provides abundant foods at different seasons of the year, requiring careful timing for harvest and a strategy for seasonal movement. None of the staple foods are available during the winter; people had to process and store goods throughout the year to avoid winter famine. The need for seasonal movement coupled with the tie to a home base where goods were stored resulted in a subsistence regime which required careful planning and coordination of procurement tasks, as well as a lot of labor to process and transport a surplus of goods during the warmer months for use during the cold season. This planning and coordination resulted in a predictable annual routine, which resembled that of the people in southwest Oregon. Winter was a season free from the rigors of direct subsistence tasks, and devoted to other pursuits. It was a time for making and mending tools, making rope and netting, visiting, myth-telling, and exchanging goods and information. Families congregated in villages along the Columbia, beginning in about October and remaining until early spring, when the first plant harvests become available. 70 Root crops formed a significant part of the Yakima diet. Gathering parties would leave the village for short periods of time as these became available at low elevations, as early as February. During late April and early May, root digging was suspended for the peak run of spring Chinook salmon, when all available labor was needed to catch, clean, and dry the fish. As the season warmed and root species became available at higher elevations, parties would dismantle the winter lodges and move to a series of camps at increasingly higher elevations, spending perhaps a week's time in each camp. Women would gather roots and other foods, which they would process back at the camp. The dried provender was then hauled back to storage facilities in the main village. In early summer, families would move in loose association with one another to the camas meadows in the mountain uplands, harvesting staple crops as they became ripe at higher altitudes. Where crops were plentiful but dispersed, camps were small, but where summer crops such as camas or huckleberries were dense and plentiful, congregations could be large. These larger camps provided opportunities for socializing, gambling, politicking, and match-making. Stays at these camps might last from one to several weeks, sufficient to provision a family with camas or berries for a year. The summer runs of salmon pulled families back to the river to harvest and process the fish. During slack periods between fishing peaks women would gather berries and fruits, with all departing for the berry camps in late summer. b ~_....._-~ 71 Fall was a busy season. Huckleberries were prime in the uplands, and the most important fish run of the year, of fall chinook, occurred in September. Where distances were not too great, men departed the berry camps to fish; in later times, with the horse, fish were hauled back to the berry camps for women to process. Huckleberry season also coincided with the prime time to hunt deer and elk in the uplands. By October, the winter village was replenished and re-occupied, with final preparations for the winter taking place. The Collector Model The ethnographic evidence for the people living in the study area illustrates the collector regime, and provides evidence for the types of sites which characterized that regime. All these people lived in stable, permanent, winter villages where goods were stored, and participated in an economy dependent on processing surplus foods during the warmer parts of the year. The village provided a firm geographic locus; tethered as they were to the winter village, the seasonal movements of its inhabitants were likely to be relatively predictable and systematic. Warmer seasons of the year saw people moving to temporary shelters at seasonal camps; these places were often re-occupied year to year. Small groups of people departed both village and seasonal camps for short forays into the countryside for speci'fic~, such as to hunt, gather specific plants, or participate in a ritual activity. "..- .... "r ; I "( 72 Archaeology: Hypotheses for a Mobile Subsistence/Settlement Regime There are three recent works which present interpretations of the archaeological record in southwest Oregon (Connolly 1986; Hannon 1992; Pettigrew and Lebow 1987). The works by Connolly and Hannon provide analyses which permit the hypothesis of a mobile subsistence/settlement regime prior to the inception of the collector pattern; Pettigrew and Lebow suggest a variant of the collector regime which may be long-lived in this region. A dissertation by Connolly (1986) sets forth an argument for a series of significantly different patterns of land-use throughout the long prehistory of southwest Oregon. He statistically groups artifact assemblages from a number of sites, and identifies three distinct patterns based on the artifact types present. He argues that the earliest pattern, called the Glade Tradition, represents an extremely long-lived and stable cultural tradition which persisted from the beginning of the Holocene, but was gradually replaced after about 1500 years ago. He hypothesizes that the Glade Tradition was characterized by a "generalized hunting and collecting strategy oriented toward terrestrial resources." Small, mobile bands of foragers are inferred as part of this pattern, and "occupation sites appear to be predominantly temporary camps," frequently located at valley edges (Connolly 1986:214). The two later patterns, called Siskiyou and Gunther Patterns, are similar to the ethnographically known way of life. These later patterns are "i ---..-,--- 73 characterized by settlement in river-side villages with fishing and intensive use of other foods as an important part of the subsistence regime. The transition from the Glade Tradition to these later patterns thus represents a significant shift in subsistence and settlement practices, from highly nomadic foraging groups to those living a more settled existence in semi-permanent villages located along major rivers and streams. A part of this change is a shift to a way of life in which resources are collected and stored in central villages, coincident with somewhat greater groupings of people. Connolly bases his argument on archaeological data for a broad region, encompassing northern California and southwestern Oregon. His sample consists of 32 cultural components from 25 sites; sites were compared based on culturally diagnostic elements (e.g., projectile point types, pottery, oil lamps, bell-shaped mauls, and other distinctive artifacts). Once cultural groups were segregated, characteristics of the sites, such as site location, were noted to provide clues regarding the way of life followed by their inhabitants. Connolly's conclusions thus remain as hypotheses to be investigated by further work, as in the present study. In another model, Pettigrew and Lebow (1987) argue that local variations in resource availability account for differences in prehistoric settlement regimes. In their work along Elk Creek, these authors note the existence of small, residential hamlets within the foothills of the Cascades. Drawing upon data from the Rogue Valley as a whole, they argue that the regional settlement system involved two kinds of habitation sites: large riverside villages (With multiple houses"and extended families) on large 74 streams where salmon could be relied upon as a staple, and small homesteads or hamlets (with one extended family group and one to three houses on average) fairly evenly scattered across the landscape on smaller streams. Though similar food staples would be used by these groups (Le., salmon, deer, acorns), emphasis would vary depending on availability within any group's territory. Salmon, for instance, were of "paramount importance to the large riverside villages while acorns and deer were more important for the homesteads ... the resource distribution largely determined the settlement distribution." They further suggest that a Iifeway involving small housepit settlements as central bases and wintertime habitation sites is of considerable antiquity (Pettigrew and Lebow 1987:12.11). A third study, already mentioned in Chapter II, relates subsistence and settlement patterns to presumed changes in the environments of the Rogue Valley during the course of the Holocene (Hannon 1992). As previously noted, Hannon argues that 'the xeric interval in the early to mid-Holocene would have affected the resources available to the hunters and gatherers of the study area. Fisheries were probably less abundant and annual runs confined to the winter season, especially along the upper reaches of the major rivers and their tributaries. Oak was more prevalent, but drought stress may have meant that crops were not always predictable. resulting in a patchy distribution of annual crops. Small game and deer were more abundant, but more dispersed for much of the year, given a shorter winter season. Hannon argues that this constellation of resources provided dispersed foods which fluctuated annually, with acorns abundant in one place "",.-- H I" ,i' • i ... sAn---- 75 one year but not the next, or deer abundant in one place one year but another place the next. If this scenario is correct, she argues, the dispersed and unstable nature of the resources dUring the early Holocene would foster a highly mobile way of life on the part of the people who depended upon them. Campsites, rather than sedentary or semi-sedentary settlements, would have characterized this period. People probably did not re-inhabit the same sites every year, since the critical resources fluctuated annually, or were depleted in one area and allowed to regenerate. This consideration would have operated for both winter and summer residential bases. That is, if an area was hunted, fished, or gathered one year, it may have been several years before the group returned. This would contribute to considerable mobility, with large "catchment" areas necessary for each group. Population densities were probably lower than later on, and maximum group sizes smaller. Under this scenario, the basic social unit would be a small group that wintered and summered together, with some splitting off at certain times for special tasks. Following the xeric period of the early Holocene, according to Hannon's model, the valley resources improved; a more mesic climate meant that staple crops such as camas and acorns were more abundant, as were anadromous fish. The winter habitation became the primary settlement focus, with groups returning to the same location annually. These places were located on anadromous fish-bearing streams, frequently where annual runs were plentiful and predictable. With harsher winters and shorter spring and fall seasons, the wintering spot was inhabited longer, and resources r:". , .. , , I ." __.Arn _ 76 were stored to accommodate this period. Social units were flexible; where winter habitations were along major streams, comparatively large groups inhabited them. These groups would split into smaller units for the warmer seasons in the uplands and special task groups would make forays from either winter or summer habitations. At other places, smaller groups would constitute a winter village, perhaps consisting of a single family (Le., the "homesteads" defined by Pettigrew and Lebow 1987). The Mobile Model The work accomplished by Connolly and Hannon helps develop the mobile model postulated as preceding the collector way of life. In this mobile regime, people lived in small groups, occupying home territories but not tethered to specific, stable, winter villages. Seasonal camps provided the main habitation sites; these were occupied by the entire group and moved when necessary. They might be reoccupied annually, or occasionally, but would be located near specific resources as such resources came available. Necessary short-term tasks, such as hunting, butchering, or quarrying, might be accomplished by a part of the group away from the camp. Subsistence/Settlement Systems and Site Types The two subsistence/settlement models used in this study, therefore, are contrasted with one another on the basis of sedentism/mobility, intensive use of resources, and the presence/absence of significant food storage as a critical element of the subsistence regime. The ethnographic record portrays ~.!r·! ,";, r i ..azn__~ 77 a semi-sedentary regime, in which foods are processed and stored for over- winter consumption. Archaeologists working in this area postulate a different, earlier pattern, in which people followed a more mobile way of life, and did not rely on significant amounts of food processing and storage to cope with the winter months. In order to identify these patterns archaeologically, it is necessary to define the types of sites which constituted them. Archaeologists in this region have long worked with three site types: the village, the seasonal camp, and the task site (e.g., Beckham, Minor, and Toepel 1981). These types were initially derived from ethnographic information, and have been used by many archaeologists as descriptive terms for the sites they have investigated. These three types are sufficient to describe the mobile and collector regimes hypothesized and demonstrated for this region, and to note the differences between them. Since these site types are in wide use, and since most of the sites used in this study have been initially described in these terms, these types are used in this study. The collector pattern produced all three types of sites; the more mobile pattern did not include the village. Although these three site types have been in wide use, there is no definition of these types specific to the archaeological record for this region, nor is there a description of the archaeological correlates associated with these types of sites. Hence, these three site types are defined below, in terms of the types of activities accomplished at these sites and the relationships of these site types to one another. The following chapter gives b -_.~-~ 78 specific archaeological content to these types, which is arrived at through the multi-step analyses which form the core of this dissertation. Villages The village was the geographic locus of the social group, the place which focused the annual round and the place where people spent the most extended period of time. As described in the ethnographic works, people spent up to five or six months a year at these places, returning to them at other times for various purposes. In some cases, the villages may have been inhabited for all of the year by some members of the group, such as the elderly. The larger winter Villages were located along those rivers and streams which produced abundant fish; smaller settlements were located along less productive streams but all were at comparatively low elevations to avoid the harsh winters of the uplands. Villages are the most functionally complex of all the site types. Numerous activities were accomplished at villages, by people of every age and status, and of both sexes. Permanent habitation, even on a semi-annual basis, made investment in substantial architecture--such as pithouses--worth the effort. The village's function as the focal point for storage made artifacts and facilities for storage necessary, such as baskets and pits. The variety of tasks at these sites as well as their stable locations also called for a variety of tools and implements, including many which were heavy and relatively non-portable, as well as those--such as pottery--which were fragile. Middens *' .Atn __ 79 and cemeteries are associated with such sites, as places for long-term accumulation of refuse and burial of the dead. Archaeological evidence defines variants of the village type. Small hamlets or homesteads, consisting of as few as one house for an extended family, provide a variant to the nucleated village settlements described more commonly in the ethnographies. The social implications of these differences in settlement size are surely significant; however, both types served a similar purpose in the settlement system. Large or small, the winter habitation was the locus of a group's territory and subsistence, providing the focus for the annual subsistence regime and the place for long-term storage. Hence, in this analysis, these two variants are included within the "village" category. Seasonal Camps Throughout the warmer months of the year, most people from the winter village moved to seasonal camps in the countryside, shifting these camps as different resources became available. Family groups moved together, though sometimes old people remained in the village, as noted above for the Shasta. The seasonal camps usually had a particular focus, such as berrying, root gathering, or hunting, and represented more specialized locations than the winter village. Yet these were also places where families camped and engaged in normal everyday maintenance tasks; tools and materials left from these camps would also reflect this more generalized focus. ~"~f!?~'" .. ,.... i .....-._- 80 These seasonal camps lasted from a week to perhaps a month or more, and provided temporary bases from which work parties could direct subsistence tasks. Temporary shelters often were erected at these seasonal camps. For collectors, these camps were also places where crops were processed, to reduce the bulk for transport to the home base. Those spots which were reliable producers of annual foods were visited on an annual basis, and heavier tools, if needed, were stored there. Other places for summer encampment may have changed from time to time within a given territory as resources fluctuated in response to various conditions. As noted in the ethnographic review above, some sites were occupied by only a small group, while others--such as the huckleberry fields--attracted large congregations of people. Regardless of the size of the group, however, the temporary and semi-specialized nature of the seasonal camp, complemented by a short-term, generalized activity focus, characterized these locations. For the more mobile subsistence pattern, hypothesized for the earlier period, the seasonal camp was the main habitation site. These camps would have been similar to the seasonal camps of the collector regime. These camps were occupied by family groups, and were moved with the availability of seasonal foods. They thus reflected both the specialized focus on a particular resource, or constellation of resources, and the everyday activities of a diverse group of people. These camps were not stable home places, however, and their locations might shift annually. This pattern did not support substantial architecture, nor accommodate long-term storage. Winter camps would have characteristics simi-Iar to summer seasonal camps, except ~I , 81 that they would be expected to occur at low elevations, probably along waterways productive of winter fish, and in areas of winter forage for deer and elk. Task Specific Sites Such sites result from focused and specialized activities accomplished by limited groups of people. Huntinglbutchering sites, hunting blinds, fishing stations, quarries, spiritual quest sites, and short-term encampments when traveling are examples of such sites. These sites differed from the seasonal camps in two important respects: they were occupied for shorter periods of time, sufficient to accomplish the purpose generating the stay, and they were more specialized. Such sites reflect a single purpose, accomplished by a specialized group of people, such as a few male hunters, or a few adults quarrying stone material, or a few women and children gathering certain plants. Task sites were tied to seasonal camps and village sites, and were generated throughout the year and at all elevations. It is predictable that task sites were more frequently associated with village sites and the logistically organized collector regime, for a number of reasons: the larger villages, at least, were better able to produce specialized work parties than small family groups; the longer residence at a single location required more forays to supply the resident group; and the emphasis on collected and stored foods would promote specialization dUring the warmer months, to optimize gathering of concurrently available foods. ~""""... '.!, 82 In summary, the site types used in this study, and the subsistence/settlement regimes of which they are part, are as follows: Site Types 1. Task-specific sites: located in a variety of environments, but related to either of the following two types. 2. Seasonal camps: (a) warm-season camps tied to a collector regime, (b) summer and winter camps postulated for a mobile subsistence pattern. 3. Semi-sedentary winter villages: includes both larger villages and small "homesteads" of one or a few houses. Subsistence/Settlement Systems 1. The Collector Model: composed of all three site types. 2. The Mobile Model: composed of seasonal camps and task sites without the winter village/homestead component. Although the three site types discussed here have been widely referred to by archaeologists in the region, there is no standard definition of the archaeological correlates of these sites. In order to place sites into these functional categories, it is thus necessary to define such correlates, and identify them in the sample of sites used in this analysis. This is the task of the next three chapters of this dissertation. 