Characteristics of the Cayuse Phase

INTRODUCTION

[38] The Cayuse Phase is represented in the archaeological record throughout the Columbia Plateau as well as in many immediately adjacent areas, including the eastern foothills of the Cascade Mountains, the Okanogan Highlands, and the westernmost foothills of the Rocky Mountains. Although it appears to have began approximately at the beginning of the Christian era everywhere it occurs, it is not a monolithic archaeological entity devoid of significant internal variation. On the contrary, it is divisible into a series of regional geographic variants each of which is composed of more or less unique temporal subphases. While regional variation is still poorly understood, the general characteristics of the Cayuse Phase can be specified in some detail.

GENERAL CHARACTERISTICS OF THE CAYUSE PHASE

The Site Complex. The emergence of the Cayuse Phase is not characterized as much by the appearance of an obviously different material culture as it is by the apparently abrupt emergence of a dramatically new configuration of site types which occur in particular relationships both to one another and to the topographic and ecological zones of the Columbia Plateau. The most basic feature of this change is the appearance of identifiable site complexes in topographically protected locations in the peripheral portions of the Columbia Plateau.

"Site complex" is a term which has been utilized by Nelson and Rice (1966) to describe geographically associated sites which are of the same general age, which are functionally diversified, and which could have been utilized concomitantly for different purposes by a single group of people. The late prehistoric, Cayuse Phase site complexes which were located on the Snake River at the southeastern margin of the Columbia Plateau by Nelson and Rice (1966) are composed of extensive open sites in association with well defined burial areas. The open sites frequently contain the remains of house structures which indicate they were used as winter villages or camps, and storage shelters, storage pits, pictographs, petroglyphs, and fish walls frequently occur at sites within the site complex. This basic configuration, having a nucleus of large open sites and burial areas containing or closely associated with a variety of other special features, is characteristic of the Cayuse Phase throughout the Columbia Plateau. [38]

previous | next

The distribution of such site complexes is difficult to define with absolute certainty primarily because Plateau archaeologists have never before dealt with such a unit or specifically attempted to demonstrate its presence or absence. Nevertheless, there is sufficient information contained in survey and site reports to state that site complexes are distributed along the fringes of the Columbia Plateau and in the adjacent Canadian Plateau.

A survey in the east Kootenay region of the Canadian Plateau has revealed several concentrations of sites, especially at the northern ends of Columbia and Windermere lakes (Borden 1956). Although the sites are not individually described, structural remains are said to be common and there is at least one large burial area present at each one of the large site clusters. Further to the west, a survey of the Okanogan Highlands (Caldwell 1953-54) revealed large numbers of sites along the shores of Lake Okanogan and in the Similkameen Valley. Similar concentrations were found [38/39] somewhat further to the south on the Okanogan River. Although descriptions of the individual sites have not been published, the high frequency of structural remains and the presence of burial areas, rockshelters, and pictograph sites indicate that site complexes are very common in the sheltered valleys of the area. Moreover, along the southern skirts of the Okanogan Highlands and in the northernmost portion of the Columbia Plateau, both survey and excavation have established the presence of site complexes (Osborne 1949; Osborne, Crabtree, and Bryan 1952; Sloan 1963; Grabert 1966; Collier, Hudson and Ford 1942).

On the western margin of the Plateau, surveys of the Vantage locale have revealed abundant evidence for the presence of site complexes (Shiner 1951b; Lee 1955). Excavated examples include 45KT28 with its adjacent burial areas and rockshelters, a series of pit house, rockshelter, burial, and petroglyph sites in the vicinity of Schaake Village (Swanson 1962a; Kidd 1964; Osborne 1956-57), and a small association of pit house sites, burial areas and petroglyph sites south of Vantage (Massey and Nelson 1958).

Survey work (Smith 1910; Weeks 1962) and the excavation of a village and adjacent burial yard at the mouth of Wenas Creek make it clear that site complexes are common in and around the Yakima Valley (see Warren 1959).

Both site survey and excavation documents the presence of numerous site complexes along the Middle Columbia and on the Snake River near its mouth (e.g., Shiner 1950a, 1952b; Cole and Newman 1958; Cole and Leonhardy 1964). Outstanding examples where there has been large scale excavation include the Wildcat Canyon site complex (Cole 1963; 1964; 1965) and the Wakemap Mound site complex (Strong, Schenck, and Steward 1930; Caldwell 1956; Butler 1958a).