83 CHAPTER IV RESEARCH METHODS: FUNCTIONAL ASSESSMENT OF SITES The site types introduced in the previous chapter provide the framework for organizing the archaeological materials into functional categories. In order to analyze the archaeological materials, it is necessary to identify those characteristics of each site type which are likely to leave an archaeological imprint. This chapter reviews those characteristics, and introduces the methods used to assign archaeological sites to functional types. Functional Types and Archaeological Correlates Mobility, or the degree of sedentism, is one of the main characteristics distinguishing villages from seasonal camps, and both of these from task sites. The length of time a site was inhabited is linked to the number of people present during the period of occupation, and the degree to which activities at the site were specialized or generalized. These differences in turn are reflected in certain characteristics of the archaeological assemblages. The density of artifacts at a site is assumed loosely to reflect the length of time spent at a site, the size of the group present, and the extent of periodic reoccupation. Very simply, this proposition assumes that the more • --_....._---- 84 people present and the longer the length of stay at a site, and the more frequently a site was revisited, the more artifacts were used and discarded at that place. Village sites should therefore have greater densities than seasonal camps, which in tum should have higher densities than task sites. This presumes fairly homogeneous depositional environments among the sites being compared, since density is measured in terms of site matrix excavated. That is, given a similar number of artifacts, slow or rapid rates of sedimentation will render different densities. In fact, most of the sites in this sample come from similar environments; they are open-air sites in the foothills and mountains of the Cascades and soil deposition is assumed to be relatively uniform among the sites. Where such environments differ, and this difference appears to be reflected in the density measures, this condition is noted in the discussion of site density. The diversity of a site's assemblage should reflect the degree to which activities, and probably also the social group, were specialized or generalized at a site. A generalized assemblage has a high diversity of tools; it contains lots of different tools representing a multiplicity of tasks. A specialized assemblage has a low diversity of tools; it contains few tool types, re'f1ecting only a few--or even just a single--task(s). Those sites which are low mobility (i.e., more sedentary) sites, occupied by a diverse group of people, produce the most generalized assemblages. In this analysis, the low mobility sites are village sites. Task sites represent the opposite extreme, having the most specialized and least diverse assemblages. Seasonal camps are --_...._-._- 85 intermediate in terms of mobility, and are assumed to have assemblages less generalized than village sites, but not as specialized as task sites. The diversity of an artifact assemblage is measured both by the number of categories of artifacts present (richness) and the uniformity of their specimen distribution among those categories (evenness). Measurement of artifact diversity has engendered considerable debate in archaeology, since the diversity of an assemblage is confounded statistically by the size of the sample analyzed. This is a distinct problem, and is treated separately in Chapter VIII. The methods used in this analysis, however, are designed to mitigate the effects of sample size differences. Differences in mobility are manifest in other ways, besides artifact density and diversity, in the archaeological record. Sedentary hunting- gathering communities are generally associated with substantial architecture, cemeteries, storage features and other permanent facilities, as described in the ethnographic summary above (see also Kelly 1992:56; Price and Brown 1985:13, 438). In this study, the presence of these features helps define village sites, and provides comparative data as a check on the density and diversity measures. The three types of sites, and their distinguishing archaeological manifestations, are defined as follows: • Village sites were the most sedentary communities, and had the most people of all ages, statuses, and both sexes; were re-occupied; had permanent architecture; and were the locations of a diversity of different activities. Such sites produced generalized, unspecialized assemblages which were both comparatively rich and comparatively even, a high density of artifacts, and habitation features. Hence assemblages are both dense and diverse, and associated with significant archaeological features. 86 ~--------------------- .Alnn __ • SeasQnal camps were occupied by smaller, heterogeneous groups for shorter periods of time than the villages. Like village sites, these sites produce assemblages reflecting the generalized range of activities coincident with a mixed group Qf peQple spending a significant amount of time at a place. These sites may not have been re-occupied annually, however, and were more fQcused Qn certain resources (e.g., meadow plants, game, acorns, or fish) than were village sites. They were frequently locations for collecting and prQcessing resources for over-winter storage. Hence, the assemblages still reflect a range of daily activities, but are more specialized than those for village sites; the assemblages are less diverse--Iess rich and less even--than village sites, but more diverse than task sites. Site assemblages are not likely to be as dense as the annually re-occupied, more densely populated, and longer-term habitatiQn sites, but are likely tQ be mQre dense than assemblages from shorter-term task sites. • Task sites are the most specialized sites, occupied for the shortest amQunts Qf time. Specialized grQups, such as a few hunters, WQuid depart from the village or seasonal camp for forays into the cQuntryside fQr a particular purpQse. Sites were nQt Qccupied for long; a diversity of tasks is not represented. Task sites might or might nQt be annually re-Qccupied. AlthQugh the basic tQQI-kit might be represented at a site, the dominant task would generate an assemblage which was mQre specialized than that fQund at the Qther two types. Site assemblages would be the least rich and even, and probably the least dense1, Qf the three types Qf sites. In the analysis, it is assumed that the dQminant use Qf a site is that represented by the diversity and density of its assemblage as just defined. It is probable that predominantly seasonal camps were occasionally used as task sites, or that Villages were once seasonal camps. In cases where the assemblages cannot be separated into stratigraphic or spatial components (which most often is the case) it is assumed that the function represented by the assemblage diversity (richness and evenness) and density measures is the main function of that site. Since the most intensive uses--such as village lit is possible that certain short-term, specialized tasks, such as quarrying, would produce a high density of materials. Such sites would appear as high density, low diversity sites in the archaeological record. " I I III I: I': I !I ,r".'i":rl 87 habitation--are more likely to drown evidence for less intensive use, it is possible that the more intensively used site types will be somewhat over- represented in the sample as a whole. Functional Analyses: Methods The intent of this analysis is to determine groupings of archaeological sites based on function. Several different analyses have been performed in the hope of finding concurrence among the results, which would strengthen the findings of anyone analysis. A secondary purpose to this endeavor is to experiment with different quantitative measures, to see which might prove useful for analyzing artifactual data from hunter-gatherer sites. Four different procedures constitute this effort. The first consists of a qualitative assessment of a site's function; the next three are based on quantitative data. Qualitative Assessment The first analysis draws upon the data presented in the site reports, including the excavator's opinion, to define the site type. These data are not generally SUbject to quantification; site function is assessed on the basis of an archaeologist's previous experience in the area, the types and abundance of various artifacts, site location, site size, site features, reports concerning the site from local residents, ethnographic or historic references, and other sources of information. Though not subject to quantification, this qualitative assessment is made on the fullest information available and is an important pr-.,."". , l~ ,'., , b 88 contributor to site function analysis. However, the method is subject to personal bias; different investigators with different experiences will interpret the record from varying perspectives. Though these subjective assessments provide valuable insights, they must be checked by more rigorous and objective quantitative metho'ds. Quantitative Assessment The three quantitative methods employed here are built upon analysis of the archaeological record from each site. The first defines groups based upon the density of the stone artifacts from each site (per cubic meter); the second defines groups based upon the proportions of various stone artifact classes at each site (Le., the "richness" and "evenness" of the assemblage). The third uses cobble and groundstone density data compared with feature data to sort sites into functional groups. These methods rely on specimens which are common to the sites in this study and characteristic of sites in this region. Almost all of the sites in this study have artifact assemblages primarily of stone, and only non- perishable items are considered in comparing artifact assemblages. At many sites the refined specimens are so few in the assemblages at hand that measures of statistical significance cannot be meaningfully applied to their presence/ absence. In order to circumvent this small-sample problem, specific tool types were combined into broader categories which are common to sites in this region. For example, various projectile point types are all subsumed under "projectile points," and various specimens exhibiting I I ,I ! r i 89 bifacial workmanship are grouped together as "bifaces." This lumping made it possible to compare the assemblage diversity among all sites in the sample. Grouping tools in this fashion also makes it possible to compare sites where tools for specific tasks may have changed, though overall site function remained the same. That is, even though hafted scrapers may have replaced hand-held flakes for hide-working, the overall characteristics of the assemblage should remain the same if the overall site function remained stable. Artifacts from site assemblages were therefore divided into seven tool categories, plus debitage. The tool classes used are: projectile points, bifaces, edge-modified flakes, cores, groundstone, battered cobbles, and other cobble tools. These categories are broadly recognized classes of stone tools in this area. Although finer distinctions are frequently noted in the site reports, it was necessary (as just noted) to assign artifacts to these general categories in order to make assemblages comparable and deal with analytical units of adequate sample size. These tool classes are defined as follows: • Projectile Point: Artifacts used to tip spears, atlatls, and arrows. • Bifaces: Drills, knives, blanks, preforms, and other chipped stone implements which are usually formed by working both sides of a flake. • Edge-Modified Flakes: Edge-modified flakes form the largest class of artifacts. They include scrapers, utilized flakes, unifaces, burins, and other tools which have one or more edges modified for or by use. 90 • Cores: Chunks of rock from which material has been removed in the process of tool manufacture. • Groundstone: Shaped/utilized implements generally associated with plant food processing: manos, pestles, metates, mortars, bowls and grinding stones. • Battered Cobbles: Hammerstones, anvil stones and other cobbles damaged from heavy use but otherwise unshaped or modified. In the last quantitative analysis, relating the density of cobble tools to habitation features, this group of artifacts is combined with the cobble tool category. • Cobble tools: Flaked cobbles, choppers, cobble flakes, and other such implements of heavy work as well as smaller cobble/pebble tools such as netsinkers and abraders. Density Measures In order to group sites according to assemblage density, it was necessary to devise a means for comparing sites. An earlier experiment with data from the Elk Creek sites prOVided a model (Nilsson and Kelly 1991 :375). In the Elk Creek analysis, the density of projectile points (per cubic meter of excavated site matrix) was plotted against the density of other chipped stone tools, for each site. The resulting scatterplot showed a strong correlation between the two measures (projectile point density and chipped stone tool density); that is, sites with many projectile points were also likely to have many other chipped stone tools. The plot not only illustrated this correlation, but also visually distinguished the high density from the low density sites. Nilsson and Kelly found that those in the high density range corresponded to the sites considered possible winter villages, with those at the low density end corresponding to sites considered task sites. In assessing the site density data for the sample of sites in this analysis, I decided to use Nilsson b,....•. ' ... ~ ., --_....._---- and Kelly's technique, and to expand it to include another, similar measure for illustrating site density. These techniques are frankly experimental and are used here based on the success of the Nilsson and Kelly procedure. Since I did not know at the outset which density measures would be most useful in distinguishing sites, I chose to look at four measures of density: density of projectile points, density of other chipped stone tools, density of all stone tools, and density of debitage. While it is possible to arrange sites in order from the most to least dense based on anyone of these measures, or based on the total artifact density, such an arrangement either loses possibly significant distinctions by combining all types of density on one measure (e.g., total artifact density), or produces a series of density measures with no demonstrable relationship to one another. The use of the scatterplots helps mitigate these problems. The scatterplots permit two types of artifact density per site to be expressed relative to other sites in the sample. For example, the density of projectile points for each site is plotted a.long one axis and the density of other types of stone tools is plotted along the other axis in the first density measure. Each point on the scatterplot represents a specific site, and the density of projectile points and stone tools at that site compared to other sites is immediately evident. In using the scatterplots, I was attempting to combine different measures of density, such as projectile point density and total tool density, in order to use data efficiently. At the same time, use of various different measures of artifact density, such as total tool density and 91 ~.r':"I !,', *= --~--~ 92 debitage density, permits the expression of possible differences among sites in terms of these measures. Density Measure 1 plots the density of projectile points per cubic meter against the density of other chipped stone artifacts per cubic meter (Nilsson and Kelly 1991). Density Measure 2 plots the density of chipped stone debitage against the density of all chipped stone, cobble, and groLindstone tools and other tools (occasionally sites would have stone artifacts--such as pipes--which did not fall easily into the chipped stone or cobble category; these were added into the overall tool density measure). Density Measure 1 measures the density of all chipped stone tools at a site, and compares the sites on that basis. Density Measure 2 uses all the data available for a site's stone tool assemblage, since it includes all stone tools as well as debitage density. The two measures were employed in order to permit possible differences in density measures to be expressed, but also to take advantage of the data available. A number of sites did not have debitage density data, and use of only the second measure would have left a these sites out of the density analysis. Use of only the first measure, however, would have precluded using all the information available for many sites which did have debitage density data. Once the sites were plotted, the resulting scattergram was divided into three groups of low, medium, and high density sites. The breaks were determined based on visual inspection of the scatterplot, with lines drawn where I distinguished breaks in the plot. These groups are taken to __. {'~ r-----~------- 1"'1\ .,.".. , , 93 represent functional types, with low-density task sites at the low end of the plot, seasonal camps in the middle, and village sites at the high end. Not all site reports recorded volume of material excavated, nor was the volume of material excavated given for distinct components at certain sites. Since density is measured in terms of the volume of material excavated per sits or per site component, it was therefore not possible to derive density statistics for some sites or site components. Multidimensional Scaling (MDS) The MDS analysis was used to place sites into functional types based on the diversity of a site's assemblage. Diversity consists of the number of different categories of tools present in a site (assemblage "richness") and the distribution of artifacts among those types (uniformity or "evenness" of the assemblage). The components of a site's tool assemblages (projectile points, bifaces, edge-modified flakes, cores, cobble tools, groundstone) were given proportional (percentage) definitions for each site, in order to make assemblages of unequal size comparable. Use of the MDS method has the advantage of being applicable to all sites which provide data on the number and types of tools collected. Hence, it was also applicable to most of the sites and site components which did not have density data. The MDS analysis proceeded as follows. First, all pairs of sites were assigned a measure of dissimilarity using Euclidean distance. Euclidean distance is a measure in which the difference between two units is expressed numerically as a measure of distance, yielding a measure of dissimilarity. ".:~ . "1 , ,; , 94 That is, two sites with assemblages whose proportions of tool types are the same would have a distance measure of 0; those whose proportions were increasingly different would have increasingly greater measures (Aldenderfer and Blashfield 1984:25). This exercise was performed on a computer using a program from ANTHROPAC (Borgatti 1990), which generated a large matrix. Next, the dissimilarity matrix was entered into a non-metric MDS program, again using ANTHROPAC. The MDS procedure then takes the matrix and creates from it a rank-ordering of sites, in which each site is ordered depending on its distance from each other site (Doran and Hodson 1975:214). Thus sites which are similar to one another but different from other sites would have low rank orders relative to one another, and high rank orders compared to the different sites. Finally, the similarities among sites, derived from the rank orders, are expressed graphically. The computer program arranges the sites in conceptual space so that the relationships among all sites (expressed as rank-orders) is preserved. In plotting these arrangements on a piece of paper, the relationships are necessarily compressed into two-dimensional space (Kachigan 1986:413-420). The program calculates a statistic, known as Kruskal's formula 1 stress coefficient, which is a measure of how good a fit the two-dimensional plot is of the original multi-dimensional arrangement. Although there are no objective standards for a "good-fit," a stress value of .15 or less is generally considered satisfactory (Kachigan 1986:418). In the ~.. '."I." ', , > 95 following analysis, the stress values were less than .15 for both the Rogue River Basin and Umpqua Basin data. Interpretation of the resulting scatterplots is based on the following considerations. Sites with the least specialized tool kits would produce assemblages which had tools in every category, and the tools would be distributed relatively evenly among the various categories. These site assemblages would be both rich and uniform; as noted above, village sites should characteristically produce rich and uniform assemblages. These assemblages would therefore be very similar to one another and the village sites should clump together in the scatterplot. Seasonal camps would be less similar to the village sites than village sites are to each other, but would nonetheless share some of the characteristics of the village sites. Seasonal camps, like village sites, were occupied by non-specialized groups and would have moderately rich assemblages. These sites would not necessarily be very similar to one another, however, since they would probably represent different specializations ref'lected in less uniform (even) assemblages. In terms of rank orders, such sites would be relatively close to the village sites, but dispersed about them depending on the degree of specialization manifested in the assemblage. Thus, in the scatterplot, the seasonal camps should form a ring around the central clump of habitation sites. Finally, the most specialized sites--the task sites--would have assemblages which are neither rich nor uniform, but rather consist of high proportions of specific artifact classes reflecting the special purposes of these ______.k_______ ,I 'I ii Ii, I :1 ' I ,I I' , 96 sites. These sites would not be very similar to either of the two other classes, and would not necessarily be similar to one another. On the scatterplot, they would be distant from the other sites and generally dispersed about the central clump of habitation and seasonal camp sites. Those highly specialized sites with similar specializations and tool kits would group together. The interpretation of the scatterplot drew upon these considerations and proceeded as follows. First, those sites which seemed firmly identified from other analyses were distinguished on the scatterplot. This exercise corroborated the assumptions outlined above, since the readily identified village sites clumped in the middle, with seasonal camps in a ring about them and tasks sites dispersed beyond both types. Those sites which had not been subjected to other analyses, or which had equivocal designations, were 'then given a functional designation based on their location within the matrix, i.e., whether clumped with other village sites in the center, or in the secondary ring of seasonal camps, or dispersed beyond the central clump with the task sites. The MDS analysis offers another way of grouping sites, based on data--assemblage richness and uniformity--which are sometimes difficult to compare. This analysis also provides a check on the other tests, and gives a way to incorporate data from sites which lack information for some of the other measures. It proved a useful exercise from these perspectives. IF.,,··""; ,\ > 97 gobble tool - Groundstone/Feature Analyses Both of the above analyses depend heavily on data from chipped stone tool assemblages. Though groundstone and cobble tools are included in the analyses, they generally comprise such small proportions of an assemblage that their effect on the overall density or diversity measures is slight. This last analysis looks at data from these categories of artifacts. The cobble and groundstone densities are computed for each site, and the sites arranged in order of increasing density for these artifacts. Again, three groups based on increasing density are distinguished. This last analysis incorporates feature data as a test. Feature data are compared to the groups derived from the cobble and groundstone density analyses. Feature data provide an outside check on the quantitative analyses based on artifacts. Sites with habitation architecture and features can by definition be considered villages. In this analysis, housepits, middens, and burials are considered as indicators of village sites. Other features present in the sample of sites include hearths and miscellaneous (buried) rock features. These features imply at least some degree of sedentism, and may be more frequently associated with seasonal camps (or village sites) than task sites. Final Assignment to Functional Type In each of the methods described above, I use the data to place each site into one of the three functional groups: ---".." II ! I: I: , I I i i 98 Group 1 = Task sites Group 2 = Seasonal camps Group 3 = Villages In many cases, all measures used produce mutually consistent results, and the assignment of a site to a particular functional category is unambiguous. Where the different tests yield different results for the same site, however, it was necessary to decide which category best represented the data. Generally, in making the final assessment in these cases, J followed the original excavator's judgement or relied upon information concerning the site which was not represented in the quantitative analyses. In these cases, where results of the various tests are ambiguous, I have stated the reasons for the final assignment. The Site Database The sites in this study are divided into two groups, those from the upper and middle Rogue River drainage basin (Rogue Basin sites) and those from the North and South Umpqua River drainage basins (Umpqua Basin sites). Figures 5 and 6 show the locations of these sites. The keys to these figures list the sites and their identifying numbers, and provide the report references for the sites. In order to keep the text less cumbersome, these reports are only referenced here, rather than every time a site is discussed in this study. The site data used in this study are presented in Tables 1-4. These data provide the raw material from which the analyses in the next two chapters are derived. , ~ ROG-UESJCS Tr (ll;AfU lAXI C""'TIR L AIt ( I i I i i NATIONAL i P.R' i I ! ._._._._._.1 (llAU' Of ~ THl WOOD$ ~ '''YIE,. MIL E S 5 0 10 I , : 3 , , , , FIGURE 5. Rogue Basin Sites (map and reference key). co co r --. ..- -~.. -r . Ref. No. Site No. Site Name Reference 1 35CU84 Marial Clark 1988; Griffen 1983; Schreindorfer 1985 2 35JA1 Gold Hill Cressman 1933a,b; Hughes 1990 3 35J04 Ritsch Wilson 1979 4 35J016 Marthaller Steele 1984; Deich 1982 5 35JA21 Saltsgaver Prouty 1988 6 35JA25 Far Hills Davis 1983 7 35JA42 Applegate Brauner 1983 8-10 35JA47 Applegate Brauner and McDonald 1981 11 35JA77 Salt Creek Satler n.d. 12 35JA133 RRNF lalande 1983 13 35JA189 Trail Connolly 1988 14 35JA190 Trail Connolly 1988 15 35JA191 Reeder lalande 1987 16 35JA197 Uttle Butte Winthrop and Gray 1991 17 35JA10 Elk Creek Davis 1983; Nilsson and Kelly 1991 18 35JA11 Elk Creek Nilsson and Kelly 1991 19-21 35JA27A Elk Creek Pettigrew and Lebow 1987 22 35JA27B Elk Creek Pettigrew and Lebow 1987 23 35JA59 Elk Creek Pettigrew and Lebow 1987 24 35JA100 Elk Creek Pettigrew and Lebow 1987 25 35JA101 Elk Creek Nilsson and Kelly 1991 26 35JA102 Elk Creek Budy et al. 1986 27 35JA103 Elk Creek Nilsson and Kelly 1991 28 35JA105 Elk Creek Nilsson and Kelly 1991 29 35JA107 Elk Creek Budy et al. 1986 FIGURE 5. Continued ~ 0 0 Ref. No. Site No. Site Name Reference 30 35JA110 Elk Creek Nilsson and Kelly 1991 31 35JA112 Elk Creek Nilsson and Kelly 1991 32 EC-2 Elk Creek Nilsson and Kelly 1991 33 Island Elk Creek Nilsson and Kelly 1991 34 Winningham Elk Creek Nilsson and Kelly 1991 35 Zimmerly Elk Creek Nilsson and Kelly 1991 36 35JA5 Lost Creek Davis 1983 37 35JA6 Lost Creek Davis 1974, 1983 38 35JA7 Lost Creek Davis 1974, 1983 39 35JA8 Lost Creek Davis 1983 40 35JA12 Lost Creek Davis 1983 41 35JA14 Lost Creek Davis 1970, 1983 42 35JA16 Lost Creek Davis 1974, 1983 43 35JA18 Lost Creek Davis 1974, 1983 44 35JA19 Lost Creek Davis 1974, 1983 45 35JA20 Lost Creek Davis 1983 46 35JA23 Fawn Butte Nilsson and Kelly 1991 FIGURE 5. Continued ..... o ..... r"-, r'-'~'-'-'-'" ! CR"TER j I LAKE Ii@a,e'lLake I i N" 110,.."l i . P" RK • I IL._._.__._ ....i 4 BAS INUMPQUA MILES 5 0 10, , , , , , . , / I /. / / I / .... -- - /)'cl'~ ~ J / -- // ----- ....... .",.,.--- ....... _ ..... """u ... DIVIOE -Q"ou E - (J4/p O FIGURE 6. Umpqua Basin Sites (map and reference key). ......o I\) Ref. No. Site No. Site Name Reference 48 3500275 Sylmon Lyman 1985 49 3500274 Orchard Simmons and Gallaghar 1985 50 350036 Crispen Baxter and Minor 1987 51 3500412 Coffee Creek Musil and Minor 1989; Baxter 1988 52 3500413 Coffee Creek Baxter 1988; Musil and Minor 1989 53 Tiller 1 Tiller Snyder 1979 54 Tiller 6 Tiller Snyder 1978 55 3500205 S.Ump.RS-U Minor 1987 56 3500205 S.Ump.RS-L Minor 1987 57 3500209 Hughes I Keyser and Carlson 1987 58 3500212 Time Sq. RS Minor and Connolly 1987 59 3500396 Sprint Baxter and Minor 1987 60-63 3500219 Section Crk O'Neill 1991b 64 3500395 Grubbe Ranch O'Neill 1989b 65 350058 Glide Churchill and Jenkins 1985 66 350061 Whistler's Connolly 1982 67 350067 Winchester O'Neill 1989a 68 3500252 Gatchel Ottis and West 1984 69-73 3500153 Narrows O'Neill 1989b 74 3500359 Swiftwater O'Neill 1990 75 35D0383 Susan Crk Musil 1992 76 3500278 Bogus Winthrop 1989 77 3500126 Steamboat Brauner and Honey 1977 78 350011 Lower Rhody Minor 1976 79 350040 Cavitt Crk Snyder and Honey 1979 80-82 3500401 Dry Creek O'Neill 1991a, 1992 FIGURE 6. Continued ~ 0 to) Ref. No. Site No. Site Name Reference 83 3500372 Reynolds Churchill 1986 84-86 3500422 Island O'Neill 1991 a 87 3500418 Apple Ck. Bnch O'Neill 1991 a 88 3500265 Apple Creek O'Neill 1991a 89 3500421 Copeland Crk O'Neill 1991 a 90 3500161 Medicine Crk Snyder 1981b 91 3500187 Powerful 1 Winthrop and Gray 1987 92 3500227 Powerful 2 Winthrop and Gray 1987 93 3500379 Snuffout Jenkins 1988 94 3500397 Shivigny East O'Neill 1988a 95 3500289 Uttle Oak Berryman 1987b 96 35D0399 Snowbird Jenkins and King 1988 97 3500160 Muddy Snyder 1981a 98 3500398 Powerline O'Neill 1988c 99 3500389 Umpy RS Baxter 1987 FIGURE 6. Continued -'" o ,J::o. TABLE 1. Rogue Basin Site Data 105 No. of No. of No. of No. of No. of Total No. of proJec- No. E~rr- Battered Other Ground- Artifacts Total No. OtherRecord Site Site tie of Mo lied No. of Cobble Cobble Stone (non- Total Total No. of Bone Other No. No. Name Points Bifaces Rakes Cores Tools Tools Tools penshable) Debitage of Bone Artifacts Artifacts HOLisepit Midden Burial Hearth Features 1 35CU84 Maria! 251 86 2357 33 204 49 23 3003 14276 2327 13 2 JA01 Gold Hill 3 J04 Ritsch 52 35 288 12 38 18 7 450 7 2 Y 4 J016 Marthaller 222 228 154 106 128 838 16375 322 Y Y Y Y 5 JA21 saltsgaver 18 9 97 4 32 3 17 180 942 40 2 Y Y 6 JA25 Far Hills 91 43 233 30 64 10 41 512 1 20 V y y 7 JA42 Applegate 204 243 494 35 88 39 79 1182 2000 17 25 Y 8 JA47 Applegate 348 587 990 94 39 5 40 2103 29365 4 Y 9 JA47-1 Applegate 44 113 160 14 12 13 356 10 JA47·2 Applegate 304 474 830 80 26 6 27 1747 4 Y 11 Sail Creek Salt Creek 17 59 101 18 10 2 6 213 3901 491 12 JA133 RRNF 1 3 11 3 1 19 152 1 13 JA189 Trail 12 22 49 9 5 9 106 1821 14 JA190 Trail 7 4 27 4 4 2 7 55 875 15 JA191 Reeder 17 3 10 4 4 1 39 168 162 Y Y 16 JA197 UtlJe Butte 23 59 37 13 9 3 2 146 3185 199 17 JA10 Elk Creek 17 13 28 24 11 3 96 18 JA11 Elk Creek 1 14 13 11 2 1 42 855 30 19 JA27A Joham 1 190 289 731 147 108 6 87 1558 24267 1870 5 18 y y 20 JA27A·1 Joham 1 172 191 459 101 81 3 51 1058 5 8 y y y 21 JA27A·2 Joham 1 18 98 272 46 36 3 36 509 10 22. JA27B Elk Creek 38 76 245 33 22 1 22 437 7500 3 Y 23 JAS9 Elk Creek 447 570 982 147 61 1 43 2251 22301 3 72 y y y 24 JA100 Elk Creek 936 1072 1896 535 147 9 128 4723 53800 17 100 Y y y y 25 JA101 Elk Creek 56 79 115 26 27 5 37 345 7622 1193 2 31 Y y 26 JA102 Elk Creek 29 18 22 7 2 3 81 1784 15 27 JA103 Elk Creek 5 13 20 5 43 11n 104 28 JA105 Elk Creek 2 3 1 1 7 383 18 29 JA107 Elk Creek 93 105 138 55 54 26 471 10287 16 y 30 JA110 Elk Creek 2 4 3 9 209 31 JA112 Elk Creek 4 11 6 1 1 23 2132 32 EC-2 Elk Creek 43 78 92 21 9 3 246 6243 478 Y 33 Island Elk Creek 3 6 6 1 16 561 13 34 Winningham Elk Creek 12 18 7 2 2 2 43 1086 24 35 Zimmerly Elk Creek 4 19 37 2 1 63 2150 20 36 JAS Lost Creek 27 12 17 34 6 e 104 37 JA6 Lost Creek 64 28 131 24 24 ~I 276 y 38 JA7 Lost Creek 4 2 15 22 39 JAB Lost Creek 10 10 23 10 6 7 66 2 40 JA12 Lost Creek 16 14 24 13 6 fJ 81 y 41 JA14 Lost Creek 7 9 43 17 7 4, 87 786 42 JA16 Lost Creek 47 33 111 13 10 ti 220 2 y 43 JA18 Lost Creek 32 28 74 10 21 E; 173 1 Y 44 JA19 Lost'Creek 60 15 83 4 6 1 170 1 Y 45 JA20 Lost Creek 1 7 47 19 3 3 81 46 JA23 Fawn Bulle 74 149 141 23 7 S 398 17992 1081 3 2 Y 47 JA29 Lost Creek 1 8 2 11 1 • '. t I, (' TABLE 2. Rogue BaBin Site Data, Computations 106 Percent Percent Percent Perc3nt Total Total Non· pro~ectileAmount Prolec- ~- Battered Other Percent Chipped Poln Chip- p~ectile Total Ground- Total CobbleRecord Site Site Exca· lie Percent M lied Percent Cobble Cobble- Ground- Stone DebitaW ~Stone oint Tool stone Cobble ToolNo. No. Name vated' Points Bifaces Rakes Cores Tools stones stone Tools Densi 001 Density Density Density Density Tools Density 1 35CU84 Marial .083 .028 .784 .011 .067 .016 .007 2443 253 2 JA01 Gold HiD 738.00 3 J04 Rilsch 144.00 .115 .077 .640 .026 .084 .040 .015 323 2.24 .3H 3.13 .0486 56 .3889 4 J016 Marthaller 98.00 .264 .272 .183 .126 .152 382 167.09 3.90 2.2'7 8.55 1.3061 106 1.0816 5 JA21 SaI~aver 10.00 .100 .050 .538 .022 .177 .0113 .094 106 94.20 10.60 1.80 18.00 1.7000 35 3.5000 6 JA25 Far Ills 34.00 .177 .084 .455 .058 .125 .01!} .080 276 8.12 2.61) 15.06 1.2059 74 2.1765 7 JA42 Applegate 42.00 .172 .205 .417 .029 .074 .03:3 .066 737 17.55 4.813 28.14 1.8810 127 3.0238 8 JA47 Applegate 107.00 .165 .279 .470 .044 .018 .019 1577 274.44 14.74 3.2!5 19.65 .3738 44 0.4112 9 JA47-1 Applegate .123 .317 .449 .039 .033 .036 273 12 10 JA47·2 ~~egate .174 .271 .475 .045 .014 .015 1304 32 11 Salt Creek Sa t Creek 15.00 .079 .277 .474 .084 .046 .028 160 260.07 10.67 1.1:J 14.20 .4000 12 .8000 12 JA133 RRNF 0.60 .052 .157 .578 .157 .052 4 253.33 6.67 31.67 1.6667 3 5.000 13 JA189 Trail 5.50 .113 .207 .462 .084 .047 .084 71 331.09 12.91 2.m 19.27 1.6364 5 .9091 14 JA190 Trail 6.50 .127 .072 .490 .072 .072 .036 .127 31 134.62 4.77 1.0a 8.46 1.0769 6 .9231 15 JA191 Reeder 1.50 .435 .076 .256 .102 .102 .025 13 112.00 8.67 11.3:3 26.00 .6667 4 2.6667 16 JA197 Uttle Butte 6.00 .157 .404 .253 .089 .061 .020 .013 96 530.83 16.00 3.83 24.33 .3333 12 2.0000 17 JA10 Elk Creek 3.00 .177 .135 .291 .250 .114 .031 41 13.67 5.6:7 32.00 1.0000 11 3.6667 18 JA11 Elk Creek 3.10 .023 .333 .309 .261 .047 .023 27 275.81 8.71 O.$~ 13.55 .3226 2 .6452 19 JA27A Joham 1 24.00 .122 .185 .469 .094 .069 .055 1020 1011.13 42.50 7.9:~ 64.92 3.625 114 4.7500 20 JA27A-1 Joham 1 .162 .180 .433 .095 .076 .048 650 84 21 JA27A-2 Joham 1 .035 .192 .534 .090 .070 .070 370 39 22 JA278 Elk Creek 26.00 .087 .173 .560 .075 .050 .050 321 288.46 12.35 1.4fi 16.81 .8462 23 .8846 23 JAS9 Elk Creek 70.00 .198 .253 .436 .065 .027 .019 1552 318.59 22.17 6.39 32.16 .6143 62 .8857 24 JA100 Elk Creek 159.00 .198 .227 .401 .113 .031 .027 2968 338.36 18.67 5.8!) 29.70. .8050 156 .9811 25 JA101 Elk Creek 21.10 .162 .229 .333 .075 .078 .014 .107 194 361.23 9.19 2.6!; 16.35 1.7536 32 1.5166 26 JA102 Elk Creek 40.40 .358 .222 .271 .086 .024 .037 40 44.16 .99 .7'.? 2.00 .0743 2 .0495 27 JA103 Elk Creek 4.40 .116 .302 .465 .116 33 267.50 7.50 1.1'~ 9.77 28 JA105 Elk Creek 12.20 .285 .428 .142 .142 5 31.39 .41 0.57 .0820 29 JA107 Elk Creek 107.70 .197 .222 .293 .116 .114 .055 243 95.52 2.26 .81:. 4.37 .2414 54 .5014 30 JA110 Elk Creek 2.00 .222 .444 .333 6 104.50 3.00 4.50 31 JA112 Elk Creek 12.19 .173 .478 .260 .043 .043 17 174.90 1.39 .3:1 1.89 .0820 32 EC·2 Elk Creek 6.00 .174 .317 .374 .085 .036 .012 170 1040.50 28.33 7.n 41.00 .5000 9 1.5000 33 ISLAND Elk Creek 4.50 .187 .375 .375 .062 12 124.67 2.67 .6'7 3.56 34 WiMi~ham Elk Creek 8.55 .279 .418 .162 .046 .046 .046 25 127.02 2.92 1.40 5.03 .2339 2 .233935 Zimme y Elk Creek 10.60 .063 .301 .587 .031 .015 56 202.83 5.28 .313 5.94 .0943 2 .1887 36 JAS Lost Creek .259 .115 .163 .326 .057 .076 29 6 37 JA6 Lost Creek 62.00 .231 .101 .474 .087 .087 .018 159 2.56 1.0:3 4.45 .0806 24 .3871 38 JA7 Lost Creek 6.00 .181 .090 .681 .045 17 2.83 .6'7 3.67 1 .1667 39 JAB Lost Creek .151 .151 .348 .151 .090 ..106 33 6 40 JA12 Lost Creek .197 .172 .296 .160 .074 .098 38 6 41 JA14 Lost Creek 5.00 .080 .103 .494 .195 .080 .046 52 157.20 10.40 1.40 17.40 .8000 7 1.4000 42 JA16 Lost Creek 13.00 .213 .150 .504 .059 .045 .027 144 11.08 3.6:! 16.92 .4615 10 .7692 43 JA18 Lost Creek 6.50 .185 .161 .427 .057 .121 .046 102 15.69 4.92 26.62 1.2308 21 3.2308 44 JA19 Lost Creek 12.40 .352 .088 .488 .023 .035 .005 98 7.90 4.8'~ 13.71 .0806 7 .5645 45 JA20 Lost Creek 11.50 .012 .086 .580 .234 .037 .012 .037 54 4.70 .O!J 7.04 .2609 4 .3478 46 JA23 Fawn Bulle 14.94 .185 .374 .354 .057 .017 .007 290 1204.28 19.41 4.91; 26.64 .2008 8 .5355 47 JA29 Lost Creek 1.20 .090 .727 .181 9 7.50 9.17 1.6667 ·Cubic meters 107 TABLE 3. Umpqua Basin Site Data --= No. of No. of No. of No. of No. of Tolal No. of Pro/ec- No. ~e- Battered Other Ground- Artifacts Total No. Record Site Site tie of M llied No. of Cobble Cobble Stone: (non- Total Total No. of Bone Other Other No. No. Name Points Bifaces Rakes Cores Tools Tools Tools penshable) Debitage of Bone Artifacts Artifacts Housepit Midden Burial Hearth Features 48 00275 Sylmon 14 7 90 49 105 20 13 298 Y 49 00274 Orchard 3 3 8 6 2 4 26 293 1 Y 50 0036 Crispen 33 28 59 6 18 2 5 151 5300 1693 Y Y 51 00412 Coffee Crk 2 3 11 4 1 21 667 32 52 00413 Coffee Ok 5 3 8 1 3 21 126 38 53 Tiller 1 Til 1 1 1 34 36 51 1 54 Till 6 Till 6 3 1 52 6 62 124 6 55 00205 S.Umprs-U 45 42 104 18 6 215 7657 51000 105 Y 56 00205 S.Umprs-L 105 42 100 11 3 2 2 265 3821 23000 13 3 Y 57 00209 Hughes 15 15 11 2 43 896 1724 1 1 58 00212 Time Sq Rs 26 3 5 4 38 414 30000 1 3 Y 59 00396 Sprint 13 1 11 1 2 28 834 81 60 00219 Section CK 111 201 487 42 30 7 11 889 70182 4662 2 Y Y 61 00219 Section-I 44 88 225 20 13 4 6 400 62 00219 Section-III 25 42 89 6 5 1 2 170 63 00219 Section-II 42 71 173 16 12 2 3 319 64 00395 Grubbe 5 2 24 2 12 1 23 69 5n y ..~ 65 0058 Glide 9 36 67 18 18 3 151 3572 55 y 66 0061 Whistlers 3 1 1 5 6 17 48 i 67 0067 Winchester 2 1 2 5 83 68 0052 Gatchel 7 24 86 14 4 4 8 147 2939 Y Y Y 69 00153 Narrows 66 80 196 44 167 17 11 581 8913 55 15 Y Y Y Y Y :1 70 00153 Narrows-I 38 32 58 16 50 2 1 197 13 eli" 71 00153 Narrows-II 12 24 53 6 29 6 130@ ;W 72 00153 Narrow-III 13 23 73 17 89 2 4 221 2 ~i 73 00153 Narrow-IV 2 1 12 5 10 30 .,"' 74 00359 Swiftwatr 1 14 15 85 ~4 75 00383 Susan Ck 4 9 11 2 1 3 30 1836 Y[ ~~ 76 00278 Bogus 22 27 70 5 5 130 3654 165 ~;' n 00126 Steamboat 10 1 13 4 281" ! 78 0011 Rhody79 0040 Cavitt Creek 1 11 12ll' 80 00401 OryCk 5 13 34 3 56 2760 81 00401 Ory Ck-E 2 7 18 2 30 1790 82 00401 Ory Ck·L 3 6 18 1 28 970 83 00372 ReynoldS 5 23 24 2 1 55 1835 84 00422 Island CMP 5 12 17 8 42 2174 85 00422 Island-E 3 12 13 6 34 1906 86 00422 Island-L 2 4 2 8 268 87 00418 Apple Bnch 4 3 11 18 1502 3 88 00265 Apple Ck 2 4 18 4 2 30 799 89 00421 Copeeland 1 8 19 28 3293 51 90 00161 Medicine 12 10 17 1 1 41 4742 91 00187 Pwrfl-1 11 13 26 3 2 8 63 1326 31 Y 92 00227 Pwrfl-2 17 30 47 9 4 2 109 2558 349 Y 93 00379 Snuff Out 6 9 4 19 381 94 00397 Shivigny 16 15 21 12 2 66 2369 4 95 00289 Little oak 4 21 1 26 415 96 00399 Snowbd 2 18 6 1 2 29 351 97 00160 Muddy 5 7 10 22 98 00398 Pwrline 1 1 55 99 00389 Umpy 96 22 89 3 11 221 3100 3294 --------------- - 108 TABLE 4. Umpqua Basin Site Data, Computations ::- Percent Percent Percent Percent Total Total Non-prolectileAmount Prolec- E~t Battered Other Percent Chipped Poin Chip- p~etile Total Ground- Total CobbleRecord Site Site Exca- tie Percent Mo lied Percent Cobble Cobblo· Ground- Stone Debita~e ":d Stone int Tool stone Cobble ToolNo. No. Name vated* Points Bifaces Flakes Cores Tools stone:l stone Tools Densi 001 Density Density Density Density Tools Density 48 00275 Sylmon 23.00 .047 .023 .302 .164 .352 .06i .043 97 4.22 .61 12.96 .5652 125 5.4348 49 00274 Orchard 3.60 .115 .115 .307 .230 .076 .153 11 81.39 3.06 .83 7.22 6 1.6667 50 0036 Cri n 5.40 .218 .185 .390 .039 .119 .013 .033 87 981.48 16.11 6.11 27.96 .9259 20 3.7037 51 00412 Co~Crk 8.80 .095 .142 .523 .190 .047 14 75.80 1.59 .23 2.39 1 .1136 52 00413 Coffee Crk 7.10 .238 .142 .381 .047 .142 .047 11 17.75 1.55 .70 2.96 .1408 3 .4225 53 TILLER 1 Til 1 3.30 .027 .027 .944 35 15.45 10.61 .30 10.91 54 TILL 6 Till 6 2.20 .048 .016 .838 .096 53 56.36 24.09 1.36 28.18 55 00205 S.Umprs-U 4.90 .209 .195 .483 .083 .027 146 1562.65 29.8 9.18 43.88 6 1.2245 56 00205 S.Umprs-L 3.50 .396 .158 .3n .041 .011 142 1091.71 40.57 30.00 75.71 .5714 5 1.4286 57 00209 Hughes 3.40 .348 .348 .255 .046 26 263.53 7.65 4.41 12.65 .5882 58 00212 Time Sqrs 1.20 .684 .078 .131 .105 8 345.00 6.67 21.67 31.67 4 3.3333 59 00396 Sprint 1.30 .464 .035 .392 .035 .071 12 641.54 9.23 10.00 21.54 1.5385 60 00219 section Crk 22.30 .124 .226 .547 .047 .033 .012 688 3147.17 30.85 4.98 39.87 .4933 37 1.6592 61 00219 Section-I .110 .220 .562 .050 .032 .0HI .015 313 17 62 00219 Section-III .147 .247 .523 .035 .029 .011 131 6 63 00219 Section-II .131 .222 .542 .050 .037 244 14 64 00395 Grubbe 3.30 .072 .029 .347 .029 .173 .0141 .333 26 174.85 7.88 1.52 20.91 6.9697 13 3.9394 65 0058 Glide 6.40 .059 .238 .443 .119 .119 .019 103 558.13 16.09 1.41 23.59 .4688 18 2.8125 66 0061 wtistlers .30 .176 .058 .058 .294 .0511 .352 1 160.00 3.33 10.00 56.67 20.0000 6 20.0000 67 0067 WU1Chester 2.00 .400 .200 .400 3 41.50 1.50 2.50 2 1.0000 68 0052 Gatchel 6.60 .047 .163 .585 .095 .027 .02;'. .054 110 445.30 16.67 1.06 22.27 1.2121 8 1.2121 69 00153 Narrows 22.60 .113 .137 .337 .075 .287 .0211 .018 276 394.38 12.21 2.92 25.71 .4867 184 8.1416 70 00153 Narrows-I .192 .162 .294 .081 .253 .OW 90 52 71 00153 Narrows-II .092 .184 .407 .046 .223 .046 n 29 72 00153 Narrows-III .058 .104 .330 .076 .402 .018 96 91 73 00153 Narrows-IV .066 .033 .400 .166 .333 13 10 74 00359 Swiftwater 10.50 .066 .933 0 8.10 .10 1.43 14 1.3333 75 00383 Susan Crk 8.60 .133 .300 .366 .066 .033 .100 20 213.49 2.33 .47 3.49 .3488 1 .1163 76 00278 Bogus 22.50 .169 .207 .538 .038 .038 97 162.40 4.31 .98 5.78 6 .2667 n 00126 Steamboat 2.40 .357 .035 .464 .142 14 5.83 4.17 11.67 4 1.6667 78 0011 Rhody .30 0 79 0040 Cavitt Crk 2.00 .083 .916 12 6.00 6.00 80 00401 Dry Creek 2.80 .089 .232 .607 .053 .01:1 47 985.71 16.79 1.79 20.00 4 1.4286 81 00401 Dry Crk-E .066 .233 .600 .066 .03:1 25 1 82 00401 Dry Crk-L .107 .214 .642 .035 24 1 83 00372 ReynoldsS 6.80 .090 .418 .436 .036 .018 47 269.85 6.91 .74 8.09 1 .1471 84 00422 Island Cmp 6.60 .119 .285 .404 .190 29 329.39 4.39 .76 6.36 8 1.2121 85 00422 IsIand-E 4.40 .088 .352 .382 .176 25 433.18 5.68 .68 7.73 6 1.3636 86 00422 IsIand-L 2.20 .250 .500 .250 4 121.82 1.82 .91 3.64 2 .9091 87 00418 Apple Bnch 2.60 .222 .166 .611 14 5n.69 5.38 1.54 6.92 88 00265 N>Ple Crk 2.30 .066 .133 .600 .133 .066 22 347.39 9.57 .87 13.04 .8696 4 1.7391 89 00421 CO~elanD 1.10 .035 .285 .678 27 2993.64 24.55 .91 25.45 90 00161 Medicine 10.50 .292 .243 .414 .024 .024 27 451.62 2.57 1.14 3.90 .0952 1 .0952 91 00187 PwrfI-1 5.80 .174 .206 .412 .047 .031 .127 39 228.62 6.72 1.90 10.86 1.3793 2 .3448 92 00227 PwrfI-2 3.70 .156 .275 .431 .082 .036 .018 n 691.35 20.81 4.59 29.46 .5405 4 1.0811 93 00379 Snuff Out 4.80 .315 .473 .210 15 79.38 3.13 3.96 94 00397 Shivigny 3.20 .242 .227 .318 .181 .030 36 740.31 11.25 5.00 20.63 2 .625 95 00289 Little Oak 2.50 .153 .807 .038 21 166.00 8.40 1.60 10.40 96 00399 Snowbound 3.20 .069 .620 .206 .034 .069 24 109.69 7.50 .63 9.06 2 .625 97 00160 Muddy 4.30 .227 .318 .454 17 3.95 1.16 5.12 98 00398 Powemne 1.60 1.000 1 34.38 .63 .63 ttl:. 