Site complexes are present but infrequent along much of the Lower Snake, a feature which appears to be correlated with a low frequency of identified structural remains and ethnographically reported winter villages, facts which will become more significant as the discussion of the Cayuse Phase unfolds. Nevertheless, well documented examples occur at the mouth of the Palouse River (Fryxell and Daugherty 1962; Nance 1966; Osborne 1948), the mouth of the Tucannon River (C. M. Nelson 1966), and at the mouth of Dry Gulch, near Ridpath (C. M. Nelson 1965b: 45C011; Sprague and Combes 1966). Local site surveys suggest that still others exist (see Osborne 1948; C. M. Nelson 1965b).

Further to the east, the presence of site complexes has been well established for the southeastern periphery of the Columbia Plateau by Nelson and Rice (1966).

Unfortunately, our entire knowledge of the eastern margin of the Columbia Plateau comes from two preliminary surveys around Coeur d'Alene and Pend Oreille lakes (Shiner 1950b, 1953b; Miller 1959). However, several sets of associated burial areas and open sites do indicate that site complexes did exist in the Coeur d'Alene area (Miller 1959:41).

This brief discussion of the distribution of site complexes has been designed to illustrate the range of their distribution and is not meant to be an exhaustive list of known occurrences. Those interested in compiling such a list will find evidence for well over 100 site complexes in the archaeological literature of the Columbia Plateau.

From everywhere there are reasonable samples of artifacts in association with site complexes, assemblages are dominated by projectile point types known to be characteristic of the Cayuse Phase, and there is no known site complex associated with tool types which antedate the Cayuse Phase. One of the best dated Cayuse Phase site complexes is located at Wildcat Canyon, [39/40] where C14 dates (see below) indicate that it appeared in the first century B.C. and at the very beginning of the Cayuse Phase.

Structural Remains. The remains of pit houses and other semi-permanent dwellings are associated closely with distribution of site complexes and the Cayuse Phase itself. And, as the proceeding discussion has already indicated, they are very frequently constituent parts of site complexes.

There are only two known occurrences of house remains which might antedate the appearance of the Cayuse Phase in the Columbia Plateau. The earliest is at Cold Springs (Osborne and Shiner 1949; Shiner 1961), where house floors have been reported in association with projectile points characteristic of the Cold Springs Phase. The stratigraphy at Cold Springs is simple, but difficult to interpret archaeologically. Volcanic ash, probably from the eruption of Mt. Mazama about 4500 B.C. (cf. Fryxell 1965), is overlain by a light, undifferentiated midden (probably floodplain loess) the lower portion of which is said to be sandy in many areas. Within the midden local stratigraphy can be defined on the basis of lenses of fresh water mussel shells, but precise lateral correlations over large distances are impossible. The presence of Cascade and Cold Springs Side-Notched projectile points immediately above the volcanic ash demonstrate the presence of a Cold Springs Phase component. However, the bulk of the overlying midden, which is approximately four feet thick, contains so few diagnostic artifacts that it cannot be assigned to any phase. Yet the presence of historic artifacts in the upper portion of the midden make it clear that terminal occupation occurred in the early historic period.

A large number of pit houses have been excavated from the uppermost portions of the midden at Cold Springs. Very few artifacts were recovered from these structures and no diagnostic projectile points were found in direct association. However, on the basis of two large, crude, leaf-shaped knives, Shiner correlates one of the house structures with the Cold Springs Phase component at the site. More recent excavations in the southern Columbia Plateau have shown, however, that such knives are not especially diagnostic of any phase. For example, fragmentary specimens have been reported from Cayuse Phase components which are clearly stratified above components of the Cold Springs and later phases and in association with a C14 date of 230 ± 165 A.D. (C. M. Nelson 1966:43, Fig. 28, m).

On the other hand, there is a good deal of circumstantial evidence to suggest that the house structures at Cold Springs are protohistoric in age. The protohistoric and early historic structures which are best known from the Cold Springs locale are saucer-shaped with sloping side-walls, possess indistinct floor areas, contain no evidence for superstructural members, and are typically associated with very few diagnostic tools (Osborne and Shiner 1951; Osborne 1957). These are the precise characteristics of the structures at Cold Springs.