99 00389 Limov 2.25 .434 .099 .402 .013 .049 111 13n.78 49.33 42.67 98.22 4.8889 ; ,~ ·Cubic meters , . • 109 The sites in the sample consist mainly of small test excavations and larger data recovery excavations conducted in the Rogue River basin, primarily in the eastern part, and in the drainage basins of the North and South Umpqua Rivers. Not all sites excavated in this area were included in this analysis, although as large a sample as possible was desired for this study. Those which were excluded did not have reports giving the very basic information used in this analysis, such as the numbers and types of artifacts recovered. Several problems were encountered while doing the quantitative analyses. The most pervasive was the lack of congruence in report standards. The data used in this study were taken from reports for site excavations; it was sometimes difficult to determine even such essential information as numbers of artifacts recovered and amount of material excavated. The resulting raw data (Tables 1 and 2) represent my best assessment of an investigator's findings for the sites used in this study. In compiling this material, I was helped by the recent completion of the Cultural Resource Overview of the Umpqua National Forest. Southwestern Oregon (Beckham and Minor 1992). This document compiles, for the Umpqua Basin sites, several of the statistics (e.g., volume excavated) needed for this study. Other problems inherent in a study of this sort reflect the sampling biases built into using site data generated primarily by project-oriented cultural resource management work. For example, there are no well- excavated, valley floor village sites in the sample, although there are village sites from the Lipper reaches of the main rivers and their tributaries. The few ---...,,--- :!./Jt'if -t' .. ltn__ 110 sites in the sample which do qualify as riverlvalley settlements were either minimally tested, heavily disturbed, or excavated by amateurs whose reports poorly present the quantitative data needed. The sites in the Rogue Valley database are heavily dominated by dam-related project studies, hence most of the sites occur at moderate elevations along tributaries to the Rogue River. Conversely, many of the sites in the Umpqua Basin have been excavated in response to Forest SelVice and Department of Transportation road projects, and are in upland or travel corridor locations. Despite these limitations, the 83 sites examined in this study provide a healthy body of data, from which it is possible to derive useful conclusions regarding the past. ---JA.,--- 111 CHAPTER V FUNCTIONAL ANALYSES: ROGUE BASIN SITES The database from the Rogue Basin consists of 43 sites. It is heavily dominated by sites from the major dam-building projects of the last twenty years (Lost Creek Dam, Applegate Dam, Elk Creek Dam); information from a handful of other sites complements that from the dam project sites. The initial sequential numbers are those assigned to each site when entered into the database. These case numbers identity sites throughout the report. Where site data are separated into components, those components are given separate record numbers and analyzed separately. Site Function Based on Qualitative Analysis The initial qualitative analysis serves to introduce each site. In this analysis, a brief description of each site is given, as well as a functional designation. Generally the functional designation follows the original investigator's interpretation; occasionally I have rendered my own opinion where the site report does not indicate a functional type or that type seems at odds with the material presented. 1. 35CU84. Marial seasonal camp b ___ ---~---~ The site produced an abundance of chipped stone and cobble tools, as well as numerous debitage and bone fragments and one cobble paving stone feature. The assemblage covers millennia; no functional variation has been noted by the investigators, who consider it served as a seasonal encampment throughout its history. 2. Gold Hill village 112 The presence of "living areas," burials, hearths, abundant and varied artifacts including ceremonial/wealth items, and the site's location on the Rogue River contribute to its designation as a village site. 3. 35J04. Ritsch village The presence of housepits, bone, and a variety of stone tools, together with the site's location along the Rogue River contribute to its designation as a village site. 4. 35J016. Marthaller village This site is considered a village site due to its location along the Rogue River, the presence of housepits, hearths, midden deposit, and burials, as well as numerous tools and a variety of non-utilitarian implements such as a pipe, and decorated bone. 5. 35JA21. Saltsgaver seasonal camp/camas harvest site This site is distinguished among the sample of sites by having over 100 camas-roasting ovens, near a camas-bearing wet meadow. The site also produced a midden deposit and a variety of stone implements. The site functioned as an encampment where seasonal tasks (camas harvest) was undertaken. 6. 35JA25. Far Hills village This site is considered a small village based on the presence of hearths, residential floors, and (reported) numerous burials. In addition to stone tools, dentalium shells were reported; other non- utilitarian items such as beads, mineral pigment, and crystals were recovered. It is located along the Rogue River. 7. 35JA42. Applegate village ___acib ~ This is a late, contact period site which consists of several housepits, clearly defined, and their contents. The assemblage included masses of small bone fragments and debitage, and utilitarian and non- utilitarian items (pipes, beads, schist disc). The site serves as a good example of the small "homestead" type defined by Pettigrew and Lebow (1987) for this area. 113 The site produced a variety of stone tools including cobble and groundstone artifacts. It is located above a creek at a moderate elevation. The artifacts and location contribute to its designation as a seasonal (summer) base camp. a. 9. 1Q. 35JA47. Applegate village (late); task/seasonal 8 = Data from both components combined camp (early) 9 = Early component 10 = Late component This is a dual component site with housepits and a dense assemblage of artifacts and bone associated with the later component. An earlier component consists of an assemblage of stone tools indicative of a task site or a seasonal camp. seasonal camp11. 35JA77. Salt Creek 12. 35JA133 seasonal camp A very minimal amount of excavation (.6 cubic meters) produced groundstone, cobble, and chipped stone tools and debitage, leading the excavator to designate it a seasonal camp. 13. 35JA189. Trail. Casey village? The site produced a dense deposit of chipped stone, groundstone, and cobble tools, as well as a midden deposit; housepits were suspected by the investigators and it is located along the Rogue River. 14. 35JA19Q. Trail. Casey seasonal camp A high density and variety of chipped stone, groundstone, and cobble tools suggest that this site is a seasonal encampment. 15. 35JA191. Reeder Reservoir seasonal camp A minimal amount of excavation at this upland site produced midden deposit, bone, a hearth, and chipped stone, groundstone, and cobble tools. The site's location, features, and artifacts suggest a seasonal camp. 16. 35JA197. Little Butte seasonal camp The site is located on an upland meadow, and produced a variety of chipped stone, groundstone, and cobble tools characteristic of upland base camps. . 114 17. 35JA10. Duvalt (Elk Creek) seasonal camp or task site The site produced a variety of chipped stone, groundstone, and cobble tools which may indicate a seasonal camp. However, the excavator considered it a task site. 18. 35JA11. Ross (Elk Creek) task site Although housepits were reported for this site, none were located during the excavations and the investigators consider the relatively sparse assemblage of chipped stone and cobble tools to indicate a temporarily occupied task site. 19. 20. 21. 35JA27A. Joham 1 (Elk Creek) village or seasonal camp or (late); seasonal camp or task site (early) 19 = Data from both components combined 20 = Late component 21 = Early component This dual component site has midden deposit, a burial, and several rock features associated with an abundance of chipped stone, groundstone, and cobble tools from the later component. The earlier component is less dense and not associated with the features; the excavators suggest it was a seasonal, temporary use site. 22. 35JA27B. Elk Creek task site or seasonal camp The site has a moderate density of chipped stone, cobble, and groundstone artifacts, and several rock features. The investigators suggest it was a task site, though the variety of implements and presence of features may indicate a seasonal camp. 23. 35JA59. Elk Creek village Housepit, hearth, midden, postmolds, and abundance of fire-eracked rock (FCR) bone, debitage, and a variety of tools indicate that this was a village. 24. 35JA100. Elk Creek village bt__ ---~---~ The site contains housepits, midden, burials, hearths, postholes, and an abundance of bone fragments, a variety of implements, ceramics, and non-utilitarian objects such as pigment stones and crystals. It is a good example of a late period winter village, probably of the "homestead" type (i.e., small settlements). 115 25. 35JA101. Elk Creek village or seasonal camp Housepits were reported for this site, which was vandalized. Investigators found midden deposit and rock features. The site produced bone fragments, a variety of stone tools, and ceramics. The inve~tigators consider it a permanent or semi-permanent village or camp site. 26. 35JA102. Elk Creek task site The site produced a low density of artifacts, primarily chipped stone with a few cobble and groundstone tools. The investigators suggest it was used on a short-term basis for certain tasks. 27. 35JA103. Elk Creek task site Although housepits were reported, the artifact assemblage is limited to chipped stone tools and debitage; the investigators class this as a short-term task site. 28. 35JA105. Elk Creek task site This site produced a very light assemblage of chipped stone artifacts, with one groundstone implement. It is classed as a task site by its investigators. 29. 35JA107, Elk Creek seasonal camp The site produced a variety of chipped stone, cobble, and groundstone tools, some bone, and some FCR. One feature of groundstone with FCR was excavated. The excavators conclude it was a seasonal camp. 30, 35JA110. Elk Creek task site A limited amount of excavation produced a small assemblage of chipped stone tools and debitage; the investigators consider this a task site. 31. 35JA112, Elk Creek task site Excavation produced a limited amount of chipped stone tools and debitage; this is considered a task site. 32. EC-2. Elk Creek seasonal camp 116 The assemblage includes bone fragments and FCR. as well as a variety of chipped stone. cobble, and groundstone tools; one rock feature was excavated. The investigators consider this a seasonal camp. 33. Island site. Elk Creek task site A light assemblage of chipped stone tools and a small amount of bone indicates this was a task site. 34. Winningham. Elk Creek task site A small assemblage of chipped stone tools with a few cobble tools and groundstone led the investigators to classify this as a task site. 35. Zimmerly. Elk Creek task site An assemblage of predominantly chipped stone tools and debitage. with a few cobble and groundstone tools. prompted classification of this site as a task site. 36. 35JA5. Lost Creek task site An assemblage of chipped stone. cobble and groundstone tools on a the lowest terrace above the Rogue River led the investigator to consider this a fishing campltask site. 37. 35JA6. Lost Creek task site or seasonal camp? The site produced a possible hearth, with chipped stone. cobble. and groundstone tools. The investigator considers it a task site/fishing site. since it is on the Rogue River; the assemblage and possible feature may indicate a seasonal camp. 38. 35JA7. Lost Creek task site This light lithic scatter produced a few chipped stone tools and was classified as a task site. 39. 35JA8. Lost Creek seasonal camp? b -_...._-- The assemblage of chipped stone. cobble and groundstone artifacts. with two stone discs. appears to represent a task site or seasonal camp. However. the investigator mentions that housepits were reported, although these features were not noted during the excavation. 40. 35JA12, Lost Creek village? 117 The report notes one housepit, and records a variety of chipped stone, cobble, and groundstone tools. The assemblage and feature suggest a village site. 41. 35JA14, Lost Creek seasonal camp The site contains chipped stone as well as cobble and groundstone tools, and a firedrill hearth. It could be a seasonal camp. 42, 35JA16, Lost Creek seasonal camp The site contains a variety of chipped stone tools, cobble tools, groundstone, and two stone discs, and one hearth. A mortar was cached at the site. The artifacts and feature suggest a seasonal camp. 43, 35JA18, Lost Creek seasonal camplvillage? The presence of midden soils and a variety of tools including a stone disc, and a fairly dense deposit of materials indicates a seasonal camp or village site (no housepits were noted or reported). 44. 35JA19, Lost Creek seasonal camp The site has hearths and a variety of chipped stone and cobble tools, with one groundstone artifact; the features and tools suggest a seasonal camp. 45, 35JA20. Lost Creek task site b __ ---_--._-- A light recovery of chipped stone tools with a few cobble and groundstone artifacts suggests that this was a task site. 46. 25JA23, Fawn Butte (Lost Creek) seasonal camp? village? Previous investigations noted possible housepits and burials, though these were vandalized. A dense and varied assemblage suggests that it was a village or seasonal camp. .' Mt__. 118 Density Measures Density Measure 1: Projectile Points vs. Other Chipped Stone Tools Figure 7 illustrates the density of site assemblages compared with one another. Density is computed as the number of items per cubic meter. In Density Measure 1, the density of projectile points for each site's assemblage is plotted against the density of other chipped stone implements for that same site. Each point on the scatterplot represents one site. There is a good correlation between these two types of density represented for each site; that is, those sites with a high density of projectile points generally also have a high density of other chipped stone tools, and vice versa. The sites have been divided into three groups, based on my interpretation of breaks in the distribution of sites on the scatterplot. Group 1 includes those with low densities (Le., task sites), Group 2 includes those with medium densities (Le., seasonal camps), and Group 3 includes those with the highest densities (Le., villages). The sites in these groups (reading generally from the lower left of the plot to the upper right) are presented in Table 5. The table lists the record number, site number and name, the Group to which it was assigned in the qualitative analysis, and the presence or absence of habitation (housepit, midden, burial) or other types of features (e.g., hearths, miscellaneous rock features). ---~_-._- r~ « ~; 12 GROUP 3 u ~ u i I ~ Cl ~ >: !;i ~ ~ ~ I 9 J . 3 IOROUP 1~ J J J J J "'- JJ JJ JJ JH ! JI, J J , 0 11.3 22.5 33.8 45.0 CHIPPED STONE TOOL DENSITY: NUMBER OF CHIPPED STONE! CUBIC METER EXCAVATED Projectile point density, measured as number of projectile points/cubic meter,plotted against chipped stone tool density, measured as number of chipped stone tools/cubic meter. Each cross represents a site. FIGURE 7. Rogue Basin Sites: Density Measure 1. ..... ..... co TABLE 5. Rogue Basin Sites, Density Measure 1 Density Groups Based on Projectile Point/Other Chipped Stone Tools 120 IItrn -----..,.,,-- Qual. Hab Oth. Rec. No. Site No.Name Type Feat. Feat. Group 1 28 JA105, Elk Ck. task no no 31 JA112, Elk Ck. task no no 26 JA102, Elk Ck. task no no 30 JA110, Elk Ck. task no no 3 Ritsch village yes no 33 Island, Elk Ck. task no no 38 JA7, Lost Ck task no no 29 JA107, Elk Ck. seas. camp no yes 37 JA6, Lost Ck. task/seas camp no yes 34 Winningham, Elk Ck. task no no 35 Zimmerly, Elk Ck. task no no 45 JA20, Lost Ck. task no no 14 JA190, Trail seas. camp no no 12 JA133 seas. camp no no 18 JA11, Elk Ck. task ? no 27 JA103, Elk Ck. task? ? no 4 Marthaller village yes yes Group 2 11 JA77, Salt Ck. seas. camp no no 41 JA14, Lost Ck task/seas.camp no no 22 JA27B, Elk Ck. task/seas.camp no yes 5 JA21, Saltsgaver seas. camp yes yes 6 Far Hills village yes yes 25 JA102, Elk Ck seas. camp/village yes yes 13 JA189, Trail village yes no 8 35JA47, Applegate village yes no 16 35JA197 seas. camp no no 42 JA16, Lost Ck. seas. camp no no 44 JA19, Lost Ck. task/seas. camp? no yes Group 3 17 JA10, Elk Ck. seas. camp/task no no 43 JA18, Lost Ck seas. camp/village yes no 87 JA42, Apple~te village yes no 46 JA23, Fawn utte village/seas. camp yes ? 24 JA100, Elk Ck. village/seas. camp yes yes 23 JA59, Elk Ck. village yes yes 32 EC-2 seas. camp no yes 19 JA27A, Joham 1 village yes yes 15 JA191 ·seas. camp yes no 121 Density Measure 2: Debitage Density vs. Total Tool Density Figure 8 illustrates the results of the second density analysis, in which the density of debitage and the density of the total tool assemblage are shown for each site. Here, the density of all the stone tools present at each site is plotted against the density of debitage for each site. Each point represents a single site. This exercise produces a distribution similar to that of Figure 7; there is a high degree of correlation between the density of tools and the density of debitage present for each site. That is, sites with a high density of tools are likely to also have a high density of debitage; the reverse is also true. Again, the sites were divided into three groups, based on my interpretation of breaks in the scatterplot. The groups are presented in Table 6. In this table, sites are presented reading from the lower-left (lowest density) comer to the upper right (highest density comer).1 There is considerable agreement between Density Measure 1, represented in Table 5, and Density Measure 2, represented in Table 6. Group 1 (low density sites) of Table 6, for example, is entirely contained ... ---~_-...~ 1 Different excavation techniques produce different amounts of debitage; in this analysis, most debitage counts are those retrieved from screening with a 1/4" screen. Several investigators employed smaller screens for a fraction of the work, but did not report the debitage retrieved separately. In these cases the entire amount reported was used, giving those sites--several of the Elk Creek sites--a slightly ~Iigher debitage count than would be expected for only 1/4" screen. This slight skewing does not seem to have affected the analysis very much, since there is considerable agreement among site types with the different methods employed. Many sites had no report of total debitage collected, and are excluded from this analysis. rI,70.0' GROUP 3 .~ _.~0 __ . 52.5 ~ < > < .. u ~~ ~;8S:! 35.0 ~ S· IGROUP 2 ... ::Jgg ... ~g 17.5 I """- GROUP I I I I I II 1 0 I I I I II I 325 650 975 DEBITAOE DENSITY: NUMBER OF DEBITAOFl CUBIC METER EXCAVATED 1300 Total tool density, measured as number of tools/cubic meter, plotted against debitage density, measured as number of debitage/cubic meter. Each cross represents a site. FIGURE 8. Rogue Basin Sites: Density Measure 2. -'"I\)I\) 1 : TABLE 6. Rogue Basin Sites, Density Measure 2 Site Density Measured by Debitage and Total Tool Density 123 I'z __ ----------- Qual. Hab Oth. Rec. No. Site No.Name Type Feat. Feat. Group 1 28 JA105, Elk Ck task no no 26 JA102, Elk Ck. task no no 33 Island, Elk Ck. task no no 31 JA112, Elk Ck. task no no 30 JA110, Elk Ck. task no no 29 JA107, Elk Ck. seas.camp no yes 34 Winningham, Elk Ck. task no no 35 Zimmerly, Elk Ck. task no no 4 J016, Marthaller Village yes yes 14 JA190, Trail seas. camp no no Group 2 5 Saltsgaver seas. camp yes yes 27 JA103, Elk Ck. task? ? no 18 JA11, Elk Ck. task? ? no 11 JA77, Salt Creek seas. camp no no 41 JA14, Lost Ck seas. camp/task no no 15 JA191, Reader Res. seas. camp yes yes 22 JA27B, Elk Ck task/seas. camp no yes 8 JA47, Applegate village yes no 25 JA101, Elk Ck Village/seas. camp yes yes 13 JA189, Trail village yes no Group 3 16 JA197 seas. camp no no 24 JA100, Elk Ck. Village/homestead yes yes 23 JA59, Elk Ck. village/homestead yes yes 12 JA133 seas. camp no no 46 JA23, Fawn Butte village/seas. camp yes yes 32 EC-2, Elk Ck. seas. camp no yes 19 JA27A, Joham 1 village/seas. camp yes yes within Group 1 of Table 5; Group 2 in Table 6 overlaps Group 2 in Table 5 with three exceptions (#27 and #18, which are high in Group 1 in Table 5; and #15 which is in Group 3); Group 3 (high density sites) in Table 6 overlaps Group 3 in Table 5 with two exceptions (#16 which is in Group 2 of Table 5 and #12 which is in'Group 1 of Table 5). Overall, 81 percent of the sites subjected to both density measures were placed in the same group with each measure. In addition to this agreement, there is considerable correspondence between site densities and site types based on the qualitative assessment and the presence/absence of features. Those sites which may be described as village sites on the basis of habitation features (Le housepits, middens, and burials) have higher densities than those which do not have such features. As an independent check on the density method, this evidence indicates that density measures can help determine site function. That the density measures seem to work is also probably due to the fact that the sites in the sample have fairly homogeneous depositional environments; highly deflated sites are not compared with those occurring in areas subject to rapid deposition. Also, similar excavation methods make site assemblages roughly comparable. Even though the size of the excavations varied widely, at all sites only a sample of material was excavated, and excavations were usually placed in those areas deemed likely to be most productive. Only a few sites appear very out-of-place with regard to the data presented in the site reports and the investigators' assessments. Both the Marthaller (#4) and Ritsch (#5) sites appear in Group 1, which are low Irtz _ ---".,,----- 124 Itrz 125 density sites interpreted as task sites. The Marthaller and Ritsch sites are located on broad river terraces at the confluence of 'the Applegate and the Rogue Rivers, and believed to be riverside villages based on the qualitative interpretation. The reports for these sites did not give good information regarding excavation procedures. and for both of these sites it was necessary to make very crude estimates of the amount excavated. It is possible that these estimates of volume excavated are inflated, which would bring the artifact density figures down. Alternatively, as riverside sites (among the few in the sample) the excavated material may have included considerable amounts of sterile flood deposits, again lowering the density estimates. Site #8 (JA47) is a dual component site, with a housepit settlement (Le., village) clearly evident in the later component, but it appears in Group 2 (seasonal camps) in each analysis. The density figures were extracted from excavation measurements which applied to both components; however, the resultant lumping may have lowered the density measures, causing the later component to appear with those which represent seasonal camps. Multidimensional Scaling: Comparison of Assemblage Richness and Evenness The Multidimensional Scaling (MOS) analysis compares the sites across two dimensions as illustrated in Figure 9. Each site is placed in space according to its likeness to the other sites, with the most similar sites the closest to one another. (The dimensions are arbitrary measures of distance and do not have particular meaning.) In order to interpret the _____sta.n _ "~ I , 1.67- 1- Group 1 (Task) 01 2- Group 2 (Seasonal Camp) 3· Group 3 (Village) 126 Itt --_..._--- 03 38 45 ~ 0.98- II 2122 35 ~ QQ 30 ~ 0.29- ~ 28 1913 11 2043 44 07 2Z OS10 ~ 23 24 Q.S. -0040 - 12 18 11: 25 ~ ~ ~ 33 ~ 26 15 -1.09- 36 04 31 34 -2!36 -1 !S6 -o.~6 -0.104 o.bs FIGURE 9. Rogue Basin: MDS plot of sites. brr 127 pattern, those sites which were most confidently assigned to a specific type based on both the density and qualitative analyses were identified on the scatterplot. It was discovered that village sites clustered in the middle of the scatter, with seasonal camps around them and task sites dispersed about the plot. This is precisely the pattern predicted initially (see Chapter IV). The MDS plot clari'fies a number of relationships between and among certain sites in the sample. Specifically, a number of sites with equivocal designations could be assigned more definitively to functional categories based on their closeness/distance to the central cluster of village sites. Sites #27 and 39, which may be classed as either seasonal camps or task sites, based on preceding analyses, appear fairly close to the village sites and have assemblages which are more "like" those sites. Hence, they are designated seasonal camps in this MDS analysis. Sites #44 and #18 are more distant from the village and seasonal camp sites, and therefore have assemblages less "like" those types. They are designated task sites in this analysis. Sites #21 and 14 are at some distance from the village group, but close to several seasonal camps; they are therefore designated seasonal camps. Sites #40 and #32 are at some distance from the central cluster but near village site #25, and may be considered at least seasonal camps. The Ritsch (#3) and Marthaller (#4) sites appear as anomalies in this analysis. Though the qualitative assessment, based on the sites' locations, artifacts, and housepits and burials, strongly indicates that they are village sites, they do not cluster with the central group as predicted. Rather, they appear as task sites, among those dispersed about the central configuration ---_..._-._- ...._-------------- 128 of villages and seasonal camps. The reason they appear with the task sites may be because the data in the reports are not an accurate reflection of the reality of the archaeological assemblage. Neither site was professionally excavated, and systematic collection and cataloging of all artifacts may not have occurred. The MDS analysis groups sites based on similarities of artifact class percentages. Table 7 gives the range and standard deviations of the percentages of different tool classes for each type of site; these statistics are illustrated in Figure 10. The central cluster of sites, representing village sites, have the most uniform, and hence most generalized, chipped-stone assemblages. In the table, this fact is represented by the tightest range and lowest standard deviation for the tool types represented. The standard deviation is particularly instructive, since all classes have outliers which skew the ranges somewhat. The standard deviation is a better measure of how tightly clustered about a mean the distribution really is. Village sites also tend to have higher proportions of cobble and groundstone artifacts. Seasonal camps show more variability among the tool classes, but generally less than that demonstrated by the task sites. The assemblages from the task sites are the most specialized, with some sites having high proportions of projectile points and low proportions of other types of tools, for example, and others having few projectile points but high proportions of other tools. This is represented in Table 7 and Figure 10 by the wide range of tool class proportions and comparatively high standard deviations for task sites ____+Mz__~ 129 TABLE 7: Rogue Basin Sites: MDS Descriptive Statistics for Functional Groups f, Group 1 GrO~2 Group 3 Tool Class Task Season Camp Village PPP Range 0-35 3 - 43 11 - 26 Mean 15 16 17 SO 11.8 9.9 4 BFP Range 8 - 47 3 - 40 7 - 37 I Mean 24 17 22 " SO 13 11 7i EMP Range 16 - 68 25 - 78 18 - 64 Mean 39 44 42 SO 16 13 10 CRP Range 0-33 1 - 25 0-11 Mean 14 9 6 SO 11 6 3 BCBP Range 0-9 0-17.7 1.4 - 12.6 Mean 3.2 7.2 6 SO 3.2 4.2 3.8 OCBP Range 0-4.5 0-3.6 0-4 Mean 0.4 0.7 0.8 SO 1.3 1 1.3 GDSP Range 0-14 0-12 0.7 - 15 Mean 4.2 4.5 5.1 SO 4.3 3.6 4.2 SO = Standard Deviation OCBP = other cobble tools PPP = % projectile points BFP = % bifaces EMP = % edge-modified tools CRP = % cores BCBP = % battered cobbles GDSP = % groundstone ... r ---------_. __ .. ----- PERCENT OF ARTIFACTS FOR FUNCTIONAL GROUPS .'-- ''''~t",w~ o 10 20 30 40 50 60 70 80 90 100 ----===-==-==101="""'=====------ -----<====0======------- -----===101====---- 1 PPP 2 3 1 BFP 2 3 1 rI.I EHP 2 rI.I 3d 1ti CRP 2 < 3~ 1BCBP 2 3 1 OCBP 2 3 1 GDSP 2 3 -=====101=====--- --e====0===------ ---=D=--- --<=====0====:===----- -""""""'===I01===~---- ---====0==-- -===l0J====--- -==101===----- ~ =D=- ~ --=D==- =O=- =0= =0= -93=-- -=0= -==0=- o SEll I:.IiY TABU! 1 mean mean i 10 range 1 = task site; 2 • seasonal camps; 3 = village sites FIGURE 10. Rogue Basin: MDS graph of artifact class percentages. .....w o ..._------------------- 131 as a group. Task sites also tend to have lower proportions of cobble tools than the other sites. Groundstone and Cobble Densities; Comparison with Feature Data Densities for groundstone and cobble tools were computed for each site, and the sites arranged in order of increasing density for each artifact class. Feature data are added to this distribution of sites. Features consist of housepits, middens, and burials, hearths, and other rock features. In this analysis, housepits, middens, burials, are considered specifically to indicate village sites. As in the other analyses, three groups were de'fined based on the density of groundstone tools (Table 8) and based on the density of cobble tools (Table 9). Sites were arranged in order of increasing density of groundstone and cobble tools, and the presence/absence of features was compared to this arrangement (Tables 8, 9, and 10). As is readily visible from Tables 8 and 9, sites with higher densities of groundstone and cobble tools are likely to have features present. Furthermore, those with the highest densities of groundstone or cobble tools (Group 3 village sites) have the highest number of habitation features. These relationships are quantified in Table 10, which gives the density figures for each site group (Le., Groups 1, 2, and 3), and compares these density figures to the number and types of features present for each group. Thus, for the cobble density measure in Table 10, Group 1 task sites have cobble densities which range from 0 to .35 (items per cubic meter). ----"".",-- rTABLE 8. Rogue Basin Sites: Groundstone Density and Features ~rd ~te Site .&gount Ground-stone Othero. Name vat~~ Density" Housepit Midden Burial Hearth Features Group 1 33 Island Elk Creek 4.50 27 JA103 Elk Creek 4.40 38 JA7 Lost Creek 6.00 30 JA110 Elk Creek 2.00 3 J04 Ritsch 144.00 .04860 Y 26 JA102 Elk Creek 40.40 .07430 37 JA6 Lost Creek 62.00 .08060 Y 44 JA19 Lost Creek 12.40 .08060 Y 28 JA105 Elk Creek 12.20 .08200 31 JA112 Elk Creek 12.19 .08200 35 Zimmerly Elk Creek 10.60 .09430 46 JA23 Fawn Butte 14.94 .20080 Y 34 Winningham Elk Creek 8.55 .23390 29 JA107 Elk Creek 107.70 45 JA20 Lost Creek 11.50 .26090 18 JA11 Elk Creek 3.10 .32260 16 JA197 Uttle Butte 6.00 .33330 Group 2 8 JA47 Applegate 107.00 .37380 Y 11 Salt Creek Satt Creek 15.00 .40000 42 JA16 Lost Creek 13.00 .46150 y 32 Ee-2 Elk Creek 6.00 .50000 y ...... (.,) I\) rTABLE 8. Continued ---~-_._- ~---~ ----':""" -C-'o\!~~ '!r"':" _ ¥ - -. <~ -- - -~---- ~ -:.~ - -~~ - ~te Site ~ount Ground- Other~rd x~- stone Housepito. o. Name vate • Density·· Midden Burial Hearth Features Group 3 23 JA59 Elk Creek 70.00 .61430 Y Y Y 15 JA191 Reeder 1.50 .66670 Y Y 41 JA14 Lost Creek 5.00 .80000 24 JA100 Elk Creek 159.00 .80500 Y Y Y Y 22 JA27B Elk Creek 26.00 .84620 Y 17 JA10 Elk Creek 3.00 1.0000 14 JA190 Trail 6.50 1.0769 6 JA25 Far Hills 34.00 1.2059 Y Y Y 43 JA18 Lost Creek 6.50 1.2308 Y 4 J016 Marthaller 98.00 1.3061 Y Y Y Y 13 JA189 Trail 5.50 1.6364 12 JA133 RRNF 0.60 1.6667 5 JA21 Saltsgaver 10.00 1.7000 Y Y 25 JA101 Elk Creek 21.10 1.7536 Y Y 7 JM2 Applegate 42.00 1.8810 Y 19 JA27A Joham 1 24.00 3.6250 Y Y * Cubic meters ** Arranged in order of increasing density -A. W W -.... TABLE 9. Rogue Basin Sites: Cobble Tool Density and Features .."~~!=-~--~= -~ Amount Cobble Record Site Site Exca- Tool Olher No. No. Name vated* Density" Housepit Midden Burial Hearth Features Group 1 33 Island Elk Creek 4.50 27 JA103 Elk Creek 4.40 30 JA110 Elk Creek 2.00 28 JA105 Elk Creek 12.20 31 JA112 Elk Creek 12.19 26 JA102 Elk Creek 40.40 .0495 38 JA7 Lost Creek 6.00 .1667 35 Zimmerly Elk Creek 10.60 .1887 34 Winningham Elk Creek 8.55 .2339 45 JA20 Lost Creek 11.50 .3478 Group 2 37 JA6 Lost Creek 62.00 .3871 Y 3 J04 Ritsch 144.00 .3889 Y 8 JA47 Applegate 107.00 .4112 Y 29 JA107 Elk Creek 107.70 .5014 Y 46 JA23 Fawn Butte 14.94 .5355 Y 44 JA19 Lost Creek 12.40 .5645 Y 18 JA11 Elk Creek 3.10 .6452 42 JA16 Lost Creek 13.00 .7692 Y 11 SALTCRK Sa~ Creek 15.00 .8000 22 JA27B Elk Creek 26.00 .8846 Y ...... w ~ TABLE 9. Continued ..~~=z~====-_u·~~~:u Amount Cobble Record Site Site Exca· Tool Other No. No. Name vated· Density·· Housepit Midden Burial Hearth Features Group 3 23 JA59 Elk Creek 70.0 .8857 Y Y Y 13 JA189 Trail 5.5 .9091 14 JA190 Trail 6.5 .9231 24 JA100 Elk Creek 159.0 .9811 Y Y Y Y 4 J016 MarthaJler 98.0 1.0816 Y Y Y Y 41 JA14 Lost Creek 5.0 1.4000 32 EG-2 Elk Creek 6.0 1.5000 Y 25 JA101 Elk Creek 21.1 1.5166 Y Y 16 JA197 Little Butte 6.0 2.0000 6 JA25 Far Hills 34.0 2.1765 Y Y Y 15 JA191 Reeder 1.5 2.6667 Y Y 7 JA42 Applegate 42.0 3.0238 Y 43 JA18 Lost Creek 6.5 3.2308 Y 5 JA21 Saltsgaver 10.0 3.5000 Y Y 17 JA10 Elk Creek 3.0 3.6667 19 JA27A Joham 1 24.0 4.7500 Y Y 12 JA133 RRNF 0.6 5.0000 • Cubic meters •• Sites arranged in order of increasing density ...... Co!) 0'1 ; i ~. < iii ,: I ~f ' ~. , '.ii I TABLE 10. Rogue Basin Sites: Groundstonel Cobble Densities and Features 136 I I " ", hz, Site Density Type N sites Range %SF %MF %HF % All Cobble Density Group 1 10 0-.35 0 0 0 0 Group 2 10 .38 - .88 70% 0 30% 70% Group 3 17 .86 - 5.0 18% 47% 59% 65% Groundstone Density Group 1 17 0-.33 29% 0% 6% 29% Group 2 4 .37 -.5 75% 0% 50% 75% Group 3 16 .61 - 3.6 19% 50% 62010 69% % SF = Percent of sites in the group with only one type of feature present (Le., either housepits, middens, burials, hearths or other rock features). % MF = Percent of sites in the group with multiple types of features present (Le., some combination of housepit, midden, burial, hearth, rock feature) % HF = Percent of sites in the group with habitation features (Le., housepit/living 'floor, midden, burial) present. % All = The total percentage of sites in the group with any type features (%SF plus %MF). i . ; , I I ti I I I 137 These task sites, as a group, have no features of any type associated with them. Group 3 sites, however, have cobble densities which range from .86 to 5.0 (items per cubic meter). Within this group, 18 percent of the sites have a single feature (of any type), and 47 percent have mUltiple types of features (e.g., housepits and middens, burials and hearths). A total of 65 percent of the sites in Group 3 therefore have features of some type; and a total of 59 percent of the sites in this group have features which are specifically associated with village sites (Le., housepits, middens, burials). There is very good correlation between the presence/absence of features and the density of cobble and groundstone tools. Generally, sites with higher densities of cobbles and/or groundstone artifacts are more likely to include some type of feature. Specifically, those sites with multiple types of features and with habitation features (housepits, burials, middens) are more likely to occur in the group containing the highest density of cobble and/or groundstone (Group 3, village sites). Sites with the lowest densities of cobbles and/or groundstone have fewer features. This trend is even more apparent for cobble tools than for groundstone. Site Function Table 11 lists the sites examined in this study, and each site's functional type based on the different analyses. In most cases, there is considerable agreement among the various measures, and the final type assignment is unequivocal. The agreement among the various methods is, in fact, astonishing (see Figure 11): 50· percent of the sites are placed in b_'-. --".,.,,-- r -.,.;;c.. TABLE 11. Rogue Basin, Functional Site Types Multi- Ground- S~erDensity Density Dimensional Cobble stone Habitation Other A , Record S~e Site Qualitative Measure 1 Measure 2 Scaling Density Density Features Features Desig· No. No. Name Assessment Group Group Group Group Group Present Present nation 1 35CU84 Maria! 2 2 2 2 JA01 Gold Hill 3 3 3 J04 Ritsch 3 1 2 2 1 Y 3 4 J016 Marthaller 3 1 1 1 3 3 Y Y 3 5 JA21 Sallsgaver 2 2 2 2 3 3 Y Y 2 6 JA25 Far Hills 2,3 2 2 3 3 Y Y 2 7 JA42 Applegate 3 3 3 3 3 Y 3 8 JA47 Applegate 3 2 2 3 2 2,3 Y 9 JA47·1 Applegate 2 2 2 10 JA47·2 Applegate 3 3 Y 3 11 35JAn Sail Creek 2 2 2 2 2 2 2 12 JA133 RRNF 2 1 3 1 3 3 2 13 JA189 Trai 3 2 2 3 3 3 3 14 JA190 Trai 2 1 1 2 3 3 2 15 JA191 Reeder 2 3 2 1 3 3 Y Y 2 16 JA197 Litle Butte 2 2 3 2 3 2 2 17 JA10 Elk Creek 1,2 3 2 3 3 2 18 JA11 Elk Creek 1 1 2 1 2 1 1 19 JA27A Joham 1 3 3 3 3 3 3 Y 20 JA27A-1 Joharn 1 3 3 Y 3 21 JA27A-2 Joham 1 2 2 2 22 JA27B Elk Creek 2,3 2 2 2 2 3 Y 2 23 JA59 Elk Creek 3 3 3 3 3 3 Y Y 3 ..... CIJ co r ,,-- ~._.. -,-.~ ~'~~'=i~ =_ ..... :<~~ TABLE 11. Continued Multi- Ground- s~e~Density Densky Dimensional Cobble slone Hab~ation Other Fila, Record S~e S~e OuaJ~ative Measure 1 Measure 2 Scaling Dens~ Dens~ Features Features Des· .Jg-No. No. Name Assessment Group Group Group Group Group Present Present nation 24 JA100 Elk Creek 3 3 3 3 3 3 Y Y 3 25 JA101 Elk Creek 2,3 2 2 3 3 3 Y Y 3 26 JA102 Elk Creek 1 1 1 1 1 1 1 27 JA103 Elk Creek 1 1 2 2 1 1 2 28 JA105 Elk Creek 1 1 1 1 1 1 1 29 JA107 Elk Creek 2 1 1 2 2 1 2 30 JA110 Elk Creek 1 1 1 1 1 1 1 31 JA112 Elk Creek 1 1 1 1 1 1 1 32 Ee-2 Elk Creek 2 3 3 2 3 2 Y 2 33 Island Elk Creek 1 1 1 1 1 1 1 34 Winningham Elk Creek 1 1 1 1 1 1 1 35 Z'mmerly Elk Creek 1 1 1 1 1 1 1 36 JA5 lost Creek 1 1 1 37 JA6 lost Creek 1 1 2 1,2 1 Y 2 38 JA7 lost Creek 1 1 1 1 1 1 39 JA8 lost Creek 1,2 2 2 40 JA12 lost Creek 3 2 Y 3 41 JA14 lost Creek 1,2 2 2 2 3 3 2 42 JA16 lost Creek 2 2 2 2 2 Y 2 43 JA18 lost Creek 2,3 3 3 3 3 Y 3 44 JA19 lost Creek 1,2 2 1 2 1 Y 2 45 JA20 lost Creek 1 1 1 1 1 1 46 JA23 Fawn Butte 2,3 3 3 2 2 1 Y 2 -.A. W U) 140 c [S} UMpqua Bas in B AgreeMent AMong Anal~ses A ~ Rogue Basin 0.0 45.0 UJ c.::) a:: t-- 30.0ffi C,.) c:: LU a.. .15.0 60.0 A = Percent of sites which were placed into the same group by all functional tests. B = Percent of sites which were placed in two adjacent groups (Groups 1 and 2, or Groups 2 and 3) by all functional tests. C = Percent of sites which were placed in all three groups, or in Groups 1 and 3, by all functional tests. FIGURE 11. Rogue and Umpqua Basin, agreement among site function classifications. \ 'I 11 ,I 1 I' I I :1 I: :., .'1 -: ~ ': '. I r I bz ----c4Ib-~ . , 1 ,I II ~I !I I 141 the same category by~ measure employed. An additional 36 percent differ by only one step among the measures employed. That is, for all measures, 36 percent of the sites are represented in no more than two groups, and those groups are adjacent groups (Le., Groups 1 and 2, Groups 2 and 3). The remaining 14 percent of the sites appear in all three groups, or Groups 1 and 3, for the various measures, making interpretation of their function less straightforward. The final category in Table 11 is the functional designation decided for use in the rest of this study. In most cases, as just noted, assignment to this category is unambiguous. Where the various methods used have placed a site in different categories, the final designation represents my best interpretation of the information available for that site. In making these designations, I have frequently relied on the original excavator's assessment of the site, taking into account the other data presented above. Final site designations, along with explanation when needed, are presented below. Site Functional Assessments 1. CU 84 Marial 2. Gold Hill 3. J04. Ritsch seasonal camp village village i I ! b ' ____rtIIhz__~ The Ritsch site falls outside the parameters defining the other habitation sites; its placement in the MDS scatterplot suggests a seasonal camp or task site. The most likely explanation for this anomaly is the fact that it was not professionally excavated, and systematic collection of all materials may not have taken place. It is classed as a village site because of the presence of habitation features and a variety of artifacts characteristic of village sites. Like the Ritsch site, this site falls outside the parameters for artifact density and diversity defining other habitation sites, probably for similar reasons. It is classed as a village site on the basis of habitation features, location, and the variety of artifacts recovered. jl I ,; I 1! I I I 4. JQ15 Marthaller 5. JA21 Saltsgaver village seasonal camp 142 The site's features indicate that it was used primarily to process camas. It is a specialized seasonal camp. 5. JA25 Far Hills seasonal camp This possible village is removed from the tight cluster of village sites in the MDS data: possibly the reported burials and features are not associated with the portion of the site excavated. Although the high densities of cobble and groundstone artifacts argue for this as a village location, it is considered a seasonal camp on the basis of the MDS plot and the excavator's report. 7. JA42. Applegate 8. JA47. Applegate 9. JA47-1 Applegate. early component 10. JA47-2 Applegate. late component 11. JA77. Salt Creek 12. JA133 RRNF village seasonal camp village seasonal camp seasonal camp The very small amount of excavation may account for this site's placement in various groups for different functional measures. The excavator's opinion is relied on for the site's classi'fication in Group 2. 13. JA189 Trail 14. JA190 Trail village seasonal camp I t I ___stIIIIInz _ The site occurs with other Group 2 sites in the MDS analysis, in concurrence with the original investigator's assessment. 143 seasonal camp village seasonal camp seasonal camp 15. JA191 Reeder 16. JA197 Little Butte 17. JA1O. Elk Ck. 18. JA11. Elk Ck. 19. JA27A, Joham 20. JA27A-1 Joham, Late component 21. JA27A-2 Joham, Early component 22. JA27B Elk Ck. This site produced high densities of artifacts, including groundstone and cobble tools, but the MDS plot places it at some distance from the habitation sites. The very small amount of excavation may have skewed the assemblage. It is classed here as a seasonal camp, due to its upland location and the presence of midden and cobble tools. seasonal camp seasonal camp task I fd' ti"t i, i I I ,1 I ! i, , ,I -I 23. JA59 Elk Crk. 24. JA100 Elk Ck. 25. JA101 Elk Ck. 26. JA102 Elk Ck. 27. JA103 Elk Ck. village village village task seasonal camp The site was classed as a task site by excavators, and fell within the task group for one of the density measures and for the groundstone and cobble density analyses. However, one of the density measures placed the site in the seasonal camp group, and assemblage is close to those from seasonal camps and village sites in the MDS plot. Housepits were originally reported for this site (though not confirmed in the excavation). On the basis of the MDS plot and the second density measure, this site is placed with Group 2 seasonal camps. 28. JA105 Elk Ck. 29. JA107 Elk Ck. 30. JA11Q Elk Ck. 31. JA112 Elk Ck. task seasonal camp task task fi 1 - I' irtrrl ---",.",--~ 144 32. EC-2 ElkCk. seasonal camp The MDS plot places this site near other seasonal camps, rather than within the village cluster; it is placed in this group on that basis. 33. Island. Elk Ck. task r [! 34. Winningham. Elk Ck. task 35. Zimmerly. Elk Ck. task 36. JA5 Lost Ck. task i 37. JA6 Lost Ck. seasonal camp,.;; l' 38. JA7 Lost Ck. task ,:'J :~ ';; 1 tJ 39. JA8 Lost Ck. seasonal campj ~ l' 40. JA12 Lost Ck. village •. ! II r ~j\1 41. JA14 Lost Ck. seasonal campUf il:.1 f, 11 Cl 42. JA16 Lost Ck. seasonal campI' I ! 43. JA18 Lost Ck. villageI 44. JA19 Lost Ck. seasonal camp 45. JA20 Lost Ck. task 46. JA23. Fawn Butte seasonal camp i ~ I ----"..----- ~ Iif; i'I; 11 ! ~ i " 01 II, 'I i I i i rl 11 11 I i 145 CHAPTER VI FUNCTIONAL ANALYSIS: UMPQUA BASIN SITES There are 40 sites in the Umpqua Basin sample; 12 of these are in the South Umpqua drainage. and the remainder in the North Umpqua drainage basin. Several sites have multiple components. which are treated separately. resulting in 52 cases. As above. the qualitative analysis serves to introduce each site. and is followed by the quantitative measures. The initial sequential numbers are those assigned to each site when entered into the database. These case numbers identify sites throughout the report. Where site data are separated into components. those components are given separate record numbers and analyzed separately. Explanations of the different analyses are more fully discussed in the preceding chapter on Rogue Basin sites. Site Function Based on Qualitative Analyses 48. D0275. Sylmon Valley School seasonal camplvillage? I f I I· brt Beckham and Minor (1992) consider this a potential village site. due to . its location on the South Umpqua in the Umpqua Valley. Lyman. the investigator. considers this a seasonally occupied site (Le.• a seasonal camp). One possible earth oven was excavated. consisting of a concentration of fire-eracked rock which may indicate a camas oven. A comparatively large amount of cobble tools. including netsinkers. were recovered. My assessment is that it resembles more a seasonal camp site. occupied for a specific purpose. such as a fishing station. The investigators consider this possibly a village due to its location in the Umpqua Valley along the South Umpqua River. However, the minimal amount of excavation produced a light assemblage of chipped stone and cobble tools; not a "village"-Iike assemblage. 