A more substantial case can be made for pre-Cayuse structural remains at the Shalkop Site, a small open site in the Vantage locale. Here there are two house pits associated with a total of five diagnostic projectile points, including four simple stemmed examples and one leaf-shaped point or knife, the presence of which prompted Swanson (1962a) to interpret the site as a Frenchman Springs Phase component. The projectile point types found at the Shalkop Site are also characteristic of the protohistoric Cayuse III Subphase, where they occur in association with other late prehistoric point types. Although the absence of these later projectile point types suggests a Frenchman Springs Phase association, certain features of the five stemmed projectile points are not typical of Rabbit Island Stemmed points, the type most closely associated with the Frenchman Springs Phase. For example, two of the specimens are slightly barbed and two are serrated, characteristics which do not appear in the published Rabbit Island Stemmed specimens (see Type 3; Swanson [40/41] 1962a: Figs. 45, m-n 36, a-f; Swanson 1962b; Crabtree 1957: Plate 8), but which are common features in the Cayuse III Subphase. Therefore, the possible association of house structures with the Frenchman Springs Phase should be treated with skepticism until corroborative evidence is found.

In contrast to their dubious association with earlier phases, the structural remains of semi-permanent dwellings are very numerous during the Cayuse Phase. They first appear at the very beginning of the phase and remain common features until well into the historic period (see discussion of the age of the Cayuse Phase).

The distribution of house remains coincides with the distribution of site complexes. Each occur in areas marginal to the Columbia Plateau and in the adjoining Canadian Plateau.

In the Canadian Plateau, Borden (1956) reported numerous house structures from the east Kootenay region and linked them with the known ethnographic settlement pattern of the area. To the west Caldwell reports the presence of structural remains at at least 24 sites along the Okanogan and Similkameen valleys (Caldwell 1953-54). They include circular and rectangular subterranean structures as well as stone and log lined structures. Although the exact age of these dwellings is not known, the presence of wooden structural members at some and the absence of trees from the depressions of many others suggest that a substantial percentage are protohistoric in age. Structural house remains are also common along the southern fringes of the Okanogan Highlands and include the remains of several early historic mat lodges (Osborne 1949; Osborne, Crabtree, and Bryan 1952; Sloan 1963; Grabert 1966; Swanson 1958-59). It is interesting to note, however, that Collier, Hudson, and Ford (1942) do not note the presence of structural remains in the Grand Coulee Dam reservoir. Nevertheless they (1942:37) state that a large number of saucer-like depressions were tested on the Columbia and lower Spokane rivers, and that none proved to be house remains. However, since closed, circular depressions are not likely to occur naturally along the floodplain of a river such as the Columbia, it is possible that they were pit house depressions and that testing was simply not extensive enough to verify this fact. Along much of the Upper Columbia pit house depressions have been filled with up to four feet of sterile overburden through post-occupational flooding. Many of the largest and most recent house structures have very little occupational fill and floors which are very difficult to define in the absence of extensive testing. In such cases, simple shovel testing to a depth of two or three feet will provide no evidence of structural remains.

Along the western margin of the Columbia Plateau, house remains are extremely common both along the Columbia River (Shiner 1951b; Lee 1955; Swanson 1958, 1962a; Osborne 1956-57; Kidd 1964) and in the Yakima Valley (Smith 1910; Warren 1959; Weeks 1962). Excavation has consistently revealed that they are associated with Cayuse Phase components.

This pattern is continued along the Middle Columbia, where surveys and site excavations have revealed large numbers of structural remains in association with Cayuse Phase components and tool assemblages (Strong, Schenck, and Steward 1930; Shiner 1950a, 1952b, 1961; Cole 1963, 1964, 1965; Cole and Cressman 1959, 1960, 1961; Cole and Leonhardy 1964; Cole and Newman 1958; Caldwell 1956; Cressman and Cole 1962; Butler 1958a; Osborne 1957; Osborne and Shiner 1949, 1951).