49. 00274. Orchard village/task? 146 50. 0036. Crispen village/seasonal camp This is a well-known use area of the Cow Creek Band of Umpqua Indians for processing riverine resources. It is assumed to be a village location. The site has been heavily distUrbed. The site is located at the confluence of Coffee Creek and the South Umpqua River, in a lowland setting, suggesting to the investigators that it was probably a village type of site. However, it has been very disturbed, and only a remnant remains. The test excavations produced a light assemblage of artifacts which the investigators considered more indicative of a seasonal camp. J ! i '.f I , . . , I I • 51. 00412 Coffee Creek 52. 00413, Coffee Creek seasonal camp? seasonal camp? The site is located across from 00412, and the same situation applies. 53. Tiller #1 task Located on a ridgetop, the site produced chipped stone tools and debitage. It appears to be a travel stop/task area. 54. Tiller #6 task The site is located on a ridgetop, and consists of a light lithic scatter. It appears to be a taskltravel stop for refurbishing tools, or performing some immediate task. No cobble tools were recovered. 55. 00205 S. Umpqua Falls RS, upper seasonal camp I, hi This site is the upper of two rockshelters located above South Umpqua Falls. The falls were and still are an important food-gathering place for the Cow Creek Indians. The presence of burials and abundant chipped stone artifacts suggest at least a seasonal camp. The limited number of cobble tools together with the rock-shelter location suggest this did not serve as a winter village habitation. The investigators consider this a seasonal camp. The lower South Umpqua Falls rockshelter produced an abundance of chipped stone and bone. A Cow Creek informant camped there as a boy; it is known to the Cow Creek people for its proximity to the falls and the good summer and winter runs of fish. Fish were smoked at the falls then taken home. The artifacts recovered. informant testimony, and the site's location all suggest to the investigators that it was used as a seasonal base camp. The comparatively light, predominantly chipped stone assemblage at this rockshelter indicates this was a task site. The investigators consider this rockshelter a task site, since the non- perishable items consist mainly of chipped stone--especially projectile point--artifacts. The assemblage includes perishable materials and hearths. It appears to be a hunting site with lots of points and bone; six hearths were also preserved. :11 'I' ,~j I•., , 56. 00205 S. Umpqua Falls RS. lower 57. 00209. Hughes IRS 58. 00212 Time Sg. RS 59. 00396. Sprint seasonal camp task task seasonal camp 147 village (late); seasonal camp (early?) The site is located at a good fishing spot for the Cow Creek Indians, near an important food gathering area. It has been heavily disturbed, and only a small portion of the site was tested. The comparatively dense assemblage of mainly chipped stone tools and the site's location suggest a seasonal camp. 60. 61. 62. 63. 00219. Section Creek 60 = Combined data for the whole site 61 = Component I 62 = Component 11/ 63 = Component" The site is reportedly a Cow Creek village, with pit houses and sweat lodges in the vicinity. Three components were identified by the investigators. Cases 61 and 62 are the two later components; these are similar to one another and include an abundant and varied assemblage of chipped and flaked stone tools. Case 63 is the earliest component and has a lighter assemblage of artifacts, perhaps reflecting use of the area as a seasonal camp. The investigator suggests this site was a seasonal camp. A limited amount of excavation produced artifacts including abundant cobble tools. Large, well-formed cobble tools, such as mauls, pestles, and a "hammer" were found at the site by the property owner. Two features, a mussel shell lens and a cobble pavement, indicate at least a seasonal camp. The amount of cobble tools and the site's location along the North Umpqua at a low elevation suggest this may have been a village type of site. A very small amount of excavation yielded a variety of artifacts and the indication of a deep, stratified deposit. The location along the North Umpqua River in the Umpqua valley suggests a village or seasonal camp site. Housepits and burials were reported for this location (late); it is located near a prime fishing spot on the North Umpqua River. The small amount of excavation produced a variety of artifacts consistent with the interpretation of the site as a village. 64. 00395. Grubbe Ranch 65. 00528. Glide 66. 0061. Whistler's Bend 67. 0067. Winchester Bridge 148 village/seasonal camp village/seasonal camp village/seasonal camp village? The site is along the North Umpqua River in the Umpqua valley, and is reportedly a winter village site. Burials have also been reported in the vicinity, and artifacts recovered. Only a-small, undisturbed portion of the site was tested. Minimal excavations produced only a few chipped stone and cobble tools. The site may be the remnants of a village site, or a task site associated with a village nearby. 68. 00 52. Gatchel Site village/seasonal camp? Ethnographic testimony and the presence of "housefloors" at the site indicate it was a habitation site, possibly a village. A village is noted nearby at the connuence of Little River and the North Umpqua, and the site is on historically known native trails. The site produced a variety of chipped stone and cobble tools. ----....,---~ 69. 70. 71. 72, 73. 74. 00153. Narrows 69 = Site data for the whole site. 70 = Component I 71 = Component" village/seasonal camp 149 72 =Component III 73 = Component IV There are four components, two early (III & IV) and two late (I & II). The later components include a burial and housepit (shallow), as well as a midden. The earlier components have hearths. The site was known as a fishing spot for the Cow Creeks, and a "kind of village" was noted for the opposite side of the river. The site is interpreted as a village type for the later two components, and seasonal camp in the earlier. 74. 00359, Swiftwater task .I," I'II r, , ~, i ' This is a unique site, consisting of little chipped stone but an abundance of cobble tools. It is located on river terrace along the North Umpqua. The investigator suggests it may have been a fish processing site. i I 75. 00383. Susan Creek task/seasonal camp , , Test excavations at this site along the North Umpqua River produced a light assemblage of chipped and some groundstone tools, as well as a rock feature associated with groundstone. The comparatively light assemblage suggests a task site; the variety of artifacts and feature indicate possibly a seasonal camp site. 76. 00278. Bogus Creek task The site is located along the North Umpqua River and produced an assemblage of primarily chipped stone artifacts, suggesting the site was used as a hunting site. The low elevation may indicate a winter task site or seasonal camp. 77. 00126 Steamboat task/seasonal camp? I " Chipped stone and a few cobble tools at this site along Steamboat Creek suggest a task site or possibly a seasonal campsite. 78. 0011, Lower Rhod't task --_.-.""-"'----- The site consists of numerous rock cairns, presumably associated with vision quest activities. A small amount of excavation did not produce any artifacts. The site is a task-specific site. Ii , , , 150 The site is a small, light lithic scatter, near a known Cow Creek occupation site/fishery, Small bands camped in the area after contact. The minimal assemblage suggests this was a task site. task/seasonal camp task79. 0040, Cavitt Creek 80, 81. 82. 00401, Dry Creek 80 = Combined data for the site. 81 = Early component 82 = Late component The site is located on a terrace along the North Umpqua River, and has both a pre- and post-Mazama component. Test excavations suggest that the later component is sparser than the pre-Mazama one; it appears to reflect use as a hunting camp. The earlier component may represent a winter seasonal camp for the Early Archaic, based on recovery of cobble stones/groundstone and the low elevation location. As yet unpUblished data recovery excavations, however, indicate that the upper component may also have served as a seasonal camp (O'Neill 1992). I' The primarily chipped stone assemblage suggests this site was an early huntingltask specific site at the confluence of several streams tributary to the North Umpqua. ~ I I 83. 00372. Reynolds task task/seasonal camp84. 85. 86. 00422. Island Campground 84 =All data for the site 85 = Early component 86 = Late component Test excavations of this site along the North Umpqua River yielded chipped and cobble tools. The higher density and greater variety of materials from the earlier component suggest it was a seasonal camp, whereas the later occupation was a task site. 87. 00418. Apple Creek Bench task This site produced an assemblage of chipped stone tools, indicating a task site. It is located along the North Umpqua River, not far from other seasonal camps sites. --....-- This early site along the North Umpqua produced an assemblage of chipped stone and cobble tools, indicating use as a seasonal camp or task site. Two components are present at this site: pre- and post-Mazama. The pre-Mazama component has chipped stone and cobble/groundstone tools, possibly reflecting use as a seasonal camp. The post-Mazama component is less dense, with fewer cobble tools. Both components are similar to the components at Dry Creek. 151 task/seasonal camp task/seasonal camp? seasonal camp task/seasonal camp? 88. 00265. Apple Creek 90. 00161. Medicine Creek 89. 00421, Copeland Creek 91. 00187. Powerful 1 A small amount of excavation produced a high amount of debitage and a few chipped stone tools. The site is located at the juncture of trails at the confluence of the North Umpqua and Copeland Creek, and at a good fishing spot. The limited tool inventory suggests a task site; the location and high density of debitage indicate a seasonal camp. Peeled ponderosa pine trees, bedrock mortars/stone bowls, and chipped stone tools in a large oak covered flat above the North Umpqua indicate a seasonal base camp. 92. 00227. Powerful 2 seasonal campltask? This site produced a high density of chipped stone artifacts along with some groundstone. Peeled trees and vision quest cairns occur adjacent to the site, which is not far 'from Powerful 1. The location, on a small knoll above a pine/oak covered flat, suggests a task/hunting camp setting; the high density and variety of tools indicate a seasonal camp. 93. 00379. Snuff Out Site task Located along a ridge, the site produced an assemblage of chipped stone tools (no points), suggesting use as a task site or a travel stop. 94. 00397, Shivigny East task This site along a ridge produced a dense assemblage of chipped stone tools with a few cobble tools. It appears to be a hunting site. 1 ) ...L ---------------------------- ----_...._-._- This light lithic scatter in the uplands is considered a task site. "·"'1 95. 00289, Little Oak Flat 96, 00399, Snowbird task task 152 , i This ridge site produced mainly a chipped stone assemblage and is interpreted as a task site. 97, 00160. Mudd~ task This upland ridge site produced only chipped stone artifacts and is considered a taskltravel site. This ridge crest site produced a light assemblage of chipped stone tools, indicating a taskltravel site. This upland rock shelter produced a high density of predominantly chipped stone tools and debitage. A large number of projectile points suggest it was a hunting task site. The high density of materials reflects either a special depositional environment or use of the site as a seasonal camp, possibly for the purpose of hunting. task/seasonal camp task 99, 00389. Limp~ RS 98. 00398. Powerline Site Oensit~ Measures Oensity Measure 1: Projectile Points vs. Other Chipped Stone Tools The first measure of density employed (Figure 12) is the comparison of projectile point density with the density of other chipped stone tools. This measure follows the same procedures as those explained for the Rogue Basin sites in Chapter VI. The plotted sites were divided into three groups based on apparent breaks in the scatterplot: Group 1 are low density sites, interpreted as task sites, Group 2 sites have intermediate densities and are .i II .J.. _ -~--....--- ~~=-- __."--,,_.,~_-_o~__00- _,~ ~'~~ ::..:~'t~~~~;,::~.-~-"""""'. --- '7'i"""".-,~..............., 5-"" @ 12 IGROUP 3 « > « u ~ ~ u 9~ u : Ii ffi 3 0 ~ • I~ •~ I I f Iff f.f IJ fu I f~ fI I f 0 11.3 22.S 33.8 45.0 CHIPPED STONE TOOL DENSITY: NUMBER OF CHIPPED STONFl CUBIC METER EXCAVATED Projectile point density. measured as number of projectile points/cubic meter. plotted against chipped stone tool density. measured as number of chipped stone tools/cubic meter. Each cross represents a site. FIGURE 12. Umpqua Basin Sites: Density Measure 1. .....0'1 to) , J ! " )1 .L-~----------- 154 considered seasonal camps. and Group 3 sites have the highest densities and are considered villages (Table 12). Density Measure 2: Debitage Density vs. Total Tool Density In Figure 13 the density of debitage is plotted against the density of the total tool assemblage. Again. the procedures follow those outlined for the Rogue Basin sites in Chapter VI. The scatterplot was divided into three groups of increasing density. as above. Several sites (cases #55. 60. 89. and 99) have debitage densities too high to show on the scale used for the graph and are not represented in the scatterplot. There is considerable overlap between functional groups for both measures. Nearly two thirds (63 percent) of the sites analysed for both measures appear in the same group for each measure. Also. as in the density analysis for the Rogue Basin sites. sites which are considered task sites by the investigators occur in Group 1 of both density measures. seasonal camps occur in Group 2. and village sites in Group 3. Sites with features are more likely to appear in Group 3 of both measures (Table 13). ----".,,---- TABLE 12. Umpqua Basin Sites: Density Measure 1 Density Groups Based on Projectile Point/Other Chipped Stone Tools 155 Rec. No Site Name/No. Qual. Type Hab. Feat. Oth. Feat. Group 1 74 00359 Swiftwater task no no , 98 00398 Powerline task no no I ' 93 00379 Snuff out task no no;r) 67 0067 Winchester village? no no 79 0040 Cavitt Creek task no no 51 00412 Coffee Creek task no no 75 00383 Susan Creek task/seas.camp no no " ; 52 00413 Coffee Creek task? no no .. 86 00422 Island-Late task? no noi 1 90 00161 Medicine task/seas.camp no no 49 00274 Orchard season campi no yes Village or task? 48 00275 Sylmon seas.campM!.? no yes 97 00160 Muddy task no no 76 00278 Bogus task/seas.camp? no no 84 00422 Island Cmpgrd task/seas.camp? no no 85 00422 Island-Early seas. camp no no 83 00372 Reynolds task no no 96 00399 Snowbird task no no 87 00418 Apple Crk Bnch task no no 91 00187 Powerful 1 seas.camp no no 64 00395 Grubbe Ranch vill/seas.camp no yes 95 00289 Little Oak task no no 88 00265 Apple Crk task no no 53 Tiller 1 task no no I' Group 2 1 77 00126 Steamboat task/seas.camp? no no 1 57 00209 Hughes RS task no no 69 00153 Narrows seas.camplvil? yes yes 65 0053 Glide seas.camplvil? yes yes 80 00401 Ory Crk task/seas.camp no no 68 0052 Gatchel vil./seas.camp yes yes 1 .', ~! l ""., 1~ i' ~j,; I , ~ : 156 TABLE 12. Continued Site Name/No. Qual. Type Hab. Feat Oth. Feat. ---~"""'--'-- 0061 Whistlers village? no no 00212 Time Sq. RS task? no yes 00396 Sprint task/seas.camp no no 0036 Crispen vil./seas.camp no yes 00227 Powerful 2 task/seas.camp no yes Tiller 6 task no no 00421 Copeland task/seas.camp no no 00219 Section Crk village/task no yes 00205 S. Ump RS-Upper seas. camp yes no 00205 S. Ump RS-Lower seas. camp no yes 00389 Limpy seas. camp no no Multidimensional Scaling Analysis: Comparison of Assemblage Evenness and Richness Figure 14 presents the results of the multidimensional scaling procedures. In interpreting this scatterplot, I proceeded in the same manner as for the Rogue River basin sites. Those sites which could be confidently assigned a functional type were identified, to see if there were any readily discernable groupings. Then those sites which were more ambiguous were assigned a type based on their proximity to other sites in the scatterplot. Generally, the same pattern observed for the Rogue River drainage was also apparent here. The most likely village sites, (with exceptions noted below) clumped neatly in the center of the scatterplot, with seasonal camps closely associated on the periphery. Task sites were scattered beyond this central grouping. ~c,'" - ,c~--~-"'J_-h~_",'V '~"'=-'E' '. ' ..• :::4+~~._''-==-" -~-;-- ..., '~~~-.-. ..;..~ if . - ·'3;:..-'.... 10dr'------.........---~------------'--"---.-------~ I ;IGROUP 3 75 ~ < > ---- ~=I0l==~- ======I0l========---------------- ---===O==~-- -=0=- ====I0l=======~------------------­ ===IOF=====~----- ---==========101===========-------- --====101=====---- =OF==------ =Ol==~--­ =OF==---- SEE KEY TABLE 14 KEY mean mean t 10 range o u=====---- =D=--- =D!=~-- =0= =O====~-------- =OF=~-­ =0=-- 1 1:12 EMP 2 1:12 3:s u 1b CRP 2 < 3~ 1 < BCBP 2 3 1 DCBP 2 3 1 GDSP 2 3 1 - task sites; 2 = seasonal camps; 3 = village sites FIGURE 15. Umpqua Basin: MDS graph of artifact class percentages. ~ en w j' I i - ~ i..1 fJ r J 1, Ii I II :L i , . ,j I j; i . 164 cobble tools") for the task sites, as would be expected for sites representing a variety of single purposes. Like the Rogue Basin sites, the greatest variability is represented in the task-site assemblages and the least in the village sites. Groundstone and Cobble Tool Densities; Comparison with Feature Data Tables 15 and 16 list the sites in order of increasing density of cobble and groundstone tools. As in the analysis for the Rogue Basin sites, three groups are distinguished on the basis of increasing density, and compared to the presence/absence of habitation and other features. Table 17 shows the strong association of cobble and groundstone artifacts with sites having features. Like the Rogue Basin sites, habitation features (Le., housepits, middens, and burials) and multiple types of features occur only with the most dense assemblages of these heavy artifacts. Site Function .__..-_c4tt.__~ Table 18 lists the sites examined in this study, and each site's functional type based on the different analyses. Overall there is good agreement among the various methods employed, although agreement is not as consistent as for the Rogue Basin sites (see Figure 11). Slightly over one-third (37%) of the sites (or site components) fell into the same group using every measure employed; of these, many are site components which were subjected only to the MDS analyses, not the density measures. Another 21 percent of the sites differed by only one step among the -:--~' :':,-C~;' ;;C':-'"_,,. • ~"~"~-'~-=--"~" _, .J-.,,";;t_.,£Q2L~: , ." ...:~ ~.", "-;:;!-~ TABLE 15. Umpqua Basin Sites: Groundstone Density and Features Amount Ground- Record Site Site Exca- stone Other No. No. Name vated· Density·· Housepit Midden Burial Hearth Features Group 1 58 00212 Time Sq Rs 1.2 Y 49 00274 Orchard 3.6 Y 51 00412 Coffee Creek 8.8 53 Tiller 1 Til 1 3.3 54 THI6 Till 6 2.2 55 00205 S.Umprs-U 4.9 Y 67 0067 Winchester 2.0 79 0040 cavitt Creek 2.0 80 00401 Dry Creek 2.8 74 00359 Swiftwater 10.5 86 00422 Island-L 2.2 76 00278 Bogus 22.5 77 00126 Steamboat 2.4 78 0011 Rhody 0.3 85 00422 Island-E 4.4 97 00160 Muddy 4.3 83 00372 Reynolds 6.8 89 00421 Copeeland 1.1 84 00422 Island Camp 6.6 96 00399 Snowbird 3.2 87 00418 Apple Bunch 2.6 93 00379 Snuff Out 4.8 94 00397 Shivigny 3.2 ..... m 01 r =- - --~'~,:==~'~~~~~-:'- -:-:;::::3Pi$~ TABLE 15. Continued Amount Ground- Record Site Site Exca· stone Other No. No. Name vated· Density·· Housepit Midden Burial Hearth Features 95 00289 Little Oak 2.50 98 00398 Powerline 1.60 Group 2 90 00161 Medicine 10.50 .0952 52 00413 Coffee Ok 7.10 .1408 75 00383 Susan Creek 8.60 .3488 Y Group 3 65 0058 Glide 6.40 .4688 Y 69 00153 Narrows 22.60 .4867 Y y y y y 60 00219 Section Creek 22.30 .4933 Y Y 92 00227 PwrfI-2 3.70 .5405 y 48 00275 Sylmon 23.00 .5652 y 56 00205 S.Umprs-L 3.50 .5714 57 00209 Hughes 3.40 .5882 88 00265 Apple Creek 2.30 .8696 50 0036 Crispen 5.40 .9259 Y Y 68 0052 Gatchel 6.60 1.2121 Y y y 91 00187 PwrfI·1 5.80 1.3793 Y 59 00396 Sprint 1.30 1.5385 99 00389 Limpy 2.25 4.8889 64 00395 Grubbe 3.30 6.9697 Y *Cubic meters **Arranged in order of increasing density .... en en ------- '-= ~ -_. "L:::~~-.:·~' .. --- . '- :!.~ ".:;==-:,~ .. -""~""-:;"',:","==-",:!",=" ' ',. ..... "'~" . - ··~1~~~ - I TABLE 16. Umpqua Basin Cobble Tool Density and Features Amount Cobble Record Site Site Exca· Tool Other No. No. Name vated· Density·· Housepit Midden Burial Hearth Features Group 1 89 00421 CopelandD 1.1 95 00289 UttIe Oak 2.5 97 00160 Muddy 4.3 87 00418 Apple Bnch 2.6 53 Tiller 1 Til 1 3.3 54 Till 6 Till 6 2.2 98 00398 Powerline 1.6 57 00209 Hughes 3.4 79 0040 Cavitt Creek 2 78 0011 Rhody 0.3 59 00396 Sprint 1.3 93 00379 Snuff Out 4.8 99 00389 Umpy 2.25 90 00161 Medicine 10.5 .0952 51 00412 Coffee Creek 8.8 .1136 75 00383 Susan Creek 8.6 .1163 Y 83 00372 Reynolds 6.8 .1471 76 00278 Bogus 22.5 .2667 Group 2 91 00187 PwrfI-1 5.8 .3448 y 52 00413 Coffee Ok 7.1 .4225 96 00399 Snowbound 3.2 .6250 94 00397 Shivigny 3.2 .6250 86 00422 Island-L 2.2 .9091 ..... m ....... r - ."'..--- -;::: .. ~:~..,~"~..'