Both survey (Osborne 1948; Drucker 1948a) and excavation (Kenaston 1966) demonstrate that the downstream portion of the Lower Snake River also contains a number of sites at which structural remains occur. Excavations at the Harder (Kenaston 1966) and Three Springs Bar (Daugherty 1965: personal communication) sites have revealed such dwellings are associated with Cayuse Phase projectile point assemblages. East of Three Springs Bar site surveys have revealed very occasional pit house sites as far upstream as the mouth of the Palouse River. A recent resurvey of [41/42] the Snake River between the mouths of the Palouse and Clearwater rivers has revealed an almost total absence pit house depressions. The single site reported by C. M. Nelson (1965b: 45GA5) has been found to be a series of pumping pits excavated by a local farmer. The absence of identifiable structural remains in this portion of the Lower Snake correlates with low site densities, a general lack of well defined site complexes, and a low density of ethnographically reported winter villages.

Further up the Snake River, near the town of Lewiston, over twenty sites have been located which contain more than 100 house depressions (Nelson and Rice 1966). To the east of this area Butler (1966) has also reported on a pit house site which was first occupied in the first century B.C.

Our archaeological knowledge of the eastern margin of the Columbia Plateau is very limited. A site survey which concentrated on Coeur d'Alene and surrounding lakes revealed no house depressions (Miller 1959). However, Miller notes that local collectors have reported house depressions from neighboring river valleys, an observation which agrees well with the ethnographic distribution of winter villages in the area (see Ray 1936: Fig. 3).

Further to the north there was some limited testing and survey work done by the Smithsonian near the northern end of Lake Pend Oreille, but no house remains were found (Shiner 1950b; 1953b).

Although site complexes appear to be rare or absent in the coulee area of the west-central Columbia Plateau, pit house remains are relatively common (Drucker 1948b; Osborne 1959; Daugherty 1952). Excavations at five of these sites have shown that the house structures are associated with projectile point assemblages characteristic of the Cayuse III Subphase. The fact many site complexes have not been identified from this area is partially a function of site size and the protohistoric nature of the winter occupation sites. It will be more fully examined in the discussion of the Cayuse III Subphase.

Increase in the Number and Density of Sites. In those geographic areas in which site complexes are distributed, both site surveys and excavations demonstrate that there is a dramatic increase in the frequency of sites at the beginning of the Cayuse Phase. For example, a thorough survey of one area at the southeastern margin of the Columbia Plateau revealed 155 archaeological sites (Nelson and Rice 1966). Of these, 37 are Cayuse Phase burial areas, 17 are Cayuse Phase storage shelters, and 97 are open sites of which 91 contain late prehistoric, Cayuse Phase components. Only 9 sites contain components which are known to antedate the Cayuse Phase, and while excavations will undoubtedly produce more pre-Cayuse components, the topographic position of most of the known sites suggests that they never will be found nearly as numerous as Cayuse Phase components. The extremely lopsided distribution of Cayuse and pre-Cayuse components in this portion of the Columbia Plateau may be partially due to the high local density of site complexes. However, if we examine the situation further westward on the Snake River, where site complexes are uncommon, similar situation will be found (e.g., C. M. Nelson 1965b; Sprague and Combes 1966). Known Cayuse Phase components always outnumber the combined totals for all other proceeding phases. This is not only true for the Lower Snake, it also holds for the Middle Columbia, the Yakima Valley, the Upper Columbia, the coulee area of the west-central Columbia Plateau, the Okanogan Highlands (viz. Caldwell 1953-54; Swanson 1958-59), and the eastern margin of the Columbia Plateau (viz. Shiner 1950b, 1953b; Miller 1959).

Although conditions of preservation and sampling procedures generally favor a more comprehensive knowledge of recent prehistoric events, these factors do not entirely account for the disparity between the numbers and densities of Cayuse Phase and pre-Cayuse Phase components throughout the Columbia Plateau. This disparity looms ever larger as we consider (1) that the [42/43] percentage of pre-Cayuse Phase components has not risen greatly as increased research has provided more primary data, (2) that pre-Cayuse Phase components do not outnumber Cayuse Phase components in areas where there has been intensive survey and excavation, (3) that Plateau prehistorians have searched for early sites more diligently than they have for recent ones, and (4) that the pre-Cayuse occupation of the Columbia spans at least 8,000 years, while the Cayuse Phase persists something on the order of 2,000 years. Therefore it seems reasonable to conclude that larger numbers of sites and greater site densities are actually characteristics of the Cayuse Phase. These conclusions apply particularly to the topographically protected areas where winter camps and villages were maintained. Cayuse Phase sites lying outside these areas also may be more numerous, but there are not yet enough archaeological data to state this with certainty.