~ ";~:;"':~=-m'~ TABLE 16. Continued Amount Cobble Record Site Site Exca- Tool Other No. No. Name vated· Density·· Housepit Midden Burial Hearth Features 67 0067 Winchester 2.0 1.0000 92 00227 Pwrfl-2 3.7 1.0811 Y Group 3 84 00422 Island Camp 6.6 1.2121 68 0052 Gatchel 6.6 1.2121 Y Y Y 55 00205 S.Umprs-U 4.9 1.2245 Y 74 00359 Swiftwater 10.5 1.3333 85 00422 Island-E 4.4 1.3636 80 00401 Dry Creek 2.8 1.4286 56 00205 S.Umprs-l 3.5 1.4286 1 60 00219 Section Creek 22.3 1.6592 Y y 49 00274 Orchard 3.6 1.6667 Y 77 00126 Steamboat 2.4 1.6667 88 00265 Apple Creek 2.3 1.7391 65 0058 Glide 6.4 2.8125 Y 58 00212 Time Sq Rs 1.2 3.3333 Y 50 0036 Crispen 5.4 3.7037 Y y 64 00395 Grubbe 3.3 3.9394 Y 48 00275 Sylmon 23.0 5.4348 y *Cubic meters **Arranged in order of increasing density -'" en CD TABLE 17. Umpqua Basin Sites: Groundstone/ Cobble Densities and Features Site Type N sites Density Range %SF %MF 169 % HF % All Cobble pensny Group 1 18 Group 2 7 28% Group 3 17 65% Groundstone Density , 0-.27 6% 0 0 6% l .34 - 1.1 28% 0 0 1.2-20 41% 24% 18% Group 1 Group 2 Group 3 24 3 15 0-0 .09-.35 .46-20 12% 33% 33% o o 27% 4% 12% o 33% 13% 60% % SF = Percent of sites in the group with only one type of feature present (Le., either housepits, middens, bUl;als, hearths of other rock features). % MF = Percent of sites in the group with multiple types of features present (Le., some combination of housepit, midden, burial, hearth, and rock features). % HF = Percent of sites in the group with habitation features (Le., housepit/living floor, midden, burials) present. % All = The total percentage of sites in the group with features (%SF plus %MF). , , I i TABLE 18. Umpqua Basin Sites: Functional Site Types Concordance of Functional Measures and Final Site Type Designations Mulli- Ground- Sit~Dens.y Density Dimensional Cobble stone Habitation Other F Record Site Site QuaJitalive Measure 1 Measure 2 Scaling Density Density Features Features Des' , .IQ- No. No. Name Assessment Group Group Group Group Group Present Present nallon 48 00275 Sylmon 2,3 1 2 3 3 Y 2 49 00274 Orchard 3 1 1 2 3 1 Y 2 50 D036 Crispen 3 3 3 3 3 3 Y 3 51 00412 CoffeeCK ? 1 1 3 1 1 3 52 00413 Coffee OK ? 1 1 3 2 2 3 53 Tiller 1 Til 1 1 1 1 1 1 1 1 54 Till 6 TiO 6 1 3 2 1 1 1 1 55 D0205 S.Umprs-U 2 3 3 2,3 3 1 Y 2 56 D0205 S.Umprs-l 2 3 3 1 3 3 Y 2 57 D0209 Hughes 1 2 2 1 1 3 1 58 00212 TimeSq As 1,2 3 3 1 3 1 Y 2 59 00396 Sprint 2 3 3 1 1 3 2 60 00219 Sectopm Ck 3 3 3 3 3 3 Y 61 00219 Section-I 3 3 3 62 00219 Section-III 2 2 2 63 00219 Section-II 3 3 3 64 00395 Grubbe 2,3 1 2 1 3 3 Y 2 65 D058 Glide 3 2 3 3 3 3 Y 3 66 0061 WhistleRS 3? 3 1 1 3 3 3 67 0067 Winchester 2,3 1 1 1 2 1 1 68 D052 Gatchel 3 2 3 3 3 3 Y Y 3 69 00153 Narrows 2,3 2 3 3 3 Y Y 70 00153 Narrows-I 3 3 3 71 00153 Narrows-II 3 3 3 .... ...... 0 - -~,."-.-- " ...~ ..,"...... ,,""'-'''''.''',.'''-' _._,..._--- .. _,,,.. ,. TABLE 18. Continued • ......_..""'''''#'¥ .. 7'WjIiiF"""-' '~'-"l'"'O""'''''''~''''' Mufti- Ground- Sit~Density Densly Dimensional Cobble stone Habitation Other A • Record S~e Site Qualitative Measure 1 Measure 2 Scaling Density Dendy Features Features Desig- No. No. Name Assessment Group Group Group Group Group Present Present nation 72 00153 Narrow-III 2 2 2 73 00153 Narrow-IV 2 2 2 74 00359 Swiftwater 1 1 1 3 1 1 75 00383 Susan Creek 1 1 1 2 1 2 Y 2 76 00278 Bogus 1 1 1 2 1 1 2 77 00126 Steamboat 1,2 2 1 3 1 1 78 0011 Rhody 1 1 Y 1 79 0040 Cavitt Creek 1 1 1 1 1 1 80 00401 Dry Creek 2,3 2 3 3 81 00401 Dry Creek-E 2 2 2 82 00401 Dry Creek-l 1 2 2 83 00372 Reynolds 1 1 2 1 1 1 1 84 00422 Island Camp 1,2 2 2 3 85 00422 Island-E 2 2 2 86 00422 Island-l 1 1 1 87 00418 AppleBnch 1,2 1 3 2 1 1 2 88 D0265 Apple Creek 1,2 1 2 2 3 3 2 89 00421 Copeland 1,2 3 3 2 1 1 2 90 00161 Medicile 1,2 1 2 2 1 2 2 91 00187 PwrfI-1 2 1 2 2,3 2 3 Y 2 92 00227 Pwrfl-2 2 3 3 2,3 2 3 Y 2 93 00379 Snuff Out 1 1 1 1 1 1 1 94 00397 Shivigny 1 2 3 2 2 1 2 95 00289 Litle Oak 1 1 1 1 1 1 1 96 D0399 Snowbird 1 1 1 2 1 1 97 00160 Muddy 1 1 2 1 1 1 98 00398 Powerine 1 1 1 1 1 1 1 99 00389 Linpy 1,2 3 3 1 1 3 2 ...... -....I ...... 172 The excavator considers this a seasonal camp (Lyman 1985); Beckham and Minor (1992) consider it a possible village due to its location. The comparatively low density of chipped stone tools is matched by a high density of cobble tools. It does not cluster with other habitation sites in the sample. I suspect that Lyman's assessment is correct; it is a seasonal fishing camp. seasonal camp Site Functional Assessments 48. D0275 Sylmon measures employed. That is, with every measure employed 21 percent of the sites are represented in no more than two groups, and those groups are adjacent groups (Groups 1 and 2, Groups 2 and 3). The remaining 42 percent of the sites fell within all three groups, or in Groups 1 and 3, for various tests. This variability makes the interpretation of site function somewhat more difficult than for the Rogue Basin sites. Specific site designations, based on the qualitative and quantitative data, are discussed below. This final designation represents my best interpretation of the information available for each site, including qualitative data not represented in the quantitative analyses. 49. D0274 Orchard seasonal camp Although the location is good for a village, the assemblage does not support that designation; it has a low density of chipped stone and groundstone tools, and no habitation features. It is more likely a seasonal camp; the high density of cobble tools may indicate fishing as a focus. 50. D036 Crispen 51. D0412 Coffee Creek village village Despite the minimal assemblage, the site's location and close association with other habitation· sites in the MDS analysis argues for its function as a village. ---_.-."",.,.,,--- The same reasoning applies to this site as to 00412. This is a high density site with a narrow range of artifacts; it is considered a task site. 173 village task task 52. 00413 Coffee Creek 53. Tiller 1 54. Tiller 6~i t'Ii, [- , i ' ! j seasonal camp seasonal camp55. p0205 South Umpqua Rockshelter Upper Shelter This site has a high density of chipped and cobble tools, as well as habitation features (bUrials) and a strong association with village sites in the MOS analysis. It is classed as a seasonal camp because its location seems to preclude a village habitation. However, it is worth noting that, compared to other sites in the sample, it is very similar to village sites. 56. 00205 South Umpqua Rockshelter Lower Shelter This rockshelter has a high density of materials, including both chipped and cobble tools. It also has a high proportion of projectile points, which places it with the more specialized task sites in the MOS analysis. The presence of a feature and the cobble/groundstone densities suggest that it was a seasonal encampment, with a focus on hunting. 57. 00209 Hughes Rockshelter task Artifact densities may be somewhat inflated for this site, due to preservation conditions in rockshelters compared to open-air sites. The MOS analysis places it in Group 1, which seems the best classification based on site location and assemblage. 58. 00212 Time SQuare Rockshelter task/seasonal camp This is a high density site with a specialized assemblage dominated by projectile points. The presence of bone, hearths, and groundstone suggests this was a seasonal camp focused on hunting. j . ! ----".,--- This is at/near a known fishing spot; and may be a seasonal camp for fishing. It is a fairly high density site with an assemblage dominated by projectile points; however groundstone also occurs. It is considered a seasonal camp, focused on hunting/fishing. 60. 00210 Section Ck This site has a comparatively low density of chipped stone and high density of cobble/groundstone artifacts. The MOS analysis places it at some distance from other habitation sites. Since it has features and cobble tools, it is classed as a seasonal camp. The very small amount of excavation may be responsible for the unlikely "Group 1" placement in Table 11 and in the MOS analysis. The site's location and the investigator's report, as well as high densities of cobble and groundstone artifacts, provides a justification for placing it in Group 3. 59. 00396 Sprint 61 Component 1 62 Component III 63 Component II 64. 00395 Grubbe 65. 0058 Glide 66. 0061 Whistlers 67. 0067 Winchester seasonal camp village seasonal camp village seasonal camp village Village? task 174 The site's location suggests a village but the other indicators place it in the task site group. It may have been associated with a village nearby. 68. 0052 Gatchel 69. 00153 Narrows 70 Component I 71 Component II 72 Component III 73 Component IV 74. 00359 Swiftwater village village village seasonal camp seasonal camp task ----",..,,--- The high density of cobble tools "at this site is thought to reflect a task focus on fishing. 75. 00383 Susan Ck seasonal camp 175 i 1 I This is a low-density site with a generalized assemblage, grouped near other seasonal camp sites. This site is designated a seasonal camp primarily on the basis of its strong association with other sites having more generalized, less specialized assemblages. I I I I I seasonal camp task task task seasonal camp task seasonal camp seasonal camp task 76. 00278 Bogus 77. 00126 Steamboat 78. 0011 Lower Rhody 79. 0040 Cavitt Ck 80. 00410 Dry Ck 81. Early component 82. Late component 83. 00375 Reynolds 84. 00422 Island 85. Early component 86. Late component 87. 00418 Apple Bench seasonal camp This site has a high density of debitage and a generalized tool assemblage, and is associated with other seasonal camp sites along the North Umpqua in the MOS analysis. .t, t', i 88. 00265 Apple Creek seasonal c. camp This is a low density site with a generalized tool assemblage, associated with other such sites in the MOS analysis, also located along the North Umpqua. 89. 00421 Copeland seasonal camp The site is designated a seasonal camp based on the high density of chipped stone artifacts and its association with other Group 2 sites in the MOS analysis. __.-crIIte... 90. 00161 Medicine seasonal camp 176 The same considerations apply to this site as to the previous one. This site is considered a seasonal camp due to its high density and close association to Group 2 and 3 sites in the MOS analysis. This is a mixed assemblage of pre- and post-Mazama artifacts; the earlier component, at least, seems to resemble that of a Group 2 site and may dominate the materials from the site. seasonal camp seasonal camp seasonal camp task 92. 00227 Powerful 93. 00397 Snuff Out 94. 00387 Shivigny 91. 00187 Powerful 95. 00289 Little Oak Flat task 96. 00399 Snowbird task 97. 00160 Muddy task 98. 00398 Powerline task 99. 00389 Limpy Rockshelter seasonal camp This site is closely associated with village sites in the MOS analysis, but is considered a seasonal camp due to density and location considerations. It is worth noting, however, that the assemblage proportions are similar to lowland village type sites. , j This high density site has an assemblage dominated by projectile points, as well as some groundstone artifacts. The specialized focus of the assemblage plus the other artifacts indicate that this was a seasonal camp focused on hunting. I. ____cA.__ -_.-_"""-~ 177 CHAPTER VII SUBSISTENCE AND SEn-LEMENT PATIERNS The purpose of this chapter is to define the subsistence and settlement patterns represented by the sites in this study. These sites have each been assigned to a specific functional category, based on the analyses in the previous chapters. Now it is necessary to group them chronologically, in order to define the patterns extant dUring different periods and to address the question of culture change. In this chapter, sites are placed into two chronological periods: the Middle Archaic, from about 6,000 to 2,000 BP, and the Late Archaic, from about 2,000 BP to the time of historic settlement (Beckham and Minor 1992).1 Several sites also have components from the Early Archaic, about 8,500 BP to 6,000 BP, although this time period is not considered in this analysis due to the sparsity of data. The site types present in each period are used to infer the subsistence/settlement regime which characterized those times. 1Tighter timeframes were precluded by the desire to include as many sites as possible. Since the chronological data available for this body of sites is highly variable, it was necessary to use the broadest categories in order to include them all. I II 11 ! {' I-~ :~I:df r.;i "~ , '!ij'I.'·.II., ~i ..~ f i ..~ 1; i : i .,' t \ --_...._-- 178 Site Chronology The two time periods used, the Middle and Late Archaic, are broad chronological categories. In both the Umpqua and Rogue Basins recent studies have produced finer-grained chronologies. Many of the sites in this study were used to formulate these chronologies. and hence have already been assigned to various local sequences. Table 19 presents the local sequences used for the Umpqua and Rogue Basins, and the correspondence of these sequences to the two broad periods used in this study. In making chronological assignments, I used the previous work done for sites in the Umpqua and Rogue Basins. For the Umpqua Basin, the recent cultural resource overview for the Umpqua National Forest summarizes the temporal information available for many of the sites included here (Beckham and Minor 1992:64-70). The overview draws upon statistical analyses done by O'Neill (O'Neill 1989b; 1991 b), which cluster sites into temporal groups based on projectile points. An example of the point types used in that statistical analysis is presented in Figure 16. I have followed Beckham and Minor's (1992) cultural resource overview to date many of the Umpqua Basin sites, based on the cluster analysis. In the Rogue Basin. work at the Elk Creek sites has produced considerable information regarding chronology pertinent to the area. Pettigrew and Lebow (1987) first defined a chronological sequence based on point types, which was then refined by Nilsson and Kelly (1991). The Elk Creek sites in this study are placed into the Middle and Late periods based on a correlation between this Rogue Basin chronology and the broader TABLE 19. Chronological Periods in Southwest Oregon 179 ARCHAIC 2000-------------------------------------------------------------------------- Rogue (Nilsson and Kelly 1991) ROGUE 2 SUBPHASE COQUILLE ROGUE 1 SUBPHASE 2250------------------ 1650------------------LATE PROTOHISTORIC Umpqua (Beckham and Minor 1992) MIDDLE ARCHAIC 500---------------------------------- FORMATIVE 1000--------------------------------- Archaic (Beckham. Minor and Toepe11981) LATE ARCHAIC MIDDLE ARCHAIC Years BP o 4000i,., ' I· 4500------------------ 6000---------------------------------------------------------------------------- MARIAL EARLY ARCHAIC EARLY ARCHAIC 8500------------------ 1O.000--------------------------------------------------------------------------- i 1" _____.z1LL_~ LATE ARCHAlC 180 I c , J ~QQ BP MIDDLE ARCHAlC -----_..._-- FIGURE 16. Umpqua Basin projectile points (after O'Neill 1989b). ! , I , · '( · i, · 1 ----_...._-- 181 Archaic sequence (Table 19). The chronology defines projectile point sequences for the Rogue Basin; the point types are illustrated in Figure 17 for reference. Work at Elk Creek has also defined a hydration curve applicable to sites in this region (Figure 18). The hydration dates derived from this curve have proven consistent with dates derived from radiocarbon studies, and provide another useful means for dating sites in the Rogue Basin. The three most common types of obsidian used in the Rogue Basin appear to hydrate at the same rate, making it possible to compare hydration readings among these sites. All the hydration readings given below (expressed in microns) are from one of these three sources. Pottery is another important chronological indicator for the Rogue Basin. It was produced between about 1100 and 400 BP (Mack 1989). Sites with pottery, therefore, have a Late Archaic component. Although many sites in this study have multiple chronological indicators, some only have projectile points present. Many of the site reports, however, were produced before the point chronologies discussed above were developed. Where this was the case, I assigned the site to a time period based on correlations with point styles defined in the Rogue or Umpqua Basin chronology, as illustrated in Figures 16 and 17, or based on the gross characteristics distinguishing Middle Archaic atlatl dart points (large, broad-stemmed) and Late Archaic arrowpoints (small, narrow- stemmed). 7~,->r ·-·41·"",~-·· • -~""'- .~,~"'.!i"'::'.'.~" - -,-" .~""'ti,~"L.:t-",.,__ "_"~~~~',i;~st::;~~~~ RRB PREFORM UCN WLM t WLL 't ECSB CCN RRB CCN RRB Lebow (1987)Pettigrew and AA A. " I ~ "". .. RRDC RRDS.. -.I: c RRBa. 0I> 0 .;: I " A.. .. ~ I. .. l~ .f ~ '--- . TC8TSBTSN - GIJfA-~~-I I • ". ,~ ~-~-ar - • As CSNECSB .Wl. :.~, CSB .. 0 .I: .. t ... .. • I~8g .. 0 .. WlM .~ 4 CCH RRB 2_tr_ CCN Nilsson and Kelly (1991) RRB .8 " CCN RRB CSB , RRB CSN CSH CSB 4A~~ I IIt\" t~I··· I I I iii !" . . . . A F1P1 .. '., .t:",., .RRDS SHCB ..,. & >Q" ~ I"~ ., '( T I I • ~ 'I' RRCN RRCN. USS No Marlal Ph••• Equivalent. 1Iul\IpIo~_Ile-'R_"_1y I I I I SNSB I I Wll I WLXLDSB :~ .I: <>0 1>0~ .. .... .. I ;; 0 '5 ~ :I .. • 8 ~.... • .= "u ":' 11 I illit 8 '" .... \WIow .....s-a \vm_ u.r ........ w· LarLarp \\,u Lar £:lin ...,. u-.e.---"" Una....51e..d ICI'id WLS • \YU. • \VlJ. • \VUL· UCIf IJSS "'rrt-pJlIr 51-1.Ie: "'eCc..~tI TtIa......5lnlpl ... TtIa_pIar C Bue: c~.UIc Scria tI..-c..e.. C......Sain~ e...1IhftPoa__.... ~S.......l ..... 1'$11 1'$1 tea CSI eer< OSI SK.~. .... JU .......... .... JU 0btaU1 CeuIridc:4 .... JU DI'I"I"I"IIIIISk. __ "'-C__"" mCrodo.~_ ~e-caft"" C.....lk sm- "................U4 UI • I.I.DC • laos . 1.1.(.,," • Eal Sl'C11 all FIGURE 17. Rogue Basin Points (Nilsson and Kelly 1991; reduced 50% of original). .....CD I\) 183 / / / / / / / / / / / / / / / / / / / / / / V f SOLID AND DASHED LINES REPRESENT ALTERNATIVE HYDRATION RATES ~ooo 1000 3000 4000 5000 .... Cl c( !' .1, , ! t Ii: ., . ., 2 3 RIM I.lEASUREMENT l~l --_...._-- o·lo'--------r------'T""""------T"""'""..:.--- I FIGURE 18. Rogue Basin hydration curve: plotted regression line from 1986 Elk Creek Obsidian hydration (from Pettigrew and Lebow 1987). 184 The site summaries which follow briefly present the chronological information provided in each site's report which was used to place that site into either the Middle or Late Archaic period. As just noted, sites which have already been placed into chronological sequences based on previous research were kept in those sequences, and related to the Middle/Late Archaic periods based on the correspondences expressed in Table 19. Rogue Basin Site Chronology 1. 35CU84. Marial Early, Middle, Late Marial is a well-stratified site with radiocarbon-dated components from 2810 +50 years BP at 60 centimeters below the surface to 8560 ±190 BP at 430 centimeters below the surface. The site produced projectile point styles characteristic of Early, Middle, and Late Archaic sites elsewhere in western Oregon. Occupation during all three major periods is attested by the projectile point styles, stratigraphy, and radiocarbon dates. The site is dated mainly on the basis of projectile point styles. Points from the site include ovate styles characteristic of the Middle Archaic as well as arrowpoints from the Late Archaic (Figure 17). .. I , 2. Gold Hill 3. J04, Ritsch Site Middle, Late Late , i, i, The site produced two radiocarbon dated components, both in the Late Archaic period. The later component was dated 460 ± 90 BP and the earlier at 1150 ±100 and 1400 ±80 BP. Both components were associated with the smaller stemmed, triangular, or barbed points characteristic of the Late Archaic in the Rogue Valley (Figure 17). 4. JA16. Marthaller Middle, Late q , ---_.... Projectile points include large ovate and broad-stemmed types characteristic of the Middle Archaic, as well as small arrowpoints typical of Late Archaic types in the Rogue Valley (Figure 17). 5. JA21. Saltsgaver Middle, Late 185 The site yielded two radiocarbon dates of 5310 ±140 BP and 1900 ±90 BP; the dates were obtained on charred wood and nut or camas fragments from the bottom of two of the oven features. The earlier date was from material collected during preliminary investigations in the 1960s. A lengthy period of use is also indicated by obsidian hydration readings, which range from 2.2 to 5.5 microns; using the hydration rate established for the Rogue Basin this suggests use of the site from about 5,000 BP. Arrowpoints characteristic of the Late Archaic were found on the surtaceof the site (Figure 17). 6. JA25. Far Hills Middle This site is dated on the basis of its projectile point assemblage, which includes mainly willow-leaf and corner-notched types characteristic of the Coquille Phase and transitional Rogue River 1 phase as defined by Nilsson and Kelly (1991). with a few arrowpoints from the Late Archaic period (Figure 17). The site assemblage analyzed here appears to derive primarily from an occupation during the later part of the Middle Archaic. 7. JA42. Applegate Late This is a single-component site from the latter part of the Late Archaic, and the only site in the whole study project to have documented historic trade goods, which place the site between about A.D. 1750 and 1850. 8. JA47. Applegate 9. JA47. Applegate see 9, 10 Middle This component is stratigraphically earlier than the later one (see below), and is dated to the Middle Archaic on the basis of very large, broad-stemmed points characteristic of the earlier part of 'this period and similar to points from about 5,000 BP at Marial (Figure 17). 10. JA47. Applegate Late This component is stratigraphically later than the one noted above. It contains small arrowpoints similar to those identified for the Rogue 2 subphase at Elk Creek (Figure 17). 11. JA77. Salt Creek Site Late -----",.,,-- The site is dated on the basis of the projectile points. These consist mainly of Late Archaic (Rogue 2 subphase) points, with possibly a few i· I ~1~1 p, 1\ ~, i'; ,to! " l"\'i ~~~' t; "~ ~ i ,r r ~; j; j t j: ~j t. t~ , " Il I 186 I I , , J Late No date Late Late Late Late see 20,21 Late 13. JA189. Trail 12. JA133. RRNF 15. JA191. Reeder 16. JA197. Little Butte The projectile points are characteristic of the Rogue 2 sUbphase (Figure 17). 14. JA190. Trail willow-leaf forms from the earlier period (Nilsson and Kelly 1991). The late period is clearly indicated; the earlier "points" illustrated may in fact be knives. No datable material was recovered from this site. Radiocarbon dates from the site place it at about 1700 to 650 BP. Pottery found at the site dates it to between about 1100 and 400 BP. Projectile points are those from the Rogue Phase (Figure 17). These chronological indicators place this site in the Late Archaic. Radiocarbon dates of about 750 - 310 BP (Connolly 1990) place this site in the Late Archaic, though a few projectile points may date from a slightly earlier period, either Rogue 1 or the later part of the Coquille period (Figure 17). The majority of the projectile points are Rogue 2 sUbphase types; obsidian hydration readings between 1.0 and 2.4 microns also place the site in the Rogue 2 subphase and hence the Late Archaic (Table 19). An earlier, Coquille occupation is lightly expressed by a few points and hydration readings; however the later period is better expressed and this site is categorized as a Late Archaic site on that basis (Nilsson and Kelly 1991). The site is dated to the Late Archaic on the basis of projectile points characteristic of the Rogue period (Figure 17). 