Increase in Site Size. Throughout the Columbia Plateau the areal extent of Cayuse Phase components tends to be greater than that of pre-Cayuse Phase components. Along the Upper Columbia at sites such as 45KT28, pre-Cayuse Phase components frequently cover less than 25 square yards and rarely have areas in excess of 500 or 1,000 square yards, while Cayuse Phase components commonly cover areas ranging from 5,000 to 30,000 square yards and reach maximum dimensions in excess of 50,000 square yards. Although this pattern generally seems to be consistent throughout the Columbia Plateau, small numbers of early components have been reported which may cover considerable areas. One such early terrace site on the Lower Snake River (C. M. Nelson 1965a; 1965b: 45GA3) traverses over 100,000 square yards. However, it is important to note that this site, and others like it, are composed of localized concentrations of camp debris separated by extensive areas in which only occasional flakes and tools are found. These local concentrations normally cover areas ranging from 50 to 1,200 square yards. Perhaps one of the largest well documented early occupations which is characterized by continuous concentrations of debris in the Cold Springs Phase component at 45C01 (C. M. Nelson 1966: Assemblage 2A), which is estimated to extend over an area of between 1,600 and 2,500 square yards. Early components of comparable dimensions also may exist at Thorn Thicket (Sprague and Combes 1966).

Thus, although there are probably a few pre-Cayuse Phase components which approach the magnitude of the more representative Cayuse Phase components, their geographic density is extremely low both in absolute terms and in comparison with the density of large Cayuse Phase components. Therefore, both the tendency towards increased site size and the far greater density of large components may be considered characteristics of the Cayuse Phase.

The Topographic Setting of Sites. Along lakes and trunk streams and their major tributaries, and other topographically sheltered areas, archaeological sites tend to be concentrated in such places as the mouths of small streams and canyons, the ends of large floodplain bars, places where the floodplains of rivers are narrow, and other highly protected portions of the local landscape (e.g., Nelson and Rice 1966: Fig. 1). This feature appears to be related to the necessity of establishing winter camps and villages in climatically favorable microenvironments near reasonably large winter food supplies, especially game animals such as deer and elk. Moreover, it is a pattern that apparently existed during all periods of Plateau prehistory, although it is difficult to generalize for periods antedating the Cayuse Phase.

However, there is a subtle change within this pattern that corresponds to the beginning of the Cayuse Phase. It is recorded only at floodplain sites, such as 45KT28 and 45C01 (C. M. Nelson 1966), where there is more or less continuous occupation across the boundary marking the emergence of the Cayuse Phase, and it involves the specific part of the site being utilized prior to and after the beginning of the phase. [43]

[44] At such floodplain sites it is possible to distinguish two major geomorphic subdivisions: (1) the active beach zone that is seasonally inundated at periods of normal high water, and (2) the adjacent floodplain that is inundated only at times of serious flooding. The subdivisions are usually separated by a cut bank approximately two meters high. The floodplain is characterized by very fine fluvial sands and silts deposited during occasional periodic floods, floodplain loess, and fine-grained alluvial deposits around the margins of alluvial fans. On the other hand, the active beach zone is characterized by a variety of fluvial and aeolian deposits that are dominated by medium sands.

Although ethnographic information makes it clear that both geomorphic zones were utilized during the Cayuse Phase, recent erosion and active back-cutting of the cut bank make it difficult to archaeologically assess the specific roles that the beach zone may have played. On the other hand, both the ethnographic and archaeological records clearly indicate that the floodplain was utilized for the construction of semi-permanent dwellings, including semi-subterranean pit houses, mat lodges, and a variety of typologically intermediate structures. Houses were built in this section of floodplain sites for a variety of reasons.

1. The loess and other fine-grained deposits of the floodplain provided reasonable structural support for the open faces of pit house walls. Such open faces soon collapse when excavated into the medium sands of the beach zone. The more consistent, fine-grained deposits of the floodplain also provided greater support for the structural members of houses. These would have to be anchored with much greater care in the beach zone.

3. Late winter thaws frequently produce episodes of high water that reach the foot of the cut bank at the very margin of the floodplain. Since house structures of the types associated with floodplain sites were primarily winter abodes, their construction in the beach zone might entail abandonment and reconstruction during the winter months at times when food supplies might be short and the necessity for continuous hunting and food gathering at its maximum.

In contrast to the Cayuse Phase, earlier phases are characterized by in situ components in the active beach areas of many floodplain sites and only very sparse evidence for occupation in the floodplain facies of the same deposits (e.g., C. M. Nelson 1966). This indicates that almost all the activities conducted at such sites were carried out in the beach zone and that semi-permanent dwellings were not in use in the floodplain context where they are so characteristic of the Cayuse Phase. Furthermore, this change which occurs the primary living area, from the active beach zone to the adjacent floodplain, may be interpreted as an accommodation to the construction of semi-permanent houses.

Population Growth. Because of the severe winters which occur in and adjacent to the Columbia Plateau, the winter months must be spent in topographically protected areas and at relatively low altitudes. This means that river valleys, deep canyons, protected lake basins, and other sheltered areas must have been the major winter population centers during all periods of Plateau prehistory. Therefore, the size and density of sites in such areas should at least very roughly reflect relative population sizes. As noted above, the evidence indicates that the Cayuse Phase is [44/45] characterized by larger site areas, greater numbers of sites, far greater densities of sites, and the appearance of large, readily identifiable site complexes. The appearance of these features in topographically protected areas suggest (1) that the Cayuse Phase was characterized by larger, denser populations than proceeding phases, and (2) that there was a relatively sudden population increase at the beginning of the Cayuse Phase.

Coastal Trade. Coastal-interior trade apparently became more intensive and shifted in a northerly direction at the beginning of the Cayuse Phase. The evidence in support of this view may be summarized as follows.

1. Although ornaments made from local shells, stone, and bone are relatively common prior to the Cayuse Phase in much of the western Columbia Plateau (Crabtree 1957: Rabbit Island I; Daugherty 1952; 45GR27, Feature 1; 45KT28). ornaments manufactured from shells traded inland from the coast are rare, being limited to 15 specimens of Schizotherus nutally in two of the Rabbit Island I burials (Crabtree 1957:39-43).

2. Although reported data other than those acquired by Crabtree (1957:39-43) do not indicate the presence of coastal trade items along the Middle Columbia in pre-Cayuse times, information from the Lower Snake River indicates that Olivella biplicata were relatively common in at least some portions of the southern Columbia Plateau (Fryxell and Daugherty 1962; C. M. Nelson 1966: Assemblages 2A and 3B). Although this shell has a spotty distribution as far north as the Strait of Juan de Fuca, it is far more abundant further to the south along the coasts of Oregon and northern California, facts which suggest trade up the Columbia and Snake rivers by reason of proximity. The absence of shell trade items in the northern Columbia Plateau also suggest this route. However, it is possible that their apparent absence in the north is due to sampling error instead of relative abundance.

3. Recently Kidd (1964:8-11) has reported the recovery of a small nephrite adze in a component at the Hole-in-the-Wall Canyon site near Vantage, Washington, which he has tentatively correlated with Swanson's Frenchman Springs Phase at a time depth of ca. 1000 B.C. If Kidd's estimate is correct, this adze is the earliest northerly trade item currently reported from the Columbia Plateau and is nearly as old as the earliest reports of such tools from the Fraser River drainage itself. However, the sample size from the earlier Hole-in-the-Wall Canyon component is very small, readily diagnostic items being limited to a single rectangular-stemmed projectile point (Kidd 1964:11, Fig. 3, b), a form which may be characteristic of the local manifestations of both the Frenchman Springs and Cayuse phases. The stratigraphic position of the component is also such that it could be considered much more recent than 1000 B.C. Therefore, in the absence of corroborative evidence, the age and phase association of this component and the adze which it contains must be considered questionable.