18. JA11. Elk Creek 19. 35JA27A. Elk Creek Only one Rogue 2-style point was found (Figure 17); the site is assigned to the Late Archaic on the basis of that point. 17. JA10. Elk Creek I iif A ~\M , i1 I I ~ .. ip 1 1'; .. 11 Radiocarbon dates of 1210 ±120 and 680 ±90 BP, together with hydration data (1.3 - 1.9 microns) and many Rogue 2 subphase points place the main period of site use in the Late Archaic. Some use during the Middle Archaic period is also indicated, however, by the 187 Late Middle Late Middle Late Late 20. JA27A-1 Joham 1 21. JA271-2 Joham 1 24. JA100. Elk Creek This component is dominated by points from the Rogue 2 subphase (Figure 17) and contains ceramics; the most intensive period of use occurs after about 1000 BP. Obsidian hydration data from this component include a strong cluster of readings from about 1.0 to 2.5 microns, also consistent with the Late Archaic date. 22. JA27B. Elk Creek This component contains broad-necked and willow-leaf point styles characteristic of the Coquille and possibly Marial phases, which place it in the Middle Archaic (Table 19). Obsidian hydration data range from about 3.0 microns to 5.0 microns, consistent with the Middle Archaic (Figure 18). 23. JA59. Elk Creek 25. JA101, Elk Creek Projectile points fit the Coquille and possibly Marial types, with a medium-sized willow-leaf type dominant. Obsidian hydration data range from about 2.0 microns to 4.0 microns, with most use indicated between 3.0 and 4.0 microns. The point and obsidian data place this primarily in the Middle Archaic (Figure 17 and Figure 18). A series of radiocarbon dates ranging from 1070 ± 110 to 50 ± 60 BP for the main occupation period, plus projectile points, the presence of ceramics, and obsidian hydration data (1.2 - 2.0 microns) all place the main period of site use in the Late Archaic. A few hydration readings, however, and a small sample of projectile points indicate use of the site at an earlier period, during the Middle Archaic. The projectile points and obsidian hydration data place the main period of use of this site in the Rogue 2 subphase, with an indication of some use during earlier periods (Figure 17). The hydration data ranges from 1.0 to about 5.0 microns, with the great majority of . readings between 1.0 and 3.0 microns (Figure 18). -----_..._-- I ------~ ,.,. 188 hydration data (3.1 - 4.9 microns) and one point characteristic of the Coquille phase (Figure 17 and Figure 18). Middle Middle Middle Middle, Late Late Middle 26. JA102. Elk Creek 27. JA103. Elk Creek 28. JA105. Elk Creek 29. JA10. Elk Creek Projectile points 'from this site resemble those from the Coquille, and possibly Marial, phases (Figure 17). These place its period of primary use in the Middle Archaic, though a few Late Archaic points were also recovered. This site produced hydration readings clustered between 3.5 and 4.2 microns, and is dated to the Middle Archaic on that basis (Figure 18). This task site produced one Coquille phase point and hydration values ranging 'from 1.1 to 6.1 microns, suggesting intermittent use over a long period of time, covering the Middle and Late Archaic. 30. JA110. Elk Creek This site is dated to the Rogue 2 subphase on the basis of a small sample of hydration readings, which range from 1.2 - 3.0 microns (Figure 17). 31. JA112. Elk Creek Projectile points from this site place the period of greatest use in the Coquille phase. Obsidian hydration data, with an average of 3.3 microns and standard deviation of 1.2 microns, also indicate Middle Archaic use, with some use during the following period (Figure 17 and Figure 18). This site has hydration readings with a range of 2.4 to 5.9 microns, with most clustered around 3.5, which place it in the Coquille phase time span. Coquille type projectile points confirm this assessment; however, two Rogue 2 subphase points attest some use at a later period (Figure 17 and Figure 18). 32. EC-2. Elk Creek Middle, Late The site produced an abundance of Rogue phase projectile points and a few Coquille phase points. HydraUon readings range from 1.2 to 6.7 b i ------------------ _. -------- ._---- ____"""6" 189 microns, however, indicating that the site was occupied during both periods; it is assigned to both time periods on that basis. Middle, Late Late 33. Island. Elk Creek 34. Winningham. Elk Creek This site produced projectile points and hydration readings from both the Coquille and Rogue phases (2.7 - 3.6 microns and 1.8 - 2.0 microns, respectively) (Figure 17 and Figure 18). It was probably used intermittently throughout that time. Late style projectile points and hydration readings generally smaller than 2.7 microns place this site in the Rogue phase (Table 19 and Figure 17 and Figure 18). " i i~ 35. Zimmerly. Elk Creek Middle Davis (1983) terms this a late period site, though no specific chronological indicators are provided in his report. Projectile point types and hydration readings place the main period of use at this site in the Coquille phase, although a few hydration readings smaller than 2.6 suggest occasional use during the Late Archaic (Table 19 and Figure 18). , I I 36. JA5. Lost Creek 37. JA6. Lost Creek Late? Late A radiocarbon date from a hearth of 550 ± 80 BP, plus arrow points place the main period of use in the Late Archaic, though larger side- notched and willow-leaf points indicate some use at an earlier period. 38. JA7. Lost Creek 39. JAB. Lost Creek No date Late The site is dated to the Rogue 2 subphase on the basis of projectile point styles (Figure 17). 40. JA12. Lost Creek Late The site produced Rogue 2 subphase style points, and is dated to the Late Archaic on that basis (Figure 17). ----- ----- 41. JA14, Lost Creek Middle 190 ... The site produced broad-necked and lanceolate points characteristic of the Coquille and Rogue 1 subphase, and is dated to the Middle Archaic (probably the latter part) on that basis (Figure 17). 42. JA16. Lost Creek Late Two radiocarbon dates, from unspecified samples, date the site to 1120 ±75 and 1660 ±80 BP. The points are primarily Rogue 2 phase types, confirming a Late Archaic period of use. However, the occurrence of broad-stemmed points also suggests some use at an earlier period. 43. JA28. Lost Creek Middle, Late The projectile points illustrated for this site suggest a time span of use from about 3000 - 1500 BP, with the best fit in the Rogue 1 subphase (Figure 17). Since this subphase is transitional between the Middle and Late Archaic, the site is classed with each period. 44. JA19, Lost Creek Late A radiocarbon date of 1120 ±75 BP plus Rogue 2 subphase style points date the main period of occupation to the Late Archaic, though finds of broad-necked side-notched and willow-leaf points indicate earlier use. 45. JA20. Lost Creek 46. JA23. Fawn Butte No date Late The Rogue 2 subphase is strongly represented at the site by numerous projectile points, hydration data, ceramics, and a radiocarbon date of 260 +60 from a concentration of fire-cracked rock with bone and charcoal. -Earlier use is lightly represented by a few Coquille points and hydration readings. Umpqua Basin Site Chronology 48. 00275. Sylmon Late ----_...._- The site produced Late style projectile points, and clusters with sites placed in the Late Archaic (Beckham and Minor 1992:68). i 'I 191 49. 00274. Orchard Middle This site produced stemmed and broad-necked points which are characteristic of the Middle period, and is classed in this time 'frame on that basis (Figure 16). 50. 0036. Crispen Late A radiocarbon date of 620 ±60 BP, plus the cluster analysis (Beckham and Minor 1992:68) place this site in the Late Archaic. The occurrence of broad-necked points, however, indicates the possibility of earlier use of the site. . 51. 00412. Coffee Creek Late A radiocarbon date of 1500 ±60 BP dates the site to the Late Archaic, though broad-necked dart points indicate possible use during an earlier period. 52. 00413. Coffee Creek Late A radiocarbon date of 1050 ±60 BP dates this site to the Late Archaic, as do the small barbed and corner-notched projectile points (Figure 16). 53. Tiller 1 54. Tiller 6 No date No date 55. 00205. S. Umpqua Rockshelter. Upper Middle A radiocarbon date of 3190 ±50 BP as well as a predominance of stemmed atlatl dart points places this site in the Middle Archaic. It also clusters with other Middle Archaic sites (Beckham and Minor 1992:68). 56. 00205. S. Umpqua Rockshelter. Lower Late A radiocarbon date of 600 ±50 BP as well as numerous arrowpoints place this site in the Late Archaic, as does cluster analysis (Beckham and Minor 1992:68). 57. 00209. Hughes Rockshelter Middle --_CrrIILZ This site clusters with other Middle Archaic sites (Beckham and Minor 1992:68). It is placed in the Middle Archaic on that basis. Some later occupation is indicated, however, by barbed arrowpoints and a radiocarbon date of 1025±110 Bf> from the upper levels. 192 58. 00212. Time Square Rockshelter Late This site clusters with other Late period sites (Beckham and Minor 1992:68). A series of radiocarbon dates between 3240 ± 60 and 800 ± 80 BP and suggest that use began during the Middle Archaic, however. Numerous Late period points indicate that the dominant use was during the Late Archaic. 59. 00396. Sprint Late Late period points and ethnographic evidence place this site in the Late Archaic. 60. 00219. Section Creek 61. 00219. Section Creek-1 see 61, 62, 63 Late A radiocarbon date of about 150±50 BP, plus Late period points, place this component in the Late Archaic; it also clusters with other Late Archaic sites (Beckham and Minor 1992:68). 62. 00219. Section Creek-III Middle Projectile points and cluster analysis place this component in the Middle Archaic (Beckham and Minor 1992:68). Obsidian hydration data from the site confirm this date. 63. 00219. Section Creek-II Late Radiocarbon dates of about 520+50 to 1540±70 BP plus cluster analysis place this site in the Late Archaic (Beckham and Minor 1992:68). 64. 00395. Grubbe Ranch Late The site produced evidence for several undated occupation episodes; however the Late Archaic, is represented by narrow-necked arrow points. 65. 0058. Glide Middle The assemblage clusters with the Middle Archaic group (Beckham and Minor 1992:68) and is placed in the Middle Archaic on that basis. A few Late period points suggest use at a later date. _____.rdIIIL._~. 66. 0061, Whistlers Bend No date · ; 67. 0067, Winchester Bridge Late 193 The site is dated to the Late Archaic on the basis of historic accounts. 68. 005, Gatchel Late Historic references place use of this site in the Late Archaic, although a small number of projectile points and an atlatl weight indicate use during an earlier period. 69. 00153, Narrows 70. 00153. Narrows-I see 70,71,72,73 Late Radiocarbon dates of 330 ± 80 to 90 + 70 BP place this component in the Late Archaic. It also clusters with other Late period sites (Beckham and Minor 1992:68). 71. 00153. Narrows-II Late Radiocarbon dates of 1020 ± 60 and 450 ± 70 BP place this component in the Late Archaic. 72. 00153, Narrows-III Middle A radiocarbon date of 5090 ±80 BP places this site in the Middle Archaic; it also clusters with other sites in this time frame (Beckham and Minor 1992:68). 73. 00153. Narrows-IV Middle, Early? The radiocarbon date of 6270 ±130 BP places this site in the Middle/Early Archaic time frame. It clusters with other assemblages from the early sites along the North Umpqua (Beckham and Minor 1992:68). 74. 00359. Swiftwater 75. 00383, Susan Creek No date Late A radiocarbon date of 660 ±70 BP from a hearth places this site in the Late Archaic; small, Late Archaic arrowpoints support this evidence (Figure 16). 76. 00278, Bogus Creek Middle, Early? ____s1IL. Cluster analysis places this site in" the earliest group (Beckham and Minor 1992:68); however projectile points, hydration data, and a post- , (; 194 Mazama deposition of artifacts suggest the main period of use was dUring the Middle Archaic. 77. 00126. Steamboat Middle Large lanceolate and side-notched points indicate a Middle Archaic occupation (Figure 16). 78. 0011. Rhody Late Lichen growth on the rock cairns at this site suggest use at some time during the last few hundred years, during the Late Archaic. 79. 0040, Cavitt Creek 80. 00401, Dry Creek 81. 00401, Dry Creek-E No date See 81, 82 Early This component lies under Mazama ash and is radiocarbon dated at about the time of the eruption 6800 years ago. The assemblage includes andesite bifaces and projectile points similar to the "Borax Lake" assemblage in California. The points, stratigraphy, and C14 place the site in the Early Archaic. 82. 00401. Dry Creek-L Middle Few projectile points were recovered from the site. The main clue to dating the post-Mazama component comes from hydration data. Two samples of 3.8 and 4.2 microns indicate a Middle Archaic occupation, on analogy with Rogue Basin obsidian dates. 83. 00372, Revnolds Middle The site clusters with the early sites (Beckham and Minor 1992:68). Points from the site indicate a Middle Archaic occupation, probably beginning early in that period. Two broad-stemmed projectile points date this earlier component to the Middle Archaic. i . , 84. 00422. Island 85. 00422. Island-E See 85, 86 Middle i 86. 00422. Island-L Late 195 The Late Archaic component is indicated by two composite charcoal samples of about 1210 ± 70 and 1040 ± 90 BP and two barbed arrowpoints. 87. 00418, Apple Bench Middle The site is dated to the Middle Archaic on the basis of a few fragmentary lanceolate points. 88. 00265. Apple Creek Middle The site is dated to the Middle Archaic on the basis of a radiocarbon date of 3500 ± 110 BP on a feature, as well as on the presence of broad-necked points and obsidian hydration readings which cluster between 3.8 and 5.1 microns. On analogy with the Rogue Basin hydration curve, these hydration readings would place the site in the Middle Archaic, or earlier. 89, 00421, Copeland Middle The site is dated to the Middle Archaic on the basis of one CoqUille style projectile point. 90. 00161. Medicine Creek Early, Middle, Late The pre-Mazama component of this site is similar to the Dry Creek site and dates from the Early Archaic, The post-Mazama component includes both atlatl and arrow points, and was used during both the Middle and Late Archaic, 91. 00187, Powerful 1 Late Arrowpoints from the site date it to the Late Archaic period. 92. 00227, Powerful 2 Middle, Late Both narrow-stemmed arrow points and broad-necked atlatl dart points date this site to the Middle and Late Archaic. 93, 00379, Snuff Out 94. 00397, Shivigny East No date Late This site clusters with other Late Archaic sites (Beckham and Minor 1992:68). 95. 00289. Little Oak Flat Late 196 One arrowpoint and hydration readings ranging from .9 to 1.4 microns place this site in the Late Archaic. 96. 00399. Snowbird Middle One broad-necked and one lanceolate point indicate a Middle Archaic occupation (Figure 16). 97. 00160. Muddy Middle Corner-notched and broad-necked points indicate a Middle Archaic occupation. 98. 00398. Powerline 99. 00389. Limpy No date Late -----.-.,-- Numerous arrowpoints and a radiocarbon date of 430 +60 BP place this site in the Late Archaic. It also clusters with other Late period sites (Beckham and Minor 1992:68). Subsistence and Settlement Change in Prehistoric Southwest Oregon Two potential subsistence/settlement regimes were identified for this area. The first consists of a collector regime, exemplified by the aboriginal cultural patterns extant at the time of contact. The second is a more mobile regime, hypothesized for an earlier period in the region, and possibly correlated with different environmental conditions. The collector pattern is manifest on the landscape through the existence of three broad classes of sites: villages, seasonal camps, and task sites. The mobile regime, however, produces only two broad classes of sites: seasonal camps and task sites. In order to consider the possibility of change in the subsistence/ settlement patterns, it was first necessary to place each site in the database r· ,rr: ,II , i'!' , , . I ----_...,,--- 197 into one of the three categories of sites just mentioned. The foregoing analyses utilized a variety of techniques, with a high degree of agreement among them, to place sites into these functional categories. These functional categories are the components used to reconstruct past settlement systems. In order to derive these past systems, it is necessary to relate the sites to one another in time; the preceding chronological analysis provides the information necessary to accomplish this task. Table 20 presents the functional and chronological information for sites from the Rogue Basin and the Umpqua Basin; this information is summarized in Table 21. The collector regime is well represented during the Late Archaic in both the Umpqua Basin and the Rogue Basin. The settlement pattern for both these areas during this period includes villages, seasonal camps, and task sites. As discussed in Chapter III, these three site types together form the core of the collector settlement pattern. For the Late Archaic in both regions, more than one third of the sites are vlllage sites; almost half the sites are seasonal camps; the remainder are task sites. There is a profound difference between the subsistence/settlement patterns of the Middle and Late Archaic periods in both the Umpqua and Rogue Basin samples. In both areas--and most dramatically in the Umpqua sample--there is a much lower percentage of village sites and higher percentage of seasonal camp sites during the Middle Archaic. For the Rogue Basin, only 17 percent of the sites are villages; 56 percent are seasonal camps, and 27 percent are task sites. In the Umpqua Basin sample, less than one percent of the sites are villages, almost three-quarters ----_....-._- TABLE 20. Site Functional Designations and Temporal Period Ref. Site Functional No. No. Elev. Site Name Type Period Middle Archaic Sites Rogue Basin Sites 1 35CU84 950 Marial seas. camp E,M,L 2 35JA1 1000 Gold Hill Village M,L 4 35J016 950 Marthaller village M,L 5 35JA21 1200 Saltsgaver seas. camp M,L 6 35JA25 1450 Far Hills seas. camp M 19 35JA27A 1700 Elk Creek 21 JA27-2 1700 seas. camp M 22 35JA27B 1700 Elk Creek seas. camp M 26 35JA102 1750 Elk Creek task M 27 35JA103 1800 Elk Creek seas. camp M,L 28 35JA105 1600 Elk Creek task M 29 35JA107 1600 Elk Creek seas. camp M 31 35JA112 1700 Elk Creek task M 32 EC-2 1900 Elk Creek seas. camp M,L 33 Island 1700 Elk Creek task M,L 35 Zimmerly 1700 Elk Creek task M 41 35JA14 2000 Lost Creek seas. camp M 43 35JA18 2000 Lost Creek village M,L Umpqua Basin Sites 49 3500274 440 Orchard seas. camp M 55 3500205 1700 S.Ump.RS-U seas. camp M 57 3500209 2150 Hughes I task M 60 3500219 1840 Section Crk 62 00219-3 1840 seas. camp M 65 350058 700 Glide village M 72 00153-3 800 seas. camp M 73 00153-4 800 seas. camp M 76 3500278 1050 Bogus seas. camp E?,M 77 3500126 1600 Steamboat task M 80 3500401 1500 Ory Creek 81 00401-E 1500 seas. camp E 82 00401-L 1500 seas. camp M 83 3500372 1600 Reynolds task M 84 3500422 1300 Island 198 ! . , 1',\ WI.:. _____szIIt, __ TABLE 20. Continued 199 r';:','," TABLE 20. Continued 200 'I I: --_."..,-- Ref. Site Functional No. No. Elev. Site Name Type Period 39 35JA8 1500 Lost Creek seas. camp L 40 35JA12 1500 Lost Creek village L 42 35JA16 1500 Lost Creek seas. camp L 43 35JA18 2000 Lost Creek village M,L 44 35JA19 2000 Lost Creek seas. camp L 46 35JA23 1950 Fawn Butte seas. camp L Umpqua Basin Sites 48 3500275 450 Sylmon seas. camp L 50 350036 900 Crispen village L 51 3500412 900 Coffee Creek village L 52 3500413 900 Coffee Creek village L 56 3500205 1700 S.Ump.RS-L seas. camp L 58 3500212 2600 Time Sq. RS seas. camp L 59 3500396 900 Sprint seas. camp L Ii: 61 00219-1 1840 village L I 63 00219-2 1840 village L 'I'i64 3500395 550 Grubbe seas. camp L 67 350067 480 Winchester task L 68 3500252 1000 Gatchel village L 69 3500153 800 Narrows 70 00153-1 800 village L I 71 00153-2 800 village L I;i 75 3500383 900 Susan Crk seas. camp L I 78 350011 3400 Lower Rhody task L 84 3500422 1300 Island 86 00422-L 1300 task L 90 3500161 2200 Medicine Crk seas. camp E,M,L 91 3500187 2400 Powerful 1 seas. camp L 92 3500227 2200 Powerful 2 seas. camp M,L 94 3500397 3280 Shivigny seas. camp L 95 3500289 3200 Little Oak task L 99 3500389 3000 Limpy RS seas. camp L Notes: Occupation Periods (sites not listed were not datable): E = Early Archaic (8.000 - 6,000 BP) M = Middle Archaic (6,000 - 2,000 BP) L = Late Archaic (2,000 - 150 BP) "" '11 , r :: ,1'1' : 201 TABLE 21. Functional Site Types by Chronological Period Rogue Basin Umpqua Basin Mid Late Mid Late Site Type N % N % N % N % Task 5 27% 5 16% 5 25% 4 17% Seas. camp 10 56% 14 45% 14 70% 11 48% Village 3 17% 12 39% 1 5.9% 8 35% ....---- ------- ----- ------- 18 100% 31 100% 20 100% 23 100% of the sites are seasonal camps, and the remaining 26 percent are task sites (see Figure 19 and Figure 20). The mobile model predicts a high percentage of seasonal camps, complemented by task sites. The distribution of site types for the Middle Archaic in both the Umpqua and Rogue Basins corresponds well to the predictions of the mobile model. Though the mobile pattern appears characteristic of 'the Middle Archaic, the transition to a more sedentary regime may have begun in some places dUring this period. Two of the three Middle Archaic village sites in the Rogue sample (the Marthaller site [#4] and the Gold Hill site [#2]) are from lower elevations and further down the Rogue than most of the other sites in the sample. Neither of these sites has a well-dated assemblage of materials; review of the projectile points suggests that intensive occupation began about 3,000 years ago. but this estimate needs corroboration from further stUdies. Possibly the village pattern appeared earlier along the mainstem of -.u'."' ~__ ---..L, __ ..... 1, -. . ~ -- "3ij .;_. ·C.. -- . . .--~c~-~c Rogue Basin Sites 80.0 60.0 UJ t=I