4. Excavations at 450K78, an open site on the northern fringe of the Columbia Plateau, have revealed an early Cayuse Phase component which has been C14 dated to 580 ± 160 B.C. (Grabert 1966:29). The assemblage from this component contains an adze (Grabert 1966: Fig. 13, f), the earliest verified northerly trade item in the Columbia Plateau. Since the Cayuse Phase is thought to have begun four or five centuries later further to the south, the initial appearance of northerly trade items should be expected in late pre-Cayuse Phase components throughout much of the northern Columbia Plateau. Although this tends to obviate possible conflict over Kidd's tentative interpretation of the Hole-in-the-Wall Canyon site, it does not provide conformation of Kidd's interpretation. [45]

[46] The following observations support the conclusion that the beginning of the Cayuse Phase marks the introduction of several new trade items into the Columbia Plateau and the emergence of a well developed pattern of trade with the Canadian Plateau and the adjacent coasts of northern Washington and southern British Columbia.

1. Dentalium trade begins with and is relatively heavy at the beginning of the Cayuse Phase, The evidence for this comes from sites such as 45KT28 (Subcomponent VIIA), 45C01 (C. M. Nelson 1966: Assemblages 4A and 4E), and Wildcat Canyon (Cole and Cressman 1960), where Dentalia first appear at the lower boundary of the Cayuse Phase. Although burial practices make it difficult to measure the relative geographic abundance of Dentalia, private collections give the impression that it is somewhat more common in the northern Plateau during the first part of the Cayuse Phase. In this regard, it is interesting to note that Dentalia were available for trade from the Fraser River basin as early as the first millennium B.C. (Sanger 1966: Lochnore Creek site. Zone I).

2. In the Columbia Plateau, the earliest confirmed occurrences of ground stone and mussel shell (Mytilus californicus) adzes date from the first half of the Cayuse Phase.

3. Trade in Olivella biplicata and Schizotherus nutally continues but does not appear to increase in volume. In general, Dentalium is a more common trade item.

4. Other marine shells, which are unknown from pre-Cayuse Phase components, make their first appearance in the first portion of the Cayuse Phase (e.g., C. M. Nelson 1966:44, Odostomia tenuisculpta).

It might be suggested that the apparent increase in coastal trade which occurs at the beginning of the Cayuse Phase is due to the expansion of Chinookan commercial activity up the Columbia River to the vicinity of The Dalles and Wishram. Although the archaeological record does not provide a final answer to this question, the patterns of prehistoric trade, as they are presently understood, do not support this view. If the trade characteristic of the early portion of the Cayuse Phase were simply a product of Chinookan expansion, one would expect an increase in Olivella traffic and a high concentration of trade goods in the southern Columbia Plateau. In fact, however, the abundance of Olivella does not appear to increase greatly at the beginning of the Cayuse Phase. Moreover, the introduction of Dentalium beads in greater concentrations in the northern Columbia Plateau contrasts sharply with the more frequent occurrence of Olivella beads in the southern Columbia Plateau and suggests that there are now two separate avenues of coastal trade, one along the Columbia River in the south and the other through the Canadian Plateau in the north. The Columbia River route represented a continuation of an earlier pattern, while the northerly route represented a new pattern of trading.

As mentioned in the discussion of the Cayuse III Subphase, the vigorous Chinookan trade observable in the early historic period is reflected in the archaeological record and does not begin until the onset of the protohistoric period when the whole trading network of the Plateau was greatly extended.

The Increased Importance of Fishing. Although the archaeological record is more tenuous on this point, the temporal distribution of fish remains and fishing implements seems to indicate that fishing became more important at the beginning of the Cayuse Phase. Since this posited characteristic of the Cayuse Phase is extremely important to the interpretation of the significance of the Cayuse Phase, its discussion will be deferred until the presentation of the Salishan expansion hypothesis. [46]

[47] Patterned Circumperipheral Diffusion in the Columbia Plateau. The emergence of the Cayuse Phase is marked by a curious pattern of diffusion in which projectile point types are locally diffused southward and eastward along the western and southern margins of the Columbia Plateau. Since this phenomenon is important to the understanding of the beginning of the Cayuse Phase, its discussion will be deferred until the presentation of the Salishan expansion hypothesis. [47]

TOP

LAST REVISED: 02 JAN 2015